From bugzilla-daemon at cs.uiuc.edu  Wed Apr  1 13:01:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Apr 2009 13:01:25 -0500
Subject: [LLVMbugs] [Bug 3862] Invalid operand for inline asm constraint 'I'!
In-Reply-To: <bug-3862-206@http.llvm.org/bugs/>
Message-ID: <200904011801.n31I1P7c020277@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3862


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-04-01 13:01:23 ---
Fixed in svn revision 68218.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  1 15:39:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Apr 2009 15:39:00 -0500
Subject: [LLVMbugs] [Bug 3923] New: Undef store doesn't clobber earlier
 stores: overaggressive instcombine
Message-ID: <bug-3923-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3923

           Summary: Undef store doesn't clobber earlier stores:
                    overaggressive instcombine
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


In some admittedly naive stack-based code, I'm trying to store undef into
popped stack locations so that LLVM will optimize away unneeded stores.
However, the instcombine pass seems to eliminate undef stores before other
passes can use them.

Here's the original code:

$ llvm-dis <~/tmp/dead_store.bc 
; ModuleID = '<stdin>'

declare i32 @rand()

define i32 @foo(i32* %stack) {
entry:
        %0 = call i32 @rand()           ; <i32> [#uses=1]
        store i32 %0, i32* %stack
        %1 = load i32* %stack           ; <i32> [#uses=1]
        store i32 undef, i32* %stack
        ret i32 %1
}

GVN + DSE does the right thing:

$ opt -gvn -dse ~/tmp/dead_store.bc |llvm-dis
; ModuleID = '<stdin>'

declare i32 @rand()

define i32 @foo(i32* %stack) {
entry:
        %0 = call i32 @rand()           ; <i32> [#uses=1]
        store i32 undef, i32* %stack
        ret i32 %0
}

But instcombine alone just removes the "store i32 undef...", which means no
subsequent pass can optimize away the store I was trying to clobber.

$ opt -instcombine ~/tmp/dead_store.bc |llvm-dis
; ModuleID = '<stdin>'

declare i32 @rand()

define i32 @foo(i32* %stack) {
entry:
        %0 = call i32 @rand()           ; <i32> [#uses=2]
        store i32 %0, i32* %stack
        ret i32 %0
}

This causes -O3 and -std-compile-opts to fail to optimize the code too:

$ opt -O3 ~/tmp/dead_store.bc |llvm-dis
; ModuleID = '<stdin>'

declare i32 @rand()

define i32 @foo(i32* %stack) {
entry:
        %0 = tail call i32 @rand()              ; <i32> [#uses=2]
        store i32 %0, i32* %stack
        ret i32 %0
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  1 16:50:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Apr 2009 16:50:48 -0500
Subject: [LLVMbugs] [Bug 3924] New: -load-vn and -gcse still mentioned by
 AliasAnalysis.html
Message-ID: <bug-3924-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3924

           Summary: -load-vn and -gcse still mentioned by AliasAnalysis.html
           Product: Documentation
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


http://llvm.org/docs/AliasAnalysis.html#loadvn recommends -load-vn
and
http://llvm.org/docs/AliasAnalysis.html#aliasanalysis-xforms describes -load-vn
and -gcse

Neither of those passes exist today.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  1 20:54:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Apr 2009 20:54:25 -0500
Subject: [LLVMbugs] [Bug 3342] Distinguish pointer incompatibilities where
 only sign differs from other incompatibilities
In-Reply-To: <bug-3342-206@http.llvm.org/bugs/>
Message-ID: <200904020154.n321sP6H009936@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3342


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-01 20:54:25 ---
Option added to driver in r68062.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  1 22:40:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 1 Apr 2009 22:40:04 -0500
Subject: [LLVMbugs] [Bug 3923] Undef store doesn't clobber earlier stores:
 overaggressive instcombine
In-Reply-To: <bug-3923-206@http.llvm.org/bugs/>
Message-ID: <200904020340.n323e46n013704@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3923


Jeffrey Yasskin <jyasskin at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #4 from Jeffrey Yasskin <jyasskin at google.com>  2009-04-01 22:40:00 ---
I've marked it as WONTFIX (would have gone for NOTABUG, but that doesn't exist)
since afaik you see the same behavior, just don't consider it wrong. Feel free
to change the resolution if I've misunderstood the labels or something. I'd
like to add a note to the documentation somewhere so the next person doesn't
make the same mistake. Is http://llvm.org/docs/LangRef.html#undefvalues the
right place, or is somewhere in the wiki better?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 00:19:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 00:19:34 -0500
Subject: [LLVMbugs] [Bug 3903] error: cannot compile this regparm attribute
	yet
In-Reply-To: <bug-3903-206@http.llvm.org/bugs/>
Message-ID: <200904020519.n325JY9F017375@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3903


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-02 00:19:33 ---


*** This bug has been marked as a duplicate of bug 3782 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 07:44:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 07:44:00 -0500
Subject: [LLVMbugs] [Bug 3925] New: clang-checker: detect undefined division
	behaviour
Message-ID: <bug-3925-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3925

           Summary: clang-checker: detect undefined division behaviour
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Clang checker already provides very useful warnings about potential divison by
zero.
It would be useful if clang checker could also provide warnings for other cases
where a division/remainder can cause a crash, like:
 INT_MIN / -1 (when both values are signed ints, ditto for long and long long
with appropriate corner cases)
 INT_MIN % -1 (as above)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 08:29:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 08:29:07 -0500
Subject: [LLVMbugs] [Bug 3926] New:
 ../../../svn/llvm-gcc-4.2/trunk/gcc/crtstuff.c:1: internal compiler error:
 Aborted
Message-ID: <bug-3926-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3926

           Summary: ../../../svn/llvm-gcc-4.2/trunk/gcc/crtstuff.c:1:
                    internal compiler error: Aborted
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: slavitch at gmail.com
                CC: llvmbugs at cs.uiuc.edu


At revision 68304.


creating as
echo timestamp > stamp-as
creating collect-ld
echo timestamp > stamp-collect-ld
creating nm
echo timestamp > stamp-nm
objext='.o' \
        LIB1ASMFUNCS='' \
        LIB2FUNCS_ST='_eprintf __gcc_bcmp' \
        LIB2FUNCS_EXCLUDE='' \
        LIBGCOV='_gcov _gcov_merge_add _gcov_merge_single _gcov_merge_delta
_gcov_fork _gcov_execl _gcov_execlp _gcov_execle _gcov_execv _gcov_execvp
_gcov_execve _gcov_interval_profiler _gcov_pow2_profiler
_gcov_one_value_profiler' \
        LIB2ADD='' \
        LIB2ADD_ST='' \
        LIB2ADDEH='../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2-fde-glibc.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-sjlj.c
../../../svn/llvm-gcc-4.2/trunk/gcc/gthr-gnat.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-c.c' \
        LIB2ADDEHSTATIC='../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2-fde-glibc.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-sjlj.c
../../../svn/llvm-gcc-4.2/trunk/gcc/gthr-gnat.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-c.c' \
        LIB2ADDEHSHARED='../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-dw2-fde-glibc.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-sjlj.c
../../../svn/llvm-gcc-4.2/trunk/gcc/gthr-gnat.c
../../../svn/llvm-gcc-4.2/trunk/gcc/unwind-c.c' \
        LIB2ADDEHDEP='unwind.inc unwind-dw2-fde.h unwind-dw2-fde.c' \
        LIB2_SIDITI_CONV_FUNCS='' \
        LIBUNWIND='' \
        LIBUNWINDDEP='' \
        SHLIBUNWIND_LINK='' \
        SHLIBUNWIND_INSTALL='' \
        FPBIT='' \
        FPBIT_FUNCS='_pack_sf _unpack_sf _addsub_sf _mul_sf _div_sf
_fpcmp_parts_sf _compare_sf _eq_sf _ne_sf _gt_sf _ge_sf _lt_sf _le_sf _unord_sf
_si_to_sf _sf_to_si _negate_sf _make_sf _sf_to_df _sf_to_tf _thenan_sf
_sf_to_usi _usi_to_sf' \
        LIB2_DIVMOD_FUNCS='_divdi3 _moddi3 _udivdi3 _umoddi3 _udiv_w_sdiv
_udivmoddi4' \
        DPBIT='' \
        DPBIT_FUNCS='_pack_df _unpack_df _addsub_df _mul_df _div_df
_fpcmp_parts_df _compare_df _eq_df _ne_df _gt_df _ge_df _lt_df _le_df _unord_df
_si_to_df _df_to_si _negate_df _make_df _df_to_sf _df_to_tf _thenan_df
_df_to_usi _usi_to_df' \
        TPBIT='' \
        TPBIT_FUNCS='_pack_tf _unpack_tf _addsub_tf _mul_tf _div_tf
_fpcmp_parts_tf _compare_tf _eq_tf _ne_tf _gt_tf _ge_tf _lt_tf _le_tf _unord_tf
_si_to_tf _tf_to_si _negate_tf _make_tf _tf_to_df _tf_to_sf _thenan_tf
_tf_to_usi _usi_to_tf' \
        DFP_ENABLE='true' \
        DFP_CFLAGS='' \
        D32PBIT='' \
        D32PBIT_FUNCS='_addsub_sd _div_sd _mul_sd _plus_sd _minus_sd _eq_sd
_ne_sd _lt_sd _gt_sd _le_sd _ge_sd _sd_to_si _sd_to_di _sd_to_usi _sd_to_udi
_si_to_sd _di_to_sd _usi_to_sd _udi_to_sd _sd_to_sf _sd_to_df _sd_to_xf
_sf_to_sd _df_to_sd _xf_to_sd _sd_to_dd _sd_to_td _unord_sd _conv_sd' \
        D64PBIT='' \
        D64PBIT_FUNCS='_addsub_dd _div_dd _mul_dd _plus_dd _minus_dd _eq_dd
_ne_dd _lt_dd _gt_dd _le_dd _ge_dd _dd_to_si _dd_to_di _dd_to_usi _dd_to_udi
_si_to_dd _di_to_dd _usi_to_dd _udi_to_dd _dd_to_sf _dd_to_df _dd_to_xf
_sf_to_dd _df_to_dd _xf_to_dd _dd_to_sd _dd_to_td _unord_dd _conv_dd' \
        D128PBIT='' \
        D128PBIT_FUNCS='_addsub_td _div_td _mul_td _plus_td _minus_td _eq_td
_ne_td _lt_td _gt_td _le_td _ge_td _td_to_si _td_to_di _td_to_usi _td_to_udi
_si_to_td _di_to_td _usi_to_td _udi_to_td _td_to_sf _td_to_df _td_to_xf
_sf_to_td _df_to_td _xf_to_td _td_to_sd _td_to_dd _unord_td _conv_td' \
        APPLE_LOCAL='APPLE LOCAL libcc_kext' \
        MULTILIBS="`/Src/build/llvm-gcc-4.2/./gcc/xgcc
-B/Src/build/llvm-gcc-4.2/./gcc/ -B/usr/local/i686-pc-linux-gnu/bin/
-B/usr/local/i686-pc-linux-gnu/lib/ -isystem
/usr/local/i686-pc-linux-gnu/include -isystem
/usr/local/i686-pc-linux-gnu/sys-include --print-multi-lib` " \
        EXTRA_MULTILIB_PARTS='' \
        SHLIB_LINK='/Src/build/llvm-gcc-4.2/./gcc/xgcc
-B/Src/build/llvm-gcc-4.2/./gcc/ -B/usr/local/i686-pc-linux-gnu/bin/
-B/usr/local/i686-pc-linux-gnu/lib/ -isystem
/usr/local/i686-pc-linux-gnu/include -isystem
/usr/local/i686-pc-linux-gnu/sys-include -O2  -O2 -g -O2  -DIN_GCC    -W -Wall
-Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition
 -isystem ./include  -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2
-D__GCC_FLOAT_NOT_NEEDED  -shared -nodefaultlibs
-Wl,--soname=@shlib_base_name at .so.1 -Wl,--version-script=@shlib_map_file@ -o
@multilib_dir@/@shlib_base_name at .so.1.tmp @multilib_flags@ @shlib_objs@ -lc &&
rm -f @multilib_dir@/@shlib_base_name at .so && if [ -f
@multilib_dir@/@shlib_base_name at .so.1 ]; then mv -f
@multilib_dir@/@shlib_base_name at .so.1
@multilib_dir@/@shlib_base_name at .so.1.backup; else true; fi && mv
@multilib_dir@/@shlib_base_name at .so.1.tmp @multilib_dir@/@shlib_base_name at .so.1
&& ln -s @shlib_base_name at .so.1 @multilib_dir@/@shlib_base_name at .so' \
        SHLIB_INSTALL='$(mkinstalldirs)
$(DESTDIR)$(slibdir)@shlib_slibdir_qual@; /usr/bin/install -c -m 644
@multilib_dir@/@shlib_base_name at .so.1
$(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so.1; rm -f
$(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so; ln -s
@shlib_base_name at .so.1
$(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so' \
        SHLIB_EXT='.so' \
        SHLIB_MULTILIB='' \
        SHLIB_MKMAP='../../../svn/llvm-gcc-4.2/trunk/gcc/mkmap-symver.awk' \
        SHLIB_MKMAP_OPTS='' \
        SHLIB_MAPFILES='../../../svn/llvm-gcc-4.2/trunk/gcc/libgcc-std.ver
../../../svn/llvm-gcc-4.2/trunk/gcc/config/libgcc-glibc.ver' \
        SHLIB_NM_FLAGS='-pg' \
        MULTILIB_OSDIRNAMES='' \
        ASM_HIDDEN_OP='' \
        GCC_FOR_TARGET='/Src/build/llvm-gcc-4.2/./gcc/xgcc
-B/Src/build/llvm-gcc-4.2/./gcc/ -B/usr/local/i686-pc-linux-gnu/bin/
-B/usr/local/i686-pc-linux-gnu/lib/ -isystem
/usr/local/i686-pc-linux-gnu/include -isystem
/usr/local/i686-pc-linux-gnu/sys-include' \
        mkinstalldirs='/bin/sh
../../../svn/llvm-gcc-4.2/trunk/gcc/../mkinstalldirs' \
          /bin/sh mklibgcc > tmp-libgcc.mk
mv tmp-libgcc.mk libgcc.mk
TARGET_CPU_DEFAULT="" \
        HEADERS="auto-host.h ansidecl.h" DEFINES="USED_FOR_TARGET " \
        /bin/sh ../../../svn/llvm-gcc-4.2/trunk/gcc/mkconfig.sh tconfig.h
/Src/build/llvm-gcc-4.2/./gcc/xgcc -B/Src/build/llvm-gcc-4.2/./gcc/
-B/usr/local/i686-pc-linux-gnu/bin/ -B/usr/local/i686-pc-linux-gnu/lib/
-isystem /usr/local/i686-pc-linux-gnu/include -isystem
/usr/local/i686-pc-linux-gnu/sys-include -O2 -O2 -g -O2  -DIN_GCC    -W -Wall
-Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition
 -isystem ./include  -I. -I. -I../../../svn/llvm-gcc-4.2/trunk/gcc
-I../../../svn/llvm-gcc-4.2/trunk/gcc/.
-I../../../svn/llvm-gcc-4.2/trunk/gcc/../include
-I../../../svn/llvm-gcc-4.2/trunk/gcc/../libcpp/include 
-I../../../svn/llvm-gcc-4.2/trunk/gcc/../libdecnumber -I../libdecnumber
-I/Src/build/llvm//include -I/Src/svn/llvm/include  -g0
-finhibit-size-directive -fno-inline-functions -fno-exceptions
-fno-zero-initialized-in-bss -fno-toplevel-reorder  -fno-omit-frame-pointer \
          -c ../../../svn/llvm-gcc-4.2/trunk/gcc/crtstuff.c -DCRT_BEGIN \
          -o crtbegin.o
cc1: ../../../svn/llvm-gcc-4.2/trunk/gcc/llvm-backend.cpp:434: void
llvm_initialize_backend(): Assertion `TheTarget->getTargetData()->isBigEndian()
== 0' failed.
../../../svn/llvm-gcc-4.2/trunk/gcc/crtstuff.c:1: internal compiler error:
Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.
make[3]: *** [crtbegin.o] Error 1
make[3]: Leaving directory `/Src/build/llvm-gcc-4.2/gcc'
make[2]: *** [all-stage1-gcc] Error 2
make[2]: Leaving directory `/Src/build/llvm-gcc-4.2'
make[1]: *** [stage1-bubble] Error 2
make[1]: Leaving directory `/Src/build/llvm-gcc-4.2'
make: *** [all] Error 2


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 09:56:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 09:56:40 -0500
Subject: [LLVMbugs] [Bug 3927] New: Clang expands the same macro differently
 depending on its identifier
Message-ID: <bug-3927-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3927

           Summary: Clang expands the same macro differently depending on
                    its identifier
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


I submitted this before, but it was abruptly closed with "resubmit". 
Anyway...last time.

Observe the output of -E for the following; both lines output should be the
same:

#define f(x) h(x
#define for(x) h(x    // for is the same macro as f .....
#define h(x) x(void) 
extern int f(f));
extern int for(for)); // but it exapnds differently....

The problem is rooted in Clang's representation of "is this token an
identifier?".  It is tempting to answer that question by testing the token
type, because this is the correct decision for all token types but identifiers.

Identifiers are treated differently and the correct test for identifiers is
only to test for identifier info being non-null (I believe).

This non-obvious situation led to this bug; which I noticed perusing the source
code (rather than dreaming up obscene tests, in case the reader wondered).

I suggest all of CPP is audited for this mistake, assuming this design decision
(efficiency perhaps?) is still deemed the right one.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 12:04:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 12:04:27 -0500
Subject: [LLVMbugs] [Bug 3928] New: Grammatical error in first sentence of
 main Clang page
Message-ID: <bug-3928-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3928

           Summary: Grammatical error in first sentence of main Clang page
           Product: Website
           Version: unspecified
          Platform: All
               URL: http://clang.llvm.org/StaticAnalysis.html
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: General Website
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: flash at pobox.com
                CC: llvmbugs at cs.uiuc.edu


The first sentence of the main web page for Clang,
http://clang.llvm.org/StaticAnalysis.html, says ??? The LLVM/Clang static
analyzer is a standalone tool that find bugs??????  ???find??? should be
???finds???.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 13:20:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 13:20:24 -0500
Subject: [LLVMbugs] [Bug 3928] Grammatical error in first sentence of main
	Clang page
In-Reply-To: <bug-3928-206@http.llvm.org/bugs/>
Message-ID: <200904021820.n32IKOib027849@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3928


Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kremenek at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Ted Kremenek <kremenek at apple.com>  2009-04-02 13:20:24 ---
Fixed!  Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 13:40:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 13:40:01 -0500
Subject: [LLVMbugs] [Bug 3907] Incorrect parsing of Class.property.property
	syntax
In-Reply-To: <bug-3907-206@http.llvm.org/bugs/>
Message-ID: <200904021840.n32Ie1lq028740@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3907


snaroff at apple.com <snaroff at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from snaroff at apple.com  2009-04-02 13:40:00 ---
Fixed in http://llvm.org/viewvc/llvm-project?rev=68338&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 21:04:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 21:04:06 -0500
Subject: [LLVMbugs] [Bug 3929] New: Rounding modes are ignored in gcc front
 end constant folding
Message-ID: <bug-3929-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3929

           Summary: Rounding modes are ignored in gcc front end constant
                    folding
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: llvm at nraynaud.com
                CC: llvmbugs at cs.uiuc.edu


I toyed around with http://llvm.org/demo/index.cgi

and found that this code : 

#include <stdio.h>
#include <stdlib.h>
#include <fenv.h>

#include <math.h>

int main(int argc, char **argv) {
#pragma STDC FENV_ACCESS ON
  fesetround(FE_UPWARD);
  double g = sin(0.9);
  printf("%A\n", g);

  fesetround(FE_DOWNWARD);
  double h = sin(0.9);
  printf("%A\n", h);
  return 0;
}

produces the same bit patterns for g and h, which is obviously false
define i32 @main(i32 %argc, i8** nocapture %argv) nounwind {
entry:
        %0 = tail call i32 @fesetround(i32 2048) nounwind               ; <i32>
[#uses=0]
        %1 = tail call i32 (i8*, ...)* @printf(i8* noalias getelementptr ([4 x
i8]* @.str, i32 0, i32 0), double 0x3FE91103985DA841) nounwind           ;
<i32> [#uses=0]
        %2 = tail call i32 @fesetround(i32 1024) nounwind               ; <i32>
[#uses=0]
        %3 = tail call i32 (i8*, ...)* @printf(i8* noalias getelementptr ([4 x
i8]* @.str, i32 0, i32 0), double 0x3FE91103985DA841) nounwind           ;
<i32> [#uses=0]
        ret i32 0
}


here is the result on x86_32 :
nraynaud at macaron-2:~/Desktop/geek/fp$ gcc  test2.c && ./a.out
0X1.91103985DA842P-1
0X1.91103985DA858P-1


This is a common IEEE 754 mistake.

Either, fesetround should be interpreted before folding the constant, or IEEE
754 shouldn't be folded in the presence of the pragma.

It should be noted that only the pragma allows for toying with rounding mode
(and flag bits), careless folding is free if the pragma is not there. So people
who don't care about ieee stuff get their usual speed, those that know ieee
stuff get what they bargained for, and speed is often the price to pay. Those
who try to play with rounding modes without setting the pragma spend hours
finding why it doesn't work reliably.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 21:23:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 21:23:31 -0500
Subject: [LLVMbugs] [Bug 3930] New: floating point exceptions flags are
 ignored in constant folding
Message-ID: <bug-3930-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3930

           Summary: floating point exceptions flags are ignored in constant
                    folding
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: llvm at nraynaud.com
                CC: llvmbugs at cs.uiuc.edu


I toyed around with http://llvm.org/demo/index.cgi

and found that this code : 
#include <stdio.h>
#include <fenv.h>

int main(int argc, char **argv) {
#pragma STDC FENV_ACCESS ON
  feclearexcept(FE_DIVBYZERO);
  double d = 1.0/0.0;
  if (fetestexcept(FE_DIVBYZERO))
    printf("error\n");
  else
    printf("%e\n", d);
  return 0;
}

is compiled this way :

define i32 @main(i32 %argc, i8** nocapture %argv) nounwind {
entry:
        %0 = tail call i32 @feclearexcept(i32 4) nounwind               ; <i32>
[#uses=0]
        %1 = tail call i32 @fetestexcept(i32 4) nounwind                ; <i32>
[#uses=1]
        %2 = icmp eq i32 %1, 0          ; <i1> [#uses=1]
        br i1 %2, label %bb1, label %bb

bb:             ; preds = %entry
        %3 = tail call i32 @puts(i8* getelementptr ([6 x i8]* @.str, i32 0, i32
0)) nounwind            ; <i32> [#uses=0]
        ret i32 0

bb1:            ; preds = %entry
        %4 = tail call i32 (i8*, ...)* @printf(i8* noalias getelementptr ([4 x
i8]* @.str1, i32 0, i32 0), double 0x7FF0000000000000) nounwind          ;
<i32> [#uses=0]
        ret i32 0
}

it as basically no chance to print "error" like it does under a normal C
compiler (tested on normal gcc x86_32).

the pragma is here to tell the compiler maker to be careful in floating point
handling because there will be some tampering with the FP subsystem.

exactly the same kind of bug as Bug 3929


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 21:55:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 21:55:23 -0500
Subject: [LLVMbugs] [Bug 3929] Rounding modes are ignored in gcc front end
 constant folding
In-Reply-To: <bug-3929-206@http.llvm.org/bugs/>
Message-ID: <200904030255.n332tNRI015680@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3929


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-02 21:55:23 ---
This is a GCC bug or missing feature, please report it at
gcc.gnu.org/bugzilla/.  I'm sure this is a known problem though.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 21:55:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 21:55:58 -0500
Subject: [LLVMbugs] [Bug 3930] floating point exceptions flags are ignored
 in constant folding
In-Reply-To: <bug-3930-206@http.llvm.org/bugs/>
Message-ID: <200904030255.n332tww4015709@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3930


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-02 21:55:58 ---
GCC ignores #pragma STDC FENV_ACCESS.  This is a known issue.  Please search
the gcc bugzilla for details.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 22:01:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 22:01:33 -0500
Subject: [LLVMbugs] [Bug 3930] floating point exceptions flags are ignored
 in constant folding
In-Reply-To: <bug-3930-206@http.llvm.org/bugs/>
Message-ID: <200904030301.n3331Xre015875@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3930


Nicolas Raynaud <llvm at nraynaud.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #2 from Nicolas Raynaud <llvm at nraynaud.com>  2009-04-02 22:01:32 ---
(In reply to comment #1)
> GCC ignores #pragma STDC FENV_ACCESS.  This is a known issue.  Please search
> the gcc bugzilla for details.
> 

This code works under gcc (mac) anyways, I was careful to test it and say so in
the report.
If they ignore the pragma, the use the real code to decide if they can optimise
or not. Which clearly LLVM doesn't.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  2 22:14:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 2 Apr 2009 22:14:43 -0500
Subject: [LLVMbugs] [Bug 3930] floating point exceptions flags are ignored
 in constant folding
In-Reply-To: <bug-3930-206@http.llvm.org/bugs/>
Message-ID: <200904030314.n333Eh5W016301@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3930


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |INVALID




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-02 22:14:38 ---
No, they just get lucky.  Try compiling at -O3 or some other high optimization
level.  GCC does not respect this.  If this one test case happens to work, it
is by chance, not design.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr  3 05:56:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Apr 2009 05:56:06 -0500
Subject: [LLVMbugs] [Bug 3804] llvm 2.5 fails to build lli on ia64 and mips
In-Reply-To: <bug-3804-206@http.llvm.org/bugs/>
Message-ID: <200904031056.n33Au6Vb015941@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3804


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Anton Korobeynikov <asl at math.spbu.ru>  2009-04-03 05:55:54 ---
Should be fixed in:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090330/076001.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090330/076002.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090330/076003.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090330/076004.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr  3 06:09:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Apr 2009 06:09:36 -0500
Subject: [LLVMbugs] [Bug 3932] New: [ObjC]Type defined as 'id' is not
 recognized as a valid object type.
Message-ID: <bug-3932-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3932

           Summary: [ObjC]Type defined as 'id' is not recognized as a valid
                    object type.
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: devlists at shadowlab.org
                CC: llvmbugs at cs.uiuc.edu


When defining a new type as an id, this new type is not recognized as a valid
id (objc object) by clang.
This is particulary annoying when this type is used where an objc object is
required, as clang will report an error where gcc does not.

------------- test.m  --------------
typedef id BYObjectIdentifier;

@interface Foo  {
  void *isa;
}

@property(copy) BYObjectIdentifier identifier;

@end

------------------------

clang -c test.m
error: property with 'copy' attribute must be of object type
    @property(copy) BYObjectIdentifier identifier;


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr  3 08:07:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 3 Apr 2009 08:07:13 -0500
Subject: [LLVMbugs] [Bug 3933] New: ruby 1.9. 1 does not compile anymore due
 to unimplemented __asm__
Message-ID: <bug-3933-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3933

           Summary: ruby 1.9.1 does not compile anymore due to unimplemented
                    __asm__
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: roberto at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2781)
 --> (http://llvm.org/bugs/attachment.cgi?id=2781)
vm_exec.c

At around r68030, ruby was compiling mostly fine (see other ticket on tests
failing), now it does not anymore (I'm now using r68372 on OSX).

-----
ruby -Ks ../tool/insns2vm.rb --srcdir=".." vmtc.inc
ruby -Ks ../tool/insns2vm.rb --srcdir=".." vm.inc
clang -O -pipe -m64 -g -Wall -Wno-unused-parameter -Wno-parentheses
-Wno-missing-field-initializers -Wshorten-64-to-32 -Wpointer-arith
-Wwrite-strings -I. -I.ext/include/i386-darwin9.6.0 -I../include -I..
-DRUBY_EXPORT -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE   -o vm.o -c ../vm.c
clang: warning: argument unused during compilation: '-Wno-parentheses'
...
<lots of "uninteresting" warnings skipped>
In file included from ../vm.c:23:
../vm_exec.c:58:5: error: cannot compile this __asm__ yet
    DECL_SC_REG(VALUE *, pc, "14");
    ^
<scratch space>:183:1: note: instantiated from:
reg_pc
^
../vm_exec.c:59:5: error: cannot compile this __asm__ yet
    DECL_SC_REG(rb_control_frame_t *, cfp, "15");
    ^
<scratch space>:184:1: note: instantiated from:
reg_cfp
^
-----


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 07:33:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 07:33:14 -0500
Subject: [LLVMbugs] [Bug 3934] New: Assertion failed: (isNew && "Can' t map
 to two src physical registers!"), function ProcessCopy
Message-ID: <bug-3934-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3934

           Summary: Assertion failed: (isNew && "Can't map to two src
                    physical registers!"), function ProcessCopy
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=2782)
 --> (http://llvm.org/bugs/attachment.cgi?id=2782)
Testcase

When compiling the attached code with Clang on AMD64 with at least -O1 enabled,
I'm seeing the following failed assertion:

Assertion failed: (isNew && "Can't map to two src physical registers!"),
function ProcessCopy, file .../llvm/lib/CodeGen/TwoAddressInstructionPass.cpp,
line 591.
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free --relocation-model static
--disable-fp-elim --unwind-tables=1 --fmath-errno=1 -O1 -o /tmp/cc-9EhLJ2.s -x
c /home/ed/llvm-crash.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Two-Address instruction pass' on function '@__bswap16'

When passing -O2, the following crash is observed:

Assertion failed: (isNew && "Can't map to two src physical registers!"),
function ProcessCopy, file .../llvm/lib/CodeGen/TwoAddressInstructionPass.cpp,
line 591.
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free --relocation-model static
--disable-fp-elim --unwind-tables=1 --fmath-errno=1 -O2 -o /tmp/cc-PZNza1.s -x
c /home/ed/llvm-crash.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Two-Address instruction pass' on function
'@config_get_hosts'

This prevents the FreeBSD source tree from being built.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 09:20:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 09:20:58 -0500
Subject: [LLVMbugs] [Bug 3935] New: Invalid addresses at compile time not
	diagnosed
Message-ID: <bug-3935-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3935

           Summary: Invalid addresses at compile time not diagnosed
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


Clang should diagnose invalid address compuatations; e.g. 

    int x[2], *y = &x[3], *z = x + 3;

Clang accepts this without comment.  Comeau gives

"ComeauTest.c", line 1: warning: subscript out of range
  int x[2], *y = &x[3], *z = x + 3;
                   ^

"ComeauTest.c", line 1: warning: pointer points outside of underlying object
  int x[2], *y = &x[3], *z = x + 3;
                               ^

This catches many bugs at compile time.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 17:48:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 17:48:41 -0500
Subject: [LLVMbugs] [Bug 3936] New: Instcombine quadtratic in GEP chain
	length
Message-ID: <bug-3936-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3936

           Summary: Instcombine quadtratic in GEP chain length
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: nicholas at mxc.ca, llvmbugs at cs.uiuc.edu


Created an attachment (id=2783)
 --> (http://llvm.org/bugs/attachment.cgi?id=2783)
Script to generate quadratic assembly

The attached script, when passed an integer argument N, generates an LLVM
assembly file containing a chain of 2N-1 GEP instructions:

$ ./instcombine_test.py 2
declare void @use(i32)

define void @foo(i32* %stack) {
  %B0 = getelementptr i32* %stack, i32 0

  %A1 = getelementptr i32* %B0, i32 1
  %B1 = getelementptr i32* %A1, i32 -1
  %C1 = load i32* %B1
  call void @use(i32 %C1)

  ret void
}

The instcombine pass is quadratic when trying to optimize the chain:

$ for n in 1000 2000 3000 4000 5000 6000; do ./instcombine_test.py
$n|$LLVM/llvm-as|time $LLVM/opt -instcombine >/dev/null;done
        1.03 real         0.85 user         0.01 sys
        3.22 real         2.89 user         0.03 sys
        6.55 real         5.94 user         0.05 sys
       11.40 real        10.47 user         0.10 sys
       17.82 real        16.52 user         0.16 sys
       24.73 real        23.02 user         0.21 sys


Shark reports that a very high percentage of the time is spent in
Value::getUnderlyingObject, resolving loads like %C1 down to %stack. It doesn't
appear to combine any GEPs before calling getUnderlyingObject on every load in
the function, causing quadratic work.

Thanks to Nick for the form of the testcase. I've seen the behavior in both 2.5
and trunk.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 20:42:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 20:42:05 -0500
Subject: [LLVMbugs] [Bug 3937] New: Clang isn't checking macro argument
	counts properly
Message-ID: <bug-3937-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3937

           Summary: Clang isn't checking macro argument counts properly
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


Just one example:

#define zero() 0
int y = zero(1);


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 20:49:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 20:49:33 -0500
Subject: [LLVMbugs] [Bug 3938] New: Clang rejects valid TU with
	-pedantic-errors
Message-ID: <bug-3938-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3938

           Summary: Clang rejects valid TU with -pedantic-errors
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


extern int x;
#define D

#ifndef D
#ifdef D
#else 1
#endif
#endif


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 20:51:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 20:51:49 -0500
Subject: [LLVMbugs] [Bug 3939] New: CPP grammar violation not diagnosed
Message-ID: <bug-3939-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3939

           Summary: CPP grammar violation not diagnosed
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


The following violates the CPP grammar with #elif.  (GCC was only recently
fixed to reject this).

extern int x;
#if 1
#elif
#endif


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 21:00:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 21:00:26 -0500
Subject: [LLVMbugs] [Bug 3940] New: #line numbers not fully checked
Message-ID: <bug-3940-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3940

           Summary: #line numbers not fully checked
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


The std says the line number is decimal.  Since this is not a constraint
diagnosing this is not required; just a QOI issue.

#line 0xf
#line 02U


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 21:03:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 21:03:40 -0500
Subject: [LLVMbugs] [Bug 3941] New: Invalid use of inline not diagnosed
Message-ID: <bug-3941-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3941

           Summary: Invalid use of inline not diagnosed
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


e.g.:

int f(void)
{
  inline int x = 0;

  return x;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 21:10:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 21:10:25 -0500
Subject: [LLVMbugs] [Bug 3942] New: Clang complains unsigned integer
 arithmetic overflows
Message-ID: <bug-3942-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3942

           Summary: Clang complains unsigned integer arithmetic overflows
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


one e.g.

#if 5U << 63
#endif


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr  4 21:15:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 4 Apr 2009 21:15:11 -0500
Subject: [LLVMbugs] [Bug 3943] New: Defining/undefining builtin macros
Message-ID: <bug-3943-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3943

           Summary: Defining/undefining builtin macros
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


The standard forbids this, but it is not a constraint.  Clang diagnoses some
but not others, e.g.

#undef __LINE__   // diagnosed
#undef __STDC_HOSTED__ // undiagnosed


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr  5 01:00:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Apr 2009 01:00:07 -0500
Subject: [LLVMbugs] [Bug 3944] New: mem2reg quadratic in number of in-edges
 to block receiving a new phi node
Message-ID: <bug-3944-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3944

           Summary: mem2reg quadratic in number of in-edges to block
                    receiving a new phi node
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=2784)
 --> (http://llvm.org/bugs/attachment.cgi?id=2784)
Script to make mem2reg quadratic

The attached script generates LLVM asm that triggers quadratic behavior in
mem2reg. The code looks like:

$ ./mem2reg_test.py 1
declare i1 @test()

define i32 @foo() {
entry:
  %retval_addr = alloca i32
  store i32 1, i32* %retval_addr
  %cond = call i1 @test()
  br i1 %cond, label %return_block, label %block0

block0:
  store i32 0, i32* %retval_addr
  %cond0 = call i1 @test()
  br i1 %cond0, label %return_block, label %block1

block1: br label %return_block
return_block:
  %retval = load i32* %retval_addr
  ret i32 %retval
}


After mem2reg, that's:

$ ./mem2reg_test.py 1|$LLVM/llvm-as|$LLVM/opt -mem2reg|$LLVM/llvm-dis
; ModuleID = '<stdin>'

declare i1 @test()

define i32 @foo() {
entry:
        %cond = call i1 @test()         ; <i1> [#uses=1]
        br i1 %cond, label %return_block, label %block0

block0:         ; preds = %entry
        %cond0 = call i1 @test()                ; <i1> [#uses=1]
        br i1 %cond0, label %return_block, label %block1

block1:         ; preds = %block0
        br label %return_block

return_block:           ; preds = %block1, %block0, %entry
        %retval_addr.0 = phi i32 [ 0, %block1 ], [ 0, %block0 ], [ 1, %entry ] 
        ; <i32> [#uses=1]
        ret i32 %retval_addr.0
}


mem2reg is quadratic in the number of arguments to the %retval_addr.0 phi node,
I believe because of the loop at the top of PromoteMem2Reg::RenamePass() that
starts with "for (unsigned i = 0, e = APN->getNumIncomingValues(); i != e;
++i)"

$ for N in 5000 10000 15000 20000 25000 30000; do ./mem2reg_test.py
$N|$LLVM/llvm-as|time $LLVM/opt -mem2reg > /dev/null; done
        1.09 real         0.66 user         0.01 sys
        2.83 real         1.96 user         0.03 sys
        5.38 real         3.97 user         0.05 sys
        8.54 real         6.62 user         0.08 sys
       12.22 real         9.92 user         0.11 sys
       16.67 real        13.90 user         0.11 sys

For comparison, llvm-as takes 0.33s for the 5000-block case, and goes up
linearly from there.

I'm marking this as "minor" because I only ran into it in the context of return
blocks, and I can work around it by manipulating the phi node myself.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr  5 15:54:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Apr 2009 15:54:49 -0500
Subject: [LLVMbugs] [Bug 3945] New: The ARM AsmPrinter doesn't recognize a
	modifier.
Message-ID: <bug-3945-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3945

           Summary: The ARM AsmPrinter doesn't recognize a modifier.
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rich at pennware.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2789)
 --> (http://llvm.org/bugs/attachment.cgi?id=2789)
Patch to handle ${0:a}

The ARM AsmPrinter flagged %a0 ${0:a} as an error. I believe that it means
"don't output an '#' in front of an immediate address value".
Could someone take a look at the enclosed patch and test case and tell me
if it looks OK?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr  5 17:11:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Apr 2009 17:11:49 -0500
Subject: [LLVMbugs] [Bug 3946] New: SmallVector::insert should use
	copy_backward
Message-ID: <bug-3946-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3946

           Summary: SmallVector::insert should use copy_backward
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: johnson.peter at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2791)
 --> (http://llvm.org/bugs/attachment.cgi?id=2791)
Testcase demonstrating the issue

SmallVector::insert() (for multiple elements) uses std::copy to copy the
elements that get replaced:
      // Copy the existing elements that get replaced.
      std::copy(I, OldEnd-NumToInsert, I+NumToInsert);

Unfortunately, I+NumToInsert can easily be within the (I, OldEnd-NumToInsert)
range.  Using std::copy_backward() fixes the problem:
      // Copy the existing elements that get replaced.
      std::copy_backward(I, OldEnd-NumToInsert, OldEnd);

We don't need to decide at runtime between these versions as OldEnd is always
after OldEnd-NumToInsert.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr  5 17:22:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Apr 2009 17:22:48 -0500
Subject: [LLVMbugs] [Bug 3947] New: SmallVector::insert overload ambiguous
	when T=int
Message-ID: <bug-3947-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3947

           Summary: SmallVector::insert overload ambiguous when T=int
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: johnson.peter at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2793)
 --> (http://llvm.org/bugs/attachment.cgi?id=2793)
Testcase demonstrating the issue

Two of the SmallVector::insert function overloads are ambiguous when T=int (or
something like size_type):

iterator insert(iterator I, size_type NumToInsert, const T &Elt)

and

template<typename ItTy>
iterator insert(iterator I, ItTy From, ItTy To)

SmallVector<int, X>::insert(..., 2, 2) results in G++ 4.3 picking the templated
overload (as ItTy=int) and ending up with a lot of errors.

Not sure how to fix this, other than untemplating the ItTy version.  Note that
std::vector does not have this limitation.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr  5 19:41:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 5 Apr 2009 19:41:21 -0500
Subject: [LLVMbugs] [Bug 3948] New: clang misparses parameter with same name
 as containing class
Message-ID: <bug-3948-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3948

           Summary: clang misparses parameter with same name as containing
                    class
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
class a {int b(int a[10]);};
<stdin>:1:21: error: array has incomplete element type 'void'
class a {int b(int a[10]);};
                    ^

I think there's something wrong with the usage of isCurrentClassName, but I'm
not sure whether Parser or Sema is at fault here.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 08:09:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 08:09:50 -0500
Subject: [LLVMbugs] [Bug 3661] ccc -O1 miscompiles git
In-Reply-To: <bug-3661-206@http.llvm.org/bugs/>
Message-ID: <200904061309.n36D9o55032501@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3661


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #10 from Ed Schouten <ed at 80386.nl>  2009-04-06 08:08:32 ---
Fixed in r68443. Benjamin, thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 10:47:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 10:47:27 -0500
Subject: [LLVMbugs] [Bug 3464] clang doesn't search current directory for
	-include
In-Reply-To: <bug-3464-206@http.llvm.org/bugs/>
Message-ID: <200904061547.n36FlRYq006052@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3464


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #5 from Daniel Dunbar <daniel at zuster.org>  2009-04-06 10:47:26 ---


*** This bug has been marked as a duplicate of bug 3395 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 12:21:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 12:21:01 -0500
Subject: [LLVMbugs] [Bug 3678] Invalid input/output constraint when
 compiling FreeBSD's libm/ OpenSSL on AMD64
In-Reply-To: <bug-3678-206@http.llvm.org/bugs/>
Message-ID: <200904061721.n36HL1Hj010269@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3678


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #17 from Ed Schouten <ed at 80386.nl>  2009-04-06 12:20:59 ---
Fixed in r68449.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 12:22:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 12:22:54 -0500
Subject: [LLVMbugs] [Bug 3678] Invalid input/output constraint when
 compiling FreeBSD's libm/ OpenSSL on AMD64
In-Reply-To: <bug-3678-206@http.llvm.org/bugs/>
Message-ID: <200904061722.n36HMsEw010410@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3678


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #18 from Ed Schouten <ed at 80386.nl>  2009-04-06 12:22:53 ---
Oomf. I meant to close 3889. Not this one.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 14:59:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 14:59:43 -0500
Subject: [LLVMbugs] [Bug 3939] CPP grammar violation not diagnosed
In-Reply-To: <bug-3939-206@http.llvm.org/bugs/>
Message-ID: <200904061959.n36Jxhsd016892@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3939


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-06 14:59:42 ---


*** This bug has been marked as a duplicate of bug 2291 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 16:10:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 16:10:20 -0500
Subject: [LLVMbugs] [Bug 3953] New: Instcombine quadratic in add chain length
Message-ID: <bug-3953-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3953

           Summary: Instcombine quadratic in add chain length
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=2799)
 --> (http://llvm.org/bugs/attachment.cgi?id=2799)
Generator for add chains that make instcombine quadratic

The attached script generates .ll files of the form:

$  ./instcombine_test3.py 3
declare void @use(i32)

define void @foo(i32 %B0) {

  %A1 = add i32 %B0, 1
  %B1 = add i32 %A1, -1
  call void @use(i32 %B1)

  %A2 = add i32 %B1, 1
  %B2 = add i32 %A2, -1
  call void @use(i32 %B2)

  ret void
}

instcombine takes approximately quadratic time to simplify these chains.

$ for N in 500 1000 1500 2000 2500; do ./instcombine_test3.py
$N|$LLVM/llvm-as|time $LLVM/opt -instcombine >/dev/null; done
        0.62 real         0.56 user         0.00 sys
        2.28 real         2.16 user         0.01 sys
        4.86 real         4.69 user         0.02 sys
        8.70 real         8.38 user         0.04 sys
       13.46 real        12.99 user         0.08 sys

This prevents me from applying the obvious workaround to bug 3936.

A time profile doesn't show an obvious culprit, spending 38% in
isInstructionTriviallyDead, 25% in visitCall, and 11% in visitAdd.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 16:14:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 16:14:52 -0500
Subject: [LLVMbugs] [Bug 3954] New: [ARM] An asm causes an assert.
Message-ID: <bug-3954-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3954

           Summary: [ARM] An asm causes an assert.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rich at pennware.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2800)
 --> (http://llvm.org/bugs/attachment.cgi?id=2800)
The source .ll file.

Compiling the attached .ll file with the command lines:

[~/ellcc/ellcc] main% llvm-as -f t0064.ll
[~/ellcc/ellcc] main% llc -march arm t0064.bc

results in:

llc: /home/rich/llvm-trunk-new/lib/CodeGen/RegisterScavenging.cpp:273: void
llvm::RegScavenger::forward(): Assertion `(isReserved(Reg) || isUnused(Reg) ||
IsImpDef || isImplicitlyDefined(Reg) || isLiveInButUnusedBefore(Reg, MI, MBB,
TRI, MRI)) && "Re-defining a live register!"' failed.                           
0   llc       0x08d5616d                                                        
1   llc       0x08d5669f                                                        
2             0x0077e400 __kernel_sigreturn + 0                                 
3   libc.so.6 0x00bcce28 abort + 392                                            
4   libc.so.6 0x00bc440e __assert_fail + 238
5   llc       0x08b1a6d8 llvm::RegScavenger::forward() + 1778
6   llc       0x08645554
llvm::RegScavenger::forward(llvm::ilist_iterator<llvm::MachineInstr>) + 126
7   llc       0x08ae8f12
8   llc       0x08aed977
9   llc       0x0845fc26
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 52
10  llc       0x08cd6ae2 llvm::FPPassManager::runOnFunction(llvm::Function&) +
288
11  llc       0x08cd75d0 llvm::FunctionPassManagerImpl::run(llvm::Function&) +
124
12  llc       0x08cd77a1 llvm::FunctionPassManager::run(llvm::Function&) + 159
13  llc       0x083d1662 main + 2687
14  libc.so.6 0x00bb66e5 __libc_start_main + 229
15  llc       0x083cfb21
Stack dump:
0.      Program arguments: llc -march arm t0064.bc
1.      Running pass 'Prolog/Epilog Insertion & Frame Finalization' on function
'@foo'
Abort

[~/ellcc/ellcc] main%


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 16:44:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 16:44:52 -0500
Subject: [LLVMbugs] [Bug 3955] New: llc -fast asserts on first-class-struct
	parameter
Message-ID: <bug-3955-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3955

           Summary: llc -fast asserts on first-class-struct parameter
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2801)
 --> (http://llvm.org/bugs/attachment.cgi?id=2801)
bitcode that causes assertion failure in llc -fast 

When I run llc -fast on the attached file, this is the result:
-----
$ llc test.bc -f -fast
llc: ValueTypes.cpp:167: static llvm::MVT llvm::MVT::getMVT(const llvm::Type*,
bool): Assertion `0 && "Unknown type!"' failed.
0   llc             0x0000000000ce5aef
1   llc             0x0000000000ce5ef9
2   libpthread.so.0 0x00007f0087e9b0f0
3   libc.so.6       0x00007f008719f015 gsignal + 53
4   libc.so.6       0x00007f00871a0b83 abort + 387
5   libc.so.6       0x00007f0087197d89 __assert_fail + 233
6   llc             0x0000000000c9a274 llvm::MVT::getMVT(llvm::Type const*,
bool) + 708
7   llc             0x000000000093f5d7
llvm::FastISel::getRegForValue(llvm::Value*) + 119
8   llc             0x00000000007b09b5
9   llc             0x00000000007b25d5
10  llc             0x00000000009bbb7f
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function&,
llvm::MachineFunction&, llvm::MachineModuleInfo*, llvm::DwarfWriter*,
llvm::TargetInstrInfo const&) + 1263
11  llc             0x00000000009bcc96
llvm::SelectionDAGISel::runOnFunction(llvm::Function&) + 1286
12  llc             0x0000000000c82159
llvm::FPPassManager::runOnFunction(llvm::Function&) + 473
13  llc             0x0000000000c823fe
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 142
14  llc             0x0000000000c8260b
llvm::FunctionPassManager::run(llvm::Function&) + 75
15  llc             0x0000000000509a19 main + 4473
16  libc.so.6       0x00007f008718a466 __libc_start_main + 230
17  llc             0x0000000000507ce9
Stack dump:
0.      Program arguments: llc test.bc -f -fast 
1.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@foo'
Aborted
-----
llc without -fast works fine.

The problem seems to be the struct type. If the call is changed to pass the i64
and i8* directly (and the declaration is changed to match) there's no assertion
failure.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 16:48:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 16:48:00 -0500
Subject: [LLVMbugs] [Bug 3945] The ARM AsmPrinter doesn't recognize a
	modifier.
In-Reply-To: <bug-3945-206@http.llvm.org/bugs/>
Message-ID: <200904062148.n36Lm0pd021073@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3945


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Bob Wilson <bob.wilson at apple.com>  2009-04-06 16:47:59 ---
I have applied your revised patch.  Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 18:36:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 18:36:12 -0500
Subject: [LLVMbugs] [Bug 3957] New: sync builtins should be overloaded
Message-ID: <bug-3957-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3957

           Summary: sync builtins should be overloaded
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


The gcc atomic synchronization builtins are intended to be overloaded for 1, 2,
4, and 8 bytes. clang is always treating them as having int type.
--
cat t.c
char f0(char *a, char b) {
  return __sync_fetch_and_add(a, b);
}
ddunbar at lordcrumb:tmp$ llvm-gcc -m32 -O3 -emit-llvm -S -o - t.c
; ModuleID = 't.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i386-apple-darwin10.0"

define signext i8 @f0(i8* nocapture %a, i8 signext %b) nounwind ssp {
entry:
        %0 = tail call i8 @llvm.atomic.load.add.i8.p0i8(i8* %a, i8 %b)         
; <i8> [#uses=1]
        ret i8 %0
}

declare i8 @llvm.atomic.load.add.i8.p0i8(i8* nocapture, i8) nounwind
ddunbar at lordcrumb:tmp$ clang -m32 -O3 -emit-llvm -S -o - t.c
t.c:2:31: warning: incompatible pointer types passing 'char *', expected 'int
*'
  return __sync_fetch_and_add(a, b);
                              ^
; ModuleID = 't.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i386-apple-darwin10.0"

define signext i8 @f0(i8* nocapture %a, i8 signext %b) nounwind {
entry:
        %conv = sext i8 %b to i32               ; <i32> [#uses=1]
        %conv2 = bitcast i8* %a to i32*         ; <i32*> [#uses=1]
        %0 = tail call i32 @llvm.atomic.load.add.i32.p0i32(i32* %conv2, i32
%conv)              ; <i32> [#uses=1]
        %conv3 = trunc i32 %0 to i8             ; <i8> [#uses=1]
        ret i8 %conv3
}

declare i32 @llvm.atomic.load.add.i32.p0i32(i32* nocapture, i32) nounwind
1 diagnostic generated.
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr  6 21:20:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 6 Apr 2009 21:20:12 -0500
Subject: [LLVMbugs] [Bug 3365] getSourceRange() doesn't contain full DeclStmt
In-Reply-To: <bug-3365-206@http.llvm.org/bugs/>
Message-ID: <200904070220.n372KCd6031987@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3365


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-04-06 21:20:11 ---
I fixed this last week.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 08:08:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 08:08:35 -0500
Subject: [LLVMbugs] [Bug 3959] New: [Mips] Return of a double constant
	causes an assert.
Message-ID: <bug-3959-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3959

           Summary: [Mips] Return of a double constant causes an assert.
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rich at pennware.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2802)
 --> (http://llvm.org/bugs/attachment.cgi?id=2802)
Test case.

The attached test case causes an assert with the Mips code generator:

[~/ellcc/ellcc] main% llvm-as -f t0065.ll
[~/ellcc/ellcc] main% llc -f -march mips t0065.bc
Cannot yet select: 0xa564fc8: f64,ch = load 0xa564c3c, 0xa5651d8, 0xa564f44
<0xa579da8:0> <anyext f32> alignment=4
0   llc       0x08d5616d
1   llc       0x08d5669f
2             0x00c27400 __kernel_sigreturn + 0
3   libc.so.6 0x0033ee28 abort + 392
4   llc       0x0859d1eb
5   llc       0x085a0228
6   llc       0x085a30ff
7   llc       0x085a46c4
8   llc       0x085a4a75
9   llc       0x085a4be3
10  llc       0x089b7101 llvm::SelectionDAGISel::CodeGenAndEmitDAG() + 2905
11  llc       0x089b941a
llvm::SelectionDAGISel::SelectBasicBlock(llvm::BasicBlock*,
llvm::ilist_iterator<llvm::Instruction>,
llvm::ilist_iterator<llvm::Instruction>) + 612
12  llc       0x089b9e29
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function&,
llvm::MachineFunction&, llvm::MachineModuleInfo*, llvm::DwarfWriter*,
llvm::TargetInstrInfo const&) + 2295
13  llc       0x089babdb llvm::SelectionDAGISel::runOnFunction(llvm::Function&)
+ 885
14  llc       0x08cd6ae2 llvm::FPPassManager::runOnFunction(llvm::Function&) +
288
15  llc       0x08cd75d0 llvm::FunctionPassManagerImpl::run(llvm::Function&) +
124
16  llc       0x08cd77a1 llvm::FunctionPassManager::run(llvm::Function&) + 159
17  llc       0x083d1662 main + 2687
18  libc.so.6 0x003286e5 __libc_start_main + 229
19  llc       0x083cfb21
Stack dump:
0.      Program arguments: llc -f -march mips t0065.bc
1.      Running pass 'MIPS DAG->DAG Pattern Instruction Selection' on function
'@foo'
Abort
[~/ellcc/ellcc] main%


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 14:42:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 14:42:09 -0500
Subject: [LLVMbugs] [Bug 3941] Invalid use of inline not diagnosed
In-Reply-To: <bug-3941-206@http.llvm.org/bugs/>
Message-ID: <200904071942.n37Jg9G1024724@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3941


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-07 14:42:09 ---
Fixed in r68541.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 15:24:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 15:24:37 -0500
Subject: [LLVMbugs] [Bug 3961] New: C output error compiling with TCC
Message-ID: <bug-3961-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3961

           Summary: C output error compiling with TCC
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bottiger1 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I tried compiling the LLVM C output with TCC http://bellard.org/tcc/.

It complained about a missing header "alloca.h".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 15:40:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 15:40:20 -0500
Subject: [LLVMbugs] [Bug 2985] Invalid code generated for cpu2000/gcc with
	-mthumb
In-Reply-To: <bug-2985-206@http.llvm.org/bugs/>
Message-ID: <200904072040.n37KeKCP026904@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2985


Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 15:41:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 15:41:05 -0500
Subject: [LLVMbugs] [Bug 3955] llc -fast asserts on first-class-struct
	parameter
In-Reply-To: <bug-3955-206@http.llvm.org/bugs/>
Message-ID: <200904072041.n37Kf5OL026973@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3955


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr  7 17:35:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 7 Apr 2009 17:35:51 -0500
Subject: [LLVMbugs] [Bug 3961] C output error compiling with TCC
In-Reply-To: <bug-3961-206@http.llvm.org/bugs/>
Message-ID: <200904072235.n37MZpUg031428@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3961


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-07 17:35:51 ---
this sounds like a bug in tcc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  8 06:21:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 8 Apr 2009 06:21:00 -0500
Subject: [LLVMbugs] [Bug 3873] The check/view/write-graph commands need some
 Bug repair / Documentation
In-Reply-To: <bug-3873-206@http.llvm.org/bugs/>
Message-ID: <200904081121.n38BL0lk007207@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3873


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru>  2009-04-08 06:20:49 ---
This is not a problem of LLVM itself anymore.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  8 13:01:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 8 Apr 2009 13:01:40 -0500
Subject: [LLVMbugs] [Bug 3795] ARM assertion failure for load of first-class
 aggregate from variable GEP
In-Reply-To: <bug-3795-206@http.llvm.org/bugs/>
Message-ID: <200904081801.n38I1esN022895@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3795


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Bob Wilson <bob.wilson at apple.com>  2009-04-08 13:01:40 ---
I applied Dan's suggested fix.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  8 15:48:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 8 Apr 2009 15:48:47 -0500
Subject: [LLVMbugs] [Bug 3941] Invalid use of inline not diagnosed
In-Reply-To: <bug-3941-206@http.llvm.org/bugs/>
Message-ID: <200904082048.n38Kml3r031027@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3941


Neil Booth <neil at daikokuya.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #2 from Neil Booth <neil at daikokuya.co.uk>  2009-04-08 15:48:46 ---
Hi Eli,  clang is still accepting erroneous inlines, such as

   inline enum { e = 2 };


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  8 15:59:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 8 Apr 2009 15:59:00 -0500
Subject: [LLVMbugs] [Bug 3962] New: Invalid uses of restrict not diagnosed
Message-ID: <bug-3962-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3962

           Summary: Invalid uses of restrict not diagnosed
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


This violates C99 constraint 6.7.3p2.

restrict struct s;


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr  8 19:52:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 8 Apr 2009 19:52:22 -0500
Subject: [LLVMbugs] [Bug 3395] clang: '-include' Path Search Incorrect
In-Reply-To: <bug-3395-206@http.llvm.org/bugs/>
Message-ID: <200904090052.n390qMt4013084@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3395


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 00:44:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 00:44:38 -0500
Subject: [LLVMbugs] [Bug 3963] New: improve error recovery for mistyped
	"types"
Message-ID: <bug-3963-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3963

           Summary: improve error recovery for mistyped "types"
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


Consider something like this where clang doesn't know foo_t is a type:

static foo_t x = 4;

we emit:

t2.c:2:14: error: invalid token after top level declarator
static foo_t x = 4;
             ^
and then errors at every use of X because it wasn't declared.  Similarly for
things like "static foo_t *X etc.

In this case, the parser thinks it is seeing a declaration of a variable named
foo_t that is implicit int.  In this case (when it is in implicit int mode) it
should do lookahead to see if the non-type token (which it thinks is the
identifier) is followed immediately by something that is not valid, e.g.
another identifier, a *, a &, etc.  In this case, it should recover by parsing
foo_t as a broken type.  This will lead to a much better diagnostic and will
squeltch the chain of bogus diagnostics at uses of X.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 08:22:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 08:22:51 -0500
Subject: [LLVMbugs] [Bug 3964] New: clang: undefined reference to
	__builtin_strcmp
Message-ID: <bug-3964-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3964

           Summary: clang: undefined reference to __builtin_strcmp
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Using SVN r68706 on Linux x86-64:

int main(int argc, char *argv[])
{
    return __builtin_strcmp(argv[0], argv[1]);
}

$ clang x.c
/tmp/cc-84cdXS.o: In function `main':
x.c:(.text+0x2e): undefined reference to `__builtin_strcmp'
collect2: ld returned 1 exit status


This has worked until now: the above happens when expanding string.h at -O2.
So either something broke in clang, or something causes another part of
string.h to be expanded.


FWIW, strcmp(argv[0], argv[1]) is expanded to this at -O2:

extern int strcmp (__const char *__s1, __const char *__s2)
  return __extension__ ({ size_t __s1_len, __s2_len; (__builtin_constant_p
(argv[0]) && __builtin_constant_p (argv[1]) && (__s1_len = strlen (argv[0]),
__s2_len = strlen (argv[1]), (!((size_t)(const void *)((argv[0]) + 1) -
(size_t)(const void *)(argv[0]) == 1)|| __s1_len >= 4) && (!((size_t)(const
void *)((argv[1]) + 1) - (size_t)(const void *)(argv[1]) == 1) || __s2_len >=
4)) ? __builtin_strcmp (argv[0], argv[1]) : (__builtin_constant_p (argv[0]) &&
((size_t)(const void *)((argv[0]) + 1) - (size_t)(const void *)(argv[0]) == 1)
&& (__s1_len = strlen (argv[0]), __s1_len < 4) ? (__builtin_constant_p
(argv[1]) && ((size_t)(const void *)((argv[1]) + 1) - (size_t)(const void
*)(argv[1]) == 1) ? __builtin_strcmp (argv[0], argv[1]) : (__extension__ ({
__const unsigned char *__s2 = (__const unsigned char *) (__const char *)
(argv[1]); register int __result = (((__const unsigned char *) (__const char *)
(argv[0]))[0] - __s2[0]); if (__s1_len > 0 && __result == 0) { __result =
(((__const unsigned char *) (__const char *) (argv[0]))[1] - __s2[1]); if
(__s1_len > 1 && __result == 0) { __result = (((__const unsigned char *)
(__const char *) (argv[0]))[2] - __s2[2]); if (__s1_len > 2 && __result == 0)
__result = (((__const unsigned char *) (__const char *) (argv[0]))[3] -
__s2[3]); } } __result; }))) : (__builtin_constant_p (argv[1]) &&
((size_t)(const void *)((argv[1]) + 1) - (size_t)(const void *)(argv[1]) == 1)
&& (__s2_len = strlen (argv[1]), __s2_len < 4) ? (__builtin_constant_p
(argv[0]) && ((size_t)(const void *)((argv[0]) + 1) - (size_t)(const void
*)(argv[0]) == 1) ? __builtin_strcmp (argv[0], argv[1]) : (__extension__ ({
__const unsigned char *__s1 = (__const unsigned char *) (__const char *)
(argv[0]); register int __result = __s1[0] - ((__const unsigned char *)
(__const char *) (argv[1]))[0]; if (__s2_len > 0 && __result == 0) { __result =
(__s1[1] - ((__const unsigned char *) (__const char *) (argv[1]))[1]); if
(__s2_len > 1 && __result == 0) { __result = (__s1[2] - ((__const unsigned char
*) (__const char *) (argv[1]))[2]); if (__s2_len > 2 && __result == 0) __result
= (__s1[3] - ((__const unsigned char *) (__const char *) (argv[1]))[3]); } }
__result; }))) : __builtin_strcmp (argv[0], argv[1])))); });


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:31:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:31:20 -0500
Subject: [LLVMbugs] [Bug 3731] clang x86-32 doesn't implement ABI properly
	on Linux
In-Reply-To: <bug-3731-206@http.llvm.org/bugs/>
Message-ID: <200904091631.n39GVKKb025001@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3731


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:31:19 ---
r67577 fixed all the failures for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:43:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:43:13 -0500
Subject: [LLVMbugs] [Bug 3964] clang: undefined reference to __builtin_strcmp
In-Reply-To: <bug-3964-206@http.llvm.org/bugs/>
Message-ID: <200904091643.n39GhD03025519@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3964


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:43:12 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090406/015260.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:55:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:55:00 -0500
Subject: [LLVMbugs] [Bug 2836] Objc: Extended setter property not synthesized
In-Reply-To: <bug-2836-206@http.llvm.org/bugs/>
Message-ID: <200904091655.n39Gt08d026288@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2836


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #10 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:55:00 ---
Fariborz fixed this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:55:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:55:48 -0500
Subject: [LLVMbugs] [Bug 2837] Objc fsyntax-only synthesized properties and
 missing implementation warnings
In-Reply-To: <bug-2837-206@http.llvm.org/bugs/>
Message-ID: <200904091655.n39GtmTZ026362@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2837


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:55:48 ---
This has been fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:56:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:56:37 -0500
Subject: [LLVMbugs] [Bug 2877] Clang parser ignores protocol information of
 referenced properties
In-Reply-To: <bug-2877-206@http.llvm.org/bugs/>
Message-ID: <200904091656.n39GubKa026436@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2877


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:56:36 ---
This has been fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 11:58:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 11:58:44 -0500
Subject: [LLVMbugs] [Bug 3012] ActOnClassMessage assertion with some invalid
	obj-c file
In-Reply-To: <bug-3012-206@http.llvm.org/bugs/>
Message-ID: <200904091658.n39GwiwZ026627@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3012


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 11:58:43 ---
This has been fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 12:05:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 12:05:18 -0500
Subject: [LLVMbugs] [Bug 3205] [clang] always_inline doesn't always
In-Reply-To: <bug-3205-206@http.llvm.org/bugs/>
Message-ID: <200904091705.n39H5Ine026893@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3205


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Daniel Dunbar <daniel at zuster.org>  2009-04-09 12:05:18 ---
This is because the inline definitions are being emitted with as weak. This
amounts to a duplicate of PR2933.

*** This bug has been marked as a duplicate of bug 2933 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 12:15:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 12:15:48 -0500
Subject: [LLVMbugs] [Bug 3114] emit-html messes up some tr entries
In-Reply-To: <bug-3114-206@http.llvm.org/bugs/>
Message-ID: <200904091715.n39HFmVl027403@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3114


Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Ted Kremenek <kremenek at apple.com>  2009-04-09 12:15:48 ---


*** This bug has been marked as a duplicate of bug 2386 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 12:25:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 12:25:14 -0500
Subject: [LLVMbugs] [Bug 3954] register scavenger asserts for inline
 assembly with tied destination register
In-Reply-To: <bug-3954-206@http.llvm.org/bugs/>
Message-ID: <200904091725.n39HPEVP027809@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3954


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Bob Wilson <bob.wilson at apple.com>  2009-04-09 12:25:14 ---
This is fixed in svn revision 68714.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 16:09:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 16:09:55 -0500
Subject: [LLVMbugs] [Bug 3718] LLC crashes with the following bitcode file
In-Reply-To: <bug-3718-206@http.llvm.org/bugs/>
Message-ID: <200904092109.n39L9tfp004229@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3718


Devang Patel <dpatel at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #11 from Devang Patel <dpatel at apple.com>  2009-04-09 16:09:54 ---


*** This bug has been marked as a duplicate of bug 3494 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 16:14:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 16:14:29 -0500
Subject: [LLVMbugs] [Bug 3965] New: SIGINT handler not restored after
	calling ParseAST()
Message-ID: <bug-3965-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3965

           Summary: SIGINT handler not restored after calling ParseAST()
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2812)
 --> (http://llvm.org/bugs/attachment.cgi?id=2812)
exit(1); -> raise(Sig); patch

This is similar to http://llvm.org/bugs/show_bug.cgi?id=3848

However, the problem is that even when there is no custom interrupt function
set, the signal handler does an exit(1); when receiving a SIGINT, rather than
calling the default handler.

It seems the custom signal handler persists even after clang returns from
ParseAST().

The attached patch changes the behaviour of receiving a SIGINT to run the
previous handler rather than calling exit(1).

A better solution may be to actually unregister the damn handler before clang
exits from ParseAST() - but I'm not sure where this code would have to go.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 19:20:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 19:20:53 -0500
Subject: [LLVMbugs] [Bug 3966] New: Problems with addrspace(256) = GS in x86
	codegen
Message-ID: <bug-3966-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3966

           Summary: Problems with addrspace(256) = GS in x86 codegen
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: rafael.espindola at gmail.com, llvmbugs at cs.uiuc.edu


The X86 backend almost has a cool feature where pointers with addrspace(256) is
a reference off the GS register.  Unfortunately, this has several problems:

1. No stores through GS are supported at all.
2. Not all loads to other address spaces properly indicate that they require
addrspace(0).

This means that we silently miscompile all stores through GS (bad) and many
loads.  For example, this code is miscompiled:

void foo() {
  int __attribute__((address_space(256))) *P = 1234;
  P[4] = 17;
}
void * bar(__attribute__((address_space(256))) unsigned long * const P) {
        if (P[4]) 
                P[5]++;
        return (void *)P[6];
}

Into:

_foo:
LBB1_0: ## entry
        movl    $17, 1250
        ret

_bar:
LBB2_0: ## entry
        movl    4(%esp), %eax
        movl    %gs:16(%eax), %ecx
        testl   %ecx, %ecx
        je      LBB2_2  ## if.end
LBB2_1: ## if.then
        addl    $1, 20(%eax)
LBB2_2: ## if.end
        movl    %gs:24(%eax), %eax
        ret

Note that the store in foo is not GS'ified, and the addl in LBB2_1 also lacks a
GS qualifier.

My hope is that Rafael's work just magically makes this all better :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr  9 23:00:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 9 Apr 2009 23:00:54 -0500
Subject: [LLVMbugs] [Bug 3967] New: llvm-gcc handles regparm incorrectly
Message-ID: <bug-3967-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3967

           Summary: llvm-gcc handles regparm incorrectly
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: haohui.mai at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2813)
 --> (http://llvm.org/bugs/attachment.cgi?id=2813)
test case

It seems that llvm-gcc handles regparm incorreclty in the following test case.
It only marks the first parameter of the function ``kobject_uevent'' as inreg
parameter, but it ignores the second one.

in the source code:
  int kobject_uevent(struct kobject *kobj, enum kobject_action action);

in the IR:
  declare i32 @kobject_uevent(%struct.kobject* inreg, i32)


Here is the command line:

llvm-gcc -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing
-fno-common -Werror-implicit-function-declaration -O2 -m32 -msoft-float
-mregparm=3 -freg-struct-return -mpreferred-stack-boundary=2 -march=i686
-ffreestanding -pipe -Wno-sign-compare -fno-asynchronous-unwind-tables -mno-sse
-mno-mmx -mno-sse2 -mno-3dnow -fno-stack-protector -fno-omit-frame-pointer
-fno-optimize-sibling-calls -Wdeclaration-after-statement -Wno-pointer-sign 
-D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(sys)" 
-D"KBUILD_MODNAME=KBUILD_STR(sys)" -emit-llvm -S -o - sys.i

I'm using the trunk of llvm and llvm-gcc, at revision 68722.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 01:51:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 01:51:55 -0500
Subject: [LLVMbugs] [Bug 3968] New: MachineSinking:: ProcessBlock crashes by
 moving iterator off the begining of ilist
Message-ID: <bug-3968-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3968

           Summary: MachineSinking::ProcessBlock crashes by moving iterator
                    off the begining of ilist
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: stoklund at 2pi.dk
                CC: llvmbugs at cs.uiuc.edu


The attached bugpoint-generated file causes a crash in a debug build of llc:

$ bin/llvm-as < ~/machine-sinking-crash.ll | bin/llc
Stack dump:
0.      Program arguments: bin/llc 
1.      Running pass 'Machine code sinking' on function '@QRiterate'
Bus error

The crash happens in:

  ilist_iterator &operator--() {      // predecrement - Back up
    NodePtr = Traits::getPrev(NodePtr);
    assert(Traits::getNext(NodePtr) && "--'d off the beginning of an ilist!");
    return *this;
  }

MachineSinking::ProcessBlock calls operator-- on an iterator with a null prev
pointer. The assert then calls getNext with a null pointer, causing the bus
error.

There are two separate issues here:

1. The assertion should be: assert(NodePtr && "--'d off...
2. MachineSinking::ProcessBlock should not call operator-- on an invalid
iterator


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 07:46:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 07:46:12 -0500
Subject: [LLVMbugs] [Bug 3970] New: clang outputs "Unsupported asm"
Message-ID: <bug-3970-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3970

           Summary: clang outputs "Unsupported asm"
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


$ clang-c -c a.c

with a.c:

double floor(double x)
{
        register long double val;

        __asm __volatile (
                        "frndint\n"
                        :
                        "=t" (val)
                        :
                        "0" (x)
                        );
}


gives:
"Unsupported asm: input constraint with a matching output constraint of
incompatible type!"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 09:04:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 09:04:27 -0500
Subject: [LLVMbugs] [Bug 3971] New: "instructiosn" instead of "instructions"
 in docs/ GettingStarted.html
Message-ID: <bug-3971-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3971

           Summary: "instructiosn" instead of "instructions" in
                    docs/GettingStarted.html
           Product: Documentation
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: shlomif at iglu.org.il
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2815)
 --> (http://llvm.org/bugs/attachment.cgi?id=2815)
Patch to correct the problem => instructiosn to instructions.

Index: docs/GettingStarted.html
===================================================================
--- docs/GettingStarted.html    (revision 68779)
+++ docs/GettingStarted.html    (working copy)
@@ -1489,7 +1489,7 @@

 <div class="doc_text">
 <p>This section gives an example of using LLVM.  llvm-gcc3 is now obsolete,
-so we only include instructiosn for llvm-gcc4.
+so we only include instructions for llvm-gcc4.
 </p>


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 10:38:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 10:38:58 -0500
Subject: [LLVMbugs] [Bug 3971] "instructiosn" instead of "instructions" in
 docs/GettingStarted. html
In-Reply-To: <bug-3971-206@http.llvm.org/bugs/>
Message-ID: <200904101538.n3AFcwMI021733@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3971


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-10 10:38:58 ---
Thanks, fixed!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 11:39:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 11:39:14 -0500
Subject: [LLVMbugs] [Bug 3968] MachineSinking:: ProcessBlock crashes by
 moving iterator off the begining of ilist
In-Reply-To: <bug-3968-206@http.llvm.org/bugs/>
Message-ID: <200904101639.n3AGdEtF024135@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3968


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-10 11:39:14 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090406/076200.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090406/076203.html

Thanks!

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 10 15:04:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 10 Apr 2009 15:04:12 -0500
Subject: [LLVMbugs] [Bug 3972] New: Poor diagnostic with missing ')'
Message-ID: <bug-3972-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3972

           Summary: Poor diagnostic with missing ')'
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


We can do better:
--
ddunbar at 67-218-103-134:tmp$ cat t.c
int g0(int);
int f0(void) { return g0(g0(1); }
ddunbar at 67-218-103-134:tmp$ clang /tmp/t.c -fsyntax-only
/tmp/t.c:2:31: error: expected ')'
int f0(void) { return g0(g0(1); }
                              ^
/tmp/t.c:2:25: note: to match this '('
int f0(void) { return g0(g0(1); }
                        ^
/tmp/t.c:2:23: warning: incompatible pointer to integer conversion returning
'int (int)', expected 'int'
int f0(void) { return g0(g0(1); }
                      ^~
3 diagnostics generated.
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 11 03:35:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 11 Apr 2009 03:35:29 -0500
Subject: [LLVMbugs] [Bug 3973] New: Version 186 does not allow cats in path
Message-ID: <bug-3973-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3973

           Summary: Version 186 does not allow cats in path
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: gerriet at mdenkmann.de
                CC: llvmbugs at cs.uiuc.edu


When I start scan-build from:
/private/tmp/cur/checker-0.186/scan-build xcodebuild
it works.

But not any of these works:
/private/tmp/K??tzchen/checker-0.186/scan-build xcodebuild
/private/tmp/?????????/checker-0.186/scan-build xcodebuild
/private/tmp/???/checker-0.186/scan-build xcodebuild
/private/tmp/??????/checker-0.186/scan-build xcodebuild

I think that this restriction to mutts is rather unfortunate.

Same problem with version 177.
Version 144 works with all animals.

Kind regards

Gerriet.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 12 15:47:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Apr 2009 15:47:47 -0500
Subject: [LLVMbugs] [Bug 3963] improve error recovery for mistyped "types"
In-Reply-To: <bug-3963-206@http.llvm.org/bugs/>
Message-ID: <200904122047.n3CKllbF014522@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3963


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-04-12 15:47:46 ---
In fact, I think that this fixes this whole issue:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090406/015412.html

On:
static foo_t x = 4;

We now produce:

t2.c:1:8: error: unknown type name 'foo_t'
static foo_t x = 4;
       ^

FWIW, GCC 4.0 produces:
t2.c:1: error: syntax error before ???x???
t2.c:1: warning: data definition has no type or storage class

And GCC 4.2 produces:
t2.c:1: error: expected ???=???, ???,???, ???;???, ???asm??? or
???__attribute__??? before ???x???


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 12 18:33:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Apr 2009 18:33:42 -0500
Subject: [LLVMbugs] [Bug 3965] SIGINT handler not restored after calling
	ParseAST()
In-Reply-To: <bug-3965-206@http.llvm.org/bugs/>
Message-ID: <200904122333.n3CNXgj1020464@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3965


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-12 18:33:41 ---
Thanks, patch applied here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090406/076236.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 12 18:51:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Apr 2009 18:51:26 -0500
Subject: [LLVMbugs] [Bug 3932] [ObjC]Type defined as 'id' is not recognized
 as a valid object type.
In-Reply-To: <bug-3932-206@http.llvm.org/bugs/>
Message-ID: <200904122351.n3CNpQbI021181@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3932


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-12 18:51:24 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090406/015435.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 12 18:59:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Apr 2009 18:59:44 -0500
Subject: [LLVMbugs] [Bug 2375] verifier accepts bitcast of struct,
 and langref claims it is valid
In-Reply-To: <bug-2375-206@http.llvm.org/bugs/>
Message-ID: <200904122359.n3CNxieu021522@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2375


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Chris Lattner <clattner at apple.com>  2009-04-12 18:59:44 ---
This is "done enough" thanks Dan.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 12 19:11:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 12 Apr 2009 19:11:14 -0500
Subject: [LLVMbugs] [Bug 3972] Poor diagnostic with missing ')'
In-Reply-To: <bug-3972-206@http.llvm.org/bugs/>
Message-ID: <200904130011.n3D0BEgZ022069@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3972


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-12 19:11:13 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090406/015437.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 01:59:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 01:59:29 -0500
Subject: [LLVMbugs] [Bug 3975] New: llvm-gcc has some problem with int128
	arithmetic
Message-ID: <bug-3975-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3975

           Summary: llvm-gcc has some problem with int128 arithmetic
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: miscompilation
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: baldrick at free.fr, llvmbugs at cs.uiuc.edu


Consider this at -m64:
__uint128_t test2() {
  const __uint128_t c_zero = ~0;
  return c_zero;
}

This should compile to all 128 ones because the -1 signed integer sign extends
to i128. GCC produces the correct code:

__Z5test2v:
LFB8:
        pushq   %rbp
LCFI0:
        movq    %rsp, %rbp
LCFI1:
        movq    $-1, %rax
        movq    $-1, %rdx
        leave
        ret

but we compile it to 64 ones:

define i128 @_Z5test2v() nounwind readnone {
entry:
        ret i128 18446744073709551615
}

I think this is some problem in llvm-gcc, because the -O0 IR is already wrong:

        store i128 18446744073709551615, i128* %c_zero, align 16


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 08:54:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 08:54:55 -0500
Subject: [LLVMbugs] [Bug 3976] New: Assert in TargetLowering:: SimplifySetCC
 when using intergers >= 256 bits
Message-ID: <bug-3976-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3976

           Summary: Assert in TargetLowering::SimplifySetCC when using
                    intergers >= 256 bits
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ddneff at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


TargetLowering::SimplifySetCC seems to assume that operands can fit inside an
uint64_t.   If you write a program that has large integer types (256 or larger
in my case) this causes an assertion.  I don't see any reason why APInt can't
be used instead of uint64_t in this situation.  I wrote a patch that appears to
work, but I haven't put it through extensive testing.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 13:03:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 13:03:45 -0500
Subject: [LLVMbugs] [Bug 3975] llvm-gcc has some problem with int128
	arithmetic
In-Reply-To: <bug-3975-206@http.llvm.org/bugs/>
Message-ID: <200904131803.n3DI3jJW005838@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3975


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
          Component|llvm-g++                    |llvm-gcc
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-13 13:03:43 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076267.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076268.html

Maybe this will help fortran or something.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 13:43:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 13:43:14 -0500
Subject: [LLVMbugs] [Bug 3980] New: clang aborts on tentative definition
 with incomplete type
Message-ID: <bug-3980-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3980

           Summary: clang aborts on tentative definition with incomplete
                    type
           Product: clang
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: T.P.Northover at sms.ed.ac.uk
                CC: llvmbugs at cs.uiuc.edu


When processing a tentative definition with a currently incomplete type, clang
tries to emit the code immediately and fails.
E.g.
struct foo x;
struct foo { int a; };

It should be equivalent to a definition with the type as known at the end of
the file (C99 6.9.2p2 I think).

Daniel Dunbar made a brief comment on the problem at
http://llvm.org/bugs/show_bug.cgi?id=3461#c5.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 14:26:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 14:26:42 -0500
Subject: [LLVMbugs] [Bug 3981] New: Recursive structure def. could use
 better error diagnostics messages
Message-ID: <bug-3981-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3981

           Summary: Recursive structure def. could use better error
                    diagnostics messages
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Consider the following C source code:

struct foo;
struct foo { struct foo f; };

The follow error output is produced:

s.c:2:25: error: field has incomplete type 'struct foo'
struct foo { struct foo f; };
                        ^
s.c:2:8: note: definition of 'struct foo' is not complete until the closing '}'
struct foo { struct foo f; };
       ^
2 diagnostics generated.


These diagnostic messages could be improved. Particularly, the second message
is spurious. 

On an unrelated note, perhaps "Diagnostics" should be a category in Bugzilla to
report bugs against...


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 15:04:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 15:04:50 -0500
Subject: [LLVMbugs] [Bug 3934] two-addr pass crashes on copy
In-Reply-To: <bug-3934-206@http.llvm.org/bugs/>
Message-ID: <200904132004.n3DK4oiW010629@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3934


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #10 from Evan Cheng <evan.cheng at apple.com>  2009-04-13 15:04:49 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076280.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 16:23:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 16:23:05 -0500
Subject: [LLVMbugs] [Bug 3934] two-addr pass crashes on copy
In-Reply-To: <bug-3934-206@http.llvm.org/bugs/>
Message-ID: <200904132123.n3DLN5mX013617@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3934


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #11 from Ed Schouten <ed at 80386.nl>  2009-04-13 16:23:05 ---
Crash still occurs when building the following code:

void bn_sqr_comba8(unsigned long *r, unsigned long *a)
{
        unsigned long t1,t2;
        unsigned long c1,c2,c3;

        asm ("mulq %3" : "=a"(t1),"=d"(t2) : "a"((a)[1]),"m"((a)[0]) : "cc");
        asm ("addq %0,%0; adcq %2,%1" : "+d"(t2),"+r"(c1) : "g"(0) : "cc");
        asm ("addq %0,%0; adcq %2,%1" : "+a"(t1),"+d"(t2) : "g"(0) : "cc");
        asm ("addq %2,%0; adcq %3,%1" : "+r"(c2),"+d"(t2) : "a"(t1),"g"(0) :
"cc");
        asm ("addq %2,%0; adcq %3,%1" : "+r"(c3),"+r"(c1) : "d"(t2),"g"(0) :
"cc");
        asm ("mulq %2" : "=a"(t1),"=d"(t2) : "a"(a[1]) : "cc");
        asm ("addq %2,%0; adcq %3,%1" : "+r"(c3),"+d"(t2) : "a"(t1),"g"(0) :
"cc");
        asm ("addq %2,%0; adcq %3,%1" : "+r"(c1),"+r"(c2) : "d"(t2),"g"(0) :
"cc");
        asm ("mulq %3" : "=a"(t1),"=d"(t2) : "a"((a)[2]),"m"((a)[0]) : "cc");
        asm ("addq %0,%0; adcq %2,%1" : "+d"(t2),"+r"(c2) : "g"(0) : "cc");
        asm ("addq %0,%0; adcq %2,%1" : "+a"(t1),"+d"(t2) : "g"(0) : "cc");
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 16:30:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 16:30:51 -0500
Subject: [LLVMbugs] [Bug 3461] clang accepts initializer for incomplete type
In-Reply-To: <bug-3461-206@http.llvm.org/bugs/>
Message-ID: <200904132130.n3DLUpqH014097@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3461


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-13 16:30:51 ---
Fix committed in r68991.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 19:35:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 19:35:38 -0500
Subject: [LLVMbugs] [Bug 3934] two-addr pass crashes on copy
In-Reply-To: <bug-3934-206@http.llvm.org/bugs/>
Message-ID: <200904140035.n3E0ZcVm021065@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3934


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #12 from Evan Cheng <evan.cheng at apple.com>  2009-04-13 19:35:38 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076295.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 20:03:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 20:03:06 -0500
Subject: [LLVMbugs] [Bug 3982] New: reject attribute weak on internal
	variables
Message-ID: <bug-3982-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3982

           Summary: reject attribute weak on internal variables
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


--
ddunbar at 67-218-103-134:tmp$ cat x.c
static int x __attribute__((weak));
ddunbar at 67-218-103-134:tmp$ clang -fsyntax-only x.c
ddunbar at 67-218-103-134:tmp$ gcc -fsyntax-only x.c
x.c:1: error: weak declaration of 'x' must be public
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 13 21:22:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 13 Apr 2009 21:22:04 -0500
Subject: [LLVMbugs] [Bug 3983] New: Improve handling of variables with
 different storage spec
Message-ID: <bug-3983-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3983

           Summary: Improve handling of variables with different storage
                    spec
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


This seems wrong, and the diagnostics could be nicer (see gcc):
--
void f0() {
  // FIXME: Diagnose this?
  int g6;
  extern int g6;
}

void f1() {
  // FIXME: Diagnose this?
  int g7;
  __private_extern__ int g7;
}

void f2() {
  extern int g8; // expected-note{{previous definition}}
  // FIXME: Improve this diagnostic.
  int g8; // expected-error{{redefinition of 'g8'}}
}

void f3() {
  __private_extern__ int g9; // expected-note{{previous definition}}
  // FIXME: Improve this diagnostic.
  int g9; // expected-error{{redefinition of 'g9'}}
}
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 01:48:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 01:48:13 -0500
Subject: [LLVMbugs] [Bug 2933] clang should support "inline" properly.
In-Reply-To: <bug-2933-206@http.llvm.org/bugs/>
Message-ID: <200904140648.n3E6mDSB002376@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2933


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #15 from Chris Lattner <clattner at apple.com>  2009-04-14 01:48:13 ---
Inline functions now get the right linkage types, so they get inlined properly.
 We do not sema them (merge definitions) fully correctly though.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 01:50:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 01:50:49 -0500
Subject: [LLVMbugs] [Bug 3712] Clang doesn't predefine __OPTIMIZE__ and
 __PIC__ etc correctly
In-Reply-To: <bug-3712-206@http.llvm.org/bugs/>
Message-ID: <200904140650.n3E6onBv002512@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3712


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-14 01:50:49 ---
This is fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 03:57:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 03:57:40 -0500
Subject: [LLVMbugs] [Bug 2278] linux/x86-64 TLS support
In-Reply-To: <bug-2278-206@http.llvm.org/bugs/>
Message-ID: <200904140857.n3E8ve1B017705@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2278


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #11 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-04-14 03:57:26 ---
X86-64 has the same features as X86 now. The only thing missing on both is the
local dynamic mode. Will open a new bug for that.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 04:08:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 04:08:12 -0500
Subject: [LLVMbugs] [Bug 3985] New: x86 and x86-64 don't implement the local
 dynamic TLS mode.
Message-ID: <bug-3985-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3985

           Summary: x86 and x86-64 don't implement the local dynamic TLS
                    mode.
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: lli
        AssignedTo: rafael.espindola at gmail.com
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


gcc with -fPIC on x86-64 compiles

static int __thread a;
static int __thread b;
int f(void)
{
  return a + b;
}

to
        leaq    b at TLSLD(%rip), %rdi
        call    __tls_get_addr at PLT
        movq    %rax, %rdx
        movl    b at DTPOFF(%rax), %eax
        addl    a at DTPOFF(%rdx), %eax
        ret

and on X86 to

        pushl   %esi
        pushl   %ebx
        call    ___i686.get_pc_thunk.bx
        addl    $_GLOBAL_OFFSET_TABLE_, %ebx
        leal    b at TLSLDM(%ebx), %eax
        call    ___tls_get_addr at PLT
        movl    %eax, %esi
        movl    b at DTPOFF(%eax), %eax
        addl    a at DTPOFF(%esi), %eax
        popl    %ebx
        popl    %esi
        ret

note that there is only one call to ___tls_get_addr. We should implement that
on llvm.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 07:50:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 07:50:41 -0500
Subject: [LLVMbugs] [Bug 3986] New: inline may be miscompiled in recent
	revisions
Message-ID: <bug-3986-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3986

           Summary: inline may be miscompiled in recent revisions
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: roberto at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2823)
 --> (http://llvm.org/bugs/attachment.cgi?id=2823)
Virgin lex.c.blt code

I'm compiling Ruby from trunk with clang r69035 and I get undefined symbol
during link.  The missing function is "tagged" with

__inline __attribute((__gnuc_inline__))

The second argument is ignored:
-----
In file included from ../parse.y:6010:
keywords:75:17: warning: '__gnu_inline__' attribute ignored
__attribute__ ((__gnu_inline__))
                ^
-----

so I guess that "__inline" is generating the miscompilation.  Testing with
"inline" leads to the same error: 

Undefined symbols:
  "_rb_reserved_word", referenced from:
      _reg_named_capture_assign_iter in parse.o
      _parser_yylex in parse.o
ld: symbol(s) not found
collect2: ld returned 1 exit status

My guess is that "inline"/"__inline" has been broken in recent revisions.

virgin code in "lex.c", copied from "lex.c.blt" (attached) is
-----
#ifdef __GNUC__
__inline
#ifdef __GNUC_STDC_INLINE__
__attribute__ ((__gnu_inline__))
#endif
#endif
const struct kwtable *
rb_reserved_word (str, len)
     register const char *str;
     register unsigned int len;
{
  static const struct kwtable wordlist[] =
    {
-----

r68705 was handling it correctly.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 11:45:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 11:45:57 -0500
Subject: [LLVMbugs] [Bug 3986] clang doesn't support
	__attribute((__gnuc_inline__))
In-Reply-To: <bug-3986-206@http.llvm.org/bugs/>
Message-ID: <200904141645.n3EGjvtR003857@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3986


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|inline may be miscompiled in|clang doesn't support
                   |recent revisions            |__attribute((__gnuc_inline__
                   |                            |))




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-14 11:45:52 ---
This should fix it, please verify:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015525.html
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015526.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 14:27:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 14:27:00 -0500
Subject: [LLVMbugs] [Bug 3987] New: clang generates invalid assembly for
 Objective C code in 64 bit mode
Message-ID: <bug-3987-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3987

           Summary: clang generates invalid assembly for Objective C code in
                    64 bit mode
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: MacOS X
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: benny.kra at gmail.com
                CC: llvmbugs at cs.uiuc.edu


$ cat test.m
#import <Foundation/Foundation.h>

@protocol Foo
@end

@interface Test : NSObject <Foo>
@end

@implementation Test
@end

If this code is compiled with clang in 64 bit mode 'as' complains:

$ clang -m64 test.m
/tmp/cc-USvXbC.s:27:FATAL:symbol: l_OBJC_LABEL_PROTOCOL_$_Foo can't be a
weak_definition (currently only supported in section of type coalesced)

The code assembles fine in 32 bit mode.

(I don't know much Objective C so the test case might be reduced further. I've
extracted this from transmission)

I'm using OS X 10.5.6 (intel) and clang/llvm r69067.

$ as -v
Apple Inc version cctools-698~13, GNU assembler version 1.38


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 15:28:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 15:28:14 -0500
Subject: [LLVMbugs] [Bug 3987] clang generates invalid assembly for
 Objective C code in 64 bit mode
In-Reply-To: <bug-3987-206@http.llvm.org/bugs/>
Message-ID: <200904142028.n3EKSEfq012801@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3987


Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 14 15:28:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 14 Apr 2009 15:28:49 -0500
Subject: [LLVMbugs] [Bug 3989] New: C99 inline + extern function not being
	emitted
Message-ID: <bug-3989-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3989

           Summary: C99 inline + extern function not being emitted
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


Codegen should emit a strong definition for this function in c99 mode because
extern gets merged onto it:

inline int toto(void)
{
}

extern int toto(void);


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 15 01:24:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Apr 2009 01:24:08 -0500
Subject: [LLVMbugs] [Bug 3936] Instcombine quadratic in GEP chain length
In-Reply-To: <bug-3936-206@http.llvm.org/bugs/>
Message-ID: <200904150624.n3F6O82l001115@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3936


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Nick Lewycky <nicholas at mxc.ca>  2009-04-15 01:24:08 ---
Fixed in r69151.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 15 01:55:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Apr 2009 01:55:07 -0500
Subject: [LLVMbugs] [Bug 3981] Recursive structure def. could use better
 error diagnostics messages
In-Reply-To: <bug-3981-206@http.llvm.org/bugs/>
Message-ID: <200904150655.n3F6t7rZ002402@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3981


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dgregor at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #3 from Douglas Gregor <dgregor at apple.com>  2009-04-15 01:55:05 ---
The error message + note look really, really good to me: it's explaining
exactly what the problem is. I'm going to close this; if someone comes up with
a better wording for the diagnostics, please re-open with that suggestion.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 15 11:05:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Apr 2009 11:05:09 -0500
Subject: [LLVMbugs] [Bug 3992] New: -include ignores includepath
Message-ID: <bug-3992-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3992

           Summary: -include ignores includepath
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


$ cat test.c
int
main(int argc, char *argv[])
{

        foo();
}
$ cat foo/test.h 
#include <stdio.h>

void foo(void) { puts("Hi"); }
$ cc -o test -Ifoo -include test.h test.c
In file included from <built-in>:106:
<command line>:1:10: fatal error: '/home/ed/test.h' file not found
#include "/home/ed/test.h"
         ^
1 diagnostic generated.
$ gcc -o test -Ifoo -include test.h test.c
$ ./test
Hi

As you can see, the -include preprocessor option ignores the include paths.
This prevents us from building OpenSSL.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 15 15:46:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Apr 2009 15:46:08 -0500
Subject: [LLVMbugs] [Bug 3994] New: tablegen intrinsic forcing wrong
 register type on final parameter
Message-ID: <bug-3994-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3994

           Summary: tablegen intrinsic forcing wrong register type on final
                    parameter
           Product: tools
           Version: 2.5
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: TableGen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2828)
 --> (http://llvm.org/bugs/attachment.cgi?id=2828)
td test case that shows this problem

This is the "Tablegen question" thread.
This is what is generated in intrinsic.gen:
case Intrinsic::opencl_math_fdistance:          //
llvm.opencl.math.fdistance
    ResultTy = Type::FloatTy;
    ArgTys.push_back(Tys[0]);
    ArgTys.push_back(Tys[0]);
    break;

This is the intrinsic definition:
def int_opencl_math_fdistance_fast   : Intrinsic<[llvm_float_ty],
                           [llvm_anyfloat_ty, LLVMMatchType<0>]>;

The problem comes when I try to use the intrinsic. It gives me the following
error:
GPRV2F32:f32:$src1 MACRO_DISTANCE_FAST_v2f32:   (set GPRF32:f32:$dst,
(intrinsic_w_chain:f32 84:iPTR, GPRV2F32:v2f32:$src0,
GPRV2F32:f32:$src1))
TableGen.exe: In MACRO_DISTANCE_FAST_v2f32: Type inference contradiction found
in node!

I'm using the following test td file.

This is generated with putting it in include and running the command
Tablegen.exe -dag-isel -I. test.td


The pattern in question is:
ILFormat<(outs Dst:$dst), (ins Src:$src0, Src:$src1),
                  ",($dst),($src0, $src1)",
                  [(set Dst:$dst, (OpNode Src:$src0, Src:$src1))]

As the intrinsic specifies, the two source register classes should be the same,
however, Tablegen is somehow interpreting that it should be
f32 instead of v2f32.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 15 17:11:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 15 Apr 2009 17:11:09 -0500
Subject: [LLVMbugs] [Bug 3980] clang aborts on tentative definition with
	incomplete type
In-Reply-To: <bug-3980-206@http.llvm.org/bugs/>
Message-ID: <200904152211.n3FMB9do012858@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3980


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Daniel Dunbar <daniel at zuster.org>  2009-04-15 17:11:09 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015666.html

Sorry Tim, somehow I missed your patch.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 00:41:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 00:41:10 -0500
Subject: [LLVMbugs] [Bug 3995] New: assertion "invalid memory reference"
Message-ID: <bug-3995-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3995

           Summary: assertion "invalid memory reference"
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=2831)
 --> (http://llvm.org/bugs/attachment.cgi?id=2831)
testcase

This is a recent failure.

$ llvm-as < c.ll | llc 
llc: X86ATTAsmPrinter.cpp:617: void
llvm::X86ATTAsmPrinter::printMemReference(const llvm::MachineInstr*, unsigned
int, const char*): Assertion `isMem(MI, Op) && "Invalid memory reference!"'
failed.
0   llc 0x089d33ea
1   llc 0x089d391d
2       0xffffe400 __kernel_sigreturn + 0
Stack dump:
0.      Program arguments: llvm-commit/Debug/bin/llc 
1.      Running pass 'X86 AT&T-Style Assembly Printer' on function '@test'
Aborted

#0  0xffffe425 in __kernel_vsyscall ()
#1  0xf7c61640 in *__GI_raise (sig=6)
    at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#2  0xf7c63008 in *__GI_abort () at abort.c:88
#3  0xf7c5a5ce in *__GI___assert_fail (
    assertion=0x8a78824 "isMem(MI, Op) && \"Invalid memory reference!\"", 
    file=0x8a7800b "X86ATTAsmPrinter.cpp", line=617, 
    function=0x8a7c9e0 "void llvm::X86ATTAsmPrinter::printMemReference(const
llvm::MachineInstr*, unsigned int, const char*)") at assert.c:78
#4  0x0838aff4 in llvm::X86ATTAsmPrinter::printMemReference (this=0x9007af0, 
    MI=0x900d68c, Op=1, Modifier=0x0) at X86ATTAsmPrinter.cpp:617
#5  0x083939f3 in llvm::X86ATTAsmPrinter::printf64mem (this=0x9007af0, 
    MI=0x900d68c, OpNo=1) at X86ATTAsmPrinter.h:88
#6  0x0838bc35 in llvm::X86ATTAsmPrinter::printInstruction (this=0x9007af0, 
    MI=0x900d68c)
    at
/home/nicholas/llvm-commit/lib/Target/X86/AsmPrinter/../X86GenAsmWriter.inc:2281
#7  0x0838d130 in llvm::X86ATTAsmPrinter::printMachineInstruction (
    this=0x9007af0, MI=0x900d68c) at X86ATTAsmPrinter.cpp:765
#8  0x0838d43c in llvm::X86ATTAsmPrinter::runOnMachineFunction (
    this=0x9007af0, MF=@0x900cfa0) at X86ATTAsmPrinter.cpp:253
#9  0x0873d1e3 in llvm::MachineFunctionPass::runOnFunction (this=0x9007af0, 
    F=@0x8fe9b90) at MachineFunction.cpp:47
#10 0x0895360a in llvm::FPPassManager::runOnFunction (this=0x8ff0e18, 
    F=@0x8fe9b90) at PassManager.cpp:1338
#11 0x089540f8 in llvm::FunctionPassManagerImpl::run (this=0x8fe7fd0, 
    F=@0x8fe9b90) at PassManager.cpp:1291
#12 0x089542c9 in llvm::FunctionPassManager::run (this=0xffb2caa8, 
    F=@0x8fe9b90) at PassManager.cpp:1233
#13 0x08049cb4 in main (argc=3, argv=0xffb2cbd4) at llc.cpp:322


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 07:35:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 07:35:36 -0500
Subject: [LLVMbugs] [Bug 3995] assertion "invalid memory reference"
In-Reply-To: <bug-3995-206@http.llvm.org/bugs/>
Message-ID: <200904161235.n3GCZajX018861@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3995


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-04-16 07:35:35 ---
Fixed on 69284.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 11:39:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 11:39:37 -0500
Subject: [LLVMbugs] [Bug 3997] New: llc lowers intrinsics incorrectly with
 the presense of regparm attribute
Message-ID: <bug-3997-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3997

           Summary: llc lowers intrinsics incorrectly with the presense of
                    regparm attribute
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: haohui.mai at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2832)
 --> (http://llvm.org/bugs/attachment.cgi?id=2832)
test case

llc does not handle this case correctly:

        call void @llvm.memcpy.i32(i8* %3, i8* %4, i32 2052, i32 4)
        %dst1 = bitcast [128 x i8]* %dst to i8*         ; <i8*> [#uses=1]
        %src2 = bitcast [128 x i8]* %src to i8*         ; <i8*> [#uses=1]
        %5 = call i8* @memcpy(i8* inreg %dst1, i8* inreg %src2, i32 inreg 128)
nounwind         ; <i8*> [#uses=0]

when llvm.memcpy.i32 lowers to call to memcpy, it does not follow the
customized memcpy calling convention.

for the first llvm.memcpy.i32, the generated assemblies are:

        movl    %eax, (%edx)
        movl    %ecx, 4(%edx)
        movl    $2052, 8(%edx)
        call    memcpy

for the second call of memcpy, the codes are:

        leal    -140(%ebp), %eax
        leal    -268(%ebp), %ecx
        movl    $128, %edx
        movl    %edx, -276(%ebp)
        movl    %ecx, %edx
        movl    -276(%ebp), %ecx
        call    memcpy


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 12:23:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 12:23:26 -0500
Subject: [LLVMbugs] [Bug 3997] llc lowers intrinsics incorrectly with the
 presense of regparm attribute
In-Reply-To: <bug-3997-206@http.llvm.org/bugs/>
Message-ID: <200904161723.n3GHNQ4c028872@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3997


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru>  2009-04-16 12:23:25 ---
llc is doing correct thing here. The intrinsic does not have any regparm
attributes.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 16:40:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 16:40:58 -0500
Subject: [LLVMbugs] [Bug 3944] mem2reg quadratic in number of in-edges to
 block receiving a new phi node
In-Reply-To: <bug-3944-206@http.llvm.org/bugs/>
Message-ID: <200904162140.n3GLewff004503@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3944


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-16 16:40:57 ---
Committed in r69312.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 16:52:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 16:52:28 -0500
Subject: [LLVMbugs] [Bug 3994] tablegen intrinsic forcing wrong register
 type on final parameter
In-Reply-To: <bug-3994-206@http.llvm.org/bugs/>
Message-ID: <200904162152.n3GLqSQb004911@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3994


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bob.wilson at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Bob Wilson <bob.wilson at apple.com>  2009-04-16 16:52:28 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076396.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 17:12:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 17:12:49 -0500
Subject: [LLVMbugs] [Bug 3998] New: LLVM JIT generates code into a fixed
	size buffer
Message-ID: <bug-3998-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3998

           Summary: LLVM JIT generates code into a fixed size buffer
           Product: new-bugs
           Version: unspecified
          Platform: All
               URL: http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=21
                    9
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


Forwarded this bug since it affects any JIT using LLVM.
http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=219

Quoted from gbenson:
"
LLVM generates all code into a buffer of fixed size, and once this buffer is
filled it simply exits with an abort(), taking out the VM.  Clearly this needs
fixing for Shark.

There are two possible places a fix could occur:

1) In lib/ExecutionEngine/JIT/JITEmitter.cpp, in JITEmitter::finishFunction.
   This is where the abort() is; there's a FIXME saying to allocate more space
   and try again.

2) In lib/ExecutionEngine/JIT/JITMemoryManager.cpp, by extending the default
   memory manager to be able to grow the block or something.
"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 19:30:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 19:30:05 -0500
Subject: [LLVMbugs] [Bug 3502] Cross-compile llvm-gcc Linux x86 to arm-linux
In-Reply-To: <bug-3502-206@http.llvm.org/bugs/>
Message-ID: <200904170030.n3H0U5nx010313@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3502


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #18 from Evan Cheng <evan.cheng at apple.com>  2009-04-16 19:30:03 ---
So it's fixed, right?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 16 21:57:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 16 Apr 2009 21:57:30 -0500
Subject: [LLVMbugs] [Bug 3510] llvm-gcc miscompiles simple static init in
	-m64
In-Reply-To: <bug-3510-206@http.llvm.org/bugs/>
Message-ID: <200904170257.n3H2vURK014764@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3510


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #9 from Chris Lattner <clattner at apple.com>  2009-04-16 21:57:29 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076403.html

TEstcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076401.html

Duncan, I'd really appreciate a review of this change, thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 16:07:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 16:07:32 -0500
Subject: [LLVMbugs] [Bug 3901] clang: pathnames are not normalized in debug
 info -> gdb can' t find file
In-Reply-To: <bug-3901-206@http.llvm.org/bugs/>
Message-ID: <200904172107.n3HL7Wjr024372@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3901


Devang Patel <dpatel at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Devang Patel <dpatel at apple.com>  2009-04-17 16:07:31 ---
Fixed.
rev 69387.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 18:40:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 18:40:14 -0500
Subject: [LLVMbugs] [Bug 3940] #line numbers not fully checked
In-Reply-To: <bug-3940-206@http.llvm.org/bugs/>
Message-ID: <200904172340.n3HNeENE029329@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3940


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-17 18:40:11 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015763.html

Thanks Neil


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 18:57:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 18:57:53 -0500
Subject: [LLVMbugs] [Bug 3916] #include corner cases
In-Reply-To: <bug-3916-206@http.llvm.org/bugs/>
Message-ID: <200904172357.n3HNvr1W029835@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3916


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-17 18:57:52 ---
Fixed here, thanks Neil!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015764.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 20:14:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 20:14:31 -0500
Subject: [LLVMbugs] [Bug 3937] Clang isn't checking macro argument counts
	properly
In-Reply-To: <bug-3937-206@http.llvm.org/bugs/>
Message-ID: <200904180114.n3I1EVa0031937@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3937


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-17 20:14:29 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015769.html

Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 20:34:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 20:34:50 -0500
Subject: [LLVMbugs] [Bug 3938] Clang rejects valid TU with -pedantic-errors
In-Reply-To: <bug-3938-206@http.llvm.org/bugs/>
Message-ID: <200904180134.n3I1Yo4I032482@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3938


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-17 20:34:50 ---
Fixed, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015771.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 21:23:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 21:23:58 -0500
Subject: [LLVMbugs] [Bug 3919] Missed Constraint 6.10.3p3 (C99) and c90
In-Reply-To: <bug-3919-206@http.llvm.org/bugs/>
Message-ID: <200904180223.n3I2Nwpf001684@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3919


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|Missed Constraint 6.10.3p3  |Missed Constraint 6.10.3p3
                   |(C99)                       |(C99) and c90




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-17 21:23:58 ---
C90 piece fixed here, thanks Neil!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015772.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 17 22:47:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 17 Apr 2009 22:47:05 -0500
Subject: [LLVMbugs] [Bug 3987] clang generates invalid assembly for
 Objective C code in 64 bit mode
In-Reply-To: <bug-3987-206@http.llvm.org/bugs/>
Message-ID: <200904180347.n3I3l5w2003991@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3987


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|RESOLVED                    |REOPENED
         Resolution|WONTFIX                     |




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2009-04-17 22:47:02 ---
Reopening. This isn't won't fix, but it isn't a current priority. Patches
welcome!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 01:46:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 01:46:58 -0500
Subject: [LLVMbugs] [Bug 3927] Clang expands the same macro differently
 depending on its identifier
In-Reply-To: <bug-3927-206@http.llvm.org/bugs/>
Message-ID: <200904180646.n3I6kw3T008843@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3927


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-18 01:46:57 ---
Nice catch, I'm impressed that you found that by inspection.  Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015777.html

I checked the rest of liblex for uses of is(tok::identifier).  The ones that
are left are:

lib/Lex/Preprocessor.cpp:  assert(Identifier.is(tok::identifier) && "Not an
identifier!");
This is checking for identifier info lookup, it is ok.

lib/Lex/TokenLexer.cpp:    if (Tok.is(tok::identifier) &&
RHS.is(tok::identifier)) {
This is an optimization, it doesn't matter for keywords.

lib/Lex/TokenLexer.cpp:  if (Tok.is(tok::identifier)) {
This is another identifierinfo lookup case.

So I think we're ok, nice catch!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 08:21:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 08:21:06 -0500
Subject: [LLVMbugs] [Bug 3999] New: CodeGen/PowerPC/vec_splat. ll fails when
 legalize types is turned on
Message-ID: <bug-3999-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3999

           Summary: CodeGen/PowerPC/vec_splat.ll fails when legalize types
                    is turned on
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


I need some help with this one.  It is blocking the removal of
the old type legalization code.  I think it needs to be fixed
in the ppc backend.  Here's the relevant part of the test:

define void @splat_h(i16 %tmp, <16 x i8>* %dst) nounwind {
        %tmp.upgrd.1 = insertelement <8 x i16> undef, i16 %tmp, i32 0
        %tmp72 = insertelement <8 x i16> %tmp.upgrd.1, i16 %tmp, i32 1
        %tmp73 = insertelement <8 x i16> %tmp72, i16 %tmp, i32 2
        %tmp74 = insertelement <8 x i16> %tmp73, i16 %tmp, i32 3
        %tmp75 = insertelement <8 x i16> %tmp74, i16 %tmp, i32 4
        %tmp76 = insertelement <8 x i16> %tmp75, i16 %tmp, i32 5
        %tmp77 = insertelement <8 x i16> %tmp76, i16 %tmp, i32 6
        %tmp78 = insertelement <8 x i16> %tmp77, i16 %tmp, i32 7
        %tmp78.upgrd.2 = bitcast <8 x i16> %tmp78 to <16 x i8>
        store <16 x i8> %tmp78.upgrd.2, <16 x i8>* %dst
        ret void
}

Here's a diff between compiling with
 llc -march=ppc32 -mcpu=g5 vec_splat.bc -disable-legalize-types
and
 llc -march=ppc32 -mcpu=g5 vec_splat.bc

<       sth 3, -16(1)
<       addi 3, 1, -16
<       lvx 2, 0, 3
<       vsplth 2, 2, 0
---
>       rlwimi 3, 3, 16, 0, 15
>       addi 6, 1, -16
>       stw 3, -16(1)
>       lvx 2, 0, 6
>       vspltw 2, 2, 0

The DAG coming out of legalize types is now very reasonable after
Bob's BUILD_VECTOR changes (note that the vector of v8i16 is being
built from i32 operands):

Optimized type-legalized selection DAG:
SelectionDAG has 9 nodes:
  0x1970718: ch = EntryToken
    0x1970718: <multiple use>
    0x1970cb8: i32 = Register  #1024
  0x1970db0: i32,ch = CopyFromReg 0x1970718, 0x1970cb8
      0x1970718: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
        0x1970db0: <multiple use>
      0x1970ac8: v8i16 = BUILD_VECTOR 0x1970db0, 0x1970db0, 0x1970db0,
0x1970db0, 0x1970db0, 0x1970db0, 0x1970db0, 0x1970db0
        0x1970718: <multiple use>
        0x1970ea8: i32 = Register  #1025
      0x1970fa0: i32,ch = CopyFromReg 0x1970718, 0x1970ea8
      0x19957c0: i32 = undef
    0x1995aa8: ch = store 0x1970718, 0x1970ac8, 0x1970fa0, 0x19957c0
<0x1961520:0> alignment=16
  0x19959b0: ch = ret 0x1995aa8

This then gets turned into something horrible by operand legalization.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 08:33:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 08:33:35 -0500
Subject: [LLVMbugs] [Bug 4000] New: LLVM has too many bugs
Message-ID: <bug-4000-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4000

           Summary: LLVM has too many bugs
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


I suggest rejecting any bug reports that would bring the
total number to more than 3999 reports!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 11:08:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 11:08:04 -0500
Subject: [LLVMbugs] [Bug 4001] New: assertion failed in lshr <2 x i64>
	without sse2
Message-ID: <bug-4001-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4001

           Summary: assertion failed in lshr <2 x i64> without sse2
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Given test/CodeGen/X86/vshift_split.ll, which for reference reads:
  define <2 x i64> @update(<2 x i64> %val) nounwind readnone {
  entry:
          %shr = lshr <2 x i64> %val, < i64 2, i64 2 >            ; <<2 x i64>>
[#uses=1]
          ret <2 x i64> %shr
  }
running llvm-as < vshift_split.ll | llc -march=-sse2 triggers an assertion
failure:
  llc: LegalizeDAG.cpp:517:
void<unnamed>::SelectionDAGLegalize::HandleOp(llvm::SDValue): Assertion
`(TypesNeedLegalizing || getTypeAction(VT) == Legal || IsLegalizingCallArgs ||
Op.getOpcode() == ISD::TargetConstant || Op.getOpcode() == ISD::BUILD_VECTOR)
&& "Illegal type introduced after type legalization?"' failed.
  0   llc 0x089a99e2
  1   llc 0x089a9f89
  2       0xffffe400 __kernel_sigreturn + 0
  Stack dump:
  0.      Program arguments: llc -mattr=-sse2 
  1.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@update'

It'd be nice if this didn't crash.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 12:18:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 12:18:45 -0500
Subject: [LLVMbugs] [Bug 4000] LLVM has too many bugs
In-Reply-To: <bug-4000-206@http.llvm.org/bugs/>
Message-ID: <200904181718.n3IHIjn5005401@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4000


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dalej at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Dale Johannesen <dalej at apple.com>  2009-04-18 12:18:44 ---
Excellent idea.  Rejected.

(How long have you been watching for this?:)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 12:23:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 12:23:04 -0500
Subject: [LLVMbugs] [Bug 4002] New: internal compiler error while compiling
	llvm-gcc
Message-ID: <bug-4002-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4002

           Summary: internal compiler error while compiling llvm-gcc
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rajika.kumarasiri at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I am getting following compiler error while compiling the llvm-gcc revision
69400. 

cc1: /home/rajika/projects/llvm/llvm/lib/CodeGen/RegAllocLocal.cpp:848:
void<unnamed>::RALocal::AllocateBasicBlock(llvm::MachineBasicBlock&): Assertion
`(!PhysRegsUsed[PhysReg] || PhysRegsUsed[PhysReg] == -1) && "Silently clearing
a virtual register?"' failed.
../../../llvm-gcc-4.2/libgomp/team.c:341: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.
make[3]: *** [team.lo] Error 1
make[3]: Leaving directory
`/home/rajika/projects/llvm/llvm-gcc-4.2-objects/i686-pc-linux-gnu/libgomp'
make[2]: *** [install-recursive] Error 1
make[2]: Leaving directory
`/home/rajika/projects/llvm/llvm-gcc-4.2-objects/i686-pc-linux-gnu/libgomp'
make[1]: *** [install-target-libgomp] Error 2
make[1]: Leaving directory `/home/rajika/projects/llvm/llvm-gcc-4.2-objects'
make: *** [install] Error 2


System information,
Running fedora 9/kernel - 2.6.27.19-78.2.30.fc9.i686.PAE
gcc version 4.2.4


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 15:11:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 15:11:06 -0500
Subject: [LLVMbugs] [Bug 4003] New: Remove SCALAR_TO_VECTOR
Message-ID: <bug-4003-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4003

           Summary: Remove SCALAR_TO_VECTOR
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


The SCALAR_TO_VECTOR codegen node is redundant: it is
equivalent to BUILD_VECTOR(Value, undef, undef, ...).
It should probably be removed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 15:49:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 15:49:38 -0500
Subject: [LLVMbugs] [Bug 3898] Incomplete breakdown of soft-float vectors
In-Reply-To: <bug-3898-206@http.llvm.org/bugs/>
Message-ID: <200904182049.n3IKncQL013609@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3898


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-18 15:49:38 ---
Patch applied here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076455.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076456.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 16:12:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 16:12:17 -0500
Subject: [LLVMbugs] [Bug 3226] clang infinite loop on the attached code
	snippet
In-Reply-To: <bug-3226-206@http.llvm.org/bugs/>
Message-ID: <200904182112.n3ILCH2h014469@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3226


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2009-04-18 16:12:16 ---
This behaves correctly, I have decided.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 16:29:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 16:29:36 -0500
Subject: [LLVMbugs] [Bug 3226] clang infinite loop on the attached code
	snippet
In-Reply-To: <bug-3226-206@http.llvm.org/bugs/>
Message-ID: <200904182129.n3ILTadv014975@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3226


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|INVALID                     |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-18 16:29:36 ---
This is now rejected:

t.m:4:5: error: illegal goto into protected scope
    goto L;
    ^
t.m:5:5: note: jump bypasses initialization of @finally block
  } @finally {
    ^


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 16:46:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 16:46:18 -0500
Subject: [LLVMbugs] [Bug 4004] New: invalid TLS with -relocation-model=pic
 -regalloc=local
Message-ID: <bug-4004-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4004

           Summary: invalid TLS with -relocation-model=pic -regalloc=local
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: rafael.espindola at gmail.com
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Compiling


@a = thread_local global i32 0          ; <i32*> [#uses=2]

define void @f(i32* nocapture %c, i32* nocapture %d) nounwind optsize {
entry:
        %0 = load i32* @a, align 4              ; <i32> [#uses=1]
        store i32 %0, i32* %c, align 4
        %1 = load i32* @a, align 4              ; <i32> [#uses=1]
        store i32 %1, i32* %d, align 4
        ret void
}


with llc -relocation-model=pic -regalloc=local produces

        .byte   0x66; leaq      a at TLSGD(%rip), %rdi; .word      0x6666; rex64
        movq    %rax, -8(%rsp)
        movq    %rcx, -16(%rsp)
        call    __tls_get_addr at PLT

This is invalid. The call must be just after the rex64.

I think this is a more general problem with the local register allocator. It is
inserting spills between two instructions linked by a flag.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 16:49:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 16:49:27 -0500
Subject: [LLVMbugs] [Bug 3999] CodeGen/PowerPC/vec_splat. ll fails when
 legalize types is turned on
In-Reply-To: <bug-3999-206@http.llvm.org/bugs/>
Message-ID: <200904182149.n3ILnRHw015555@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3999


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-04-18 16:49:26 ---
Fixed here:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090413/076454.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 18 17:30:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 18 Apr 2009 17:30:40 -0500
Subject: [LLVMbugs] [Bug 3917] __LINE__ wrong value
In-Reply-To: <bug-3917-206@http.llvm.org/bugs/>
Message-ID: <200904182230.n3IMUejm016819@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3917


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-18 17:30:40 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015821.html

Several other patches were related to this before it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 00:12:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 00:12:36 -0500
Subject: [LLVMbugs] [Bug 4005] New: wrong arm abi selected
Message-ID: <bug-4005-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4005

           Summary: wrong arm abi selected
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


The ARM build on a little-endian ARM with EBI fails with this:

/home/nlewycky/llvm-gcc/build/./gcc/xgcc -B/home/nlewycky/llvm-gcc/build/./gcc/
-B/home/nlewycky/llvm-gcc/install/armv7l-unknown-linux-gnu/bin/ [[...]]
/usr/bin/ld: unrecognised emulation mode: armelf_linux
Supported emulations: armelf_linux_eabi armelfb_linux_eabi
collect2: ld returned 1 exit status
make[4]: *** [libgcc_s.so] Error 1

In llvm-gcc gcc/config/arm there's linux-elf.h which chooses armelf_linux and
linux-eabi.h which chooses armelf_linux_eabi. The latter is correct, but the
former is being chosen. Why?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 02:21:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 02:21:20 -0500
Subject: [LLVMbugs] [Bug 4006] New: variable argument macro argumetn count
	issues
Message-ID: <bug-4006-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4006

           Summary: variable argument macro argumetn count issues
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


#define e(...) __VA_ARGS__
#if e() 0
#error
#endif

gives

~/src/nobackup/llvm/Release-Asserts/bin/clang-cc -pedantic-errors -std=c99
/tmp/bug.c
/tmp/bug.c:1:16: error: __VA_ARGS__ can only appear in the expansion of a C99
variadic macro
#define e(...) __VA_ARGS__
               ^
/tmp/bug.c:2:7: error: varargs argument missing, but tolerated as an extension
#if e() 0
      ^
2 diagnostics generated.

both of which appear bogus.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 12:32:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 12:32:51 -0500
Subject: [LLVMbugs] [Bug 3970] clang doesn't support fp stack in inline asm
	yet
In-Reply-To: <bug-3970-206@http.llvm.org/bugs/>
Message-ID: <200904191732.n3JHWp3H016067@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3970


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-04-19 12:32:50 ---
Ok, here ya go:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015858.html
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015859.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 13:27:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 13:27:33 -0500
Subject: [LLVMbugs] [Bug 4006] variable argument macro argumetn count issues
In-Reply-To: <bug-4006-206@http.llvm.org/bugs/>
Message-ID: <200904191827.n3JIRX24017743@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4006


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-19 13:27:32 ---
First warning fixed here, it was specific to __VA_ARGS__ being the first token
of a macro definition:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015864.html

Please reopen this if you disagree with my analysis of the second issue. 
Thanks Neil!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 15:08:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 15:08:02 -0500
Subject: [LLVMbugs] [Bug 3918] Invalid use of __VA_ARGS__ not diagnosed
In-Reply-To: <bug-3918-206@http.llvm.org/bugs/>
Message-ID: <200904192008.n3JK82G1020944@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3918


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-04-19 15:08:01 ---
This was actually a general problem with all poisoning.  Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090413/015867.html

I need to enhance some diags stuff before I add a testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 17:17:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 17:17:26 -0500
Subject: [LLVMbugs] [Bug 4006] variable argument macro argumetn count issues
In-Reply-To: <bug-4006-206@http.llvm.org/bugs/>
Message-ID: <200904192217.n3JMHQ8W025205@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4006


Neil Booth <neil at daikokuya.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #3 from Neil Booth <neil at daikokuya.co.uk>  2009-04-19 17:17:25 ---
But there is one argument, it's just empty.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 17:24:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 17:24:03 -0500
Subject: [LLVMbugs] [Bug 4008] New: ".." doesn't make it through "clang -E"
	intact
Message-ID: <bug-4008-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4008

           Summary: ".." doesn't make it through "clang -E" intact
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
..

Output of "gcc -E":
# 1 "<stdin>"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "<stdin>"
..

Output of "clang -E":
# 1 "<stdin>"
# 1 "<stdin>" 1
# 1 "<built-in>" 1
# 103 "<built-in>"
# 103 "<command line>" 1
# 104 "<built-in>" 2
# 1 "<stdin>" 2
. .

This issue blocks building emacs; the emacs build system runs its makefiles
through the C preprocessor.  Relevant snippet:
ALL_CFLAGS = C_SWITCH_SYSTEM C_SWITCH_MACHINE -DHAVE_CONFIG_H \
   -I. -I../src -I${srcdir} -I${srcdir}/../src ${LDFLAGS} ${CPPFLAGS} ${CFLAGS}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 18:07:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 18:07:20 -0500
Subject: [LLVMbugs] [Bug 4009] New: Crash with -instcombine -indvars
 ("Assertion `L-> isLCSSAForm()' failed.")
Message-ID: <bug-4009-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4009

           Summary: Crash with -instcombine -indvars ("Assertion `L-
                    >isLCSSAForm()' failed.")
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2835)
 --> (http://llvm.org/bugs/attachment.cgi?id=2835)
Testcase

Testcase attached; to reproduce, run "opt bugpoint-reduced-simplified.bc
-instcombine -indvars".

Reduced from a file in emacs compiled with clang.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 22:01:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 22:01:58 -0500
Subject: [LLVMbugs] [Bug 3307] gcc compat: clang rejects pointer/ int
 mismatch in conditional expression
In-Reply-To: <bug-3307-206@http.llvm.org/bugs/>
Message-ID: <200904200301.n3K31wEt002804@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3307


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-19 22:01:57 ---
THis was fixed in r68617.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 22:04:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 22:04:57 -0500
Subject: [LLVMbugs] [Bug 3689] clang: support -dD -E
In-Reply-To: <bug-3689-206@http.llvm.org/bugs/>
Message-ID: <200904200304.n3K34vjD002918@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3689


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-19 22:04:57 ---
This was fixed in r68884.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 22:22:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 22:22:13 -0500
Subject: [LLVMbugs] [Bug 3248] Operand of sizeof(vla) not evaluated
In-Reply-To: <bug-3248-206@http.llvm.org/bugs/>
Message-ID: <200904200322.n3K3MDi8003466@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3248


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-19 22:22:12 ---
Fix committed in r69566.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 22:54:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 22:54:34 -0500
Subject: [LLVMbugs] [Bug 3247] Assertion in VLA codegen with sizeof+cast
In-Reply-To: <bug-3247-206@http.llvm.org/bugs/>
Message-ID: <200904200354.n3K3sYHm004560@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3247


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-19 22:54:34 ---
Fixed in r69567.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 19 23:50:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 19 Apr 2009 23:50:51 -0500
Subject: [LLVMbugs] [Bug 4012] New: The clang web page looks atrocious in IE
	6.0
Message-ID: <bug-4012-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4012

           Summary: The clang web page looks atrocious in IE 6.0
           Product: Documentation
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2838)
 --> (http://llvm.org/bugs/attachment.cgi?id=2838)
clang.llvm.org in IE

See screenshot.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 01:08:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 01:08:02 -0500
Subject: [LLVMbugs] [Bug 4017] New: crash in bitfield diagnostic
Message-ID: <bug-4017-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4017

           Summary: crash in bitfield diagnostic
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


clang crashes in diagnostic printing for:
--
struct s0 { char : 10; };
--
and
--
struct s0 { unsigned : -2; };
--

--
ddunbar at ozzy:tmp$ echo "struct s0 { char : 10 };" | clang-cc
<stdin>:1:13: error: 0   clang-cc          0x0000000100b09310
PrintStackTrace(void*) + 38
1   clang-cc          0x0000000100b09820 SignalHandler(int) + 288
2   libSystem.B.dylib 0x00007fff80aa58ba _sigtramp + 26
3   libSystem.B.dylib 0x00007fff5fbfa4e0 _sigtramp + 3742714944
4   clang-cc          0x00000001003bf90d
clang::DiagnosticInfo::FormatDiagnostic(llvm::SmallVectorImpl<char>&) const +
1485
5   clang-cc          0x00000001002b8d3c
clang::TextDiagnosticPrinter::HandleDiagnostic(clang::Diagnostic::Level,
clang::DiagnosticInfo const&) + 1430
6   clang-cc          0x00000001003bfba8 clang::Diagnostic::ProcessDiag() + 474
7   clang-cc          0x000000010036d12c clang::DiagnosticBuilder::Emit() + 96
8   clang-cc          0x00000001001bbd3c
clang::Sema::SemaDiagnosticBuilder::~SemaDiagnosticBuilder() + 22
9   clang-cc          0x00000001001c7673
clang::Sema::VerifyBitField(clang::SourceLocation, clang::IdentifierInfo*,
clang::QualType, clang::Expr const*) + 1031
10  clang-cc          0x00000001001cafb4
clang::Sema::CheckFieldDecl(clang::DeclarationName, clang::QualType,
clang::RecordDecl*, clang::SourceLocation, bool, clang::Expr*,
clang::AccessSpecifier, clang::NamedDecl*, clang::Declarator*) + 458
11  clang-cc          0x00000001001d2504
clang::Sema::HandleField(clang::Scope*, clang::RecordDecl*,
clang::SourceLocation, clang::Declarator&, clang::Expr*,
clang::AccessSpecifier) + 620
12  clang-cc          0x00000001001d25e3 clang::Sema::ActOnField(clang::Scope*,
clang::OpaquePtr<0>, clang::SourceLocation, clang::Declarator&, void*) + 89
13  clang-cc          0x000000010035b0ed
clang::Parser::ParseStructUnionBody(clang::SourceLocation, unsigned int,
clang::OpaquePtr<0>) + 825
14  clang-cc          0x000000010036386a
clang::Parser::ParseClassSpecifier(clang::tok::TokenKind,
clang::SourceLocation, clang::DeclSpec&, llvm::SmallVector<void*, 4u>*,
clang::AccessSpecifier) + 2290
15  clang-cc          0x0000000100358e39
clang::Parser::ParseDeclarationSpecifiers(clang::DeclSpec&,
llvm::SmallVector<void*, 4u>*, clang::AccessSpecifier) + 4723
16  clang-cc          0x000000010038bd26
clang::Parser::ParseDeclarationOrFunctionDefinition(llvm::SmallVector<void*,
4u>*, clang::AccessSpecifier) + 92
17  clang-cc          0x000000010038c9e7
clang::Parser::ParseExternalDeclaration() + 1237
18  clang-cc          0x000000010038ca98
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<1>&) + 106
19  clang-cc          0x00000001001bb1c7 clang::ParseAST(clang::Preprocessor&,
clang::ASTConsumer*, clang::ASTContext&, bool, bool) + 365
20  clang-cc          0x00000001000744f3 ProcessInputFile(clang::Preprocessor&,
clang::PreprocessorFactory&, std::string const&, ProgActions) + 2558
21  clang-cc          0x00000001000778ed main + 2149
22  clang-cc          0x00000001000272ec start + 52
Stack dump:
0.      Program arguments: clang-cc 
1.      <stdin>:1:23: current parser token '}'
2.      <stdin>:1:1: parsing struct/union body 's0'
Segmentation fault
ddunbar at ozzy:tmp$ echo "struct s0 { char : -2 };" | clang-cc
<stdin>:1:13: error: 0   clang-cc          0x0000000100b09310
PrintStackTrace(void*) + 38
1   clang-cc          0x0000000100b09820 SignalHandler(int) + 288
2   libSystem.B.dylib 0x00007fff80aa58ba _sigtramp + 26
3   libSystem.B.dylib 0x00007fff5fbfa4e0 _sigtramp + 3742714944
4   clang-cc          0x00000001003bf90d
clang::DiagnosticInfo::FormatDiagnostic(llvm::SmallVectorImpl<char>&) const +
1485
5   clang-cc          0x00000001002b8d3c
clang::TextDiagnosticPrinter::HandleDiagnostic(clang::Diagnostic::Level,
clang::DiagnosticInfo const&) + 1430
6   clang-cc          0x00000001003bfba8 clang::Diagnostic::ProcessDiag() + 474
7   clang-cc          0x000000010036d12c clang::DiagnosticBuilder::Emit() + 96
8   clang-cc          0x00000001001bbd3c
clang::Sema::SemaDiagnosticBuilder::~SemaDiagnosticBuilder() + 22
9   clang-cc          0x00000001001c759c
clang::Sema::VerifyBitField(clang::SourceLocation, clang::IdentifierInfo*,
clang::QualType, clang::Expr const*) + 816
10  clang-cc          0x00000001001cafb4
clang::Sema::CheckFieldDecl(clang::DeclarationName, clang::QualType,
clang::RecordDecl*, clang::SourceLocation, bool, clang::Expr*,
clang::AccessSpecifier, clang::NamedDecl*, clang::Declarator*) + 458
11  clang-cc          0x00000001001d2504
clang::Sema::HandleField(clang::Scope*, clang::RecordDecl*,
clang::SourceLocation, clang::Declarator&, clang::Expr*,
clang::AccessSpecifier) + 620
12  clang-cc          0x00000001001d25e3 clang::Sema::ActOnField(clang::Scope*,
clang::OpaquePtr<0>, clang::SourceLocation, clang::Declarator&, void*) + 89
13  clang-cc          0x000000010035b0ed
clang::Parser::ParseStructUnionBody(clang::SourceLocation, unsigned int,
clang::OpaquePtr<0>) + 825
14  clang-cc          0x000000010036386a
clang::Parser::ParseClassSpecifier(clang::tok::TokenKind,
clang::SourceLocation, clang::DeclSpec&, llvm::SmallVector<void*, 4u>*,
clang::AccessSpecifier) + 2290
15  clang-cc          0x0000000100358e39
clang::Parser::ParseDeclarationSpecifiers(clang::DeclSpec&,
llvm::SmallVector<void*, 4u>*, clang::AccessSpecifier) + 4723
16  clang-cc          0x000000010038bd26
clang::Parser::ParseDeclarationOrFunctionDefinition(llvm::SmallVector<void*,
4u>*, clang::AccessSpecifier) + 92
17  clang-cc          0x000000010038c9e7
clang::Parser::ParseExternalDeclaration() + 1237
18  clang-cc          0x000000010038ca98
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<1>&) + 106
19  clang-cc          0x00000001001bb1c7 clang::ParseAST(clang::Preprocessor&,
clang::ASTConsumer*, clang::ASTContext&, bool, bool) + 365
20  clang-cc          0x00000001000744f3 ProcessInputFile(clang::Preprocessor&,
clang::PreprocessorFactory&, std::string const&, ProgActions) + 2558
21  clang-cc          0x00000001000778ed main + 2149
22  clang-cc          0x00000001000272ec start + 52
Stack dump:
0.      Program arguments: clang-cc 
1.      <stdin>:1:23: current parser token '}'
2.      <stdin>:1:1: parsing struct/union body 's0'
Segmentation fault
--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 01:48:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 01:48:42 -0500
Subject: [LLVMbugs] [Bug 4018] New: llvm.used is undocumented
Message-ID: <bug-4018-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4018

           Summary: llvm.used is undocumented
           Product: Documentation
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


llvm.used is undocumented and our users will only use documented magic!

Should this just go into LangRef?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 12:30:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 12:30:58 -0500
Subject: [LLVMbugs] [Bug 4017] crash in bitfield diagnostic
In-Reply-To: <bug-4017-206@http.llvm.org/bugs/>
Message-ID: <200904201730.n3KHUwL2007406@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4017


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-20 12:30:58 ---
Fixed, thanks:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/015925.htmlbug


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 13:54:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 13:54:06 -0500
Subject: [LLVMbugs] [Bug 4005] wrong arm abi selected
In-Reply-To: <bug-4005-206@http.llvm.org/bugs/>
Message-ID: <200904201854.n3KIs6QJ010141@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4005


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bob.wilson at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-04-20 13:54:05 ---
I'm not sure what you mean by "little-endian ARM with EBI" (even assuming you
meant to type "EABI"), but judging from the path shown in your description, you
configured with a target triple of "armv7l-linux-gnu".  If you look at the
llvm-gcc sources in gcc/config.gcc, you should see that linux-eabi.h is only
included if the target matches the following pattern:

    arm*-*-linux-*eabi

If you configure llvm-gcc with a target triple that matches that pattern, it
should do what you want.  Hopefully that answers your question.  If there is
something more subtle going on and you still think it is a bug, I'd be glad to
take another look.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 16:09:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 16:09:03 -0500
Subject: [LLVMbugs] [Bug 4006] variable argument macro argumetn count issues
In-Reply-To: <bug-4006-206@http.llvm.org/bugs/>
Message-ID: <200904202109.n3KL93UJ014831@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4006


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-04-20 16:09:03 ---
You're right as usual, thanks again Neil:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/015939.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 16:51:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 16:51:45 -0500
Subject: [LLVMbugs] [Bug 4005] wrong arm abi selected
In-Reply-To: <bug-4005-206@http.llvm.org/bugs/>
Message-ID: <200904202151.n3KLpjjK016445@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4005


Nick Lewycky <nlewycky at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nlewycky at google.com
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #2 from Nick Lewycky <nlewycky at google.com>  2009-04-20 16:51:44 ---
I did mean to type EABI, sorry.

The problem is that I didn't specify a target when configuring, I ran it on a
native ARM machine. Why did configure get the wrong target?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 17:46:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 17:46:05 -0500
Subject: [LLVMbugs] [Bug 4024] New: TextDiagnosticPrinter doesn' t respect
 ShowLocation when printing include stack
Message-ID: <bug-4024-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4024

           Summary: TextDiagnosticPrinter doesn't respect ShowLocation when
                    printing include stack
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2842)
 --> (http://llvm.org/bugs/attachment.cgi?id=2842)
Fix for issue.

Currently, TextDiagnosticPrinter will print the location in the include stack
even if ShowLocation is false.

Attached patch fixes the issue.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 20:32:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 20:32:51 -0500
Subject: [LLVMbugs] [Bug 4025] New: debug info generates duplicate label
Message-ID: <bug-4025-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4025

           Summary: debug info generates duplicate label
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2844)
 --> (http://llvm.org/bugs/attachment.cgi?id=2844)
c++ testcase

The attached delta-reduced test program produces bad assembly when built with
debug info:

$ llvm-g++ y.c -g
/tmp/cc5eEOy7.s: Assembler messages:
/tmp/cc5eEOy7.s:79: Error: symbol `.Llabel11' is already defined
$ llvm-g++ y.c
/usr/lib/../lib64/crt1.o: In function `_start':
(.text+0x20): undefined reference to `main'
collect2: ld returned 1 exit status

Not sure if it's llvm-gcc's fault for producing bad debug info in the first
place or a backend bug, so I'm filing it as a backend bug.

This happens on x86-64 linux.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 22:58:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 22:58:17 -0500
Subject: [LLVMbugs] [Bug 4024] TextDiagnosticPrinter doesn' t respect
 ShowLocation when printing include stack
In-Reply-To: <bug-4024-206@http.llvm.org/bugs/>
Message-ID: <200904210358.n3L3wH3s027607@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4024


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-20 22:58:16 ---
Applied, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/015957.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 20 23:30:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 20 Apr 2009 23:30:52 -0500
Subject: [LLVMbugs] [Bug 4026] New: Clang can't codegen __func__ without
	implicit cast
Message-ID: <bug-4026-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4026

           Summary: Clang can't codegen __func__ without implicit cast
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


$ cat t.c
void foo() {
  __func__;
}
$ clang t.c 
t.c:4:3: warning: expression result unused [-Wunused-value]
  __func__;
  ^~~~~~~~
t.c:4:3: error: cannot compile this aggregate expression yet
  __func__;
  ^~~~~~~~


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 00:02:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 00:02:41 -0500
Subject: [LLVMbugs] [Bug 4027] New: compound literals rejected in strange way
Message-ID: <bug-4027-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4027

           Summary: compound literals rejected in strange way
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


This compiles fine:
struct a { int x, y };
struct a V = (struct a){ 1, 2};

This does is rejected by sema:
struct a { int x, y };
const struct a V = (struct a){ 1, 2};

t.c:4:20: error: initializer element is not a compile-time constant
const struct a V = (struct a){ 1, 2};
                   ^~~~~~~~~~~~~~~~~


Huh?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 00:19:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 00:19:30 -0500
Subject: [LLVMbugs] [Bug 4027] compound literals rejected in strange way
In-Reply-To: <bug-4027-206@http.llvm.org/bugs/>
Message-ID: <200904210519.n3L5JUCD030031@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4027


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-21 00:19:29 ---
Silly little bug, fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/015961.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 03:22:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 03:22:53 -0500
Subject: [LLVMbugs] [Bug 4002] internal compiler error while compiling
	llvm-gcc
In-Reply-To: <bug-4002-206@http.llvm.org/bugs/>
Message-ID: <200904210822.n3L8MrQx008519@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4002


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #9 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-04-21 03:22:52 ---
Fixed on revision 69672.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 10:18:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 10:18:15 -0500
Subject: [LLVMbugs] [Bug 3997] llc lowers intrinsics incorrectly with the
 presense of regparm attribute
In-Reply-To: <bug-3997-206@http.llvm.org/bugs/>
Message-ID: <200904211518.n3LFIFKw022878@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3997


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |alenhar2 at uiuc.edu
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #4 from Andrew Lenharth <alenhar2 at uiuc.edu>  2009-04-21 10:18:14 ---
The problem here is llvm-gcc is not respecting either the -ffreestanding or
-fno-builtin flags by generating an llvm intrinsic for memcpy.  The definition
in the C file (with the regparm flag) no longer matches the intrinsic and
should not be lowered to an intrinsic in the IR.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 12:22:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 12:22:58 -0500
Subject: [LLVMbugs] [Bug 4025] debug info generates duplicate label
In-Reply-To: <bug-4025-206@http.llvm.org/bugs/>
Message-ID: <200904211722.n3LHMwPV029081@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4025


Devang Patel <dpatel at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Devang Patel <dpatel at apple.com>  2009-04-21 12:22:58 ---
Debug info entries include display name and linkage name for symbols. The
llvm-gcc FE was not interpreting GCC trees enough to extract linkage (or
mangled) name.

I applied fix to use llvm value name, if it is available. revision 69683.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 12:30:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 12:30:56 -0500
Subject: [LLVMbugs] [Bug 4005] wrong arm abi selected
In-Reply-To: <bug-4005-206@http.llvm.org/bugs/>
Message-ID: <200904211730.n3LHUuqX029383@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4005


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Bob Wilson <bob.wilson at apple.com>  2009-04-21 12:30:56 ---
In revision 69685, I added a check for __ARM_EABI__ (reimplemented in a
different style so we don't contaminate this file with GPL3 licensing).  I
don't have an ARM Linux system to test it.  Please try it out and reopen the
bug if it still doesn't work.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 14:16:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 14:16:42 -0500
Subject: [LLVMbugs] [Bug 4028] New: Iterating over all uses of a Function is
 buggy on Windows
Message-ID: <bug-4028-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4028

           Summary: Iterating over all uses of a Function is buggy on
                    Windows
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Bitcode Reader
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: foldr at codedgers.com
                CC: llvmbugs at cs.uiuc.edu, foldr at codedgers.com


Created an attachment (id=2845)
 --> (http://llvm.org/bugs/attachment.cgi?id=2845)
Test pass that exhibits this behaviour

[Copied from my message to llvm-dev]

I try to iterate over all uses of a Function with the following
code (simplified):

for (Value::use_iterator UI = F->use_begin(), UE = F->use_end();
                         UI != UE; ++UI) {

    if (CallInst* I = dyn_cast<CallInst>(*UI)) {
    // do something interesting
    }
}

This works on Linux, but on Windows the dyn_cast fails, even
though the only use of F in that module is a 'call'. However,
dyn_cast<Instruction>(*UI) works, and returns a pointer
to "<Invalid operator>" instruction.

[Addendum]

I wrote a small test pass that exhibits this behaviour, and also
manages to crash opt:

main
interesting_function
<Invalid operator>
This is not a CallInst!
Stack dump:
0.      Running pass 'Function Pass Manager' on module 'simple.bc'.
1.      Running pass 'Module Verifier' on function '@main'
0053AD15 (0x00564090 0x53E58955 0x0022F860 0x005F88ED)
0053AF59 (0x0022F860 0x00564090 0x53E58955 0x00000002)
0053AFC6 (0x0022F860 0x0081D430 0x0022F5EC 0x005BE860)
0053FA01 (0x0022F7D0 0x0081D430 0x0022F848 0x00540F67)
005D9158 (0x0022F7D0 0x0081D430 0x0022F7F0 0x00000000)
00540F67 (0x0081D430 0x0022F860 0x00000000 0x0060B559)
005412DF (0x0081D430 0x003F7CF4 0x00000000 0x00638943)
006733CC (0x003F7CF4 0x0081D430 0x0022F8D8 0x003F7CD0)
005D724D (0x003F7CD0 0x0081D430 0x0022F9B8 0x0054F512)
005D7356 (0x003F7CD0 0x0022F980 0x003F698C 0x0081D430)
0054F512 (0x003F7CD0 0x003F698C 0x0022FAC8 0x0054A6DD)
0054EB97 (0x003F7CD0 0x003F698C 0x0022FA48 0x005EC980)
005ECF0C (0x003F7CD0 0x003F698C 0x0022FA48 0x00691064)
005EC980 (0x003F7CD0 0x003F698C 0x003F7C28 0x003F7BD8)
00548D19 (0x003F7CD0 0x003F698C 0x0022FAC8 0x005E5B38)
005ECCF8 (0x003F7CD0 0x003F8F04 0x003F8E78 0x006103E9)
005EC701 (0x003F7CD0 0x003F8E78 0x0022FB08 0x00663D80)
005ECCA8 (0x003F7CD0 0x003F7BF8 0x003F7BF8 0x00536896)
005ECC59 (0x003F7CD0 0x003F7BD8 0x0022FB48 0x0062FBAD)
005D7530 (0x003F7CD0 0x003F7BD8 0x003F7BD8 0x00000004)
0053804F (0x003FBE68 0x003F7BD8 0x003F9318 0x003F7BD8)
005381DC (0x003FBE68 0x003F9340 0x003F9340 0x00000005)
00538407 (0x003F6C18 0x003F9340 0x0022FC58 0x00000000)
0053879C (0x003FAC80 0x003F9340 0x0022FF78 0x00404E2E)
005389EF (0x0022FE10 0x003F9340 0x008142FC 0x00000000)
00404E2E (0x00000003 0x003F2D20 0x003F30E8 0x00821004)
004010B6 (0x00000001 0x00000009 0x0022FFF0 0x7D4E7D42)
00401148 (0x00000000 0x00000000 0x7EFDE000 0xC0000005)
7D4E7D42 (0x00401130 0x00000000 0x00000000 0x00000000),
BaseProcessInitPostImport()+0141 bytes(s)
E:\code\codedgers\llvm\Debug\bin\opt.exe: could not open file


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 18:00:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 18:00:29 -0500
Subject: [LLVMbugs] [Bug 4026] Clang can't codegen __func__ without implicit
	cast
In-Reply-To: <bug-4026-206@http.llvm.org/bugs/>
Message-ID: <200904212300.n3LN0TUE008666@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4026


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-21 18:00:27 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/016032.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 21 18:29:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 21 Apr 2009 18:29:00 -0500
Subject: [LLVMbugs] [Bug 4008] ".." doesn't make it through "clang -E" intact
In-Reply-To: <bug-4008-206@http.llvm.org/bugs/>
Message-ID: <200904212329.n3LNT0Ec009604@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4008


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-04-21 18:28:59 ---
No, you're right.  This is a great solution!  Applied here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/016034.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 02:17:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 02:17:09 -0500
Subject: [LLVMbugs] [Bug 3946] SmallVector::insert should use copy_backward
In-Reply-To: <bug-3946-206@http.llvm.org/bugs/>
Message-ID: <200904220717.n3M7H9TY025580@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3946


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Owen Anderson <resistor at mac.com>  2009-04-22 02:17:08 ---
Applied in r69785.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 02:19:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 02:19:47 -0500
Subject: [LLVMbugs] [Bug 3909] assert in SCEV
In-Reply-To: <bug-3909-206@http.llvm.org/bugs/>
Message-ID: <200904220719.n3M7JlkM025667@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3909


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #5 from Owen Anderson <resistor at mac.com>  2009-04-22 02:19:46 ---
This does not reproduce with LLVM TOT.  Please reopen if you can reproduce it
again.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 03:51:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 03:51:12 -0500
Subject: [LLVMbugs] [Bug 3549] LCSSA performance problem
In-Reply-To: <bug-3549-206@http.llvm.org/bugs/>
Message-ID: <200904220851.n3M8pCdl001575@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3549


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #12 from Owen Anderson <resistor at mac.com>  2009-04-22 03:51:02 ---
As of r69791, there's not a whole lot more LCSSA can do to improve this, so I'm
calling it closed.  Feel free to file another bug asking Gabor to make
getUser() less slow. ;-)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:06:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:06:20 -0500
Subject: [LLVMbugs] [Bug 3497] memory barriers not lowered correctly on
	x86/x86-64
In-Reply-To: <bug-3497-206@http.llvm.org/bugs/>
Message-ID: <200904220906.n3M96KwK002660@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3497


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Owen Anderson <resistor at mac.com>  2009-04-22 04:05:56 ---


*** This bug has been marked as a duplicate of bug 2829 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:06:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:06:20 -0500
Subject: [LLVMbugs] [Bug 3856] LLVM runs out of memory invoking functions
	under JIT
In-Reply-To: <bug-3856-206@http.llvm.org/bugs/>
Message-ID: <200904220906.n3M96KbF002661@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3856


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Owen Anderson <resistor at mac.com>  2009-04-22 04:05:16 ---


*** This bug has been marked as a duplicate of bug 2891 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:20:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:20:04 -0500
Subject: [LLVMbugs] [Bug 2124] llvm-dis exhausts all memory (unnamed
	recursive types)
In-Reply-To: <bug-2124-206@http.llvm.org/bugs/>
Message-ID: <200904220920.n3M9K4rR003852@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2124


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #11 from Owen Anderson <resistor at mac.com>  2009-04-22 04:18:50 ---
This seems to work now.  I can disassemble the file in a second or so, and I
can link it with some other trivial files without perceptible slowdown.  If the
linking is still an issue, please reopened with a testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:29:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:29:06 -0500
Subject: [LLVMbugs] [Bug 893] Need a machine code LICM pass
In-Reply-To: <bug-893-206@http.llvm.org/bugs/>
Message-ID: <200904220929.n3M9T6Br004260@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=893


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Owen Anderson <resistor at mac.com>  2009-04-22 04:28:25 ---
This has been done for a while.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:33:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:33:58 -0500
Subject: [LLVMbugs] [Bug 1710] No debug info for optimised code
In-Reply-To: <bug-1710-206@http.llvm.org/bugs/>
Message-ID: <200904220933.n3M9XwT3004472@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1710


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Owen Anderson <resistor at mac.com>  2009-04-22 04:33:49 ---
This is more or less working now, thanks to Devang, Bill, and Dale.  Any
particular failures should be reported in their own PRs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:35:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:35:54 -0500
Subject: [LLVMbugs] [Bug 1270] Support AP integers in the code generator
In-Reply-To: <bug-1270-206@http.llvm.org/bugs/>
Message-ID: <200904220935.n3M9ZsRx004579@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1270


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Duncan Sands <baldrick at free.fr>  2009-04-22 04:35:43 ---
This is done now that LegalizeTypes has landed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:37:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:37:56 -0500
Subject: [LLVMbugs] [Bug 452] memset/memcpy analysis/optimizations need
	improving
In-Reply-To: <bug-452-206@http.llvm.org/bugs/>
Message-ID: <200904220937.n3M9buJF004661@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=452


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Owen Anderson <resistor at mac.com>  2009-04-22 04:37:27 ---
I'm closing this open-ended bug, as we'll never catch EVERY case.  If you still
have specific cases that you'd like to see addressed, please file individual
PRs for them.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:45:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:45:08 -0500
Subject: [LLVMbugs] [Bug 1309] [meta] DSE is not aggressive enough
In-Reply-To: <bug-1309-206@http.llvm.org/bugs/>
Message-ID: <200904220945.n3M9j8Vv004951@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1309


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #12 from Owen Anderson <resistor at mac.com>  2009-04-22 04:44:52 ---
Closing a pointless meta bug.  Please file individual PRs for remaining missed
optzns.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:47:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:47:06 -0500
Subject: [LLVMbugs] [Bug 1368] Per-changeset build and regression testing
In-Reply-To: <bug-1368-206@http.llvm.org/bugs/>
Message-ID: <200904220947.n3M9l6Sr005046@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1368


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Owen Anderson <resistor at mac.com>  2009-04-22 04:46:47 ---
This is now taken care of by the buildbots.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 04:55:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 04:55:15 -0500
Subject: [LLVMbugs] [Bug 2062] build: cannot build llvm from source on
 read-only file system
In-Reply-To: <bug-2062-206@http.llvm.org/bugs/>
Message-ID: <200904220955.n3M9tEo2005409@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2062


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Owen Anderson <resistor at mac.com>  2009-04-22 04:54:44 ---
We don't use lex and yacc anymore, so this should work.  I'm aware of several
users who build from read-only directories, so this appears to be working. 
Please reopen if this is not the case.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:02:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:02:31 -0500
Subject: [LLVMbugs] [Bug 2639] opt -loop-deletion -loop-rotate triggers
 assert in domfrontier
In-Reply-To: <bug-2639-206@http.llvm.org/bugs/>
Message-ID: <200904221002.n3MA2VE6005715@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2639


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #7 from Owen Anderson <resistor at mac.com>  2009-04-22 05:02:28 ---
This works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:05:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:05:45 -0500
Subject: [LLVMbugs] [Bug 2537] -loop-reduce crashes with extremely large
	stride (i128)
In-Reply-To: <bug-2537-206@http.llvm.org/bugs/>
Message-ID: <200904221005.n3MA5jTE005820@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2537


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Owen Anderson <resistor at mac.com>  2009-04-22 05:05:44 ---
Works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:11:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:11:49 -0500
Subject: [LLVMbugs] [Bug 3085] "Instruction does not dominate all uses!"
In-Reply-To: <bug-3085-206@http.llvm.org/bugs/>
Message-ID: <200904221011.n3MABnd5006095@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3085


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #2 from Owen Anderson <resistor at mac.com>  2009-04-22 05:11:40 ---
Works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:12:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:12:56 -0500
Subject: [LLVMbugs] [Bug 3092] opt segfaults with -addreadattrs -die -inline
 -globalsmodref-aa -licm
In-Reply-To: <bug-3092-206@http.llvm.org/bugs/>
Message-ID: <200904221012.n3MACuL2006179@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3092


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #3 from Owen Anderson <resistor at mac.com>  2009-04-22 05:12:54 ---
Works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:13:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:13:48 -0500
Subject: [LLVMbugs] [Bug 3086] Segmentation fault in alias analysis
In-Reply-To: <bug-3086-206@http.llvm.org/bugs/>
Message-ID: <200904221013.n3MADmUS006236@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3086


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #4 from Owen Anderson <resistor at mac.com>  2009-04-22 05:13:45 ---
Works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:15:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:15:53 -0500
Subject: [LLVMbugs] [Bug 3409] gvnpre: segmentation fault
In-Reply-To: <bug-3409-206@http.llvm.org/bugs/>
Message-ID: <200904221015.n3MAFrIx006341@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3409


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #2 from Owen Anderson <resistor at mac.com>  2009-04-22 05:15:51 ---
Honestly, GVNPRE probably needs to be completely overhauled anyways.  This is
WONTFIX'd until someone wants to take it over.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:21:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:21:29 -0500
Subject: [LLVMbugs] [Bug 3998] LLVM JIT generates code into a fixed size
	buffer
In-Reply-To: <bug-3998-206@http.llvm.org/bugs/>
Message-ID: <200904221021.n3MALTcK006567@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3998


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |resistor at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #2 from Owen Anderson <resistor at mac.com>  2009-04-22 05:21:22 ---


*** This bug has been marked as a duplicate of bug 2891 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:26:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:26:12 -0500
Subject: [LLVMbugs] [Bug 3715] llvm-gcc fails to build on Fedora Linux x86_86
In-Reply-To: <bug-3715-206@http.llvm.org/bugs/>
Message-ID: <200904221026.n3MAQCso006878@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3715


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-04-22 05:26:10 ---
Looks like user error: should have configured with --disable-multlib.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:27:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:27:08 -0500
Subject: [LLVMbugs] [Bug 3160] llvm-gcc bootstrap internal error during pch
	generation
In-Reply-To: <bug-3160-206@http.llvm.org/bugs/>
Message-ID: <200904221027.n3MAR8qR006927@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3160


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #9 from Duncan Sands <baldrick at free.fr>  2009-04-22 05:27:05 ---
The offending patch was reverted.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:40:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:40:59 -0500
Subject: [LLVMbugs] [Bug 3632] test/FrontendC/2009-01-20-k8.c is not portable
In-Reply-To: <bug-3632-206@http.llvm.org/bugs/>
Message-ID: <200904221040.n3MAexh0007376@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3632


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-04-22 05:40:47 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076566.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:41:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:41:03 -0500
Subject: [LLVMbugs] [Bug 3631]
 test/FrontendFortran/2008-11-03-OptionOverride.f90 not portable
In-Reply-To: <bug-3631-206@http.llvm.org/bugs/>
Message-ID: <200904221041.n3MAf3R5007390@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3631


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Duncan Sands <baldrick at free.fr>  2009-04-22 05:41:00 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076566.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 05:42:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 05:42:23 -0500
Subject: [LLVMbugs] [Bug 2890] llvm-gcc doesn't support the GNU objc runtime
In-Reply-To: <bug-2890-206@http.llvm.org/bugs/>
Message-ID: <200904221042.n3MAgNHc007468@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2890


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #5 from Duncan Sands <baldrick at free.fr>  2009-04-22 05:42:21 ---
It seems there is no interest in fixing the objc runtime on linux.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 07:05:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 07:05:56 -0500
Subject: [LLVMbugs] [Bug 4029] New: CallGraph not always deleted before
 running gvn/simplifycfg
Message-ID: <bug-4029-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4029

           Summary: CallGraph not always deleted before running
                    gvn/simplifycfg
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: dpatel at apple.com, llvmbugs at cs.uiuc.edu


Created an attachment (id=2851)
 --> (http://llvm.org/bugs/attachment.cgi?id=2851)
Debugging patch that fires an assertion if the callgraph contains deleted
instructions

The test Analysis/GlobalsModRef/purecse.ll runs opt -globalsmodref-aa -gvn
-instcombine.  The callgraph constructed for globalsmodref-aa is not destroyed
before gvn is run.  Similarly, the test
Transforms/CondProp/2007-08-01-InvalidRead.ll runs opt -inline -tailduplicate
-condprop -simplifycfg, and
the callgraph constructed for -inline is not destroyed before simplifycfg is
run.  This is in spite of the fact that gvn/simplifycfg neither require nor
preserve the callgraph.

Spotted using the attached debugging patch (which would be good to commit
one day).  My guess is that this is essentially the same problem as PR3601.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 08:43:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 08:43:38 -0500
Subject: [LLVMbugs] [Bug 4030] New: nightly tester build failure emails have
 wrong URL for the build log
Message-ID: <bug-4030-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4030

           Summary: nightly tester build failure emails have wrong URL for
                    the build log
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


The machine number is missing in the url.  For example:
### BEGIN EXAMPLE EMAIL ###
[llvm-testresults] Grawp-PIC i386 nightly tester results
From: 
Apache <apache at cs.uiuc.edu>
  To: 
llvm-testresults at cs.uiuc.edu
  Date: 
Today 01:30:29 pm

http://llvm.org/nightlytest/test.php?machine=372&night=10521
Name: grawp.apple.com
Nickname: Grawp-PIC
Buildstatus: Error: compilation aborted

Buildlog available at
http://llvm.org/nightlytest/machines/2009-04-22_06:20:42-Build-Log.txt
### END EXAMPLE EMAIL ###

The URL given in the "Buildlog available at" line doesn't exist.  It should be:
http://llvm.org/nightlytest/machines/372/2009-04-22_06:20:42-Build-Log.txt
The machine number needs to be inserted in the URL.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 08:53:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 08:53:54 -0500
Subject: [LLVMbugs] [Bug 2622] movlps does not get selected for 4, 5, 2,
 3 shuffle to/from memory
In-Reply-To: <bug-2622-206@http.llvm.org/bugs/>
Message-ID: <200904221353.n3MDrsW3014075@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2622


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-04-22 08:53:50 ---
This now compiles to

        pushl   %ebp
        movl    %esp, %ebp
        andl    $-16, %esp
        movaps  __unnamed_2_1, %xmm0
        addps   __unnamed_2_2, %xmm0
        movlps  %xmm0, __unnamed_2_3
        movl    %ebp, %esp
        popl    %ebp
        ret


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 10:01:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 10:01:58 -0500
Subject: [LLVMbugs] [Bug 4031] New: no LLVM intrinsic for
 __builtin_ia32_ptestz128 SSE4. 1 instruction
Message-ID: <bug-4031-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4031

           Summary: no LLVM intrinsic for __builtin_ia32_ptestz128 SSE4.1
                    instruction
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


This causes _mm_testz_si128 to be unlowerable:
/home/edwin/llvm-bootstrap/install/bin/../lib/gcc/x86_64-unknown-linux-gnu/4.2.1/include/smmintrin.h:258:
error: unsupported targetbuiltin ???__builtin_ia32_ptestz128??? used

I see that llc has +sse41, so I assume that LLVM wants to support sse4.1, so I
looked at LLVM LangRef, but didn't find a corresponding intrinsic to lower it
to.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 10:14:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 10:14:49 -0500
Subject: [LLVMbugs] [Bug 4032] New: llc produces spurious error messages on
	Windows
Message-ID: <bug-4032-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4032

           Summary: llc produces spurious error messages on Windows
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: foldr at codedgers.com
                CC: llvmbugs at cs.uiuc.edu


On Windows, running llc (with or without arguments) produces the following
error messages:

E:\>llc
<premain>: CommandLine Error: Argument 'mattr' defined more than once!
<premain>: CommandLine Error: Argument 'mcpu' defined more than once!
<premain>: CommandLine Error: Argument 'march' defined more than once!

$ llc simple.bc
<premain>: CommandLine Error: Argument 'mattr' defined more than once!
<premain>: CommandLine Error: Argument 'mcpu' defined more than once!
<premain>: CommandLine Error: Argument 'march' defined more than once!
llc.exe: CommandLine Error: Argument 'mattr' defined more than once!
llc.exe: CommandLine Error: Argument 'mcpu' defined more than once!
llc.exe: CommandLine Error: Argument 'march' defined more than once!

The tool itself seems to work fine, however.

This is with recent trunk and a clean compile (with current MinGW).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 11:12:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 11:12:53 -0500
Subject: [LLVMbugs] [Bug 4009] -indvars produces invalid IR
In-Reply-To: <bug-4009-206@http.llvm.org/bugs/>
Message-ID: <200904221612.n3MGCrlI019566@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4009


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Dan Gohman <gohman at apple.com>  2009-04-22 11:12:52 ---
Yes, this is exposed by my recent IndVars etc. changes. Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076573.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 12:51:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 12:51:51 -0500
Subject: [LLVMbugs] [Bug 4034] New: Colescer crash
Message-ID: <bug-4034-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4034

           Summary: Colescer crash
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: evan.cheng at apple.com, gohman at apple.com,
                    llvmbugs at cs.uiuc.edu


Created an attachment (id=2854)
 --> (http://llvm.org/bugs/attachment.cgi?id=2854)
Bitcode

Consider the attached bitcode. Running llc yields:

$ ./llc bugpoint-reduced-function.bc
llc: /home/asl/proj/llvm/src/lib/CodeGen/SimpleRegisterCoalescing.cpp:115: bool
llvm::SimpleRegisterCoalescing::AdjustCopiesBackFrom(llvm::LiveInterval&,
llvm::LiveInterval&, llvm::MachineInstr*): Assertion `BValNo->def == CopyIdx &&
"Copy doesn't define the value?"' failed.
0   llc             0x0000000000e2860f
1   llc             0x0000000000e28c8a
2   libpthread.so.0 0x00007f3100592390
3   libc.so.6       0x00007f30ff8d45c5 gsignal + 53
4   libc.so.6       0x00007f30ff8d5a70 abort + 272
5   libc.so.6       0x00007f30ff8cda3f __assert_fail + 239
6   llc             0x0000000000c1f1ec
llvm::SimpleRegisterCoalescing::AdjustCopiesBackFrom(llvm::LiveInterval&,
llvm::LiveInterval&, llvm::MachineInstr*) + 972
7   llc             0x0000000000c2527f
llvm::SimpleRegisterCoalescing::JoinCopy(llvm::CopyRec&, bool&) + 11311
8   llc             0x0000000000c26b54
llvm::SimpleRegisterCoalescing::joinIntervals() + 724
9   llc             0x0000000000c287dd
llvm::SimpleRegisterCoalescing::runOnMachineFunction(llvm::MachineFunction&) +
5389
10  llc             0x0000000000bb3b45
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 53
11  llc             0x0000000000db9235
llvm::FPPassManager::runOnFunction(llvm::Function&) + 565
12  llc             0x0000000000db9658
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 136
13  llc             0x0000000000db9846
llvm::FunctionPassManager::run(llvm::Function&) + 70
14  llc             0x0000000000502363 main + 2627
15  libc.so.6       0x00007f30ff8c15e4 __libc_start_main + 244
16  llc             0x00000000004ffa29
Stack dump:
0.      Program arguments: ./llc bugpoint-reduced-function.bc 
1.      Running pass 'Simple Register Coalescing' on function
'@getAffNeighbour'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 13:20:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 13:20:27 -0500
Subject: [LLVMbugs] [Bug 4035] New: clang miscompiles global complex float
Message-ID: <bug-4035-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4035

           Summary: clang miscompiles global complex float
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


This:
_Complex float G = 4;

Gets compiled to:
        type { i32, i32 }               ; type %0
@G = global %0 { i32 4, i32 0 }, align 4                ; <%0*> [#uses=0]


4 != 4.0


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 13:56:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 13:56:18 -0500
Subject: [LLVMbugs] [Bug 2958] ICE with global initialised to address of
 incomplete static
In-Reply-To: <bug-2958-206@http.llvm.org/bugs/>
Message-ID: <200904221856.n3MIuIMn014803@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2958


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2009-04-22 13:56:18 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076580.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 14:23:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 14:23:27 -0500
Subject: [LLVMbugs] [Bug 4035] clang miscompiles global complex float
In-Reply-To: <bug-4035-206@http.llvm.org/bugs/>
Message-ID: <200904221923.n3MJNR4S021512@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4035


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-22 14:23:27 ---
Fixed in r69821.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 14:36:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 14:36:16 -0500
Subject: [LLVMbugs] [Bug 4036] New: `opt -verify` crashes on Windows
Message-ID: <bug-4036-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4036

           Summary: `opt -verify` crashes on Windows
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: foldr at codedgers.com
                CC: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


Created an attachment (id=2855)
 --> (http://llvm.org/bugs/attachment.cgi?id=2855)
Bitcode file that exhibits this behaviour

Currently on Windows, `opt -verify` crashes even on simplest
input, like the code generated from

int main() {
  return 0;
}

Example:

$ llvm-gcc --emit-llvm simple.c -o simple.bc -c

$ opt -verify simple.bc
Stack dump:
0.      Running pass 'Function Pass Manager' on module 'simple.bc'.
1.      Running pass 'Module Verifier' on function '@main'
0053A735 (0x005624E0 0x53E58955 0x0022F860 0x005F8311)
0053A979 (0x0022F860 0x005624E0 0x53E58955 0x00000002)
0053A9E6 (0x0022F860 0x0081D208 0x0022F5EC 0x005BE250)
0053F421 (0x0022F7D0 0x0081D208 0x0022F848 0x00540987)
005D8AF8 (0x0022F7D0 0x0081D208 0x0022F7F0 0x00000000)
00540987 (0x0081D208 0x0022F860 0x00000000 0x0060AF59)
00540CFF (0x0081D208 0x003FBEEC 0x00000000 0x006383B3)
00672FA4 (0x003FBEEC 0x0081D208 0x0022F8D8 0x003FBEC8)
005D6BED (0x003FBEC8 0x0081D208 0x0022F9B8 0x0054D962)
005D6CF6 (0x003FBEC8 0x0022F980 0x003F885C 0x0081D208)
0054D962 (0x003FBEC8 0x003F885C 0x0022FAC8 0x00548B2D)
0054CFE7 (0x003FBEC8 0x003F885C 0x0022FA48 0x005EC284)
005EC810 (0x003FBEC8 0x003F885C 0x0022FA48 0x00690C68)
005EC284 (0x003FBEC8 0x003F885C 0x003F7260 0x003F7210)
00547169 (0x003FBEC8 0x003F885C 0x0022FAC8 0x005E544C)
005EC5FC (0x003FBEC8 0x003F889C 0x003F8478 0x0060FE85)
005EC005 (0x003FBEC8 0x003F8478 0x0022FB08 0x006638A0)
005EC5AC (0x003FBEC8 0x003F7230 0x003F7230 0x005362B6)
005EC55D (0x003FBEC8 0x003F7210 0x0022FB48 0x0062F61D)
005D6ED0 (0x003FBEC8 0x003F7210 0x003F7210 0x00000004)
00537A6F (0x003F7D08 0x003F7210 0x003F92E0 0x003F7210)
00537BFC (0x003F7D08 0x003F9308 0x003F9308 0x00000005)
00537E27 (0x003F9070 0x003F9308 0x0022FC58 0x00000000)
005381BC (0x003F9DE0 0x003F9308 0x0022FF78 0x004048EE)
0053840F (0x0022FE10 0x003F9308 0x00813E8C 0x00000000)
004048EE (0x00000003 0x003F2D50 0x003F3110 0x00821004)
004010B6 (0x00000001 0x00000009 0x0022FFF0 0x7D4E7D42)
00401148 (0x00000000 0x00000000 0x7EFDE000 0xC0000005)
7D4E7D42 (0x00401130 0x00000000 0x00000000 0x00000000),
BaseProcessInitPostImport()+0141 bytes(s)
E:\code\codedgers\llvm\Debug\bin\opt.exe: could not open file


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 17:07:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 17:07:09 -0500
Subject: [LLVMbugs] [Bug 4037] New: independently control IR optimization
 and fast codegen
Message-ID: <bug-4037-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4037

           Summary: independently control IR optimization and fast codegen
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


For debugging, I'd like to be able to specify the IR-level optimizations (-O0,
-O2) independently from whether codegen is running in fast mode at the llvm-gcc
driver level.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 17:16:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 17:16:08 -0500
Subject: [LLVMbugs] [Bug 3992] -include ignores includepath
In-Reply-To: <bug-3992-206@http.llvm.org/bugs/>
Message-ID: <200904222216.n3MMG8Fo027811@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3992


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Chris Lattner <clattner at apple.com>  2009-04-22 17:16:07 ---
great!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From isanbard at gmail.com  Wed Apr 22 18:21:06 2009
From: isanbard at gmail.com (Bill Wendling)
Date: Wed, 22 Apr 2009 16:21:06 -0700
Subject: [LLVMbugs] Fwd: [Bug 3092] opt segfaults with -addreadattrs -die
	-inline -globalsmodref-aa -licm
In-Reply-To: <16e5fdf90904221620w1279875btc2ff354496eaef71@mail.gmail.com>
References: <bug-3092-206@http.llvm.org/bugs/>
	<200904221012.n3MACuL2006179@zion.cs.uiuc.edu>
	<16e5fdf90904221620w1279875btc2ff354496eaef71@mail.gmail.com>
Message-ID: <16e5fdf90904221621h1d45be3cv2fabd042c8d98fd2@mail.gmail.com>

---------- Forwarded message ----------
From: Bill Wendling <isanbard at gmail.com>
Date: Wed, Apr 22, 2009 at 4:20 PM
Subject: Re: [LLVMbugs] [Bug 3092] opt segfaults with -addreadattrs
-die -inline -globalsmodref-aa -licm
To: bugzilla-daemon at cs.uiuc.edu


Hi Owen,

If there are testcases attached to the bugs you're closing, please go
ahead and add them to the test suite.

Thanks!
-bw

On Wed, Apr 22, 2009 at 3:12 AM, ?<bugzilla-daemon at cs.uiuc.edu> wrote:
> http://llvm.org/bugs/show_bug.cgi?id=3092
>
>
> Owen Anderson <resistor at mac.com> changed:
>
> ? ? ? ? ? What ? ?|Removed ? ? ? ? ? ? ? ? ? ? |Added
> ----------------------------------------------------------------------------
> ? ? ? ? ? ? ? ? CC| ? ? ? ? ? ? ? ? ? ? ? ? ? ?|resistor at mac.com
> ? ? ? ? ? ? Status|NEW ? ? ? ? ? ? ? ? ? ? ? ? |RESOLVED
> ? ? ? ? Resolution| ? ? ? ? ? ? ? ? ? ? ? ? ? ?|WORKSFORME
>
>
>
>
> --- Comment #3 from Owen Anderson <resistor at mac.com> ?2009-04-22 05:12:54 ---
> Works for me.
>
>
> --
> Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug.
> _______________________________________________
> LLVMbugs mailing list
> LLVMbugs at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
>


From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 18:29:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 18:29:12 -0500
Subject: [LLVMbugs] [Bug 4038] New: Crash with -instcombine -indvars
 ("Assertion `L-> isLCSSAForm()' failed.")
Message-ID: <bug-4038-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4038

           Summary: Crash with -instcombine -indvars ("Assertion `L-
                    >isLCSSAForm()' failed.")
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2856)
 --> (http://llvm.org/bugs/attachment.cgi?id=2856)
Testcase (reproduce with opt -instcombine -indvars)

Attached crashes with "opt -instcombine -indvars
bugpoint-reduced-simplified.bc".

Note that this is reduced from the same testcase as bug 4009, but this is still
an issue with current trunk (r69841).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 22 19:36:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 22 Apr 2009 19:36:41 -0500
Subject: [LLVMbugs] [Bug 4039] New: LegalizeDAG does not allow custom
 lowering of EXTRACT_SUBVECTOR
Message-ID: <bug-4039-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4039

           Summary: LegalizeDAG does not allow custom lowering of
                    EXTRACT_SUBVECTOR
           Product: tools
           Version: 2.5
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


The original code:
 case ISD::EXTRACT_SUBVECTOR: 
   Tmp1 = Node->getOperand(0);
   Tmp2 = LegalizeOp(Node->getOperand(1));
    Result = DAG.UpdateNodeOperands(Result, Tmp1, Tmp2);
   Result = ExpandEXTRACT_SUBVECTOR(Result);
   break;


Code that allows Custom Lowering:
LegalizeDAG.cpp:1792
 case ISD::EXTRACT_SUBVECTOR: 
   Tmp1 = LegalizeOp(Node->getOperand(0));
   Tmp2 = LegalizeOp(Node->getOperand(1));
    Result = DAG.UpdateNodeOperands(Result, Tmp1, Tmp2);
// Allow targets to custom lower the SHUFFLEs they support.
    switch
(TLI.getOperationAction(ISD::EXTRACT_SUBVECTOR,Result.getValueType())) {
    default: assert(0 && "Unknown operation action!");
    case TargetLowering::Custom:
      Tmp3 = TLI.LowerOperation(Result, DAG);
      if (Tmp3.getNode()) {
        Result = Tmp3;
        break;
      }
    case TargetLowering::Expand: {
      Result = LegalizeOp(ExpandEXTRACT_SUBVECTOR(Result));
      }
    break;
        };
   break;


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 01:53:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 01:53:35 -0500
Subject: [LLVMbugs] [Bug 4040] New: APInt's string constructor is too strict
Message-ID: <bug-4040-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4040

           Summary: APInt's string constructor is too strict
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Support Libraries
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2859)
 --> (http://llvm.org/bugs/attachment.cgi?id=2859)
Overly-simple test and fix

Calls like APInt(1, "1", 1, 10) fail the assert that ((slen*64)/22 <= numbits.
I ran into this in code that llc -march=cpp emits.

I've attached a patch with a test for this kind of case and the simplest fix
that could possibly work. I just subtract one from slen before checking if it's
too big for the number of bits. You probably actually want better a test and
fix, but this should capture the idea.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 03:20:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 03:20:57 -0500
Subject: [LLVMbugs] [Bug 3041] Disabled code should only be disabled on
 certain platforms
In-Reply-To: <bug-3041-206@http.llvm.org/bugs/>
Message-ID: <200904230820.n3N8Kvhx026275@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3041


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Duncan Sands <baldrick at free.fr>  2009-04-23 03:20:56 ---
The fix reached llvm-gcc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 10:24:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 10:24:27 -0500
Subject: [LLVMbugs] [Bug 4038] Crash with -instcombine -indvars ("Assertion
 `L->isLCSSAForm()' failed.")
In-Reply-To: <bug-4038-206@http.llvm.org/bugs/>
Message-ID: <200904231524.n3NFOR4c009205@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4038


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Dan Gohman <gohman at apple.com>  2009-04-23 10:24:24 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076608.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 10:26:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 10:26:27 -0500
Subject: [LLVMbugs] [Bug 4028] Iterating over all uses of a Function is
	buggy on Windows
In-Reply-To: <bug-4028-206@http.llvm.org/bugs/>
Message-ID: <200904231526.n3NFQRKS009297@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4028


Mikhail Glushenkov <foldr at codedgers.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Mikhail Glushenkov <foldr at codedgers.com>  2009-04-23 10:26:25 ---
I verified that reverting r68156 fixes both bug 4036 and bug 4028. In the
process, though, I stumbled upon a crash in llc :-(

*** This bug has been marked as a duplicate of bug 4036 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 15:02:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 15:02:40 -0500
Subject: [LLVMbugs] [Bug 4044] New: LegalizeDAG does not promote math ops
	correctly
Message-ID: <bug-4044-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4044

           Summary: LegalizeDAG does not promote math ops correctly
           Product: tools
           Version: 2.5
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


TOT LLVM only promotes correctly AND, OR, XOR but does not promote the
following opcodes.
ADD, SUB, MUL, MULHS, MULHU, UDIV, SDIV, SHL, SRL, SRA among other.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 15:26:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 15:26:13 -0500
Subject: [LLVMbugs] [Bug 3273] GNU Emacs doesn't compile
In-Reply-To: <bug-3273-206@http.llvm.org/bugs/>
Message-ID: <200904232026.n3NKQDRV019499@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3273


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-23 15:26:12 ---
The only issue blocking building/running Emacs now is bug 4041, which is a bug
in the Emacs source code.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 15:55:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 15:55:03 -0500
Subject: [LLVMbugs] [Bug 4045] New: llvm-gcc not using collect2,
 missing paths to .o files
Message-ID: <bug-4045-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4045

           Summary: llvm-gcc not using collect2, missing paths to .o files
           Product: tools
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


On my ARM box, the llvm-gcc build fails because it runs the wrong command to
link. For example:

  $ llvm-gcc/build/prev-gcc/xgcc x.o -o x -v
[...]
ld --eh-frame-hdr -dynamic-linker /lib/ld-linux.so.3 -X -m armelf_linux_eabi -o
x /usr/lib/crt1.o /usr/lib/crti.o crtbegin.o x.o -lgcc --as-needed -lgcc_s
--no-as-needed -lc -lgcc --as-needed -lgcc_s --no-as-needed crtend.o
/usr/lib/crtn.o

vs the system compiler:

 $ gcc-4.2 x.o -o x -v
[...]
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/collect2 --eh-frame-hdr -dynamic-linker
/lib/ld-linux.so.3 -X -m armelf_linux_eabi -o x
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/../../../crt1.o
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/../../../crti.o
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/crtbegin.o
-L/usr/lib/gcc/arm-linux-gnueabi/4.2.4 -L/usr/lib/gcc/arm-linux-gnueabi/4.2.4
-L/usr/lib/gcc/arm-linux-gnueabi/4.2.4/../../.. x.o -lgcc --as-needed -lgcc_s
--no-as-needed -lc -lgcc --as-needed -lgcc_s --no-as-needed
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/crtend.o
/usr/lib/gcc/arm-linux-gnueabi/4.2.4/../../../crtn.o

To be clear, prev-gcc/collect2 does exist, I'm not sure why it isn't using it
and why it isn't passing the right paths to crtbegin.o, etc (also in prev-gcc).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 15:56:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 15:56:53 -0500
Subject: [LLVMbugs] [Bug 4046] New: Use after free in
	test/Feature/embeddedmetadata.ll
Message-ID: <bug-4046-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4046

           Summary: Use after free in test/Feature/embeddedmetadata.ll
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


$ llvm-as < embeddedmetadata.ll | valgrind llvm-dis                             
==849== Memcheck, a memory error detector.                                      
==849== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.         
==849== Using LibVEX rev 1672M, a library for dynamic binary translation.       
==849== Copyright (C) 2004-2009, and GNU GPL'd, by OpenWorks LLP.               
==849== Using valgrind-3.5.0.SVN, a dynamic binary instrumentation framework.   
==849== Copyright (C) 2000-2009, and GNU GPL'd, by Julian Seward et al.         
==849== For more details, rerun with: -v                                        
==849==                                                                         
==849== Invalid read of size 4                                                  
==849==    at 0x43484A: llvm::MDNode::Profile(llvm::FoldingSetNodeID&) (in
llvm-objects/Release/bin/llvm-dis)                                 
==849==    by 0x4D710F:
llvm::FoldingSetImpl::FindNodeOrInsertPos(llvm::FoldingSetNodeID const&,
void*&) (in llvm-objects/Release/bin/llvm-dis)
==849==    by 0x436062: llvm::MDNode::get(llvm::Constant* const*, unsigned int)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x41173E: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x412C8F: llvm::BitcodeReader::ParseFunctionBody(llvm::Function*)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x416D09:
llvm::BitcodeReader::materializeFunction(llvm::Function*, std::string*) (in
llvm-objects/Release/bin/llvm-dis)         
==849==    by 0x40BFB4: llvm::BitcodeReader::materializeModule(std::string*)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418BD0: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
==849==  Address 0x5d70d20 is 96 bytes inside a block of size 112 free'd        
==849==    at 0x4C2614D: operator delete(void*) (vg_replace_malloc.c:313)       
==849==    by 0x40CA85:
llvm::BitcodeReaderValueList::ResolveConstantForwardRefs() (in
llvm-objects/Release/bin/llvm-dis)                      
==849==    by 0x4121E9: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x417BA4: llvm::BitcodeReader::ParseModule(std::string const&)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418798: llvm::BitcodeReader::ParseBitcode() (in
llvm-objects/Release/bin/llvm-dis)                                             
==849==    by 0x418B54: llvm::getBitcodeModuleProvider(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)               
==849==    by 0x418BBC: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
==849==                                                                         
==849== Invalid read of size 8                                                  
==849==    at 0x434860: llvm::MDNode::Profile(llvm::FoldingSetNodeID&) (in
llvm-objects/Release/bin/llvm-dis)                                  
==849==    by 0x4D710F:
llvm::FoldingSetImpl::FindNodeOrInsertPos(llvm::FoldingSetNodeID const&,
void*&) (in llvm-objects/Release/bin/llvm-dis)
==849==    by 0x436062: llvm::MDNode::get(llvm::Constant* const*, unsigned int)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x41173E: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x412C8F: llvm::BitcodeReader::ParseFunctionBody(llvm::Function*)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x416D09:
llvm::BitcodeReader::materializeFunction(llvm::Function*, std::string*) (in
llvm-objects/Release/bin/llvm-dis)         
==849==    by 0x40BFB4: llvm::BitcodeReader::materializeModule(std::string*)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418BD0: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
==849==  Address 0x5d70cc0 is 0 bytes inside a block of size 112 free'd         
==849==    at 0x4C2614D: operator delete(void*) (vg_replace_malloc.c:313)       
==849==    by 0x40CA85:
llvm::BitcodeReaderValueList::ResolveConstantForwardRefs() (in
llvm-objects/Release/bin/llvm-dis)                      
==849==    by 0x4121E9: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x417BA4: llvm::BitcodeReader::ParseModule(std::string const&)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418798: llvm::BitcodeReader::ParseBitcode() (in
llvm-objects/Release/bin/llvm-dis)                                             
==849==    by 0x418B54: llvm::getBitcodeModuleProvider(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)               
==849==    by 0x418BBC: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
==849==                                                                         
==849== Invalid read of size 8                                                  
==849==    at 0x4D711F:
llvm::FoldingSetImpl::FindNodeOrInsertPos(llvm::FoldingSetNodeID const&,
void*&) (in llvm-objects/Release/bin/llvm-dis)
==849==    by 0x436062: llvm::MDNode::get(llvm::Constant* const*, unsigned int)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x41173E: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x412C8F: llvm::BitcodeReader::ParseFunctionBody(llvm::Function*)
(in llvm-objects/Release/bin/llvm-dis)                         
==849==    by 0x416D09:
llvm::BitcodeReader::materializeFunction(llvm::Function*, std::string*) (in
llvm-objects/Release/bin/llvm-dis)         
==849==    by 0x40BFB4: llvm::BitcodeReader::materializeModule(std::string*)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418BD0: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
==849==  Address 0x5d70d28 is 104 bytes inside a block of size 112 free'd       
==849==    at 0x4C2614D: operator delete(void*) (vg_replace_malloc.c:313)       
==849==    by 0x40CA85:
llvm::BitcodeReaderValueList::ResolveConstantForwardRefs() (in
llvm-objects/Release/bin/llvm-dis)                      
==849==    by 0x4121E9: llvm::BitcodeReader::ParseConstants() (in
llvm-objects/Release/bin/llvm-dis)                                           
==849==    by 0x417BA4: llvm::BitcodeReader::ParseModule(std::string const&)
(in llvm-objects/Release/bin/llvm-dis)                            
==849==    by 0x418798: llvm::BitcodeReader::ParseBitcode() (in
llvm-objects/Release/bin/llvm-dis)                                             
==849==    by 0x418B54: llvm::getBitcodeModuleProvider(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)               
==849==    by 0x418BBC: llvm::ParseBitcodeFile(llvm::MemoryBuffer*,
std::string*) (in llvm-objects/Release/bin/llvm-dis)                       
==849==    by 0x40AFC9: main (in llvm-objects/Release/bin/llvm-dis)             
; ModuleID = '<stdin>'                                                          
@llvm.foo = internal constant { } !{i17 123, { } !"foobar"}             ; <{
}*> [#uses=0]

declare i8 @llvm.something({ })

define void @foo() {
        %x = call i8 @llvm.something({ } !{{ } !"f\00oa", i42 123})            
; <i8> [#uses=0]
        ret void
}
==849==
==849== ERROR SUMMARY: 4 errors from 3 contexts (suppressed: 2 from 1)
==849== malloc/free: in use at exit: 2,544 bytes in 38 blocks.
==849== malloc/free: 355 allocs, 317 frees, 70,162 bytes allocated.
==849== For counts of detected errors, rerun with: -v
==849== searching for pointers to 38 not-freed blocks.
==849== checked 208,360 bytes.
==849==
==849== LEAK SUMMARY:
==849==    definitely lost: 552 bytes in 7 blocks.
==849==    indirectly lost: 704 bytes in 10 blocks.
==849==      possibly lost: 0 bytes in 0 blocks.
==849==    still reachable: 1,288 bytes in 21 blocks.
==849==         suppressed: 0 bytes in 0 blocks.
==849== Rerun with --leak-check=full to see details of leaked memory.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 16:20:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 16:20:33 -0500
Subject: [LLVMbugs] [Bug 3989] C99 inline + extern decl merging problems
In-Reply-To: <bug-3989-206@http.llvm.org/bugs/>
Message-ID: <200904232120.n3NLKXP8021355@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3989


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dgregor at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Douglas Gregor <dgregor at apple.com>  2009-04-23 16:20:32 ---
Fixed here:

 
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/016146.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 22:44:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 22:44:06 -0500
Subject: [LLVMbugs] [Bug 2957] Change vector_shuffle into a variadic SDNode
In-Reply-To: <bug-2957-206@http.llvm.org/bugs/>
Message-ID: <200904240344.n3O3i6Le002014@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2957


Nate Begeman <natebegeman at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |natebegeman at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Nate Begeman <natebegeman at mac.com>  2009-04-23 22:44:04 ---
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076639.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 22:45:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 22:45:53 -0500
Subject: [LLVMbugs] [Bug 2554] assertion failure in
 X86TargetLowering::LowerVECTOR_SHUFFLE
In-Reply-To: <bug-2554-206@http.llvm.org/bugs/>
Message-ID: <200904240345.n3O3jrVF002103@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2554


Nate Begeman <natebegeman at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |natebegeman at mac.com
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Nate Begeman <natebegeman at mac.com>  2009-04-23 22:45:51 ---
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076639.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 23 23:35:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 23 Apr 2009 23:35:23 -0500
Subject: [LLVMbugs] [Bug 4047] New: permit configure --enable-targets=host,
 cpp for example
Message-ID: <bug-4047-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4047

           Summary: permit configure --enable-targets=host,cpp for example
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: autoconf
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


For configure --enable-targets, the three options are "all", "host-only" or a
comma-separated list of target names. I'd like to specify "host,cpp" but that's
not supported.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 04:57:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 04:57:12 -0500
Subject: [LLVMbugs] [Bug 4048] New: Mysterious crash in pass `Simple
 Register Coalescing'
Message-ID: <bug-4048-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4048

           Summary: Mysterious crash in pass `Simple Register Coalescing'
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=2863)
 --> (http://llvm.org/bugs/attachment.cgi?id=2863)
Reduced testcase

The attached code will cause Clang to crash. Because it seems to be related to
inline asm, I'm filing it under Backend: X86.

Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free -main-file-name llvm-crash2.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --fmath-errno=1
-O2 -fdiagnostics-show-option -o /tmp/cc-R9jTht.s -x c /home/ed/llvm-crash2.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Simple Register Coalescing' on function
'@ixgbe_set_vfta_82599'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 05:00:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 05:00:17 -0500
Subject: [LLVMbugs] [Bug 4049] New: Assertion failed: (NumParams ==
 getNumTypeParams(getType()) && "Parameter count mismatch!")
Message-ID: <bug-4049-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4049

           Summary: Assertion failed: (NumParams ==
                    getNumTypeParams(getType()) && "Parameter count
                    mismatch!")
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=2864)
 --> (http://llvm.org/bugs/attachment.cgi?id=2864)
Reduced testcase

The attached piece of code causes Clang to crash. The code is probably a mess,
but still it shouldn't crash.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 08:02:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 08:02:21 -0500
Subject: [LLVMbugs] [Bug 4004] invalid TLS with -relocation-model=pic
	-regalloc=local
In-Reply-To: <bug-4004-206@http.llvm.org/bugs/>
Message-ID: <200904241302.n3OD2L5V031267@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4004


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-04-24 08:02:03 ---
Fixed on revision 69968.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 10:22:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 10:22:50 -0500
Subject: [LLVMbugs] [Bug 4050] New: ICE: Node was deleted but visit returned
	new node!
Message-ID: <bug-4050-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4050

           Summary: ICE: Node was deleted but visit returned new node!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp148$ llvm-gcc -c -Wall -O2 small.c
cc1: DAGCombiner.cpp:675: void<unnamed>::DAGCombiner::Run(llvm::CombineLevel):
Assertion `N->getOpcode() != ISD::DELETED_NODE && RV.getNode()->getOpcode() !=
ISD::DELETED_NODE && "Node was deleted but visit returned new node!"' failed.
small.c:24: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp148$ cat small.c

struct S1
{
  short int f0;
  int f1;
};

struct S1 g_10 = {
  1, 1
};

void func_28 (long long int p_30, long long int p_32);

struct S1 func_26 (struct S1 p_27)
{
  func_28 (p_27.f1, g_10.f1);
  return g_10;
}

void func93 (struct S1);

void int322 (int foo)
{
  func93 (func_26 (g_10));
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 10:34:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 10:34:57 -0500
Subject: [LLVMbugs] [Bug 4051] New: ICE: Invalid instruction!
Message-ID: <bug-4051-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4051

           Summary: ICE: Invalid instruction!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp151$ llvm-gcc -O2 small.c -c
cc1:
/home/regehr/z/tmp/llvm-r69856/include/llvm/CodeGen/LiveIntervalAnalysis.h:209:
unsigned int llvm::LiveIntervals::getInstructionIndex(llvm::MachineInstr*)
const: Assertion `it != mi2iMap_.end() && "Invalid instruction!"' failed.
small.c:19: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp151$ cat small.c

int bar (int _left, int _right)
{
  return _right || 1 >> _right ? _left : 1;
}

int g_6;

int func_9 (char p_10)
{
  return 0;
}

int foo (int);

void int163 (int p_4, int p_5)
{
  (bar (p_5, foo (1)) && p_5 & 1) != g_6 && !func_9 (1);
  foo (p_5);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 10:44:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 10:44:26 -0500
Subject: [LLVMbugs] [Bug 4052] New: ICE: This is not a truncating conversion!
Message-ID: <bug-4052-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4052

           Summary: ICE: This is not a truncating conversion!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp152$ llvm-gcc -O2 -Wall -c small.c
cc1: ScalarEvolution.cpp:651: llvm::SCEVHandle
llvm::ScalarEvolution::getTruncateExpr(const llvm::SCEVHandle&, const
llvm::Type*): Assertion `getTypeSizeInBits(Op->getType()) >
getTypeSizeInBits(Ty) && "This is not a truncating conversion!"' failed.
small.c:22: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp152$ cat small.c

char safe_sub_func_int32_t_s_s (int _si1, char _si2)
{
  return ((_si1 ^ _si2) & _si1 << ((1 * 8) - 1 - _si2)) < 0;
}

int safe_sub_func_uint64_t_u_u (int _ui1, int _ui2)
{
  return _ui1 - 1;
}

volatile int x;

void int87 (char p_48, char p_49)
{
  unsigned l_52;
  for (; l_52; l_52 = safe_sub_func_int32_t_s_s (l_52, 1)) {
    x;
    for (l_52 = 0; l_52 <= 0; l_52 = safe_sub_func_uint64_t_u_u (l_52, 1))
      {
      }
  }
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 10:54:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 10:54:45 -0500
Subject: [LLVMbugs] [Bug 4053] New: integer wrong code bug
Message-ID: <bug-4053-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4053

           Summary: integer wrong code bug
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

1 is the right answer.

regehr at john-home:~/volatile/tmp154$ llvm-gcc -O0 small.c -o small -w
regehr at john-home:~/volatile/tmp154$ ./small
checksum = 1
regehr at john-home:~/volatile/tmp154$ llvm-gcc -O1 small.c -o small -w
regehr at john-home:~/volatile/tmp154$ ./small
checksum = 0

#include <stdint.h>

#define safe_add_macro_int64_t_s_s(si1,si2) \
                ((((((int64_t)(si1))>((int64_t)0)) &&
(((int64_t)(si2))>((int64_t)0)) && (((int64_t)(si1)) >
((INT64_MAX)-((int64_t)(si2))))) \
                  || ((((int64_t)(si1))<((int64_t)0)) &&
(((int64_t)(si2))<((int64_t)0)) && (((int64_t)(si1)) <
((INT64_MIN)-((int64_t)(si2)))))) \
                 ? ((int64_t)(si1)) \
                 : (((int64_t)(si1)) + ((int64_t)(si2))) \
                 ) 

static int64_t
safe_add_func_int64_t_s_s(int64_t _si1, int64_t _si2)
{
  return safe_add_macro_int64_t_s_s(_si1,_si2);
}

#define safe_mod_macro_uint16_t_u_u(ui1,ui2) \
        ((((uint16_t)(ui2)) == ((uint16_t)0)) \
                        ? ((uint16_t)(ui1)) \
                        : (((uint16_t)(ui1)) % ((uint16_t)(ui2))))

static uint16_t
safe_mod_func_uint16_t_u_u (uint16_t _ui1, uint16_t _ui2)
{
  return safe_mod_macro_uint16_t_u_u(_ui1,_ui2);
}

uint8_t g_8;

void func_2 (int16_t p_4);
void func_2 (int16_t p_4)
{
  for (p_4 = 0; (p_4 >= 0); p_4 = safe_add_func_int64_t_s_s (p_4, 1))
    {
      g_8 = safe_mod_func_uint16_t_u_u (1, 0);
    }
}

int main (void)
{
  func_2 (1);
  printf ("checksum = %d\n", g_8);
  return 0;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 10:57:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 10:57:04 -0500
Subject: [LLVMbugs] [Bug 4054] New: ICE: bitwidth too small
Message-ID: <bug-4054-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4054

           Summary: ICE: bitwidth too small
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp153$ llvm-gcc -O2 small.c -c -Wall
small.c: In function ???int80???:
small.c:15: warning: statement with no effect
small.c:15: warning: statement with no effect
cc1: Type.cpp:913: static const llvm::IntegerType*
llvm::IntegerType::get(unsigned int): Assertion `NumBits >= MIN_INT_BITS &&
"bitwidth too small"' failed.
small.c: At top level:
small.c:31: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp153$ cat small.c

typedef int uint32_t;

int safe (int);

int foo (int _ui1, int _ui2)
{
  return _ui1 - 1;
}

char g_141;

int int80 (char p_71)
{
  char l;
  for (l; 1; l)
    {
      if (1)
        {
          if (safe (1))
            return p_71;
          for (p_71 = 1; p_71; p_71 = foo (p_71, 1))
            {
            }
          if (g_141)
            p_71 = p_71 & g_141;
          else
            return 1;
        }
    }
  return 0;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 11:08:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 11:08:13 -0500
Subject: [LLVMbugs] [Bug 4055] New: ICE: Too many bits for uint64_t
Message-ID: <bug-4055-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4055

           Summary: ICE: Too many bits for uint64_t
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp149$ llvm-gcc -O2 small.c -c -w
cc1: /home/regehr/z/tmp/llvm-r69856/include/llvm/ADT/APInt.h:1049: uint64_t
llvm::APInt::getZExtValue() const: Assertion `getActiveBits() <= 64 && "Too
many bits for uint64_t"' failed.
small.c:64: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp149$ cat small.c

struct S0
{
  int f1;
  short f2;
  short f3;
};
struct S1
{
  int f0;
};

struct S2
{
  int f0;
};

struct S3
{
  struct S2 f0;
  struct S1 f1;
  short f2;
  struct S0 f3;
  struct S1 f4;
};

struct S6
{
  short f0;
};

struct S7
{
  struct S2 f0;
  struct S3 f1;
};

struct S9
{
  int f0;
  struct S7 f1;
};

struct S1 g_149 = {
  1
};

struct S6 func_120 (struct S9 p_123, struct S1 p_125)
{
  if (p_123.f1.f1.f3.f3)
    {
      struct S6 l_129 = {
        1
      };
      return l_129;
    }
}

struct S1 func (struct S0 p_144)
{
  struct S9 l_148 = {
    1L, 0L, 1L, 0L, -2L, -1L, 0L, 8L, 1
  };
  p_144.f1 = func_58 (func_120 (l_148, g_149));
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 11:14:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 11:14:58 -0500
Subject: [LLVMbugs] [Bug 4056] New: ICE: physreg was not killed in defining
	block!
Message-ID: <bug-4056-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4056

           Summary: ICE: physreg was not killed in defining block!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r69856 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp155$ llvm-gcc -O2 small.c 
cc1: LiveIntervalAnalysis.cpp:631: void
llvm::LiveIntervals::handlePhysicalRegisterDef(llvm::MachineBasicBlock*,
llvm::ilist_iterator<llvm::MachineInstr>, unsigned int, llvm::MachineOperand&,
llvm::LiveInterval&, llvm::MachineInstr*): Assertion `!CopyMI && "physreg was
not killed in defining block!"' failed.
small.c:23: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp155$ cat small.c

int foo (int _left, int _right)
{
  return _right >= 1 * 8 || 1 >> _right ? _left : _left << 1;
}

int g_6;

int func_9 (char p_10)
{
  return 0;
}

int bar (int);

void int163 (int p_4, int p_5)
{
  (foo (p_5, bar (1)) && p_5 & 1) != g_6
    && !func_9 (1);
  func_9 ((bar (p_5), 1));
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 12:43:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 12:43:03 -0500
Subject: [LLVMbugs] [Bug 4057] New: Assertion in SelectionDAG.cpp when
 casting float to char
Message-ID: <bug-4057-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4057

           Summary: Assertion in SelectionDAG.cpp when casting float to char
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2867)
 --> (http://llvm.org/bugs/attachment.cgi?id=2867)
bitcode of test case

The following test case asserts in SelectionDAG.cpp:getConstant:869 because the
value type being passed in is i8.

The test case is:
void test_cast_float_to_char(char* result)
{
        float test = (float)(47.8363037109375 * 100);
        *result = (char)(test);
}


This fails both x86 and my custom backend.

Get constant is called from selectionDag:2283
case ISD::FP_TO_UINT: {
        integerPart x;
        bool ignored;
        assert(integerPartWidth >= 64);
        // FIXME need to be more flexible about rounding mode.
        APFloat::opStatus s = V.convertToInteger(&x, 64U,
                              Opcode==ISD::FP_TO_SINT,
                              APFloat::rmTowardZero, &ignored);
        if (s==APFloat::opInvalidOp)     // inexact is OK, in fact usual
          break;
        return getConstant(x, VT);
      } 



        llc.exe!_NMSG_WRITE(int rterrnum=10)  Line 198  C
        llc.exe!abort()  Line 59 + 0x7 bytes    C
        llc.exe!_wassert(const wchar_t * expr=0x017ec4c8, const wchar_t *
filename=0x017eb8d8, unsigned int lineno=871)  Line 163       C
        llc.exe!llvm::SelectionDAG::getConstant(unsigned __int64 Val=4783,
llvm::MVT VT={...}, bool isT=false)  Line 871 + 0x69 bytes   C++
>	llc.exe!llvm::SelectionDAG::getNode(unsigned int Opcode=110, llvm::DebugLoc DL={...}, llvm::MVT VT={...}, llvm::SDValue Operand={...})  Line 2293 + 0x20 bytes	C++
        llc.exe!`anonymous
namespace'::DAGCombiner::visitFP_TO_SINT(llvm::SDNode * N=0x03a584f0)  Line
4231 + 0x2e bytes        C++
        llc.exe!`anonymous namespace'::DAGCombiner::visit(llvm::SDNode *
N=0x03a584f0)  Line 760 + 0x10 bytes   C++
        llc.exe!`anonymous namespace'::DAGCombiner::combine(llvm::SDNode *
N=0x03a584f0)  Line 784      C++
        llc.exe!`anonymous namespace'::DAGCombiner::Run(llvm::CombineLevel
AtLevel=Unrestricted)  Line 657      C++
        llc.exe!llvm::SelectionDAG::Combine(llvm::CombineLevel
Level=Unrestricted, llvm::AliasAnalysis & AA={...}, bool Fast=false)  Line 6119
+ 0x40 bytes     C++
        llc.exe!llvm::SelectionDAGISel::CodeGenAndEmitDAG()  Line 581   C++
        llc.exe!llvm::SelectionDAGISel::SelectBasicBlock(llvm::BasicBlock *
LLVMBB=0x029ada10, llvm::ilist_iterator<llvm::Instruction> Begin={...},
llvm::ilist_iterator<llvm::Instruction> End={...})  Line 499        C++
        llc.exe!llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function &
Fn={...}, llvm::MachineFunction & MF={...}, llvm::MachineModuleInfo *
MMI=0x03a5ecf0, llvm::DwarfWriter * DW=0x03a5bde0, const llvm::TargetInstrInfo
& TII={...})  Line 856       C++
        llc.exe!llvm::SelectionDAGISel::runOnFunction(llvm::Function &
Fn={...})  Line 332      C++
        llc.exe!llvm::FPPassManager::runOnFunction(llvm::Function & F={...}) 
Line 1323 + 0x17 bytes    C++
        llc.exe!llvm::FunctionPassManagerImpl::run(llvm::Function & F={...}) 
Line 1281 + 0x10 bytes    C++
        llc.exe!llvm::FunctionPassManager::run(llvm::Function & F={...})  Line
1226 + 0xf bytes C++
        llc.exe!main(int argc=4, char * * argv=0x029a22f0)  Line 319    C++
        llc.exe!__tmainCRTStartup()  Line 266 + 0x19 bytes      C
        llc.exe!mainCRTStartup()  Line 182      C


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 13:16:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 13:16:24 -0500
Subject: [LLVMbugs] [Bug 3989] C99 inline + extern decl merging problems
In-Reply-To: <bug-3989-206@http.llvm.org/bugs/>
Message-ID: <200904241816.n3OIGO1B011048@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3989


Matthieu castet <castet.matthieu at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #8 from Matthieu castet <castet.matthieu at free.fr>  2009-04-24 13:15:27 ---
There are still case broken, for example toto should be here not externaly
visible (gnu extern inline) :
$ cat test/toto2.c 
extern int toto(void);
extern inline __attribute__ ((__gnu_inline__)) int toto(void)
{
}

inline int foo()
{
}

int main()
{
        toto();
        return foo();
}

$ clang -ccc-host-triple i686-linux-gnu -O3 test/toto2.c -c -o /tmp/pp.o
-std=c99
$ readelf -s /tmp/pp.o | grep toto
     1: 00000000     0 FILE    LOCAL  DEFAULT  ABS test/toto2.c
     6: 00000000     5 FUNC    GLOBAL DEFAULT    1 toto

FYI :
$ gcc -O3 test/toto2.c -c -o  /tmp/pp.o -std=c99
$ readelf -s /tmp/pp.o  | grep toto
     1: 00000000     0 FILE    LOCAL  DEFAULT  ABS toto2.c


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 16:36:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 16:36:22 -0500
Subject: [LLVMbugs] [Bug 4057] Assertion in SelectionDAG.cpp when casting
	float to char
In-Reply-To: <bug-4057-206@http.llvm.org/bugs/>
Message-ID: <200904242136.n3OLaMRI018823@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4057


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dale Johannesen <dalej at apple.com>  2009-04-24 16:36:22 ---
Fixed.  It would probably be good to have APFloat::convertToInteger return an
APInt, but I'm not tackling that now.
http://llvm.org/viewvc/llvm-project?view=rev&revision=69990


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 17:49:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 17:49:38 -0500
Subject: [LLVMbugs] [Bug 4058] New: ARM: i64 arguments may be split between
 registers and stack with AAPCS ABI
Message-ID: <bug-4058-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4058

           Summary: ARM: i64 arguments may be split between registers and
                    stack with AAPCS ABI
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


The AAPCS ARM ABI requires that i64 arguments be passed in even/odd registers
pairs or on the stack with 8-byte alignment.  LLVM currently splits up i64
arguments into pairs of i32 values.  The ARMCallingConv.td file attempts to
handle this by checking for i32 types with 8-byte alignment, and it does manage
to get them into even/odd registers, but it fails to prevent them from being
split with one word in r3 and the other on the stack.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Apr 24 19:43:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 24 Apr 2009 19:43:29 -0500
Subject: [LLVMbugs] [Bug 4059] New: ARM: double arguments may be split
 between registers and stack with AAPCS ABI
Message-ID: <bug-4059-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4059

           Summary: ARM: double arguments may be split between registers and
                    stack with AAPCS ABI
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2871)
 --> (http://llvm.org/bugs/attachment.cgi?id=2871)
testcase

This is similar to pr4058, which involves ia64 arguments.  Double
floating-point arguments (f64) work correctly if the target architecture
includes the VFP floating-point registers.  Otherwise, f64 arguments get split
up into i32/i32 pairs and the tests for f64 in ARMCallingConv.td fail.  The
base AAPCS ABI requires f64 arguments to be passed in even/odd register pairs
or on the stack, and the ABI applies whether you have VFP support or not. 
(There is an optional hard-float variant of AAPCS but LLVM does not support
that yet.)  f64 arguments cannot be split between r3 and the stack when using
AAPCS.  The attached testcase demonstrates the problem.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 01:12:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 01:12:54 -0500
Subject: [LLVMbugs] [Bug 4049] Assertion failed: (NumParams ==
 getNumTypeParams(getType()) && " Parameter count mismatch!")
In-Reply-To: <bug-4049-206@http.llvm.org/bugs/>
Message-ID: <200904250612.n3P6Cst6001051@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4049


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-25 01:12:53 ---
Fixed here, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090420/016233.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 05:29:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 05:29:16 -0500
Subject: [LLVMbugs] [Bug 4060] New: building with g++-4.4 causes ocaml test
	to fail
Message-ID: <bug-4060-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4060

           Summary: building with g++-4.4  causes ocaml test to fail
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I get this when building llvm with g++-4.4:
FAIL: /home/edwin/llvm-svn/llvm/test/Bindings/Ocaml/executionengine.ml
Failed with signal(SIGABRT) at line 2
while running: ./executionengine.ml.tmp executionengine.ml.tmp.bc
While deleting: i32 (i32, i32)* %plus
An asserting value handle still pointed to this value!


I don't see this failure on the llvm buildbots, so either this is  specific to
g++-4.4, or the buildbots don't have ocaml installed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 05:44:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 05:44:38 -0500
Subject: [LLVMbugs] [Bug 4061] New: clang breaks strict-aliasing rules
 according to g++4.4
Message-ID: <bug-4061-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4061

           Summary: clang breaks strict-aliasing rules according to g++4.4
           Product: new-bugs
           Version: unspecified
          Platform: PC
               URL: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39891#c2
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


/home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h: In member function ???bool
clang::Expr::isIntegerConstantExpr(llvm::APSInt&, clang::ASTContext&,
clang::SourceLocation*, bool) const???:
/home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h:99: warning: dereferencing
pointer ???<anonymous>??? does break strict-aliasing rules
/home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h:548: warning: dereferencing
pointer ???<anonymous>??? does break strict-aliasing rules
/home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h:549: warning: dereferencing
pointer ???<anonymous>??? does break strict-aliasing rules
/home/edwin/llvm-svn/llvm/include/llvm/ADT/APSInt.h:38: warning: dereferencing
pointer ???<anonymous>??? does break strict-aliasing rules
Expr.cpp:1334: note: initialized from here

The location info for the warning is not accurate (it doesn't show the very
line causing the violation), but the aliasing violation is real:
  return *(APSInt*)(void*)Data;

See URL for a possible solution.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 12:06:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 12:06:19 -0500
Subject: [LLVMbugs] [Bug 4054] ScalarEvolution::createSCEV tries to create
	type i0
In-Reply-To: <bug-4054-206@http.llvm.org/bugs/>
Message-ID: <200904251706.n3PH6JsN032083@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4054


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dan Gohman <gohman at apple.com>  2009-04-25 12:06:19 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076686.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 12:44:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 12:44:17 -0500
Subject: [LLVMbugs] [Bug 4062] New: RFE: Support "-Wp,
 -MD" option (req. for Linux kernel build)
Message-ID: <bug-4062-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4062

           Summary: RFE: Support "-Wp,-MD" option (req. for Linux kernel
                    build)
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jgarzik at pobox.com
                CC: llvmbugs at cs.uiuc.edu


The Linux kernel uses "-Wp,-MD..." for dependencies.  Example command line,
when attempting to use clang to build kernel:

  clang -Wp,-MD,kernel/.bounds.s.d  -nostdinc -isystem include -Iinclude 
-I/spare/repo/linux-2.6/arch/x86/include -include include/linux/autoconf.h
-D__KERNEL__ -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs
-fno-strict-aliasing -fno-common -Werror-implicit-function-declaration -O2 -m64
-march=core2 -mno-red-zone -mcmodel=kernel -funit-at-a-time
-maccumulate-outgoing-args -DCONFIG_AS_CFI=1 -DCONFIG_AS_CFI_SIGNAL_FRAME=1
-pipe -Wno-sign-compare -fno-asynchronous-unwind-tables -mno-sse -mno-mmx
-mno-sse2 -mno-3dnow -Wframe-larger-than=2048 -fno-stack-protector
-fno-omit-frame-pointer -fno-optimize-sibling-calls -pg
-Wdeclaration-after-statement -Wno-pointer-sign -fwrapv -fno-dwarf2-cfi-asm  
-D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(bounds)" 
-D"KBUILD_MODNAME=KBUILD_STR(bounds)"  -fverbose-asm -S -o kernel/bounds.s
kernel/bounds.c
clang: warning: the clang compiler does not yet support '-pg'
clang: warning: argument unused during compilation: '-march=core2'
clang: warning: argument unused during compilation: '-mcmodel=kernel'
clang: warning: argument unused during compilation: '-funit-at-a-time'
clang: warning: argument unused during compilation:
'-maccumulate-outgoing-args'
clang: warning: argument unused during compilation:
'-fno-asynchronous-unwind-tables'
clang: warning: argument unused during compilation: '-fno-omit-frame-pointer'
clang: warning: argument unused during compilation:
'-fno-optimize-sibling-calls'
clang: warning: argument unused during compilation: '-fwrapv'
clang: warning: argument unused during compilation: '-fno-dwarf2-cfi-asm'
clang-cc: Unknown command line argument '-MD'.  Try:
'/usr/local/bin/../libexec/clang-cc --help'
make[1]: *** [kernel/bounds.s] Error 1
make: *** [prepare0] Error 2


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 13:15:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 13:15:17 -0500
Subject: [LLVMbugs] [Bug 4063] New: clang -MD -MF doesn't create dependency
 file for . S source files
Message-ID: <bug-4063-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4063

           Summary: clang -MD -MF doesn't create dependency file for .S
                    source files
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Works fine for .c files, but not for .S files:

clang -MD -MF arch/x86/kernel/.entry_64.o.d  -nostdinc -isystem include
-Iinclude  -I/home/edwin/builds/linux-2.6/arch/x86/include -include
include/linux/autoconf.h -D__KERNEL__ -D__ASSEMBLY__ -m64 -DCONFIG_AS_CFI=1
-DCONFIG_AS_CFI_SIGNAL_FRAME=1 -Iarch/x86/include/asm/mach-default     -c -o
arch/x86/kernel/entry_64.o arch/x86/kernel/entry_64.S
fixdep: arch/x86/kernel/.entry_64.o.d: No such file or directory


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 13:34:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 13:34:30 -0500
Subject: [LLVMbugs] [Bug 4040] APInt's string constructor is too strict
In-Reply-To: <bug-4040-206@http.llvm.org/bugs/>
Message-ID: <200904251834.n3PIYUST002473@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4040


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-25 13:34:30 ---
Applied, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076692.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 13:48:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 13:48:43 -0500
Subject: [LLVMbugs] [Bug 4066] New: clang: Broken module found: Aliasee
 should be either GlobalValue or bitcast of GlobalValue
Message-ID: <bug-4066-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4066

           Summary: clang: Broken module found: Aliasee should be either
                    GlobalValue or bitcast of GlobalValue
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2878)
 --> (http://llvm.org/bugs/attachment.cgi?id=2878)
x.i

clang -MD -MF drivers/hwmon/.w83l785ts.o.d  -nostdinc -isystem include
-Iinclude  -I/home/edwin/builds/linux-2.6/arch/x86/include -include
include/linux/autoconf.h -D__KERNEL__ -Wall -Wundef -Wstrict-prototypes
-Wno-trigraphs -fno-strict-aliasing -fno-common
-Werror-implicit-function-declaration -O2 -m64 -march=core2 -mno-red-zone
-mcmodel=kernel -funit-at-a-time -maccumulate-outgoing-args -DCONFIG_AS_CFI=1
-DCONFIG_AS_CFI_SIGNAL_FRAME=1 -pipe -Wno-sign-compare
-fno-asynchronous-unwind-tables -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
-Iarch/x86/include/asm/mach-default -Wframe-larger-than=2048
-fno-stack-protector -fno-omit-frame-pointer -fno-optimize-sibling-calls
-Wdeclaration-after-statement -Wno-pointer-sign -fwrapv -fno-dwarf2-cfi-asm
-DMODULE -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(w83l785ts)" 
-D"KBUILD_MODNAME=KBUILD_STR(w83l785ts)"  -c  drivers/hwmon/w83l785ts.c
-emit-llvm 

Trying to get a reduced testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 13:58:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 13:58:51 -0500
Subject: [LLVMbugs] [Bug 4068] New: [META] Compiling the Linux kernel with
	clang
Message-ID: <bug-4068-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4068

           Summary: [META] Compiling the Linux kernel with clang
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


This is a meta-bug, other bugs that prevent building the Linux kernel with
clang will be made to depend on it.

We should keep this bug open, until someone boots a kernel built with clang :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 14:10:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 14:10:21 -0500
Subject: [LLVMbugs] [Bug 4070] New: error recovery improvement for stmtexprs
Message-ID: <bug-4070-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4070

           Summary: error recovery improvement for stmtexprs
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


In this example:

struct fooo {} ;
struct fooo x;

void test(void) {

  *({
    0;
    x/x;
    });
}

We emit two diagnostics: one about x/x (good) and one about "indirection
requires a pointer operand".  The problem is that we correctly kill the
expression tree for x/x, but that means that we build and return a  compound
statement with just { 0; } and the stmt expr gets built with type int.  We
should somehow propagate invalidity of the last statement in a stmt expr up
into the stmt expr itself.

This came from PR4065.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 16:06:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 16:06:52 -0500
Subject: [LLVMbugs] [Bug 3947] SmallVector::insert overload ambiguous when
	T=int
In-Reply-To: <bug-3947-206@http.llvm.org/bugs/>
Message-ID: <200904252106.n3PL6qPc008613@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3947


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-25 16:06:52 ---
I don't think the pain to fix this is worth the benefit, you can just use other
APIs to avoid this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 16:11:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 16:11:52 -0500
Subject: [LLVMbugs] [Bug 3924] -load-vn and -gcse still mentioned by
	AliasAnalysis.html
In-Reply-To: <bug-3924-206@http.llvm.org/bugs/>
Message-ID: <200904252111.n3PLBqrW008840@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3924


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-04-25 16:11:52 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076697.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 16:26:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 16:26:30 -0500
Subject: [LLVMbugs] [Bug 4066] clang: Broken module found: Aliasee should be
 either GlobalValue or bitcast of GlobalValue
In-Reply-To: <bug-4066-206@http.llvm.org/bugs/>
Message-ID: <200904252126.n3PLQUAr009338@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4066


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|major                       |normal
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-25 16:26:29 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076698.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076699.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 17:29:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 17:29:15 -0500
Subject: [LLVMbugs] [Bug 3412] -licm -loop-reduce crash
In-Reply-To: <bug-3412-206@http.llvm.org/bugs/>
Message-ID: <200904252229.n3PMTFa4011777@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3412


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-25 17:29:15 ---
Ah, this works with mainline.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 18:00:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 18:00:23 -0500
Subject: [LLVMbugs] [Bug 4075] New: `opt -view-cfg` can't handle function
 names with a slash
Message-ID: <bug-4075-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4075

           Summary: `opt -view-cfg` can't handle function names with a slash
           Product: tools
           Version: 2.4
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: opt
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: vext01 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


When invoking `opt -view-cfg` error messages will be printed for function names
containing slashes. For example:

Writing '/tmp/llvm_JMwD7i/cfgInteger::oper/Integer[1].dot'... error opening
file for writing!

I am assuming this is because a forward slash is the directory delimiter on
UNIX file-systems.

Thanks


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 19:54:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 19:54:02 -0500
Subject: [LLVMbugs] [Bug 4076] New: bug in inline asm constraint
Message-ID: <bug-4076-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4076

           Summary: bug in inline asm constraint
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=2886)
 --> (http://llvm.org/bugs/attachment.cgi?id=2886)
.ll testcase

The attached .ll file demonstrates a problem where LLVM is picking %eax to
serve in a ptr computation even though it's already reserved as another input
to the inline asm.

$ llvm-as < b.ll | llc -disable-fp-elim
[...]
.LBB1_16:       # 
        movl    %esi, %edx
        xorl    %eax, %eax
        movl    20(%ebp), %eax
        #APP
        lock; cmpxchgl %edx,(%eax,%edi,8)
        #NO_APP

Note that cmpxchgl takes %eax as input and output, and it needs to be zero as
per the xorl instruction. The original inline asm statement in C looks like:

inline int32_t NoBarrier_CompareAndSwap(volatile int32_t* ptr,
                                        int32_t old_value,
                                        int32_t new_value) {
   int32_t prev; 
   __asm__ __volatile__("lock; cmpxchgl %1,%2"
                        : "=a" (prev)
                        : "q" (new_value), "m" (*ptr), "0" (old_value)
                        : "memory");
  return prev;
}

which compiles fine as a standalone function. The problem only surfaces with
inlining.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Apr 25 21:11:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 25 Apr 2009 21:11:03 -0500
Subject: [LLVMbugs] [Bug 4062] RFE: Support "-Wp,
 -MD" option (req. for Linux kernel build)
In-Reply-To: <bug-4062-206@http.llvm.org/bugs/>
Message-ID: <200904260211.n3Q2B3B7018890@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4062


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-04-25 21:11:03 ---
Using -Wp is a hack for doing this. Is there any reason the kernel can't use
-MD directly?

I don't want the driver to need to get in the business of ripping apart the
-Wp, argument and reinterpreting it, unless there is a very compelling reason.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 00:27:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 00:27:05 -0500
Subject: [LLVMbugs] [Bug 4062] RFE: Support "-Wp,
 -MD" option (req. for Linux kernel build)
In-Reply-To: <bug-4062-206@http.llvm.org/bugs/>
Message-ID: <200904260527.n3Q5R5fs024547@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4062


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|WONTFIX                     |




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 02:44:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 02:44:32 -0500
Subject: [LLVMbugs] [Bug 3594] clang: ccc -fPIC/-DPIC not working on x86-64
In-Reply-To: <bug-3594-206@http.llvm.org/bugs/>
Message-ID: <200904260744.n3Q7iVnp002495@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3594


T??r??k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from T??r??k Edwin <edwintorok at gmail.com>  2009-04-26 02:44:31 ---
Yes, its fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 12:44:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 12:44:49 -0500
Subject: [LLVMbugs] [Bug 4081] New: RFE: data component of initialized
 global variables should come last
Message-ID: <bug-4081-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4081

           Summary: RFE: data component of initialized global variables
                    should come last
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM assembly language parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jgarzik at pobox.com
                CC: llvmbugs at cs.uiuc.edu


In the Linux kernel, we have a few cases where a global variable in C stores
data at compile time in a _huge_ char[] array.  An example would be where a
device firmware is compiled into the device driver, rather than being shipped
as a separate file.

As such, LLVM assembly becomes inconvent with large initialized globals,
because the 'align' attribute currently follows the initialized data:

@myarray = global [6 x i32] [i32 10, i32 20, i32 30, i32 40, i32 50, i32 69],
align 4

Large initializers can easily push the 'align 4' far, far away from the
"@myvariable" definition.  This makes life difficult for some code generators,
and also makes human review of LLVM assembly more difficult.

My suggestion is to reorder variable definitions such that initialized data
always comes last (implying movement of current 'align').


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 16:03:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 16:03:05 -0500
Subject: [LLVMbugs] [Bug 4082] New: bizarre array causes crash in target
	folder
Message-ID: <bug-4082-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4082

           Summary: bizarre array causes crash in target folder
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2898)
 --> (http://llvm.org/bugs/attachment.cgi?id=2898)
testcase (y.c, actually c++)

This crashes in llvm-g++ on 64-bit Linux but not 32-bit Linux.

$ llvm-g++ y.c
cc1plus: Constants.cpp:2239: static llvm::Constant*
llvm::ConstantExpr::getGetElementPtr(llvm::Constant*, llvm::Value* const*,
unsigned int): Assertion `Ty && "GEP indices invalid!"' failed.
y.c:30: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

(gdb) bt
#0  0x00007f1847412095 in raise () from /lib/libc.so.6
#1  0x00007f1847413af0 in abort () from /lib/libc.so.6
#2  0x00007f184740b2df in __assert_fail () from /lib/libc.so.6
#3  0x0000000000ff4978 in llvm::ConstantExpr::getGetElementPtr (C=0x19abdc8, 
    Idxs=0x19b1be0, NumIdx=2) at Constants.cpp:2239
#4  0x00000000008a1e64 in TreeConstantToLLVM::EmitLV_ARRAY_REF (exp=0x19aa8a0)
    at
/usr/local/google/home/nlewycky/llvm/include/llvm/Support/TargetFolder.h:118
#5  0x00000000008a15a5 in TreeConstantToLLVM::Convert (exp=0x7f184731c780)
    at ../../src/gcc/llvm-convert.cpp:6421
#6  0x00000000008a36b0 in TreeConstantToLLVM::ConvertNOP_EXPR (
    exp=0x7f184731c7c0) at ../../src/gcc/llvm-convert.cpp:6583
#7  0x00000000008a28bb in TreeConstantToLLVM::ConvertArrayCONSTRUCTOR (
    exp=0x7f18473314b0) at ../../src/gcc/llvm-convert.cpp:6688
#8  0x0000000000890b52 in emit_global_to_llvm (decl=0x7f1847326c00)
    at ../../src/gcc/llvm-backend.cpp:1263
#9  0x000000000085f237 in assemble_variable () at varasm.c:1781
#10 0x00000000009102bb in cgraph_varpool_assemble_decl () at cgraphunit.c:940
#11 0x00000000009104e6 in cgraph_varpool_assemble_pending_decls ()
    at cgraphunit.c:958
#12 0x0000000000911fd1 in cgraph_optimize () at cgraphunit.c:1620
#13 0x00000000004aa9d6 in cp_finish_file () at decl2.c:3343
#14 0x0000000000415d99 in finish_file () at cp-lang.c:150
---Type <return> to continue, or q <return> to quit---
#15 0x000000000056f125 in c_common_parse_file () at c-opts.c:1305
#16 0x00000000008328a4 in toplev_main () at toplev.c:2287
#17 0x000000000057dbce in main (argc=2, argv=0x7fff5071ff28)
    at ../../src/gcc/llvm-main.cpp:38
(gdb) up 3
#3  0x0000000000ff4978 in llvm::ConstantExpr::getGetElementPtr (C=0x19abdc8, 
    Idxs=0x19b1be0, NumIdx=2) at Constants.cpp:2239
2239      assert(Ty && "GEP indices invalid!");
(gdb) p Ty
$1 = (const llvm::Type *) 0x0
(gdb) p C
$2 = (class llvm::Constant *) 0x19abdc8
(gdb) p C->dump()
@_ZN17_Foo_NameInfoInit6by_posE = global %0 zeroinitializer, align 4          
; <%0*> [#uses=1]

$3 = void
(gdb) up
#4  0x00000000008a1e64 in TreeConstantToLLVM::EmitLV_ARRAY_REF (exp=0x19aa8a0)
    at
/usr/local/google/home/nlewycky/llvm/include/llvm/Support/TargetFolder.h:118
118         return Fold(ConstantExpr::getGetElementPtr(C, IdxList, NumIdx));
(gdb) p IdxList
No symbol "IdxList" in current context.
(gdb) p NumIdx
No symbol "NumIdx" in current context.
(gdb) up
#5  0x00000000008a15a5 in TreeConstantToLLVM::Convert (exp=0x7f184731c780)
    at ../../src/gcc/llvm-convert.cpp:6421
6421                                        ConvertType(TREE_TYPE(exp)));


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 17:40:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 17:40:46 -0500
Subject: [LLVMbugs] [Bug 4083] New: Using Clang on iPhone
Message-ID: <bug-4083-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4083

           Summary: Using Clang on iPhone
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: Mac System 9.x
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Documentation
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chadsspameateremail at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


When trying to use Clang on a project built for the iPhone (on Mac OS X) I
found that the instructions didn't provide the correct result.  Instead I
created alternate instructions which I have attached below.  Feel free to use
this to update your documentation for people building for the iPhone.

-----

DOWNLOADING
Step 1: Download the Mac OS X version of clang from this website:
http://clang.llvm.org/StaticAnalysis.html

Step 2: Double click on the download to unarchive it.

INSTALLING
Step 3: Place the checker-xxx directory where you want to install Clang.  For
this example I'll place it in the Documents directory which would be
~/Documents.  The folder name is checker-0.189 in my example.

Step 4: Open Terminal and add Clang to your path by typing:
echo 'export PATH=~/Documents/checker-0.189/:$PATH' >> ~/.bash_profile

Replace ~/Documents and checker-0.189 as appropriate.

Step 5: Close the current terminal window or open a new terminal window.  This
must be done for the changes in the path to take effect.

RUNNING
Step 6: At this point you may need to set the code signing identity of the
project you want to scan.  This must be setup correctly or Clang will not be
able to build your project correctly.  In XCode you can set the code signing
identity to yourself by going to "Project->Edit Project Settings" in XCode and
then under the build tab finding "Code Signing Identity" and selecting
yourself.  This is the same step which is necessary to build on the device. 
Note that Clang will often try and build the release version of the application
so be sure the "Code Signing Identity" is also set correctly for the Release
"Configuration" you can select your "Configuration" by using the
"Configuration" pop-up menu.

Step 7: In the "Edit Project Settings" window above under the build tab you
should also set the Base SDK to Simulator-iPhone OS 2.2.1.  This will allow
Clang to better analyze the project.  You should not be building for the device
here or you will see warnings on the terminal that the arm processor is not
recognized by Clang and it can't find the appropriate version of GCC.

Step 7: In Terminal Navigate to the directory where the project you want to
test is located.  Make sure you are in the directory with the project.  In my
example the iCS folder is located on the desktop so I type:
cd ~/Desktop/iCS

Step 8: Once we are in the proper directory where the xcode project file is
located we can run the scan-build program which launches Clang.  It keys off of
the current path in terminal and builds the files/project within that
directory.  You should not specify a project name.  Though you can specify and
output directory using the -o option as I do below.  Once run successfully
Clang will create a series of HTML files in the output directory you specified.
 In my example I type:

scan-build -V -o ~/Desktop/ClangOutput xcodebuild

After the program finishes it automatically launches a web browser displaying
the Clang results.


NOTES
Each time you run Clang will build the project and during the building finds
the errors it reports.  If you run Clang twice in a row the second time it will
generate no output because no files need to be built.  If you want Clang to
retest the entire project you will need to make a change that makes Clang
rebuild the entire project.  One example would be to modify the iCS_Prefix.pch
file by adding some whitespace.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 19:16:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 19:16:40 -0500
Subject: [LLVMbugs] [Bug 4084] New: fix-it diagnostics on multiple lines
	need fixin'
Message-ID: <bug-4084-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4084

           Summary: fix-it diagnostics on multiple lines need fixin'
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


This is confusing and weird:
--
ddunbar at frank:tmp$ cat t.c
int f0(char *a, char *b) {
  return ("a" != 
          "b");
}
ddunbar at frank:tmp$ clang -Wall -fsyntax-only /tmp/t.c
/tmp/t.c:2:15: warning: result of comparison against a string literal is
unspecified (use strcmp instead)
  return ("a" != 
          ~~~ ^~
          strcmp(
1 diagnostic generated.
--

Also, the strcmp is in a somewhat funny place.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 20:05:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 20:05:37 -0500
Subject: [LLVMbugs] [Bug 3482] clang miscompiles
	MultiSource/Benchmarks/VersaBench/bmm
In-Reply-To: <bug-3482-206@http.llvm.org/bugs/>
Message-ID: <200904270105.n3R15b95001081@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3482


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-04-26 20:05:36 ---
Actually, it looks like my gcc is buggy; gcc's -O0 and -O3 output differ.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 20:06:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 20:06:39 -0500
Subject: [LLVMbugs] [Bug 3483] clang miscompiles
	MultiSource/Benchmarks/VersaBench/dbms
In-Reply-To: <bug-3483-206@http.llvm.org/bugs/>
Message-ID: <200904270106.n3R16dkv001114@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3483


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2009-04-26 20:06:39 ---
This works if I compile at -O0, pretty sure I just have a buggy gcc as with
3482.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 20:41:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 20:41:51 -0500
Subject: [LLVMbugs] [Bug 4052] ICE: This is not a truncating conversion!
In-Reply-To: <bug-4052-206@http.llvm.org/bugs/>
Message-ID: <200904270141.n3R1fpt6002411@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4052


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-04-26 20:41:49 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076738.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 21:01:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 21:01:22 -0500
Subject: [LLVMbugs] [Bug 4050] ICE: Node was deleted but visit returned new
	node!
In-Reply-To: <bug-4050-206@http.llvm.org/bugs/>
Message-ID: <200904270201.n3R21Mwm003021@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4050


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-04-26 21:01:22 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090420/076739.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 21:38:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 21:38:02 -0500
Subject: [LLVMbugs] [Bug 4085] New: CppWriter::progname is not initialized
Message-ID: <bug-4085-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4085

           Summary: CppWriter::progname is not initialized
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: abbeyj at gmail.com
                CC: llvmbugs at cs.uiuc.edu


CppWriter::progname is never initialized but it is used in CppWriter::error(). 
This leads to a crash if error() is ever called.

Steps to Reproduce:
$ cat temp.ll
; x86_fp80 not supported by CppWriter
@x = common global x86_fp80 0xK00000000000000000000

$ llvm-as temp.ll
$ llc -march=cpp temp.bc


Actual Results:
Crash inside iostream

Expected Results:
Print "llc: Invalid primitive type"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Apr 26 22:48:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 26 Apr 2009 22:48:50 -0500
Subject: [LLVMbugs] [Bug 4086] New: float induction variable elimination
	issue
Message-ID: <bug-4086-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4086

           Summary: float induction variable elimination issue
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: major
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: markleone at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I think this code tickles a bug in float induction variable elimination.  

declare void @foo()

define void @test() {
entry:
        br label %loop_body

loop_body:              
        %i = phi float [ %nexti, %loop_body ], [ 0.0, %entry ]          
        tail call void @foo()
        %nexti = add float %i, 1.0              
        %less = fcmp olt float %nexti, 2.0              
        br i1 %less, label %loop_body, label %done

done:           
        ret void
}

To replicate: llvm-as test.ll ; opt -indvars test.bc | llvm-dis 
The resulting code has a bogus less-than comparison, leading to an infinite
loop.

define void @test() {
entry:
        br label %loop_body

loop_body:              
        %i.int = phi i32 [ 0, %entry ], [ %indvar.next, %loop_body ]            
        tail call void @foo()
        %less1 = icmp ult i32 0, 2                      ; XXX should be %i.int,
not 0.
        %indvar.next = add i32 %i.int, 1                
        br i1 %less1, label %loop_body, label %done

done:           
        ret void
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 03:05:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 03:05:54 -0500
Subject: [LLVMbugs] [Bug 4087] New: Headers for libffi not found in
 lib/ExecutionEngine/ Interpreter
Message-ID: <bug-4087-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4087

           Summary: Headers for libffi not found in
                    lib/ExecutionEngine/Interpreter
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: magnus.reftel at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I have libffi installed in a non-standard location (and set CFLAGS to let gcc
search there before running configure), but ExternalFunctions.cpp in
lib/ExecutionEngine/Interpreter does not find ffi.h:

make[2]: Entering directory `/c/temp/llvm-svn/lib/ExecutionEngine'
llvm[2]: Compiling ExecutionEngine.cpp for Release build 
llvm[2]: Compiling ExecutionEngineBindings.cpp for Release build 
llvm[2]: Linking Release Object Library LLVMExecutionEngine.o
make[3]: Entering directory `/c/temp/llvm-svn/lib/ExecutionEngine/Interpreter'
llvm[3]: Compiling Execution.cpp for Release build 
llvm[3]: Compiling ExternalFunctions.cpp for Release build 
ExternalFunctions.cpp:38:17: ffi.h: No such file or directory

my CFLAGS is

$ echo $CFLAGS 
-I/opt/llvm-svn/lib/libffi-3.0.8/include/

and the contents of that directory are:

$ ls /opt/llvm-svn/lib/libffi-3.0.8/include/
ffi.h  ffitarget.h

The CFLAGS was picked up by configure according to config.status:

$ ./config.status  --version
llvm config.status 2.6svn
configured by ./configure, generated by GNU Autoconf 2.60,
  with options "'--prefix=/opt/llvm-svn' '--enable-optimized' 'CC=gcc-4'
'CFLAGS=-I/opt/llvm-svn/lib/libffi-3.0.8/include/'
'LDFLAGS=-L/opt/llvm-svn/lib/'"

Copyright (C) 2006 Free Software Foundation, Inc.
This config.status script is free software; the Free Software Foundation
gives unlimited permission to copy, distribute and modify it.

I did not have this problem with llvm-2.5. The revision I'm at now is 70196:

$ svn info
Path: .
URL: http://llvm.org/svn/llvm-project/llvm/trunk
Repository Root: http://llvm.org/svn/llvm-project
Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8
Revision: 70196
Node Kind: directory
Schedule: normal
Last Changed Author: nicholas
Last Changed Rev: 70185
Last Changed Date: 2009-04-27 07:09:44 +0200 (Mon, 27 Apr 2009)

I'm running cygwin on Windows Vista.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 03:15:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 03:15:41 -0500
Subject: [LLVMbugs] [Bug 4087] Headers for libffi not found in
 lib/ExecutionEngine/Interpreter
In-Reply-To: <bug-4087-206@http.llvm.org/bugs/>
Message-ID: <200904270815.n3R8FfRu025651@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4087


Magnus Reftel <magnus.reftel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Magnus Reftel <magnus.reftel at gmail.com>  2009-04-27 03:15:41 ---
C++ compilation uses CXXFLAGS, not CFLAGS. THe problem was solved by running

export CXXFLAGS=$CFLAGS

and then make


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 03:17:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 03:17:39 -0500
Subject: [LLVMbugs] [Bug 4087] Headers for libffi not found in
 lib/ExecutionEngine/Interpreter
In-Reply-To: <bug-4087-206@http.llvm.org/bugs/>
Message-ID: <200904270817.n3R8HdRE025723@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4087


Magnus Reftel <magnus.reftel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
          Component|Makefiles                   |autoconf
         Resolution|INVALID                     |




--- Comment #2 from Magnus Reftel <magnus.reftel at gmail.com>  2009-04-27 03:17:39 ---
There is actually a bug here, though: the configure script should verify that
C++ code that uses libffi can be built.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 07:02:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 07:02:24 -0500
Subject: [LLVMbugs] [Bug 4089] New: LLVM asm has no way to specify alignment
 in structure types
Message-ID: <bug-4089-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4089

           Summary: LLVM asm has no way to specify alignment in structure
                    types
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM assembly language parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jgarzik at pobox.com
                CC: llvmbugs at cs.uiuc.edu


When writing LLVM assembly language (i.e. NOT using the LLVM libraries), there
is no way to specify alignment of a structure type for implementation of the C
feature __attribute__((__aligned__(nnn))).

The current workaround is for my asm writer to insert i8's as padding.  This
works, but doing so defeats a key attraction of the structure type:  struct
layout is completely offloaded onto LLVM.

To support __attribute__((__aligned__(nnn))), my asm writer has two options,
listed in order of increasing complexity:

(1) Perform full target ABI layout internally, and unconditionally use packed
structure type

(2) Perform full target ABI layout internally, use normal structure type where
possible.  Guess where LLVM will insert padding, and insert additional padding
as needed.

What I hope for is a much more simple third option...

(3) Perform NO target ABI layout internally, and specify alignment in LLVM
assembly language.  Something like this:

       %struct.foo = type <{i8, i32/align8, i32}>


I feel this is consistent with the rest of LLVM assembly language, which
permits specification of custom alignment in multiple places.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 11:02:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 11:02:49 -0500
Subject: [LLVMbugs] [Bug 4053] integer wrong code bug
In-Reply-To: <bug-4053-206@http.llvm.org/bugs/>
Message-ID: <200904271602.n3RG2nRl009190@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4053


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #10 from Duncan Sands <baldrick at free.fr>  2009-04-27 11:02:37 ---
This seems to have been fixed at some point.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 12:17:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 12:17:02 -0500
Subject: [LLVMbugs] [Bug 4090] New: Assertion failed: Re-defining a live
 register!", lib/  CodeGen/RegisterScavenging.cpp, line 272
Message-ID: <bug-4090-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4090

           Summary: Assertion failed: Re-defining a live register!",
                    lib/CodeGen/RegisterScavenging.cpp, line 272
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=2899)
 --> (http://llvm.org/bugs/attachment.cgi?id=2899)
Preprocessed C file

When compiling the attached source file for ARM, I'm seeing the following
failed assertion:

Assertion failed: ((isReserved(Reg) || isUnused(Reg) || IsImpDef ||
isImplicitlyDefined(Reg) || isLiveInButUnusedBefore(Reg, MI, MBB, TRI, MRI)) &&
"Re-defining a live register!"), function forward, file
/home/ed/projects/freebsd-clang/usr.bin/clang/lib/libllvmcodegen/../../../../contrib/llvm/lib/CodeGen/RegisterScavenging.cpp,
line 272.
Stack dump:
0.      Program arguments:
/usr/obj/arm/home/ed/projects/freebsd-clang/tmp/usr/bin/../libexec/clang-cc
-triple arm-undermydesk-freebsd8.0 -S -disable-free -main-file-name main.c
--relocation-model static --disable-fp-elim --unwind-tables=0 --fmath-errno=1
-DPACKAGE=sntp -DVERSION=1.6 -DSYS_FREEBSD -DPARSE -DHAVE_CONFIG_H -DOPENSSL
-I/home/ed/projects/freebsd-clang/usr.sbin/ntp/sntp/../../../contrib/ntp/include
-I/home/ed/projects/freebsd-clang/usr.sbin/ntp/sntp/../
-I/home/ed/projects/freebsd-clang/usr.sbin/ntp/sntp/../../../contrib/ntp/libopts
-O1 -std=gnu99 -fdiagnostics-show-option -o /tmp/cc-uz7idY.s -x c
/home/ed/projects/freebsd-clang/usr.sbin/ntp/sntp/../../../contrib/ntp/sntp/main.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Prolog/Epilog Insertion & Frame Finalization' on function
'@correct_drift'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 12:31:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 12:31:57 -0500
Subject: [LLVMbugs] [Bug 4091] New: ARM: Inline asm failure!
Message-ID: <bug-4091-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4091

           Summary: ARM: Inline asm failure!
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


The following C code:

void
foo(void)
{
        unsigned int i, *p;

        asm volatile("swp %0, %2, [%3]" : "=&r" (i), "=m" (*p) : "r" (i), "r"
(p), "m" (*p) : "memory");
}

or the following LLVM IR:

; ModuleID = 'arm-inline-asm.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:64:64"
target triple = "arm-undermydesk-freebsd8.0"

define void @foo() nounwind {
entry: 
        %0 = tail call i32 asm sideeffect "swp $0, $2, [$3]",
"=&r,=*m,r,r,*m,~{memory}"(i32* undef, i32 undef, i32* undef, i32* undef)
nounwind                ; <i32> [#uses=0]
        ret void
}

generates the following error message:

Could not match memory address.  Inline asm failure!

No line numbers, etc. It just dies there.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 12:33:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 12:33:16 -0500
Subject: [LLVMbugs] [Bug 4092] New: spurious diagnostic following duplicate
 field definition
Message-ID: <bug-4092-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4092

           Summary: spurious diagnostic following duplicate field definition
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


The "incomplete result type" error below should be avoided:
--
ddunbar at 67-218-106-103:tmp$ cat t.c
struct s0 {
  char a;
  char a;
};

struct s0 f0(void) {}
ddunbar at 67-218-106-103:tmp$ clang-cc t.c
t.c:3:8: error: duplicate member 'a'
  char a;
       ^
t.c:2:8: note: previous declaration is here
  char a;
       ^
t.c:6:11: error: incomplete result type 'struct s0' in function definition
struct s0 f0(void) {}
          ^
3 diagnostics generated.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 12:58:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 12:58:36 -0500
Subject: [LLVMbugs] [Bug 3863] clang should support ARM ABI
In-Reply-To: <bug-3863-206@http.llvm.org/bugs/>
Message-ID: <200904271758.n3RHwaDU014004@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3863


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Daniel Dunbar <daniel at zuster.org>  2009-04-27 12:58:35 ---
Ok. I'm closing this, my guess is things work "ok" enough that specific bug
reports are in order.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 15:43:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 15:43:13 -0500
Subject: [LLVMbugs] [Bug 4076] bug in inline asm constraint
In-Reply-To: <bug-4076-206@http.llvm.org/bugs/>
Message-ID: <200904272043.n3RKhDvp020154@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4076


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Evan Cheng <evan.cheng at apple.com>  2009-04-27 15:43:12 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076778.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 15:44:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 15:44:48 -0500
Subject: [LLVMbugs] [Bug 4056] ICE: physreg was not killed in defining block!
In-Reply-To: <bug-4056-206@http.llvm.org/bugs/>
Message-ID: <200904272044.n3RKimFt020217@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4056


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2009-04-27 15:44:46 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076758.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 16:03:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 16:03:56 -0500
Subject: [LLVMbugs] [Bug 4086] float induction variable elimination issue
In-Reply-To: <bug-4086-206@http.llvm.org/bugs/>
Message-ID: <200904272103.n3RL3uCj021046@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4086


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dalej at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dale Johannesen <dalej at apple.com>  2009-04-27 16:03:56 ---
Fixed here.
http://llvm.org/viewvc/llvm-project?view=rev&revision=70247


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 19:10:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 19:10:33 -0500
Subject: [LLVMbugs] [Bug 4093] New: LegalizeDAG does not allow expansion of
 subtraction operation
Message-ID: <bug-4093-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4093

           Summary: LegalizeDAG does not allow expansion of subtraction
                    operation
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


I ran into a problem where setOperationAction(ISD::SUB, MVT::i64, Expand) was
giving an assertion.

I wrote this code to expand subtraction into multiple instructions.

It goes at line 3352 of LegalizeDAG.cpp.

if (Node->getOpcode() == ISD::SUB &&
        TLI.isOperationLegalOrCustom(ISD::ADD, VT)) {
                if (VT.isFloatingPoint() &&
                        TLI.isOperationLegalOrCustom(ISD::MUL, VT)) {
                SDValue t1 = SDValue(DAG.getNode(ISD::MUL, dl,
                                                 VT, Tmp2,
DAG.getConstantFP(-1, VT)
                                                 ).getNode(), 0);
                Result = SDValue(DAG.getNode(ISD::ADD, dl, VT, Tmp1, 
                                                 t1).getNode(), 0);
                break;
                } else if (VT.isInteger() &&
                        TLI.isOperationLegalOrCustom(ISD::XOR, VT)) {           
                SDValue t1 = SDValue(DAG.getNode(ISD::XOR, dl,
                                                 VT, Tmp2,DAG.getConstant(
                                                 (uint64_t)-1, VT)
                                                 ).getNode(), 0);               
                SDValue t2 = SDValue(DAG.getNode(ISD::ADD, dl,
                                                 VT, t1, DAG.getConstant(1, VT)
                                                 ).getNode(), 0);               
                Result = SDValue(DAG.getNode(ISD::ADD, dl, VT, Tmp1, 
                                                 t2).getNode(), 0);
                break;
                }
          }


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Apr 27 21:12:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 27 Apr 2009 21:12:58 -0500
Subject: [LLVMbugs] [Bug 4051] ICE: Invalid instruction!
In-Reply-To: <bug-4051-206@http.llvm.org/bugs/>
Message-ID: <200904280212.n3S2CwpF031648@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4051


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2009-04-27 21:12:57 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076793.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 01:27:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 01:27:35 -0500
Subject: [LLVMbugs] [Bug 4034] Coalescer crash
In-Reply-To: <bug-4034-206@http.llvm.org/bugs/>
Message-ID: <200904280627.n3S6RZww006989@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4034


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-04-28 01:27:35 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076798.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 01:39:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 01:39:32 -0500
Subject: [LLVMbugs] [Bug 3264] Functions attributes are only in the last
	declaration.
In-Reply-To: <bug-3264-206@http.llvm.org/bugs/>
Message-ID: <200904280639.n3S6dWSM007445@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3264


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Douglas Gregor <dgregor at apple.com>  2009-04-28 01:39:32 ---
We now keep attributes on all declarations as they were written in the source.
The fix is here:

 
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090427/016401.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 01:40:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 01:40:33 -0500
Subject: [LLVMbugs] [Bug 3989] C99 inline + extern decl merging problems
In-Reply-To: <bug-3989-206@http.llvm.org/bugs/>
Message-ID: <200904280640.n3S6eXlD007542@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3989


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #9 from Douglas Gregor <dgregor at apple.com>  2009-04-28 01:40:32 ---
I believe this is fixed, here:

 
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090427/016401.html

However, this is a very tricky area. Please feel free to re-open the bug if
problems remain!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 13:13:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 13:13:37 -0500
Subject: [LLVMbugs] [Bug 4089] LLVM asm has no way to specify alignment in
 structure types
In-Reply-To: <bug-4089-206@http.llvm.org/bugs/>
Message-ID: <200904281813.n3SIDbrK010607@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4089


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-28 13:13:36 ---
This is actually by design.  LLVM supports two classes of languages:

1. Languages which do not have ABI requirements like scheme, java, etc.  These
languages can generate platform neutral LLVM IR and want to know as little of
the target as possible.  These languages should use the normal LLVM struct
type.

2. Languages like C that have specific ABIs to match.  For these, we recommend
using llvm packed structure types exclusively.  LLVM requires the front-end to
do target-specific lowering and other things (bitfield layout etc).  Since the
front-end already has to do a lot of target specific stuff, adding this feature
wouldn't make them substantially more complex.

Closing this as "not to be fixed" because I don't think that the added
complexity of supporting this is worth the benefit.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 13:22:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 13:22:29 -0500
Subject: [LLVMbugs] [Bug 4095] New: transparent_union modeled wrong
Message-ID: <bug-4095-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4095

           Summary: transparent_union modeled wrong
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


It looks like we handle transparent union arguments by claiming a function
takes a void*, which leads to sema problems.  Here's a simple testcase:

struct d_object_s {
    int x;
};
typedef union {
    struct d_object_s *_do;
} d_object_t __attribute__((transparent_union));
void test(d_object_t _head) {
  d_object_t x;
  x;
  struct d_object_s *head = _head._do;
}

---

We reject the "head =" line with:
t.c:13:35: error: member reference base type 'd_object_t' (aka 'void *const')
is not a structure or union
  struct d_object_s *head = _head._do;
                            ~~~~~ ^

Looking at the -ast-dump, it looks like dispatch_object_t is just a typedef for
void* in the AST (!):

void _dispatch_queue_push_list(dispatch_object_t _head)
(CompoundStmt 0x26037d0 <t.c:10:1, line:14:1>
  (DeclStmt 0x26040e0 <line:11:3, col:22>
    0x2604040 "dispatch_object_t x"
  (DeclRefExpr 0x2604100 <line:12:3> 'dispatch_object_t':'void *' Var='x'
0x2604040)
  (DeclStmt 0x2604190 <line:13:3, col:41>
    0x2604120 "struct dispatch_object_s *head =
      (ImplicitCastExpr 0x2604170 <col:36> 'struct dispatch_object_s *'
        (DeclRefExpr 0x2604150 <col:36> 'dispatch_object_t':'void *'
ParmVar='_head' 0x2603fd0))")


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 13:54:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 13:54:25 -0500
Subject: [LLVMbugs] [Bug 4095] transparent_union modeled wrong
In-Reply-To: <bug-4095-206@http.llvm.org/bugs/>
Message-ID: <200904281854.n3SIsPei012153@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4095


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-28 13:54:24 ---
Yes, the current transparent_union implementation is an ugly hack.

*** This bug has been marked as a duplicate of bug 2015 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 13:55:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 13:55:57 -0500
Subject: [LLVMbugs] [Bug 3957] sync builtins should be overloaded
In-Reply-To: <bug-3957-206@http.llvm.org/bugs/>
Message-ID: <200904281855.n3SItvV7012244@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3957


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-28 13:55:57 ---


*** This bug has been marked as a duplicate of bug 3824 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 15:44:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 15:44:06 -0500
Subject: [LLVMbugs] [Bug 4100] New: assertion failed,
	"Re-defining a live register!"
Message-ID: <bug-4100-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4100

           Summary: assertion failed, "Re-defining a live register!"
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


The attached testcase fails:

$ llc b.bc -fast -regalloc=local -march=arm

llc: RegisterScavenging.cpp:272: void llvm::RegScavenger::forward(): Assertion
`(isReserved(Reg) || isUnused(Reg) || IsImpDef || isImplicitlyDefined(Reg) ||
isLiveInButUnusedBefore(Reg, MI, MBB, TRI, MRI)) && "Re-defining a live
register!"' failed.
(gdb) bt
#0  0x00007f211049d095 in raise () from /lib/libc.so.6
#1  0x00007f211049eaf0 in abort () from /lib/libc.so.6
#2  0x00007f21104962df in __assert_fail () from /lib/libc.so.6
#3  0x0000000000f40019 in llvm::RegScavenger::forward (this=0x182dd50)
    at RegisterScavenging.cpp:269
#4  0x0000000000a66e01 in llvm::RegScavenger::forward (this=0x182dd50, I=
      {<bidirectional_iterator<llvm::MachineInstr,long int>> =
{<std::iterator<std::bidirectional_iterator_tag,llvm::MachineInstr,long
int,llvm::MachineInstr*,llvm::MachineInstr&>> = {<No data fields>}, <No data
fields>}, NodePtr = 0x1828d80})
    at
/usr/local/google/home/nlewycky/llvm/include/llvm/CodeGen/RegisterScavenging.h:102
#5  0x0000000000f0e0d9 in replaceFrameIndices (this=0x181bf70, Fn=@0x1828660)
    at PrologEpilogInserter.cpp:1448
#6  0x0000000000f124a3 in runOnMachineFunction (this=0x181bf70, Fn=@0x1828660)
    at PrologEpilogInserter.cpp:143
#7  0x0000000000ee97db in llvm::MachineFunctionPass::runOnFunction (
    this=0x181bf70, F=@0x180ace0) at MachineFunction.cpp:47
#8  0x00000000010e77b2 in llvm::FPPassManager::runOnFunction (this=0x18103f0, 
    F=@0x180ace0) at PassManager.cpp:1338
#9  0x00000000010e7d27 in llvm::FunctionPassManagerImpl::run (this=0x180dbe0, 
    F=@0x180ace0) at PassManager.cpp:1291
#10 0x00000000010e7ece in llvm::FunctionPassManager::run (this=0x7fff197acab0, 
    F=@0x180ace0) at PassManager.cpp:1233
#11 0x0000000000820080 in main (argc=5, argv=0x7fff197accc8) at llc.cpp:322
(gdb) up 3
#3  0x0000000000f40019 in llvm::RegScavenger::forward (this=0x182dd50)
    at RegisterScavenging.cpp:269
269         assert((isReserved(Reg) || isUnused(Reg) ||
(gdb) p Reg
$1 = 26


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 16:52:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 16:52:34 -0500
Subject: [LLVMbugs] [Bug 3379] Invalid operand found in inline asm: 'movl
	%gs:${1:P}, $0'
In-Reply-To: <bug-3379-206@http.llvm.org/bugs/>
Message-ID: <200904282152.n3SLqYDd018416@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3379


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Anton Korobeynikov <asl at math.spbu.ru>  2009-04-28 16:52:33 ---
This should fix it, I think:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076811.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Apr 28 23:30:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 28 Apr 2009 23:30:19 -0500
Subject: [LLVMbugs] [Bug 4102] New: Write after free in debug info writer
	with fast isel
Message-ID: <bug-4102-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4102

           Summary: Write after free in debug info writer with fast isel
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2911)
 --> (http://llvm.org/bugs/attachment.cgi?id=2911)
Testcase (unreduced bitcode)

Testcase attached; to reproduce, run the following:
valgrind llc x.bc -o x.s -fast

First error:
==7399== Invalid write of size 4
==7399==    at 0x8AC41C2:
llvm::DwarfDebug::getOrCreateScope(llvm::GlobalVariable*)
(DwarfWriter.cpp:2034)
==7399==    by 0x8AC44CD:
llvm::DwarfDebug::RecordRegionStart(llvm::GlobalVariable*)
(DwarfWriter.cpp:3471)
==7399==    by 0x8AAD37A:
llvm::DwarfWriter::RecordRegionStart(llvm::GlobalVariable*)
(DwarfWriter.cpp:4756)
==7399==    by 0x898311C: llvm::FastISel::SelectCall(llvm::User*)
(FastISel.cpp:349)
==7399==    by 0x8984A70: llvm::FastISel::SelectOperator(llvm::User*, unsigned)
(FastISel.cpp:721)
==7399==    by 0x8984CAE: llvm::FastISel::SelectInstruction(llvm::Instruction*)
(FastISel.cpp:626)
==7399==    by 0x8A08A61:
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function&,
llvm::MachineFunction&, llvm::MachineModuleInfo*, llvm::DwarfWriter*,
llvm::TargetInstrInfo const&) (SelectionDAGISel.cpp:810)
==7399==    by 0x8A09B88:
llvm::SelectionDAGISel::runOnFunction(llvm::Function&)
(SelectionDAGISel.cpp:331)
==7399==    by 0x8D35F71: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1338)
==7399==    by 0x8D36A5F: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1291)
==7399==    by 0x8D36C30: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1233)
==7399==    by 0x8419F89: main (llc.cpp:322)
==7399==  Address 0x4b40fa4 is 68 bytes inside a block of size 512 free'd
==7399==    at 0x40246EA: operator delete(void*) (vg_replace_malloc.c:342)
==7399==    by 0x8ABA5DE: llvm::DenseMap<llvm::GlobalVariable*,
llvm::DbgScope*, llvm::DenseMapInfo<llvm::GlobalVariable*>,
llvm::DenseMapInfo<llvm::DbgScope*> >::grow(unsigned) (DenseMap.h:469)
==7399==    by 0x8ABA64D: llvm::DenseMap<llvm::GlobalVariable*,
llvm::DbgScope*, llvm::DenseMapInfo<llvm::GlobalVariable*>,
llvm::DenseMapInfo<llvm::DbgScope*> >::InsertIntoBucket(llvm::GlobalVariable*
const&, llvm::DbgScope* const&, std::pair<llvm::GlobalVariable*,
llvm::DbgScope*>*) (DenseMap.h:351)
==7399==    by 0x8ABA73D: llvm::DenseMap<llvm::GlobalVariable*,
llvm::DbgScope*, llvm::DenseMapInfo<llvm::GlobalVariable*>,
llvm::DenseMapInfo<llvm::DbgScope*> >::FindAndConstruct(llvm::GlobalVariable*
const&) (DenseMap.h:282)
==7399==    by 0x8ABA76D: llvm::DenseMap<llvm::GlobalVariable*,
llvm::DbgScope*, llvm::DenseMapInfo<llvm::GlobalVariable*>,
llvm::DenseMapInfo<llvm::DbgScope*> >::operator[](llvm::GlobalVariable* const&)
(DenseMap.h:286)
==7399==    by 0x8AC40F4:
llvm::DwarfDebug::getOrCreateScope(llvm::GlobalVariable*)
(DwarfWriter.cpp:2024)
==7399==    by 0x8AC417F:
llvm::DwarfDebug::getOrCreateScope(llvm::GlobalVariable*)
(DwarfWriter.cpp:2032)
==7399==    by 0x8AC44CD:
llvm::DwarfDebug::RecordRegionStart(llvm::GlobalVariable*)
(DwarfWriter.cpp:3471)
==7399==    by 0x8AAD37A:
llvm::DwarfWriter::RecordRegionStart(llvm::GlobalVariable*)
(DwarfWriter.cpp:4756)
==7399==    by 0x898311C: llvm::FastISel::SelectCall(llvm::User*)
(FastISel.cpp:349)
==7399==    by 0x8984A70: llvm::FastISel::SelectOperator(llvm::User*, unsigned)
(FastISel.cpp:721)
==7399==    by 0x8984CAE: llvm::FastISel::SelectInstruction(llvm::Instruction*)
(FastISel.cpp:626)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 00:12:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 00:12:40 -0500
Subject: [LLVMbugs] [Bug 4092] spurious diagnostic following duplicate field
	definition
In-Reply-To: <bug-4092-206@http.llvm.org/bugs/>
Message-ID: <200904290512.n3T5CerP032053@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4092


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-29 00:12:39 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090427/016454.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 01:48:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 01:48:06 -0500
Subject: [LLVMbugs] [Bug 4082] bizarre array causes crash in target folder
In-Reply-To: <bug-4082-206@http.llvm.org/bugs/>
Message-ID: <200904290648.n3T6m6Bf002613@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4082


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-04-29 01:48:03 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076829.html

Testcase here:
llvm/test/FrontendC/2009-04-28-UnionArrayCrash.c


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 15:23:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 15:23:48 -0500
Subject: [LLVMbugs] [Bug 2425] merge of function declarations doesn't ignore
 __declspec()
In-Reply-To: <bug-2425-206@http.llvm.org/bugs/>
Message-ID: <200904292023.n3TKNm4e010027@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2425


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #4 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-29 15:23:47 ---
This bug doesn't make sense as filed; we don't support declspec in any
meaningful way at the moment.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 19:14:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 19:14:54 -0500
Subject: [LLVMbugs] [Bug 4108] New: can't codegen compound literal init
Message-ID: <bug-4108-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4108

           Summary: can't codegen compound literal init
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


We output "t.c:3:16: error: cannot compile this static initializer yet"

struct x { int a, b; };
const struct x Init = (struct x){0, 0};

Just a random thing I noticed


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 19:24:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 19:24:54 -0500
Subject: [LLVMbugs] [Bug 4085] CppWriter::progname is not initialized
In-Reply-To: <bug-4085-206@http.llvm.org/bugs/>
Message-ID: <200904300024.n3U0Os5B018407@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4085


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-04-29 19:24:53 ---
Fixed, thanks
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076866.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 20:42:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 20:42:16 -0500
Subject: [LLVMbugs] [Bug 4109] New: libLTO searches for "gcc" by path
	instead of sysroot
Message-ID: <bug-4109-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4109

           Summary: libLTO searches for "gcc" by path instead of sysroot
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: lto
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: rafael.espindola at gmail.com, llvmbugs at cs.uiuc.edu


I build in a clean environment with no environment variables defined, not even
PATH. This leads to failed LTO builds because libLTO expects to be able to find
the system GCC by searching the current path. Excerpt from
tools/lto/LTOCodeGenerator.cpp:216:

    // find compiler driver
    const sys::Path gcc = sys::Program::FindProgramByName("gcc");

Perhaps we could add an #ifdef SYSROOT path to this logic to mirror llvm-gcc?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Apr 29 21:50:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 29 Apr 2009 21:50:47 -0500
Subject: [LLVMbugs] [Bug 4110] New: codegen crash on i128 in 32-bit mode
Message-ID: <bug-4110-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4110

           Summary: codegen crash on i128 in 32-bit mode
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


llc crashes on this:
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i386-apple-darwin9.0"
@numer = external global i64            ; <i64*> [#uses=1]
@denom = external global i64            ; <i64*> [#uses=1]

define i64 @test(i64 %val) nounwind readonly {
entry:
        %conv = zext i64 %val to i128           ; <i128> [#uses=1]
        %tmp3 = load i64* @numer, align 8               ; <i64> [#uses=1]
        %conv4 = zext i64 %tmp3 to i128         ; <i128> [#uses=1]
        %mul = mul i128 %conv4, %conv           ; <i128> [#uses=1]
        %tmp6 = load i64* @denom, align 8               ; <i64> [#uses=1]
        %conv7 = zext i64 %tmp6 to i128         ; <i128> [#uses=1]
        %div = sdiv i128 %mul, %conv7           ; <i128> [#uses=1]
        %conv9 = trunc i128 %div to i64         ; <i64> [#uses=1]
        ret i64 %conv9
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 01:10:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 01:10:16 -0500
Subject: [LLVMbugs] [Bug 4111] New: poor error recovery for incorrect
	typename
Message-ID: <bug-4111-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4111

           Summary: poor error recovery for incorrect typename
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


This C or C++ program:
void f(sqrgl);

produces:
t.c:1:8: error: a parameter list without types is only allowed in a function
definition
void f(sqrgl);
       ^
t.c:1:6: error: variable has incomplete type 'void'
void f(sqrgl);
     ^
in C mode, and:
t.c:1:6: error: variable has incomplete type 'void'
void f(sqrgl);
     ^
t.c:1:8: error: use of undeclared identifier 'sqrgl'
void f(sqrgl);
       ^

in C++ mode.  These are both pretty bad, each should only produce one
diagnostic.  In practice, I saw this with "void foo(mytype_t);" which looks to
a human like it's a type :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 01:31:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 01:31:36 -0500
Subject: [LLVMbugs] [Bug 4055] TargetLowering::SimplifySetCC fails on large
 integer types
In-Reply-To: <bug-4055-206@http.llvm.org/bugs/>
Message-ID: <200904300631.n3U6Va76031081@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4055


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-30 01:31:35 ---
I fixed this in r70360 without realizing it:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090427/076822.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 01:32:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 01:32:34 -0500
Subject: [LLVMbugs] [Bug 4048] Mysterious crash in pass `Simple Register
	Coalescing'
In-Reply-To: <bug-4048-206@http.llvm.org/bugs/>
Message-ID: <200904300632.n3U6WYrl031149@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4048


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-04-30 01:32:33 ---
I think this was fixed along the way, it works for me.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 02:04:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 02:04:33 -0500
Subject: [LLVMbugs] [Bug 4108] can't codegen compound literal init
In-Reply-To: <bug-4108-206@http.llvm.org/bugs/>
Message-ID: <200904300704.n3U74XeR032208@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4108


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-04-30 02:04:32 ---
I knew exactly what the issue was here when I saw the testcase :)

Fixed in r70483.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 02:59:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 02:59:55 -0500
Subject: [LLVMbugs] [Bug 4110] codegen crash on i128 in 32-bit mode
In-Reply-To: <bug-4110-206@http.llvm.org/bugs/>
Message-ID: <200904300759.n3U7xteT012264@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4110


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2009-04-30 02:59:54 ---


*** This bug has been marked as a duplicate of bug 2660 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 12:46:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 12:46:41 -0500
Subject: [LLVMbugs] [Bug 4112] New: TargetData missing in Scalar Evolution
 when called from ModulePass
Message-ID: <bug-4112-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4112

           Summary: TargetData missing in Scalar Evolution when called from
                    ModulePass
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2913)
 --> (http://llvm.org/bugs/attachment.cgi?id=2913)
foo.cpp

See attached testcase:

opt: foo.cpp:28: virtual bool<unnamed>::Foo::runOnModule(llvm::Module&):
Assertion `SE.isSCEVable(IntegerType::get(32)->getPointerTo()) && "TargetData
missing"' failed.
0   opt             0x0000000000a7162c
1   opt             0x0000000000a71b0d
2   libpthread.so.0 0x00007f1a50a097b0
3   libc.so.6       0x00007f1a4fd29115 gsignal + 53
4   libc.so.6       0x00007f1a4fd2a633 abort + 387
5   libc.so.6       0x00007f1a4fd22159 __assert_fail + 233
6   LLVMModel.so    0x00007f1a4fad50fb
7   opt             0x00000000009facac
llvm::MPPassManager::runOnModule(llvm::Module&) + 206
8   opt             0x00000000009fc631
llvm::PassManagerImpl::run(llvm::Module&) + 111
9   opt             0x00000000009fc693 llvm::PassManager::run(llvm::Module&) +
33
10  opt             0x000000000070099e main + 3909
11  libc.so.6       0x00007f1a4fd155a6 __libc_start_main + 230
12  opt             0x00000000006f0fe9
Stack dump:
0.      Running pass 'test foo' on module 'x.bc'.

TargetData should be found by getAnalysisIfAvailable, because I explicitly
required it.

I tried moving the SCEV to a functionpass and requiring that functionpass from
the modulepass, but that leads to another bug (will open another bugreport).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 12:52:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 12:52:38 -0500
Subject: [LLVMbugs] [Bug 4113] New: ModulePass requires FunctionPass that
 requires TargetData: Assertion `0 && "ERROR: Bad TargetData ctor used. " "
 Tool did not specify a TargetData to use?"' failed.
Message-ID: <bug-4113-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4113

           Summary: ModulePass requires FunctionPass that requires
                    TargetData: Assertion `0 && "ERROR: Bad TargetData ctor
                    used.  " "Tool did not specify a TargetData to use?"'
                    failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2914)
 --> (http://llvm.org/bugs/attachment.cgi?id=2914)
foo2.cpp

See attached testcase:

x.bc is any module that has a targetdata.

When running functionpass only all is good, and targetdata  is found,
Scalarevolution has it too:
$ opt -load=../../../Debug/lib/Foo.so -bar x.bc -disable-output

When running modulepass it complains targetdata is not available:
$ opt -load=../../../Debug/lib/Foo.so -foobar x.bc -disable-output

opt: /home/edwin/llvm-svn/llvm/include/llvm/Target/TargetData.h:114:
llvm::TargetData::TargetData(): Assertion `0 && "ERROR: Bad TargetData ctor
used.  " "Tool did not specify a TargetData to use?"' failed.
0   opt             0x0000000000a7162c
1   opt             0x0000000000a71b0d
2   libpthread.so.0 0x00007fa3f6ca67b0
3   libc.so.6       0x00007fa3f5fc6115 gsignal + 53
4   libc.so.6       0x00007fa3f5fc7633 abort + 387
5   libc.so.6       0x00007fa3f5fbf159 __assert_fail + 233
6   opt             0x000000000095e64a llvm::Pass*
llvm::callDefaultCtor<llvm::TargetData>() + 0
7   opt             0x000000000095e668 llvm::Pass*
llvm::callDefaultCtor<llvm::TargetData>() + 30
8   opt             0x0000000000a00cf6 llvm::PassInfo::createPass() const + 128
9   opt             0x00000000009fa20a
llvm::PMTopLevelManager::schedulePass(llvm::Pass*) + 280
10  opt             0x0000000000a05948
llvm::FunctionPassManagerImpl::add(llvm::Pass*) + 36
11  opt             0x00000000009fa48a
llvm::MPPassManager::addLowerLevelRequiredPass(llvm::Pass*, llvm::Pass*) + 370
12  opt             0x00000000009fb36d llvm::PMDataManager::add(llvm::Pass*,
bool) + 845
13  opt             0x00000000009fb7a6
llvm::ModulePass::assignPassManager(llvm::PMStack&, llvm::PassManagerType) +
132
14  opt             0x0000000000a05803
llvm::PassManagerImpl::addTopLevelPass(llvm::Pass*) + 227
15  opt             0x00000000009fa2e6
llvm::PMTopLevelManager::schedulePass(llvm::Pass*) + 500
16  opt             0x0000000000a05922 llvm::PassManagerImpl::add(llvm::Pass*)
+ 36
17  opt             0x00000000009fa30f llvm::PassManager::add(llvm::Pass*) + 33
18  opt             0x00000000006feb07
19  opt             0x0000000000700613 main + 3002
20  libc.so.6       0x00007fa3f5fb25a6 __libc_start_main + 230
21  opt             0x00000000006f0fe9
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 16:02:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 16:02:08 -0500
Subject: [LLVMbugs] [Bug 4109] libLTO searches for "gcc" by path instead of
	sysroot
In-Reply-To: <bug-4109-206@http.llvm.org/bugs/>
Message-ID: <200904302102.n3UL28FC006299@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4109


Nick Lewycky <nlewycky at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 17:41:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 17:41:38 -0500
Subject: [LLVMbugs] [Bug 3948] clang misparses parameter with same name as
 containing class
In-Reply-To: <bug-3948-206@http.llvm.org/bugs/>
Message-ID: <200904302241.n3UMfcWD009436@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3948


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Anders Carlsson <andersca at mac.com>  2009-04-30 17:41:38 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090427/016547.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Apr 30 19:09:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 30 Apr 2009 19:09:10 -0500
Subject: [LLVMbugs] [Bug 4114] New: lli brings in tons of useless
	dependencies
Message-ID: <bug-4114-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4114

           Summary: lli brings in tons of useless dependencies
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: lli
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


lli now depends on LLVMAsmPrinter, LLVMScalarOpts, LLVMTransformUtils, LLVMipa
and LLVMAnalysis. None of those should be in there (except possibly
LLVMAnalysis, not sure about that one).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.