[Bug target/43703] Unexpected floating point precision loss due to ARM NEON autovectorization
--- Comment #5 from jules at gcc dot gnu dot org 2010-06-16 11:41 --- I am working on this. -- jules at gcc dot gnu dot org changed: What|Removed |Added AssignedTo|unassigned at gcc dot gnu |jules at gcc dot gnu dot org |dot org | Status|NEW |ASSIGNED Last reconfirmed|2010-04-12 09:17:11 |2010-06-16 11:41:32 date|| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43703
[Bug middle-end/44291] New: [4.6 regression] ICE in set_user_assembler_libfunc
Building (a cross compiler) for ARM Linux currently breaks whilst building
GLIBC. I believe this is due to the following patch (r159321):
http://gcc.gnu.org/ml/gcc-patches/2010-05/msg00766.html
From this commit, the ARM build initially failed in a way identical to #44197
(an ICE in varpool_remove_node). Unfortunately the patch in that PR does not
fix the issue for ARM: from r159629, I get the following error whilst building
GLIBC instead:
../include/stdlib.h:34:1: internal compiler error: in
set_user_assembler_libfunc, at optabs.c:6104
which is:
rtx
set_user_assembler_libfunc (const char *name, const char *asmspec)
{
tree id, decl;
void **slot;
hashval_t hash;
id = get_identifier (name);
hash = htab_hash_string (name);
slot = htab_find_slot_with_hash (libfunc_decls, id, hash, NO_INSERT);
gcc_assert (slot); // --- here
decl = (tree) *slot;
set_user_assembler_name (decl, asmspec);
return XEXP (DECL_RTL (decl), 0);
}
In the attached test case, if abort is renamed to abort2, the crash goes away,
which suggests the problem is related to the special handling of abort (and
friends) in builtins.c:set_builtin_user_assembler_name.
To reproduce: /path/to/cc1 -fpreprocessed init-first.i
--
Summary: [4.6 regression] ICE in set_user_assembler_libfunc
Product: gcc
Version: 4.6.0
Status: UNCONFIRMED
Keywords: ice-on-valid-code
Severity: critical
Priority: P3
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: jules at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44291
[Bug middle-end/44291] [4.6 regression] ICE in set_user_assembler_libfunc
--- Comment #1 from jules at gcc dot gnu dot org 2010-05-27 10:44 --- Created an attachment (id=20760) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20760action=view) The test case. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44291
[Bug bootstrap/39929] [4.5 Regression] Bootstrapping fails at stage 1 on powerpc-ibm-aix
--- Comment #17 from jules at gcc dot gnu dot org 2009-05-05 16:07 --- This still seems to be broken as of r147126 for ARM Linux. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39929
[Bug bootstrap/39929] [4.5 Regression] Bootstrapping fails at stage 1 on powerpc-apple-darwin9 and powerpc-ibm-aix
--- Comment #12 from jules at gcc dot gnu dot org 2009-04-28 22:31 ---
FYI: the above patch seems to have caused ARM Linux (cross) builds to break,
e.g.:
/scratch/julian/arm-mainline/src/gcc-mainline/libgcc/../gcc/config/arm/unwind-arm.c:589:
internal compiler error: Segmentation fault
Program received signal SIGSEGV, Segmentation fault.
0x08253611 in emit_insn_after_1 (first=0xf7cd7798, after=0xf7c744f8,
bb=0xf7c74540)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/emit-rtl.c:4129
4129 if (BB_END (bb) == after)
Backtrace:
#0 0x08253611 in emit_insn_after_1 (first=0xf7cd7798, after=0xf7c744f8,
bb=0xf7c74540)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/emit-rtl.c:4129
#1 0x082544ac in emit_insn_after (pattern=0x0, after=0xf7c744f8)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/emit-rtl.c:4335
#2 0x0862b3c6 in require_pic_register ()
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/config/arm/arm.c:3590
#3 legitimize_pic_address (orig=0xf7ca2ccc, mode=SImode, reg=0x0)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/config/arm/arm.c:3607
#4 0x086f32ed in gen_movsi (operand0=0x0, operand1=0xf7ca2ccc)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/config/arm/arm.md:4966
#5 0x0828ef94 in emit_move_insn_1 (x=0xf7cca0a0, y=0xf7ca2ccc)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/expr.c:3337
#6 0x0828f282 in emit_move_insn (x=0xf7cca0a0, y=0xf7ca2ccc)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/expr.c:3425
#7 0x0881e7ea in insert_value_copy_on_edge (src=value optimized out,
dest=value optimized out, e=value optimized out)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/tree-outof-ssa.c:190
#8 eliminate_phi (g=value optimized out, e=value optimized out)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/tree-outof-ssa.c:594
#9 expand_phi_nodes (sa=0x8d125b4)
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/tree-outof-ssa.c:756
#10 0x08795c2f in gimple_expand_cfg ()
at /scratch/julian/arm-mainline/src/gcc-mainline/gcc/cfgexpand.c:2506
...
(gdb) print *bb
$1 = {preds = 0xf7e57a8c, succs = 0xf7e2e390, aux = 0x0,
loop_father = 0xf7e57ab0, dom = {0x0, 0xf7c74558}, prev_bb = 0xf7e2a5b0,
next_bb = 0x0, il = {gimple = 0x0, rtl = 0x0}, count = -584888453233663636,
index = -136179944, loop_depth = -137911904, frequency = 0,
flags = -136126440}
Maybe there's some significant difference between the behaviour of
mips16_gp_pseudo_reg() and require_pic_register() which causes this, or maybe
MIPS is broken too?
--
jules at gcc dot gnu dot org changed:
What|Removed |Added
CC||jules at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39929
[Bug target/38697] gcc.target/arm/neon/neon.exp tests for vmov fail on arm-linux-eabi
--- Comment #3 from jules at gcc dot gnu dot org 2009-03-11 20:47 --- These failures show up because the tests are kind of weak. There's no particular reason that vget_low* intrinsics should generate vmov instructions as the tests are expecting: the assembly output shown with vldr/fstd works just as well, and no instructions at all (i.e. just a hint to the register allocator) would work even better. Without converting all the Neon intrinsic tests to be execution rather than compilation tests, I'm not sure if there's any way they can be made robust. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38697
[Bug target/35965] -fstack-protector produces segfaulting binaries on arm/armel
--- Comment #4 from jules at gcc dot gnu dot org 2008-06-26 10:10 --- I've posted a candidate fix: http://gcc.gnu.org/ml/gcc-patches/2008-06/msg01641.html -- jules at gcc dot gnu dot org changed: What|Removed |Added AssignedTo|unassigned at gcc dot gnu |jules at gcc dot gnu dot org |dot org | Status|UNCONFIRMED |ASSIGNED Ever Confirmed|0 |1 Last reconfirmed|-00-00 00:00:00 |2008-06-26 10:10:47 date|| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35965
[Bug middle-end/35964] ICE with -funroll-loops on arm/arm eabi
--- Comment #12 from jules at gcc dot gnu dot org 2008-05-12 15:27 --- I think this was probably fixed by one of Andrew Jenner's patches in gfortran-armel-updates: http://lists.debian.org/debian-gcc/2008/04/msg00351.html (The Unshare RTX earlier... one). Andrew and I actually did this work (ObjC and Fortran improvements for ARM EABI) primarily against Debian's version of 4.3, with the intention of then forward-porting the patches to FSF mainline -- but, that particular patch hasn't made it to FSF submission yet. If this is a regression, we'll consider submitting the patch for FSF 4.3 as well, when the time comes. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35964
[Bug testsuite/25590] [4.1/4.2 regression] FAIL: gcc.dg/tree-ssa/gen-vect-11.c scan-tree-dump-times vectorized 1 loops 1
--- Comment #10 from jules at gcc dot gnu dot org 2006-01-23 17:42 --- Subject: Bug 25590 Author: jules Date: Mon Jan 23 17:42:38 2006 New Revision: 110133 URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=110133 Log: Backport from mainline. 2006-01-23 Kazu Hirata [EMAIL PROTECTED] PR testsuite/25590 * gcc/testsuite/gcc.dg/tree-ssa/gen-vect-11.c: Add -fwrapv. Modified: branches/csl/sourcerygxx-4_1/ChangeLog.csl branches/csl/sourcerygxx-4_1/gcc/testsuite/gcc.dg/tree-ssa/gen-vect-11.c -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25590
[Bug testsuite/25590] [4.1/4.2 regression] FAIL: gcc.dg/tree-ssa/gen-vect-11.c scan-tree-dump-times vectorized 1 loops 1
--- Comment #11 from jules at gcc dot gnu dot org 2006-01-23 17:45 --- Subject: Bug 25590 Author: jules Date: Mon Jan 23 17:45:08 2006 New Revision: 110134 URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=110134 Log: Backport from mainline. 2006-01-23 Kazu Hirata [EMAIL PROTECTED] PR testsuite/25590 * gcc.dg/tree-ssa/gen-vect-11.c: Add -fwrapv. Modified: branches/csl/coldfire-4_1/ChangeLog.csl branches/csl/coldfire-4_1/gcc/testsuite/gcc.dg/tree-ssa/gen-vect-11.c -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25590
