public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
@ 2023-11-30 12:44 immoloism at googlemail dot com
  2023-11-30 12:45 ` [Bug c/112782] " immoloism at googlemail dot com
                   ` (7 more replies)
  0 siblings, 8 replies; 9+ messages in thread
From: immoloism at googlemail dot com @ 2023-11-30 12:44 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

            Bug ID: 112782
           Summary: [14 Regression] stage2/3 differs when compiled with
                    -mfix-r5900 on mips
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: immoloism at googlemail dot com
  Target Milestone: ---

While testing GCC-14 support in Gentoo I have found an interesting issue where
GCC no longer builds correctly when setting -mfix-r5900. With the help of Sam
James we have confirmed this causes the error in the stage 2 and 3 checks
however I have been unable to find where this breakage started so far and can
only confirm there is no issue using the flag with versions 12 and 13.

To reproduce: (sorry for vague steps but I normally use portage to manage this)

Setup a MIPS install (cross compiling does not hit the issue)
Set CFLAGS to "-Os -march=mips3 -mabi=32 -mplt  -pipe -mfix-r5900"
Compile GCC14 as normal

Error:

Comparing stages 2 and 3
Bootstrap comparison failure!
gcc/spellcheck.o differs
gcc/tree-data-ref.o differs
gcc/asan.o differs
gcc/pointer-query.o differs
gcc/tree-ssa-math-opts.o differs
gcc/lto/lto-common.o differs
gcc/lto/lto-partition.o differs
gcc/tree-vectorizer.o differs
gcc/gimple-range-op.o differs
gcc/dwarf2out.o differs
gcc/gimple-match-4.o differs
gcc/gimple-ssa-warn-restrict.o differs
gcc/tree-chrec.o differs
gcc/function.o differs
gcc/tree-vect-loop-manip.o differs
gcc/tree-vrp.o differs
gcc/gimple-array-bounds.o differs
gcc/generic-match-2.o differs
gcc/tree-vect-patterns.o differs
gcc/gimple-match-5.o differs
gcc/simplify-rtx.o differs
gcc/tree-vect-data-refs.o differs
gcc/tree-vect-slp.o differs
gcc/tree-ssa-alias.o differs
gcc/ipa-fnsummary.o differs
gcc/tree.o differs
gcc/tree-scalar-evolution.o differs
gcc/splay-tree-utils.o differs
gcc/lto-streamer-in.o differs
gcc/fortran/trans-array.o differs
gcc/fortran/trans-intrinsic.o differs
gcc/fortran/interface.o differs
gcc/fortran/frontend-passes.o differs
gcc/fortran/openmp.o differs
gcc/fortran/trans-common.o differs
gcc/fortran/parse.o differs
gcc/vr-values.o differs
gcc/predict.o differs
gcc/var-tracking.o differs
gcc/mips.o differs
gcc/gimple-walk.o differs
gcc/tree-switch-conversion.o differs
gcc/tree-ssa-loop-prefetch.o differs
gcc/optabs.o differs
gcc/rtl-ssa/accesses.o differs
gcc/omp-offload.o differs
gcc/tree-ssa-phiopt.o differs
gcc/symtab.o differs
gcc/gimple-match-7.o differs
gcc/value-range.o differs
gcc/tree-ssa-operands.o differs
gcc/sel-sched-ir.o differs
gcc/tree-ssa-sink.o differs
gcc/tree-ssa-reassoc.o differs
gcc/analyzer/region-model.o differs
gcc/analyzer/checker-path.o differs
gcc/analyzer/program-point.o differs
gcc/analyzer/sm-file.o differs
gcc/analyzer/region.o differs
gcc/analyzer/call-details.o differs
gcc/analyzer/diagnostic-manager.o differs
gcc/analyzer/sm-fd.o differs
gcc/analyzer/sm-malloc.o differs
gcc/analyzer/record-layout.o differs
gcc/analyzer/constraint-manager.o differs
gcc/analyzer/program-state.o differs
gcc/analyzer/access-diagram.o differs
gcc/analyzer/infinite-recursion.o differs
gcc/analyzer/call-summary.o differs
gcc/analyzer/engine.o differs
gcc/analyzer/store.o differs
gcc/analyzer/supergraph.o differs
gcc/analyzer/feasible-graph.o differs
gcc/tree-cfg.o differs
gcc/gimple-match-6.o differs
gcc/tree-inline.o differs
gcc/tree-ssa-ccp.o differs
gcc/haifa-sched.o differs
gcc/input.o differs
gcc/loop-init.o differs
gcc/gcov.o differs
gcc/cfgrtl.o differs
gcc/ipa-utils.o differs
gcc/gimple-ssa-warn-access.o differs
gcc/combine.o differs
gcc/wide-int.o differs
gcc/tree-vect-loop.o differs
gcc/gimple-match-8.o differs
gcc/passes.o differs
gcc/cfg.o differs
gcc/ipa-devirt.o differs
gcc/tree-ssa-dom.o differs
gcc/final.o differs
gcc/tree-ssa-loop-unswitch.o differs
gcc/read-md.o differs
gcc/text-art/styled-string.o differs
gcc/text-art/widget.o differs
gcc/text-art/style.o differs
gcc/text-art/table.o differs
gcc/gimple-match-2.o differs
gcc/generic-match-5.o differs
gcc/tree-ssa-threadedge.o differs
gcc/fold-const.o differs
gcc/varasm.o differs
gcc/modulo-sched.o differs
gcc/omp-low.o differs
gcc/tree-pretty-print.o differs
gcc/gimplify.o differs
gcc/gimple-range-path.o differs
gcc/cgraphunit.o differs
gcc/generic-match-4.o differs
gcc/print-rtl.o differs
gcc/sched-deps.o differs
gcc/gimple-ssa-strength-reduction.o differs
gcc/gimple-match-3.o differs
gcc/gimple-fold.o differs
gcc/df-core.o differs
gcc/dwarf2cfi.o differs
gcc/tree-ssa-loop-niter.o differs
gcc/gimple-match-10.o differs
gcc/tree-ssa-uncprop.o differs
gcc/gimple-match-9.o differs
gcc/build/print-rtl.o differs
gcc/build/genmatch.o differs
gcc/build/genrecog.o differs
gcc/build/read-md.o differs
gcc/c-family/c-common.o differs
gcc/c-family/c-omp.o differs
gcc/c-family/c-warn.o differs
gcc/value-range-storage.o differs
gcc/tree-eh.o differs
gcc/gimple-match-1.o differs
gcc/gimple-predicate-analysis.o differs
gcc/tree-ssa-loop-ch.o differs
gcc/tree-if-conv.o differs
gcc/ordered-hash-map-tests.o differs
gcc/tree-ssa-strlen.o differs
gcc/gimple-streamer-out.o differs
gcc/tree-dfa.o differs
gcc/tree-predcom.o differs
gcc/early-remat.o differs
gcc/tree-ssa-ifcombine.o differs
gcc/ipa-inline.o differs
gcc/gimple-loop-interchange.o differs
gcc/ipa-cp.o differs
gcc/cp/cp-gimplify.o differs
gcc/cp/decl.o differs
gcc/cp/parser.o differs
gcc/cp/semantics.o differs
gcc/cp/constraint.o differs
gcc/cp/tree.o differs
gcc/cp/error.o differs
gcc/cp/typeck.o differs
gcc/cp/decl2.o differs
gcc/cp/rtti.o differs
gcc/expmed.o differs
gcc/sanopt.o differs
gcc/trans-mem.o differs
gcc/c/c-decl.o differs
gcc/c/c-parser.o differs
gcc/targhooks.o differs
gcc/tree-ssa-threadbackward.o differs
gcc/gimple.o differs
gcc/ipa-prop.o differs
gcc/tree-ssa-sccvn.o differs
gcc/range-op.o differs
libcody/server.o differs
libcpp/files.o differs
libcpp/macro.o differs
libcpp/lex.o differs
libiberty/pic/simple-object-mach-o.o differs
libiberty/simple-object.o differs
libiberty/simple-object-mach-o.o differs
make[2]: *** [Makefile:22598: compare] Error 1
make[1]: *** [Makefile:22578: stage3-bubble] Error 2
make: *** [Makefile:22652: bootstrap-lean] Error 2

I'll include the stage 2 and 3 spellcheck.o files as well.

I can provide a clean qemu-user image to aid in testing this if it is helpful
so it can be reproduced on a x86_64 machine.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug c/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
@ 2023-11-30 12:45 ` immoloism at googlemail dot com
  2023-11-30 12:46 ` immoloism at googlemail dot com
                   ` (6 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: immoloism at googlemail dot com @ 2023-11-30 12:45 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

--- Comment #1 from Immolo <immoloism at googlemail dot com> ---
Created attachment 56728
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=56728&action=edit
build log

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug c/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
  2023-11-30 12:45 ` [Bug c/112782] " immoloism at googlemail dot com
@ 2023-11-30 12:46 ` immoloism at googlemail dot com
  2023-11-30 13:56 ` [Bug bootstrap/112782] " rguenth at gcc dot gnu.org
                   ` (5 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: immoloism at googlemail dot com @ 2023-11-30 12:46 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

--- Comment #2 from Immolo <immoloism at googlemail dot com> ---
Created attachment 56729
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=56729&action=edit
spellcheck.o

Output of stage2 and stage3 spellcheck.o

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
  2023-11-30 12:45 ` [Bug c/112782] " immoloism at googlemail dot com
  2023-11-30 12:46 ` immoloism at googlemail dot com
@ 2023-11-30 13:56 ` rguenth at gcc dot gnu.org
  2023-12-01 23:58 ` pinskia at gcc dot gnu.org
                   ` (4 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: rguenth at gcc dot gnu.org @ 2023-11-30 13:56 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |bootstrap
             Target|                            |mips
   Target Milestone|---                         |14.0

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
                   ` (2 preceding siblings ...)
  2023-11-30 13:56 ` [Bug bootstrap/112782] " rguenth at gcc dot gnu.org
@ 2023-12-01 23:58 ` pinskia at gcc dot gnu.org
  2023-12-02  0:03 ` pinskia at gcc dot gnu.org
                   ` (3 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-01 23:58 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I thought this was reported beforehand  ...

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
                   ` (3 preceding siblings ...)
  2023-12-01 23:58 ` pinskia at gcc dot gnu.org
@ 2023-12-02  0:03 ` pinskia at gcc dot gnu.org
  2023-12-02  0:08 ` pinskia at gcc dot gnu.org
                   ` (2 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-02  0:03 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
           Keywords|                            |compare-debug-failure
         Resolution|---                         |MOVED

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
-mfix-r5900 is all handled in gas rather than in GCC ....

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
                   ` (4 preceding siblings ...)
  2023-12-02  0:03 ` pinskia at gcc dot gnu.org
@ 2023-12-02  0:08 ` pinskia at gcc dot gnu.org
  2023-12-02  0:14 ` pinskia at gcc dot gnu.org
  2023-12-05 18:33 ` immoloism at googlemail dot com
  7 siblings, 0 replies; 9+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-02  0:08 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://sourceware.org/bugz
                   |                            |illa/show_bug.cgi?id=13509

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I can't remember exactly but I did file
https://sourceware.org/bugzilla/show_bug.cgi?id=13509 when I noticed a
bootstrap comparison failure a long time ago but I am not 100% sure if this is
the same issue or just a related one. But again I highly doubt this is not a
GCC issue, especially when -mfix-r5900 does almost nothing on the GCC side and
is only handled in gas.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
                   ` (5 preceding siblings ...)
  2023-12-02  0:08 ` pinskia at gcc dot gnu.org
@ 2023-12-02  0:14 ` pinskia at gcc dot gnu.org
  2023-12-05 18:33 ` immoloism at googlemail dot com
  7 siblings, 0 replies; 9+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-02  0:14 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
My bet is the code for -mfix-r5900 is not expecting the extra labels that is
produced for debuging ...

Anyways as I mentioned if -mfix-r5900 is the issue then it is a gas issue.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug bootstrap/112782] [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips
  2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
                   ` (6 preceding siblings ...)
  2023-12-02  0:14 ` pinskia at gcc dot gnu.org
@ 2023-12-05 18:33 ` immoloism at googlemail dot com
  7 siblings, 0 replies; 9+ messages in thread
From: immoloism at googlemail dot com @ 2023-12-05 18:33 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112782

--- Comment #7 from Immolo <immoloism at googlemail dot com> ---
Just to document my findings which pretty much confirms what Pinski has said
here.

I have gone back and doubled checked what I found and noticed portage was
removing the -mfix-r5900 option since the start and I just missed it.
Retesting this has always been broken and by pure chance and dumb luck I have
never compiled anything on a r5900 cpu to hit this.

I've tested this back as far as gcc-11.3.

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2023-12-05 18:33 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-11-30 12:44 [Bug c/112782] New: [14 Regression] stage2/3 differs when compiled with -mfix-r5900 on mips immoloism at googlemail dot com
2023-11-30 12:45 ` [Bug c/112782] " immoloism at googlemail dot com
2023-11-30 12:46 ` immoloism at googlemail dot com
2023-11-30 13:56 ` [Bug bootstrap/112782] " rguenth at gcc dot gnu.org
2023-12-01 23:58 ` pinskia at gcc dot gnu.org
2023-12-02  0:03 ` pinskia at gcc dot gnu.org
2023-12-02  0:08 ` pinskia at gcc dot gnu.org
2023-12-02  0:14 ` pinskia at gcc dot gnu.org
2023-12-05 18:33 ` immoloism at googlemail dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).