[Bug target/56091] New: gcc.target/arm/pr43137.c fails for THUMB-1

[Bug target/56091] New: gcc.target/arm/pr43137.c fails for THUMB-1

[janis at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56091

             Bug #: 56091
           Summary: gcc.target/arm/pr43137.c fails for THUMB-1
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: janis@gcc.gnu.org
            Target: arm-none-eabi


This test failure now appears for mainline arm-none-eabi for all THUMB-1
effective targets:

    FAIL: gcc.target/arm/pr43137.c scan-assembler-not movtr1, r[1-9]

This is a compile-only test so I was able to run it for 34 sets of flags
comprising all values of -march supported by GAS, with and without -mthumb as
appropriate.  The test failure begins with this patch to GCC mainline:

   r187015 | rsandifo | 2012-05-01 07:45:24 -0700 (Tue, 01 May 2012) | 33 lines

This is a regression from GCC 4.7

[Bug target/56091] gcc.target/arm/pr43137.c fails for THUMB-1

[ramana at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56091

Ramana Radhakrishnan <ramana at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ramana at gcc dot gnu.org,
                   |                            |rsandifo at gcc dot gnu.org

--- Comment #1 from Ramana Radhakrishnan <ramana at gcc dot gnu.org> 2013-02-19 23:51:21 UTC ---
Adding rsandifo to the CC list.

[Bug target/56091] gcc.target/arm/pr43137.c fails for THUMB-1

[rsandifo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56091

--- Comment #2 from rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> 2013-02-20 21:59:45 UTC ---
FWIW, target-specific fallout along these lines was expected at the time.
The patch made lower_subreg take rtx_costs into account, so some targets'
rtx_costs might need updating.

[Bug target/56091] gcc.target/arm/pr43137.c fails for THUMB-1

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56091

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |ktkachov at gcc dot gnu.org
         Resolution|---                         |WORKSFORME

--- Comment #3 from ktkachov at gcc dot gnu.org ---
I can't reproduce it with current 4.8.3, 4.9.2 or trunk on
arm-none-eabi/-march=armv5t/-mthumb

I'm closing this. If you can reproduce this with any maintained version of gcc
please reopen with the relevant details.

RTX costs for arm have been rewritten quite extensively for 4.9 btw