[Bug rtl-optimization/35044] New: resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[Bug rtl-optimization/35044] New: resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu dot org]

find_dead_or_set_registers uses mark_set_resources to determine if a register
is set, and if it is, it concludes that the register is before that set.
A register is also considered set when the set is inside a COND_EXEC, thus
delay slot scheduling can clobber values in registers that are supposed
to remain untouched when a COND_EXEC is not executed.

I have a patch for this, which changes find_dead_or_set_registers to not call
mark_set_resources for COND_EXEC patterns when the purpose is to find registers
that are killed.
I can't post this patch, or the target port we are using, at the moment
since we don't have a copyright assignment on file yet.

Another approach would be to change mark_set_resources to take a parameter
to tell it if the conservative assumption is that the register is set or
if it that the register is not set, and change all callers to provide an
appropriate value.


-- 
           Summary: resource.c:find_dead_or_set_registers doesn't grok
                    COND_EXEC
           Product: gcc
           Version: 4.2.1
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: amylaar at gcc dot gnu dot org
GCC target triplet: arc-elf32


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu dot org]

------- Comment #1 from amylaar at gcc dot gnu dot org  2008-01-31 18:45 -------
Created an attachment (id=15067)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15067&action=view)
Test case

This is a test case, extracted from an ARC linux kernel.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu dot org]

-- 

amylaar at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|39302                       |
OtherBugsDependingO|                            |39363
              nThis|                            |
         AssignedTo|unassigned at gcc dot gnu   |amylaar at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
           Keywords|                            |patch
   Last reconfirmed|0000-00-00 00:00:00         |2009-03-04 22:25:34
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu dot org]

------- Comment #2 from amylaar at gcc dot gnu dot org  2009-03-05 00:29 -------
The patch is here:
http://gcc.gnu.org/ml/gcc-patches/2009-03/msg00248.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu dot org]

-- 

amylaar at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|amylaar at gcc dot gnu dot  |unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[jsm28 at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

Joseph S. Myers <jsm28 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|arc-elf32                   |

--- Comment #3 from Joseph S. Myers <jsm28 at gcc dot gnu.org> 2011-03-22 20:18:13 UTC ---
ARC support has been removed for GCC 4.7, but please ping patches such as this
to generic code if you think they are logically right (even if the issues are
latent for other targets).

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

Jorn Wolfgang Rennecke <amylaar at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2009-03-04 22:25:34         |2011-11-27

--- Comment #4 from Jorn Wolfgang Rennecke <amylaar at gcc dot gnu.org> 2012-09-20 01:07:43 UTC ---
Another testcase for ARC700 is: gcc.c-torture/execute/builtin-bitops-1.c -O1

[Bug rtl-optimization/35044] resource.c:find_dead_or_set_registers doesn't grok COND_EXEC

[amylaar at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35044

--- Comment #5 from Jorn Wolfgang Rennecke <amylaar at gcc dot gnu.org> 2012-09-20 02:07:01 UTC ---
(In reply to comment #4)
> Another testcase for ARC700 is: gcc.c-torture/execute/builtin-bitops-1.c -O1

Sorry, that's a mixup: the above is a testcase of an analogous bug in
rtlanal.c:dead_or_set_regno_p .

Note that the policy for ambiguous codes like POST_INC, ZERO_EXTRACT or
COND_EXEC of mark_set_resources is different from dead_or_set_p:
while the former errs on the side of marking something set, the latter
(is suspposed to) err(s) on the side of something not entirely killed.