[Bug c/13585] New: Incorrect optimisation of call to sfunc

["stuart dot menefy at st dot com" <gcc-bugzilla at gcc dot gnu dot org>]

The problem occurs when building the attached code. The command line
is:
   % sh4-linux-gcc -O2 -c pcm_params.i -fPIC -o pcm_params.lo       

The generated code for line 8252:
         tmp = *params;
which is a struct assignment, is:

gcc version 3.2 20020529 (experimental) / 3.2.1 20021022 (prerelease)
and 3.3.4:
        mov.l   .L22,r0
        mov     r14,r4
        mov.w   .L32,r7
        mov     r10,r5
        mov.l   @(r0,r12),r0
        mov     #74,r6
        add     r14,r7
        mov.l   r0,@(36,r7)
        jsr     @r7
        mov     r0,r7

The problem is that the compiler has migrated the "mov r0,r7" which sets
up the address to jump to, into the delay slot.

This problem is seen in SuperH's 3.2/3.2.1 compiler, and a 3.3.2
compiler, but not in a 3.4 snapshot, however this may just be a side
effect of slightly different optimisations.

-- 
           Summary: Incorrect optimisation of call to sfunc
           Product: gcc
           Version: 3.3.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: stuart dot menefy at st dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: sh4-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13585