This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: PR target/40470: unable to find a registe r to spill in class ‘SSE_FIRST_REG’


On Thu, Jun 18, 2009 at 5:18 AM, Bernd Schmidt<bernds_cb1@t-online.de> wrote:
> Jeff Law wrote:
>>
>> Someone needs to look at what it's going to take to fix combine. ?I would
>> recommend against your patch until we know precisely what changes we'd need
>> to make to combine to fix this problem.
>
> In theory this kind of thing should be handled by cant_combine_insn_p.
>
>

There is nothing unusual passed to cant_combine_insn_p.
There are

Breakpoint 1, cant_combine_insn_p (insn=0x7ffff317f140)
    at /export/gnu/src/gcc-work/gcc/gcc/combine.c:1992
1992	  if (! INSN_P (insn))
(insn 29 26 31 4 x.i:7 (set (reg:V16QI 82)
        (unspec:V16QI [
                (reg:V16QI 83 [ frag2 ])
                (reg:V16QI 64 [ D.2973 ])
                (reg:V16QI 63 [ D.2974 ])
            ] 134)) 1797 {sse4_1_pblendvb} (expr_list:REG_DEAD
(reg:V16QI 83 [ frag2 ])
        (nil)))
(gdb)

Although operand 3 for sse4_1_pblendvb must be xmm0, it
doesn't show up here.  Any more suggestions?

Thanks.

-- 
H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]