Re: (set (reg:SI ) (scratch:SI)) unrecognizable insn

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



"Kim, Eunchan" <helpbygrace@xxxxxxxxx> writes:

> On Tue, Oct 6, 2009 at 5:10 AM, Ian Lance Taylor <iant@xxxxxxxxxx> wrote:
>> "Kim, Eunchan" <helpbygrace@xxxxxxxxx> writes:
>>
>>> (insn 81 80 0 ../../../../sdk/gcc-4.3.3/libgcc/../gcc/unwind-dw2.c:1503
>>> (parallel [
>>>             (set (mem:BLK (reg/v/f:SI 43 [ c ]) [0 S4 A8])
>>>                 (unspec:BLK [
>>>                         (reg:SI 75)
>>>                         (mem:QI (reg/v/f:SI 43 [ c ]) [0 S1 A8])
>>>                         (mem:BLK (reg/v/f:SI 43 [ c ]) [0 S4 A8])
>>>                     ] 21))
>>>             (clobber (scratch:SI))
>>>         ]) -1 (nil))
>>
>> Where did this insn come from?  (clobber (scratch:SI)) is something
>> one might see in a reload insn pattern, but it shouldn't be in a
>> normal insn patter.
>>
>> If you are trying to write an insn which does a memory copy and needs
>> a scratch register, then you need to use a define_expand which calls
>> gen_reg_rtx.
>>
>> Ian
>>
>
> That insn is from gcc's error message when i compile
> libgcc(unwind-dw2.c) with this gcc. I based on mips cross compiler.
> But i'd rather make whole new. Because i do not fully understand mips
> target description.

I don't see any define_insn like that in mips.md.  It looks like
mov_<store>r, but that has no (clobber (scratch)) in it.  So I think
it is an insn you added or modified.

Ian


[Index of Archives]     [Linux C Programming]     [Linux Kernel]     [eCos]     [Fedora Development]     [Fedora Announce]     [Autoconf]     [The DWARVES Debugging Tools]     [Yosemite Campsites]     [Yosemite News]     [Linux GCC]

  Powered by Linux