This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Unrecognizable insn
- From: Shinpei Kato <shinny at j02 dot itscom dot net>
- To: James E Wilson <wilson at specifixinc dot com>
- Cc: gcc at gcc dot gnu dot org
- Date: Fri, 6 Aug 2004 00:07:44 +0900 (JST)
- Subject: Unrecognizable insn
- References: <410FDB07.4010503@specifixinc.com>
Hi,
>You will have to spend a little time looking at the lreg and greg dumps,
>and stepping through reload, to figure out what is going on.
I'm trying to solve my problems though, I've not done yet.
I've worked on this problem for long so I need to proceed anyway. Of course I'll be back on it again later.
I bypassed, at the moment, this problem and another condition code problem by omitting "=z" in the define_insn.
Then I faced the next problem that I've also had trouble with for long.
I think I posted a message about it before and you kindly replied for it.
***********************************************************************
../../../../../gcc-3.4.0/newlib/libc/stdlib/erand48.c: In function `_erand48_r':../../../../../gcc-3.4.0/newlib/libc/stdlib/erand48.c:25: error: unrecognizable insn:
(insn 94 35 36 0 ../../../../../gcc-3.4.0/newlib/libc/stdlib/erand48.c:22 (set (subreg:HI (reg:SI 32 $f0) 2)
(mem:HI (reg/v/f:SI 17 $17 [orig:175 xseed ] [175]) [5 S2 A16])) -1 (nil)
(nil))
***********************************************************************
Although I got dump files, the above insn never appears in them.
The last insn number in the dump files is 89.
The way is only checking a md file step by step? I can't guess at all...
Could you advise anything for me? If so, what kind of information do you need?
Regards,
Shinpei