This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Regression in gcc-2.95 19990525
- To: hjl at varesearch dot com (H.J. Lu)
- Subject: Re: Regression in gcc-2.95 19990525
- From: Jeffrey A Law <law at upchuck dot cygnus dot com>
- Date: Wed, 26 May 1999 11:19:12 -0600
- cc: egcs-bugs at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <19990526133153.C2494666E@osmium.su.varesearch.com>you write:
> With gcc 2.95 19990525, I got this on x86:
>
> # gcc -S -O3 -mcpu=i686 -march=i686 foo.c
> foo.c: In function `main':
> foo.c:27: Could not split insn
> (insn 116 112 61 (set (reg/v:SI 5 %edi)
> (if_then_else:SI (ne:SI (reg/v:SI 3 %ebx)
> (mem/s:SI (plus:SI (mult:SI (reg/v:SI 4 %esi)
> (const_int 4 [0x4]))
> (reg:SI 0 %eax)) 3))
> (mem/s:SI (plus:SI (mult:SI (reg/v:SI 4 %esi)
> (const_int 4 [0x4]))
> (mem/f:SI (plus:SI (reg:SI 6 %ebp)
> (const_int 12 [0xc])) 0)) 3)
> (reg/v:SI 5 %edi))) 414 {movsicc+2} (insn_list/j/c 112 (insn_li
> st/j
> 138 (insn_list/j/c 112 (nil))))
> (expr_list:REG_DEAD (reg:SI 0 %eax)
> (nil)))
>
> egcs 1.1.2 is ok. It is a new bug.
Yes. It's a bug in the reload pass. I gave an analysis to Joern (who is most
familiar with the code in question) a day or two ago. Hopefully he'll be able
take a look before the end of this week.
It's easily the most common bug I've seen reported against the x86 port over
the last few months.
jeff