Re: 970917 on mips-sgi-irix6.2, haifa disabled: many problems

[Zack Weinberg <zack at rabi dot phys dot columbia dot edu>]

On Tue, 23 Sep 1997 01:48:51 -0600, Jeffrey A Law wrote:
>  > 2) Twice I got the machine into a state where it was impossible to log
>  > in either locally or over the network by running a build in a detached
>  > `screen' session -- once it was in `make bootstrap', once in `make
>  > check'.  I have no idea how it happened, there was nothing in syslog
>  > or the make log to indicate any trouble; but I had to hard-boot the
>  > machine.
>Probably while building insn-attrtab.c and insn-attrtab.o during
>the make bootstrap -- these take an enormous amount of memory
>to build for mips targets
>
>During the make check it was the 961203-1.c testcase which will
>suck up every byte of available VM (which will effectively lock
>the machine until the compiler dies).

It is _not_ caused by the compiler sucking VM; I've managed to
reproduce the problem with closer monitoring, and it looks like egcs
tickles a kernel bug which causes any exec() to block forever,
unkillably.  960924 does this every time I run a `make check', I think
always at c-torture/execute/961125-1.c; however, I can't duplicate the
problem by compiling that file by hand.  IRIX has historically been
plagued by bugs of this kind; I will report it to SGI and let you know
if they respond.

[In case you want to try to duplicate this: R5k Indy, Irix 6.2
w/recommended patches, as and ld from 7.1 compiler release.]

On the up side, I can confirm that the assertion failure in
dwarf2out.c is gone with this release.  The `insn does not satisfy its
constraints' error is still here:

$ ./cc1 961223-1.i -quiet -dumpbase -O0 -version
GNU C version egcs-2.90.10 970924 (gcc2-970802 experimental)
(mips-sgi-irix6.2) compiled by GNU C version egcs-2.90.10 970924
(gcc2-970802 experimental).
../../gcc/testsuite/gcc.c-torture/execute/961223-1.c: In function `sub':
../../gcc/testsuite/gcc.c-torture/execute/961223-1.c:10: internal 
error--insn does not satisfy its constraints:
(insn 31 32 4 (set (reg:DF 35 $f3)
        (reg:DF 84)) 193 {movdf_internal1} (nil)
    (nil))
Abort

Doesn't happen with -O.

I think this is an inlining problem, because I can't find an insn like
the one it complains about anywhere in a -drx RTL dump (with or
without -O), and cc1 aborts while generating RTL for `main' not `sub'.

zw

p.s. Is there a gdb anywhere that can read the debug information egcs
writes?  4.16 says "does not support mips-sgi-irix6.2" when I try to
configure it, and the 4.14 binaries from the SGI freeware cd can't
read N32 debug info.