[Qemu-devel] git commit c832e3de chokes gcc-4.5.2 (out of memory error)

[walt <w41ter@gmail.com>]

Hi qemu team,

The following commit has a strange side-effect on gcc-4.5.2:

commit c832e3de64f1069313fc0672087791cc3dd5b4d8
Author: Richard Henderson
Date:   Mon Jan 10 19:23:47 2011 -0800

     target-i386: Use deposit operation.
     
     Use this for assignment to the low byte or low word of a register.
     
     Acked-by: Aurelien Jarno
     Signed-off-by: Richard Henderson
     Signed-off-by: Edgar E. Iglesias


The new code in 'target-i386/translate.c' compiles just fine with gcc-4.4.4,
using less than 1GB of RAM, but gcc-4.5.2 uses the whole GB and then fills
up 2GB of swap space in addition while the machine grinds slowly to a halt.

I'm using the latest qemu.git, and when I revert that one commit, gcc-4.5.2
chugs along just fine with my puny 1GB of RAM.

I filed a bug against gcc-4.5.2 a few hours ago, but I notice that another
ten or fifteen bugs were filed *after* mine, so I don't have much hope that
anyone will pay much attention :(

I'm hoping someone here might know how to find the source of the problem
so I can add some helpful details to my bug report.

Thanks for any suggestions, and thanks for qemu :)