2.5.7 build breakage around blkpg.c

What hit me?

Thanks,
Bill

VERSION = 2
PATCHLEVEL = 5
SUBLEVEL = 7
EXTRAVERSION =

$ gcc --version
2.95.4

$ as --version
GNU assembler 2.11.93.0.2 20020207 Debian/GNU Linux
...

gcc -D__KERNEL__ -I/home/wli/src/linus/linux-2.5/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=i386 -DKBUILD_BASENAME=blkpg -c -o blkpg.o blkpg.c
blkpg.c: In function `blk_ioctl':
blkpg.c:311: Internal compiler error:
blkpg.c:311: internal error--unrecognizable insn:
(insn 898 1478 907 (set (reg/v:SI 3 %ebx)

        (asm_operands/v ("1:    movl %%eax,0(%2)

2: movl %%edx,4(%2)
3:
.section .fixup,"ax"
4: movl %3,%0
jmp 3b
.previous
.section __ex_table,"a"

        .align 4
        .long 1b,4b
        .long 2b,4b
.previous") ("=r") 0[ 
                (reg:DI 1 %edx)
                (reg/v:SI 4 %esi)
                (const_int -14 [0xfffffff2])
                (reg/v:SI 3 %ebx)
            ] 
            [ 
                (asm_input:DI ("A"))
                (asm_input:SI ("r"))
                (asm_input:SI ("i"))
                (asm_input:SI ("0"))
            ]  ("blkpg.c") 268)) -1 (insn_list 873 (insn_list 897 (nil)))
    (nil))

Hi!

gcc bug hit you. Workaround by adding volatile's to all local variables in affected function.
Pavel

(about SSSCA) "I don't say this lightly.  However, I really think that the U.S.
no longer is classifiable as a democracy, but rather as a plutocracy." --hpa

seems to do with generating an illegal instruction in response to high register pressure around inline assembly.

Cheers,
Bill