From RISKS: secret scrubbing code removed by optimizers

[Darren Tucker]

Ben Lindstrom wrote:
> On Fri, 8 Nov 2002, Darren Tucker wrote:
> > Dan Kaminsky wrote:
> > >     Has somebody actually verified this optimizing behavior in any build
> > > of GCC?  Does voliatile actually stop it?
> >
> > Yes (gcc-3.2 on a SPARC). Yes.
> 
> does setting -fno-builtin-memset at compile time stop gcc from
> miscompiling?

Yes. Incidentally, the dead code elimination happens with -O so sticking
to -O2 or less won't help in this case.

gcc -S -O3 -fno-builtin-memset myfunc1.c

myfunc1:
        !#PROLOGUE# 0
        save    %sp, -216, %sp
        !#PROLOGUE# 1
        add     %fp, -120, %l0
        sethi   %hi(.LLC0), %g1
        mov     %l0, %o1
        call    scanf, 0
        or      %g1, %lo(.LLC0), %o0
        mov     %l0, %o0
        mov     0, %o1
        call    memset, 0
        mov     100, %o2
        nop
        ret
        restore


-- 
Darren Tucker (dtucker at zip.com.au)
GPG key 8FF4FA69 / D9A3 86E9 7EEE AF4B B2D4  37C9 C982 80C7 8FF4 FA69
    Good judgement comes with experience. Unfortunately, the experience
usually comes from bad judgement.