[CRIU] [PATCH] compel/x86: Don't use pushq for a label

[Dmitry Safonov]

On Wed, 11 Sep 2019 at 11:20, Cyrill Gorcunov <gorcunov at gmail.com> wrote:
>
> On Wed, Sep 11, 2019 at 11:13:51AM +0100, Dmitry Safonov wrote:
> > `pushq` sign-extends the value. Which is a bummer as the label's address
> > may be higher that 2Gb, which means that the sign-bit will be set.
> >
> > As it long-jumps with ia32 selector, %r11 can be scratched.
> > Use %r11 register as a temporary to push the 32-bit address.
> >
> > Complements: a9a760278c1a ("arch/x86: push correct eip on the stack
> > before lretq")
> > Cc: Cyrill Gorcunov <gorcunov at gmail.com>
> > Reported-by: Andrei Vagin <avagin at gmail.com>
> > Signed-off-by: Dmitry Safonov <dima at arista.com>
> > ---
> >  compel/arch/x86/plugins/std/parasite-head.S | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/compel/arch/x86/plugins/std/parasite-head.S b/compel/arch/x86/plugins/std/parasite-head.S
> > index a988de9d4218..465cd887b1f9 100644
> > --- a/compel/arch/x86/plugins/std/parasite-head.S
> > +++ b/compel/arch/x86/plugins/std/parasite-head.S
> > @@ -25,7 +25,9 @@ ENTRY(__export_parasite_head_start_compat)
> >  .code64
> >       PARASITE_ENTRY 0
> >       pushq   $__USER32_CS
> > -     pushq   $2f
> > +     xor     %r11, %r11
> > +     movl    $2f, %r11d
> > +     pushq   %r11
> >       lretq
> >  2:
> >  .code32
>
> Wait, didn't Andrew already fixed it?

Yes, but as the guys who reviewed it hasn't noticed the same piece of code
for threads curing... :)

Note, there is also feature-test.mak, but we don't care about correctness there
as it's a test for compilation.

Thanks,
             Dmitry