Re: [PATCH] x86: vdso32/syscall.S: do not load __USER32_DS to %ss

[Ingo Molnar <mingo@kernel.org>]

* Denys Vlasenko <dvlasenk@redhat.com> wrote:

> On 03/24/2015 07:34 AM, Ingo Molnar wrote:
> > 
> > * Denys Vlasenko <vda.linux@googlemail.com> wrote:
> > 
> >> On Mon, Mar 23, 2015 at 9:38 PM, Andy Lutomirski <luto@amacapital.net> wrote:
> >>> Actually, I want to remove the added comment in the code.  I don't see
> >>> why we should have a specific comment about SS and not about, say, CS,
> >>> ESP, or anything else.  OK?
> >>
> >> Ok.
> > 
> > Might be nice to place a more generic description there, which 
> > registers are expected to be saved by user-space calling in here, etc.
> 
> __kernel_vsyscall entry point has the same ABI in any 32-bit vDSO,
> the good old int 0x80 calling convention:
> 
> syscall# in eax,
> params in ebx/ecx/edx/esi/edi/ebp,
> all registers are preserved by the syscall.
> 
> (I think we don't guarantee that all flags are preserved:
> I have a testcase where DF gets cleared).

I think the fact that the people developing this code are unsure about 
exactly what gets saved/restored is justification enough to document 
the circumstances a bit better.

> Each flavor of fast kernel call does necessary massaging to conform 
> to the ABI. E.g. SYSCALL-based fast call clobbers ecx, so its vDSO 
> saves/restores ecx on stack.
> 
> Do you want a patch which adds such comment into every vDSO?

Well, maybe it's better to extend the already existing descriptions at 
the syscall entry points to be a full description of all details, and 
put a reference to that description into the vDSOs?

Thanks,

	Ingo