Re: [PATCH 2/3] x86: Pass in pt_regs pointer for syscalls that need it

[Brian Gerst <brgerst@gmail.com>]

On Wed, Feb 11, 2009 at 10:05 AM, Tejun Heo <htejun@gmail.com> wrote:
> Brian Gerst wrote:
>> On Wed, Feb 11, 2009 at 9:48 AM, Tejun Heo <htejun@gmail.com> wrote:
>>> Tejun Heo wrote:
>>>>> I checked the disassembly of these functions and didn't see this
>>>>> happen on gcc 4.3.0.
>>>> Well, tracking down why run_init_process() is returning 0 with
>>>> -fstack-protector wasn't much of fun.  These breakages are very subtle
>>>> and if we're gonna pass in pointer to pt_regs anyway and thus can
>>>> guarantee such breakage can't happen at no additional cost, I think we
>>>> should do that even if it means slightly more argument fetching in a
>>>> few places.
>>> In addition, if we do that, we can remove the horrible
>>> asmlinkage_protect() thing altogether.
>>
>> Like I said before, the tail-call optimization problem isn't limited
>> to just this set of syscalls.  There are only two real ways to fix it.
>> 1) Set up a real stack frame for the syscalls instead of overalying
>> pt_regs, or 2) patch gcc to tell it not to touch the args area of the
>> stack.
>
> Right, I forgot about the generic ones.  We can pass pointer to
> pt_regs to all of them like x86_64 does but yeah we're likely to lose
> more than we gain by doing that.  :-(

x86-64 doesn't have the tail-call problem because it doesn't use the
pt_regs on stack trick for syscall args.  All the args are passed in
registers.

--
Brian Gerst