Re: [PATCH, RFC] MIPS: Implement the getcontext API

[David Daney <ddaney@caviumnetworks.com>]

Daniel Jacobowitz wrote:
> On Wed, Mar 04, 2009 at 09:19:28AM +0100, Brian Foster wrote:
>>  moving the signal trampoline to a vdso (which
>>  is(? was?) called, maybe misleadingly, ‘vsyscall’,
>>  on other architectures) is the obvious solution to
>>  that part of the puzzle.  and yes, it is possible
>>  to maintain the ABI; the signal trampoline is still
>>  also put on the stack, and modulo XI, would work if
>>  used — the trampoline-on-stack is simply not used
>>  if there is a vdso with the signal trampoline.
> 
> That won't quite retain the ABI: you need to make sure everyone
> locates it by using the stack pointer instead of the return pc.
> Fortunately, GCC uses the return PC only for instruction matching
> today.  I have a vague memory it used to use the stack pointer but
> this was more reliable.

That is correct.  Due to various errata the trampoline cannot always be 
at a fixed offset to the signal context bits.  So we had to use the 
return PC as you indicate.

David Daney