On Mar 10, 2007, at 2:18 PM, David Laight wrote:

> On Sat, Mar 10, 2007 at 02:14:41PM -0800, Jason Thorpe wrote:
>>
>> On Mar 10, 2007, at 2:11 PM, David Laight wrote:
>>
>>
>>> I've made a start on killing the 'stackgap' which is a rather nasty
>>
>>> hack used to do compat system calls.
>>
>> Um, doesn't the gap get allocated on that LWPs own stack?
>
> The calls are all:
>    sg = stackgap_init(p, 0);
>    usun = stackgap_alloc(p, &sg, sizeof(*usun));
> so are definitely all per process.

Ah, predates all of the lwp stuff.  So, that should have been made per- 
lwp when the LWP stuff went in...

...not that it would help cloned processes.

-- thorpej