Re: [PATCH] i386 __verify_write fixes

[Alan Cox]

On Sun, 2002-10-13 at 21:13, Brian Gerst wrote:
> mark the page read-only.  A 386 would continue to write to the page when 
> the first thread woke up again.  We need to prevent anyone from changing 
> the page permissions until we are done writing to userspace.  My 
> suggestion is to replace access_ok() with begin_user_{read,write}(), and 
> add end_user_{read,write}().  Thoughs?

For 2.6 with the time scale as it is I think we just block pre-emption
on 386. Its not a big deal since almost nobody uses 386 any more