Re: [uml-devel] stack and scheduler patches

2005-09-06 Thread Jeff Dike
On Sun, Sep 04, 2005 at 01:51:32PM +0200, Blaisorblade wrote: > Also, what is the difference between your "stack trace" patch, which adds a > member to thread_struct, for each process, so I'd like a minimum of > justification, and one of the sysrq options which does exactly this? What this does,

Re: [uml-devel] stack and scheduler patches

2005-09-06 Thread Allan Graves
Void casting: my compiler gave me a warning, and I don't like that. pid=-1, leftover from error checking that i didn't need, my fault. pid=0, find_task_by_pid(0) will return a valid pointer, but it doesn't seem to really be a good task. This seems to be related to the kernel idle task. If you

Re: [uml-devel] stack and scheduler patches

2005-09-04 Thread Blaisorblade
On Friday 02 September 2005 18:59, Jeff Dike wrote: > On Wed, Aug 31, 2005 at 07:06:33PM -0400, Allan Graves wrote: > > Jeff, > > Here's the patches. Hope these meet your approval! Could you please write a proper changelog? Reading a patch is possible, but needs time which I don't have in the ver

Re: [uml-devel] stack and scheduler patches

2005-09-02 Thread Jeff Dike
On Wed, Aug 31, 2005 at 07:06:33PM -0400, Allan Graves wrote: > Jeff, > Here's the patches. Hope these meet your approval! > Allan Some comments - I have these all fixed: +struct task_struct *from=(struct task_struct*)prev; +struct task_struct *to=(struct task_str