Re: [PATCH] drm/radeon/kms: fence cleanup + more reliable GPU lockup detection

[Jerome Glisse <jglisse@redhat.com>]

On Sun, Feb 28, 2010 at 12:22:52PM +0000, Alan Swanson wrote:
> On Fri, 2010-02-26 at 15:49 +0100, Jerome Glisse wrote:
> > This patch cleanup the fence code, it drops the timeout field of
> > fence as the time to complete each IB is unpredictable and shouldn't
> > be bound.
> > 
> > The fence cleanup lead to GPU lockup detection improvement, this
> > patch introduce a callback, allowing to do asic specific test for
> > lockup detection. In this patch the CP is use as a first indicator
> > of GPU lockup. If CP doesn't make progress during 1second we assume
> > we are facing a GPU lockup.
> > 
> > To avoid overhead of testing GPU lockup frequently due to fence
> > taking time to be signaled we query the lockup callback every
> > 100msec. There is plenty code comment explaining the design & choise
> > inside the code.
> 
> Every 100msec? Is this running all the time? If so, that's not very good
> for CPU power saving to lower C-states in an idle system. We could at
> least use one of the round_jiffies.
> 

This run only when userspace call bo wait thus it only happen when userspace
is waiting for something.

Cheers,
Jerome

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--