On Sat, 19 Mar 2011 07:44:24 +0000, Chris Wilson <ch...@chris-wilson.co.uk> wrote:
> I have an aversion to passing parameters around to control locking. > Considering that we modify dev_private and adjust LP_RING during enable, > wouldn't it be prudent to move that under the struct mutex? Especially in > the future if we allow it be enabled at runtime. This seems like the right plan to me as well -- make access to the renderctx always covered by the struct mutex. > Failing that: bool was_locked = mutex_is_locked(&dev->struct_mutex); Ick. Better to just stick the needed locks around the two calls to ironlake_teardown_rc6. You're just masking the issue otherwise. -- keith.pack...@intel.com
pgpRT2DIBjX84.pgp
Description: PGP signature
_______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
