On Mon, 5 Sep 2011 08:38:07 +0200 Daniel Vetter <dan...@ffwll.ch> wrote:
> On Sun, Sep 04, 2011 at 09:38:56PM +0000, Ben Widawsky wrote: > > Oops, you're totally right, I think I meant: > > - I915_WRITE(GEN6_PMIMR, pm_imr & ~pm_iir); > > + I915_WRITE(GEN6_PMIMR, dev_priv->pm_iir); > > Imo still racy without the irqsafe rps_lock around it. gcc is free to > compile that into a separate load and store which the irq handler can > get in between and change dev_priv->pm_iir and PMIMR. The race is now > only one instruction wide, though ;-) > -Daniel You are absolutely correct. The modification to GEN6_PMIMR must be within the protection of rps_lock. Ben _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
