On Tue, Oct 04, 2011 at 10:13:47PM -0700, Kenneth Graunke wrote: > On 10/04/2011 01:30 AM, Daniel Vetter wrote: > > On Mon, Oct 03, 2011 at 11:02:39PM -0700, Kenneth Graunke wrote: > >> "STALL_AT_SCOREBOARD" is much clearer than "STALL_EN" now that there are > >> several different kinds of stalls. Also, "INSTRUCTION_CACHE_FLUSH" is a > >> lot easier to understand at a glance than the terse "IS_FLUSH." > >> > >> Signed-off-by: Kenneth Graunke <kenn...@whitecape.org> > >> --- > >> drivers/gpu/drm/i915/i915_reg.h | 16 ++++++++-------- > >> drivers/gpu/drm/i915/intel_ringbuffer.c | 6 ++++-- > >> 2 files changed, 12 insertions(+), 10 deletions(-) > >> > >> diff --git a/drivers/gpu/drm/i915/i915_reg.h > >> b/drivers/gpu/drm/i915/i915_reg.h > >> index d691781..bfe8488 100644 > >> --- a/drivers/gpu/drm/i915/i915_reg.h > >> +++ b/drivers/gpu/drm/i915/i915_reg.h > >> @@ -243,15 +243,15 @@ > >> #define DISPLAY_PLANE_A (0<<20) > >> #define DISPLAY_PLANE_B (1<<20) > >> #define GFX_OP_PIPE_CONTROL ((0x3<<29)|(0x3<<27)|(0x2<<24)) > >> -#define PIPE_CONTROL_QW_WRITE (1<<14) > >> -#define PIPE_CONTROL_DEPTH_STALL (1<<13) > >> -#define PIPE_CONTROL_WC_FLUSH (1<<12) > >> -#define PIPE_CONTROL_IS_FLUSH (1<<11) /* MBZ on Ironlake */ > >> -#define PIPE_CONTROL_TC_FLUSH (1<<10) /* GM45+ only */ > >> -#define PIPE_CONTROL_ISP_DIS (1<<9) > >> -#define PIPE_CONTROL_NOTIFY (1<<8) > >> +#define PIPE_CONTROL_QW_WRITE (1<<14) > >> +#define PIPE_CONTROL_DEPTH_STALL (1<<13) > >> +#define PIPE_CONTROL_WRITE_FLUSH (1<<12) > >> +#define PIPE_CONTROL_INSTRUCTION_CACHE_FLUSH (1<<11) /* MBZ on > >> Ironlake */ > >> +#define PIPE_CONTROL_TEXTURE_CACHE_FLUSH (1<<10) /* GM45+ only */ > > > > Minor bikeshed: You retain the _FLUSH for the read-only > > instruction/texture caches, but use _INVALIDATE for the new bits for > > read-only caches. I think we want _INVALIDATE for all of them. > > > > Otherwise, these three patches are: > > Reviewed-by: Daniel Vetter <daniel.vet...@ffwll.ch> > > Yeah. The docs for pre-SNB use "Flush" in all the field names, while > the docs for SNB+ switch to "Invalidate". So I just used it for the new > ones. If we'd rather go with "invalidate" for the old bits too, I can > do that.
Invalidate agrees more with our own cache tracking code and how the legacy MI_FLUSH is defined. So I vote for that because I like my bikesheds to be painted consistenly ;-) -Daniel -- Daniel Vetter Mail: dan...@ffwll.ch Mobile: +41 (0)79 365 57 48 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
