| Message ID | 1453911003-9856-2-git-send-email-shobhit.kumar@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, Jan 27, 2016 at 09:39:59PM +0530, Shobhit Kumar wrote: > From: "Kumar, Mahesh" <mahesh1.kumar@intel.com> > > don't always use 8 ddb as minimum, instead calculate using proper > algorithm. > > v2: optimizations as per Matt's comments. > > Cc: matthew.d.roper@intel.com > Signed-off-by: Kumar, Mahesh <mahesh1.kumar@intel.com> > --- > drivers/gpu/drm/i915/intel_pm.c | 50 ++++++++++++++++++++++++++++++++++++++--- > 1 file changed, 47 insertions(+), 3 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index d55e5d0..708f329 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -2928,6 +2928,51 @@ skl_get_total_relative_data_rate(const struct intel_crtc_state *cstate) > return total_data_rate; > } > > +static uint16_t > +skl_ddb_min_alloc(const struct intel_crtc *crtc, > + const struct drm_plane *plane, int y) > +{ > + struct drm_framebuffer *fb = plane->state->fb; > + struct intel_plane_state *pstate = to_intel_plane_state(plane->state); > + uint32_t src_w, src_h; > + uint32_t min_scanlines = 8; > + uint8_t bytes_per_pixel; > + > + /* For packed formats, no y-plane, return 0 */ > + if (y && !fb && !(fb->pixel_format == DRM_FORMAT_NV12)) I think you meant if (!fb || (y && fb->pixel_format != DRM_FORMAT_NV12)) right? > + return 0; > + > + /* For Non Y-tile return 8-blocks */ > + if (!(fb->modifier[0] == I915_FORMAT_MOD_Y_TILED) && > + !(fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED)) > + return 8; Minor nitpick, but it might be more readable to use != instead of pulling the ! outside of the comparison. On my first quick read I missed the inversions. Matt > + > + src_w = drm_rect_width(&pstate->src) >> 16; > + src_h = drm_rect_height(&pstate->src) >> 16; > + > + if (intel_rotation_90_or_270(plane->state->rotation)) > + swap(src_w, src_h); > + > + bytes_per_pixel = y ? drm_format_plane_cpp(fb->pixel_format, 0) : > + drm_format_plane_cpp(fb->pixel_format, 1); > + > + if (intel_rotation_90_or_270(plane->state->rotation)) { > + switch (bytes_per_pixel) { > + case 1: > + min_scanlines = 32; > + break; > + case 2: > + min_scanlines = 16; > + break; > + case 8: > + WARN(1, "Unsupported pixel depth for rotation"); > + } > + } > + > + return DIV_ROUND_UP((4 * src_w / (y ? 1 : 2) * bytes_per_pixel), 512) * > + min_scanlines/4 + 3; > +} > + > static void > skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, > struct skl_ddb_allocation *ddb /* out */) > @@ -2964,7 +3009,6 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, > /* 1. Allocate the mininum required blocks for each active plane */ > for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) { > struct drm_plane *plane = &intel_plane->base; > - struct drm_framebuffer *fb = plane->state->fb; > int id = skl_wm_plane_id(intel_plane); > > if (!to_intel_plane_state(plane->state)->visible) > @@ -2973,9 +3017,9 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, > if (plane->type == DRM_PLANE_TYPE_CURSOR) > continue; > > - minimum[id] = 8; > + minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 0); > alloc_size -= minimum[id]; > - y_minimum[id] = (fb->pixel_format == DRM_FORMAT_NV12) ? 8 : 0; > + y_minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 1); > alloc_size -= y_minimum[id]; > } > > -- > 2.5.0 >
On 02/05/2016 07:59 PM, Matt Roper wrote: > On Wed, Jan 27, 2016 at 09:39:59PM +0530, Shobhit Kumar wrote: >> From: "Kumar, Mahesh" <mahesh1.kumar@intel.com> >> >> don't always use 8 ddb as minimum, instead calculate using proper >> algorithm. >> >> v2: optimizations as per Matt's comments. >> >> Cc: matthew.d.roper@intel.com >> Signed-off-by: Kumar, Mahesh <mahesh1.kumar@intel.com> >> --- >> drivers/gpu/drm/i915/intel_pm.c | 50 ++++++++++++++++++++++++++++++++++++++--- >> 1 file changed, 47 insertions(+), 3 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index d55e5d0..708f329 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -2928,6 +2928,51 @@ skl_get_total_relative_data_rate(const struct intel_crtc_state *cstate) >> return total_data_rate; >> } >> >> +static uint16_t >> +skl_ddb_min_alloc(const struct intel_crtc *crtc, >> + const struct drm_plane *plane, int y) >> +{ >> + struct drm_framebuffer *fb = plane->state->fb; >> + struct intel_plane_state *pstate = to_intel_plane_state(plane->state); >> + uint32_t src_w, src_h; >> + uint32_t min_scanlines = 8; >> + uint8_t bytes_per_pixel; >> + >> + /* For packed formats, no y-plane, return 0 */ >> + if (y && !fb && !(fb->pixel_format == DRM_FORMAT_NV12)) > > I think you meant > > if (!fb || (y && fb->pixel_format != DRM_FORMAT_NV12)) > > right? > Yeah, looks like I made a blunder while testing and adding some fb checks in few places. This is one place it got messed up. >> + return 0; >> + >> + /* For Non Y-tile return 8-blocks */ >> + if (!(fb->modifier[0] == I915_FORMAT_MOD_Y_TILED) && >> + !(fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED)) >> + return 8; > > Minor nitpick, but it might be more readable to use != instead of > pulling the ! outside of the comparison. On my first quick read I > missed the inversions. Yeah will do. > > > Matt > >> + >> + src_w = drm_rect_width(&pstate->src) >> 16; >> + src_h = drm_rect_height(&pstate->src) >> 16; >> + >> + if (intel_rotation_90_or_270(plane->state->rotation)) >> + swap(src_w, src_h); >> + >> + bytes_per_pixel = y ? drm_format_plane_cpp(fb->pixel_format, 0) : >> + drm_format_plane_cpp(fb->pixel_format, 1); >> + >> + if (intel_rotation_90_or_270(plane->state->rotation)) { >> + switch (bytes_per_pixel) { >> + case 1: >> + min_scanlines = 32; >> + break; >> + case 2: >> + min_scanlines = 16; >> + break; >> + case 8: >> + WARN(1, "Unsupported pixel depth for rotation"); >> + } >> + } >> + >> + return DIV_ROUND_UP((4 * src_w / (y ? 1 : 2) * bytes_per_pixel), 512) * >> + min_scanlines/4 + 3; >> +} >> + >> static void >> skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, >> struct skl_ddb_allocation *ddb /* out */) >> @@ -2964,7 +3009,6 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, >> /* 1. Allocate the mininum required blocks for each active plane */ >> for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) { >> struct drm_plane *plane = &intel_plane->base; >> - struct drm_framebuffer *fb = plane->state->fb; >> int id = skl_wm_plane_id(intel_plane); >> >> if (!to_intel_plane_state(plane->state)->visible) >> @@ -2973,9 +3017,9 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, >> if (plane->type == DRM_PLANE_TYPE_CURSOR) >> continue; >> >> - minimum[id] = 8; >> + minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 0); >> alloc_size -= minimum[id]; >> - y_minimum[id] = (fb->pixel_format == DRM_FORMAT_NV12) ? 8 : 0; >> + y_minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 1); >> alloc_size -= y_minimum[id]; >> } >> >> -- >> 2.5.0 >> >
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index d55e5d0..708f329 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -2928,6 +2928,51 @@ skl_get_total_relative_data_rate(const struct intel_crtc_state *cstate) return total_data_rate; } +static uint16_t +skl_ddb_min_alloc(const struct intel_crtc *crtc, + const struct drm_plane *plane, int y) +{ + struct drm_framebuffer *fb = plane->state->fb; + struct intel_plane_state *pstate = to_intel_plane_state(plane->state); + uint32_t src_w, src_h; + uint32_t min_scanlines = 8; + uint8_t bytes_per_pixel; + + /* For packed formats, no y-plane, return 0 */ + if (y && !fb && !(fb->pixel_format == DRM_FORMAT_NV12)) + return 0; + + /* For Non Y-tile return 8-blocks */ + if (!(fb->modifier[0] == I915_FORMAT_MOD_Y_TILED) && + !(fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED)) + return 8; + + src_w = drm_rect_width(&pstate->src) >> 16; + src_h = drm_rect_height(&pstate->src) >> 16; + + if (intel_rotation_90_or_270(plane->state->rotation)) + swap(src_w, src_h); + + bytes_per_pixel = y ? drm_format_plane_cpp(fb->pixel_format, 0) : + drm_format_plane_cpp(fb->pixel_format, 1); + + if (intel_rotation_90_or_270(plane->state->rotation)) { + switch (bytes_per_pixel) { + case 1: + min_scanlines = 32; + break; + case 2: + min_scanlines = 16; + break; + case 8: + WARN(1, "Unsupported pixel depth for rotation"); + } + } + + return DIV_ROUND_UP((4 * src_w / (y ? 1 : 2) * bytes_per_pixel), 512) * + min_scanlines/4 + 3; +} + static void skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, struct skl_ddb_allocation *ddb /* out */) @@ -2964,7 +3009,6 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, /* 1. Allocate the mininum required blocks for each active plane */ for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) { struct drm_plane *plane = &intel_plane->base; - struct drm_framebuffer *fb = plane->state->fb; int id = skl_wm_plane_id(intel_plane); if (!to_intel_plane_state(plane->state)->visible) @@ -2973,9 +3017,9 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, if (plane->type == DRM_PLANE_TYPE_CURSOR) continue; - minimum[id] = 8; + minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 0); alloc_size -= minimum[id]; - y_minimum[id] = (fb->pixel_format == DRM_FORMAT_NV12) ? 8 : 0; + y_minimum[id] = skl_ddb_min_alloc(intel_crtc, plane, 1); alloc_size -= y_minimum[id]; }