From: "Kumar, Shobhit" <shobhit.kumar@linux.intel.com>
To: Matt Roper <matthew.d.roper@intel.com>,
Shobhit Kumar <shobhit.kumar@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [v2 2/6] drm/i915/skl+: calculate ddb minimum allocation
Date: Tue, 9 Feb 2016 10:21:43 +0530 [thread overview]
Message-ID: <56B9705F.3030000@linux.intel.com> (raw)
In-Reply-To: <20160205142948.GV20935@intel.com>
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
>>
>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2016-02-09 4:51 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-27 16:09 [v2 1/6] drm/i915/skl+: Use plane size for relative data rate calculation Shobhit Kumar
2016-01-27 16:09 ` [v2 2/6] drm/i915/skl+: calculate ddb minimum allocation Shobhit Kumar
2016-02-05 14:29 ` Matt Roper
2016-02-09 4:51 ` Kumar, Shobhit [this message]
2016-01-27 16:10 ` [v2 3/6] drm/i915/skl+: calculate plane pixel rate Shobhit Kumar
2016-02-10 18:53 ` Matt Roper
2016-01-27 16:10 ` [v2 4/6] drm/i915/skl+: Use scaling amount for plane data rate calculation Shobhit Kumar
2016-02-10 19:39 ` Matt Roper
2016-02-11 8:43 ` Daniel Vetter
2016-01-27 16:10 ` [v2 5/6] drm/i915: Add support to parse DMI table and get platform memory info Shobhit Kumar
2016-02-10 22:29 ` Matt Roper
2016-01-27 16:10 ` [v2 6/6] drm/i915/skl: WA for watermark calculation based on Arbitrated Display BW Shobhit Kumar
2016-02-02 13:47 ` [v2 1/6] drm/i915/skl+: Use plane size for relative data rate calculation Kumar, Shobhit
2016-02-05 14:29 ` Matt Roper
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=56B9705F.3030000@linux.intel.com \
--to=shobhit.kumar@linux.intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=matthew.d.roper@intel.com \
--cc=shobhit.kumar@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).