From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Cc: intel-gfx@lists.freedesktop.org, intel-xe@lists.freedesktop.org,
suraj.kandpal@intel.com
Subject: Re: [PATCH 13/15] drm/i915: Compute config and mode valid changes for ultrajoiner
Date: Thu, 19 Sep 2024 21:30:09 +0300 [thread overview]
Message-ID: <ZuxtsRz00qagilSw@intel.com> (raw)
In-Reply-To: <20240918144343.2876184-14-ankit.k.nautiyal@intel.com>
On Wed, Sep 18, 2024 at 08:13:41PM +0530, Ankit Nautiyal wrote:
> From: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
>
> Implement required changes for mode validation and compute config,
> to support Ultrajoiner.
>
> v2:
> -Drop changes for HDMI.
> -Separate out DSC changes into another patch.
> v3: Fix check in can_ultrajoiner. (Ankit)
>
> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
> Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_dp.c | 44 +++++++++++++++++++++----
> 1 file changed, 37 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> index 369829ea5a12..4005700ab043 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> @@ -871,24 +871,34 @@ u32 get_max_compressed_bpp_with_joiner(struct drm_i915_private *i915,
> int num_joined_pipes)
> {
> u32 max_bpp_small_joiner_ram;
> + u32 max_bpp_joiner;
>
> /* Small Joiner Check: output bpp <= joiner RAM (bits) / Horiz. width */
> max_bpp_small_joiner_ram = small_joiner_ram_size_bits(i915) / mode_hdisplay;
> + max_bpp_joiner = max_bpp_small_joiner_ram;
>
> - if (num_joined_pipes == 2) {
> - int bigjoiner_interface_bits = DISPLAY_VER(i915) >= 14 ? 36 : 24;
> + /* if ultra joiner is enabled, we have 2 bigjoiners enabled */
> + if (num_joined_pipes == 2 ||
> + num_joined_pipes == 4) {
I guess just 'num_joined_pipes > 1' or something could be
used for all cases like this.
> + int joiner_interface_bits = DISPLAY_VER(i915) >= 14 ? 36 : 24;
Isn't this specifically about bigjoiner? If so the name should stay as
is.
> /* With bigjoiner multiple dsc engines are used in parallel so PPC is 2 */
> int ppc = 2;
> - u32 max_bpp_bigjoiner =
> - i915->display.cdclk.max_cdclk_freq * ppc * bigjoiner_interface_bits /
> + max_bpp_joiner =
> + i915->display.cdclk.max_cdclk_freq * ppc * joiner_interface_bits /
> intel_dp_mode_to_fec_clock(mode_clock);
>
> max_bpp_small_joiner_ram *= 2;
Can't we just multiply both max_* with num_joined_pipes
to take care of both bigjoiner and ultrajoiner?
>
> - return min(max_bpp_small_joiner_ram, max_bpp_bigjoiner);
> }
> + if (num_joined_pipes == 4) {
> + /* TODO: Check for ultrajoiner ram constraints */
>
> - return max_bpp_small_joiner_ram;
> + /* both get multiplied by 2, because ram bits/ppc now doubled */
> + max_bpp_small_joiner_ram *= 2;
> + max_bpp_joiner *= 2;
> + }
> +
> + return min(max_bpp_small_joiner_ram, max_bpp_joiner);
> }
>
> u16 intel_dp_dsc_get_max_compressed_bpp(struct drm_i915_private *i915,
> @@ -994,6 +1004,10 @@ u8 intel_dp_dsc_get_slice_count(const struct intel_connector *connector,
> if (num_joined_pipes == 2 && test_slice_count < 4)
> continue;
>
> + /* ultrajoiner needs 2 bigjoiners to be enabled */
> + if (num_joined_pipes == 4 && test_slice_count < 8)
> + continue;
> +
> if (min_slice_count <= test_slice_count)
> return test_slice_count;
> }
> @@ -1270,6 +1284,18 @@ intel_dp_mode_valid_downstream(struct intel_connector *connector,
> return MODE_OK;
> }
>
> +static
> +bool intel_dp_needs_ultrajoiner(struct intel_dp *dp, int clock)
> +{
> + const struct intel_encoder *encoder = &dp_to_dig_port(dp)->base;
> + struct drm_i915_private *i915 = to_i915(encoder->base.dev);
struct intel_display *display = intel_display(intel_dp);
> +
> + if (!HAS_ULTRAJOINER(i915))
> + return false;
> +
> + return clock > (i915->display.cdclk.max_dotclk_freq * 2);
Redundant parens.
> +}
> +
> static
> bool intel_dp_needs_bigjoiner(struct intel_dp *intel_dp,
> struct intel_connector *connector,
> @@ -1296,6 +1322,8 @@ int intel_dp_compute_num_pipes(struct intel_dp *intel_dp,
> MISSING_CASE(connector->force_joined_pipes);
> fallthrough;
> case 0:
> + if (intel_dp_needs_ultrajoiner(intel_dp, clock))
Hmm. Technically we should be checking the same things for both
bigjoiner and ultrajoiner.
How about something like this with parametrized number
of pipes:
bool intel_dp_needs_joiner(struct intel_dp *intel_dp,
struct intel_connector *connector,
int hdisplay, int clock,
int num_joined_pipes)
{
if (!intel_dp_has_joiner(intel_dp))
return false;
num_joined_pipes /= 2;
return clock > num_joined_pipes * max_dotclk_freq ||
hdisplay > num_joined_pipes * 5120;
}
Pair that with HAS_*JOINER(), and pass in the correct number
of joined pipes to get final answer.
> + return 4;
> if (intel_dp_needs_bigjoiner(intel_dp, connector, hdisplay, clock))
> return 2;
> }
> @@ -2542,8 +2570,10 @@ bool intel_dp_joiner_needs_dsc(struct drm_i915_private *i915,
> * Pipe joiner needs compression up to display 12 due to bandwidth
> * limitation. DG2 onwards pipe joiner can be enabled without
> * compression.
> + * Ultrajoiner always needs compression.
> */
> - return !HAS_UNCOMPRESSED_JOINER(i915) && num_joined_pipes == 2;
> + return (!HAS_UNCOMPRESSED_JOINER(i915) && num_joined_pipes == 2) ||
> + num_joined_pipes == 4;
> }
>
> static int
> --
> 2.45.2
--
Ville Syrjälä
Intel
next prev parent reply other threads:[~2024-09-19 18:30 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-18 14:43 [PATCH 00/15] Ultrajoiner basic functionality series Ankit Nautiyal
2024-09-18 14:43 ` [PATCH 01/15] drm/i915: Add some essential functionality for joiners Ankit Nautiyal
2024-09-18 14:43 ` [PATCH 02/15] drm/i915/display: Enhance iterators for modeset en/disable Ankit Nautiyal
2024-09-18 14:43 ` [PATCH 03/15] drm/i915/display_debugfs: Allow force joiner only if supported Ankit Nautiyal
2024-09-19 15:04 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 04/15] drm/i915/display: Modify debugfs for joiner to force n pipes Ankit Nautiyal
2024-09-19 15:07 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 05/15] drm/i915/dp: Add helper to compute num pipes required Ankit Nautiyal
2024-09-19 15:12 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 06/15] drm/i915/display: Add debugfs support to avoid joiner Ankit Nautiyal
2024-09-19 15:15 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 07/15] drm/i915: Split current joiner hw state readout Ankit Nautiyal
2024-09-19 15:18 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 08/15] drm/i915: Add bigjoiner and uncompressed joiner hw readout sanity checks Ankit Nautiyal
2024-09-19 15:22 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 09/15] drm/i915/display: Add macro HAS_ULTRAJOINER() Ankit Nautiyal
2024-09-19 15:24 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 10/15] drm/i915: Implement hw state readout and checks for ultrajoiner Ankit Nautiyal
2024-09-19 15:33 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 11/15] drm/i915/display: Refactor enable_joiner_pipes Ankit Nautiyal
2024-09-19 18:02 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 12/15] drm/i915/display/vdsc: Add ultrajoiner support with DSC Ankit Nautiyal
2024-09-19 18:06 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 13/15] drm/i915: Compute config and mode valid changes for ultrajoiner Ankit Nautiyal
2024-09-19 18:30 ` Ville Syrjälä [this message]
2024-09-18 14:43 ` [PATCH 14/15] drm/i915/display: Consider ultrajoiner for computing maxdotclock Ankit Nautiyal
2024-09-19 18:42 ` Ville Syrjälä
2024-09-18 14:43 ` [PATCH 15/15] drm/i915/intel_dp: Add support for forcing ultrajoiner Ankit Nautiyal
2024-09-19 18:45 ` Ville Syrjälä
2024-09-18 15:22 ` ✗ Fi.CI.CHECKPATCH: warning for Ultrajoiner basic functionality series (rev9) Patchwork
2024-09-18 15:22 ` ✗ Fi.CI.SPARSE: " Patchwork
2024-09-18 15:30 ` ✓ Fi.CI.BAT: success " Patchwork
2024-09-18 16:28 ` ✓ CI.Patch_applied: success for Ultrajoiner basic functionality series Patchwork
2024-09-18 16:29 ` ✗ CI.checkpatch: warning " Patchwork
2024-09-18 16:30 ` ✓ CI.KUnit: success " Patchwork
2024-09-18 16:42 ` ✓ CI.Build: " Patchwork
2024-09-18 16:44 ` ✓ CI.Hooks: " Patchwork
2024-09-18 16:46 ` ✗ CI.checksparse: warning " Patchwork
2024-09-18 17:10 ` ✓ CI.BAT: success " Patchwork
2024-09-19 1:52 ` ✗ CI.FULL: failure " Patchwork
2024-09-19 5:01 ` ✗ Fi.CI.IGT: failure for Ultrajoiner basic functionality series (rev9) Patchwork
-- strict thread matches above, loose matches on Subject: below --
2024-09-26 7:26 [PATCH 00/15] Ultrajoiner basic functionality series Ankit Nautiyal
2024-09-26 7:26 ` [PATCH 13/15] drm/i915: Compute config and mode valid changes for ultrajoiner Ankit Nautiyal
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=ZuxtsRz00qagilSw@intel.com \
--to=ville.syrjala@linux.intel.com \
--cc=ankit.k.nautiyal@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=suraj.kandpal@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.