Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Hogander, Jouni" <jouni.hogander@intel.com>
To: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
	"dri-devel@lists.freedesktop.org"
	<dri-devel@lists.freedesktop.org>,
	"Manna, Animesh" <animesh.manna@intel.com>,
	"intel-gfx@lists.freedesktop.org"
	<intel-gfx@lists.freedesktop.org>
Cc: "Kandpal, Suraj" <suraj.kandpal@intel.com>,
	"Nikula, Jani" <jani.nikula@intel.com>
Subject: Re: [PATCH v6 06/18] drm/i915/alpm: Auxless wake time calculation for Xe3p
Date: Tue, 29 Sep 2026 05:34:04 +0000	[thread overview]
Message-ID: <5cef6cb6f540b893730076330c1b527197ed3504.camel@intel.com> (raw)
In-Reply-To: <20260915180449.277933-7-animesh.manna@intel.com>

On Tue, 2026-09-15 at 23:34 +0530, Animesh Manna wrote:
> Add support for auxless waketime calculation for DP2.1 ALPM
> as dependent parameter got changed.
> 
> v1: Initial version.
> v2: Use intel_dp_is_uhbr(). [Jani]
> v3:
> - Separate out switch to active latency calculation.
> - Use macro definition for ML_PHY_LOCK length. [Suraj]
> v4:
> - Fix 32-bit signed integer overflow. [sashiko]
> - Fix negative return of drm_dp_lttpr_count(). [sashiko]
> - Introduce get_ml_phy_lock_len() and get_dp_link_symbol_size().
> [Jouni]

My comment was referring existing intel_dp.c:intel_dp_link_symbol_size

> - Align variable naming with bpec formula of establishment_period.
> [Jouni]
> 
> Bspec: 71477
> Cc: Jouni Högander <jouni.hogander@intel.com>
> Signed-off-by: Animesh Manna <animesh.manna@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_alpm.c | 69 +++++++++++++++++++--
> --
>  1 file changed, 58 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c
> b/drivers/gpu/drm/i915/display/intel_alpm.c
> index d5823174576a..924fa8dd4638 100644
> --- a/drivers/gpu/drm/i915/display/intel_alpm.c
> +++ b/drivers/gpu/drm/i915/display/intel_alpm.c
> @@ -102,21 +102,64 @@ static int get_lfps_half_cycle_clocks(const
> struct intel_crtc_state *crtc_state)
>  		1000 / (2 * LFPS_CYCLE_COUNT);
>  }
>  
> +#define ML_PHY_LOCK_LEN		252
> +#define ML_PHY_LOCK_LEN_UHBR	396
> +
>  static int get_tphy2_p2_to_p0(const struct intel_crtc_state
> *crtc_state)
>  {
> -	return 12 * 1000;
> +	struct intel_display *display =
> to_intel_display(crtc_state);
> +
> +	return DISPLAY_VER(display) >= 35 ? (20 * 1000) : (12 *
> 1000);

This should actually use HAS_LT_PHY().

> +}
> +
> +static int get_ml_phy_lock_len(const struct intel_crtc_state
> *crtc_state)
> +{
> +	return intel_dp_is_uhbr(crtc_state) ? ML_PHY_LOCK_LEN_UHBR :
> ML_PHY_LOCK_LEN;
> +}
> +
> +static int get_dp_link_symbol_size(const struct intel_crtc_state
> *crtc_state)
> +{
> +	return intel_dp_is_uhbr(crtc_state) ? 32 : 10;
>  }

This is duplicate of intel_dp_link_symbol_size.

>  
> -static int get_establishment_period(const struct intel_crtc_state
> *crtc_state)
> +static int get_establishment_period(struct intel_dp *intel_dp,
> +				    const struct intel_crtc_state
> *crtc_state)
>  {
>  	int t1 = 50 * 1000;
> -	int tps4 = 252;
> -	/* port_clock is link rate in 10kbit/s units */
> -	int tml_phy_lock = 1000 * 1000 * tps4 / crtc_state-
> >port_clock;
> +	int tps4 = get_ml_phy_lock_len(crtc_state) *
> get_dp_link_symbol_size(crtc_state);
> +	/*
> +	 * port_clock is link rate in 10kbit/s units, so multiplying
> by
> +	 * (1000 * 1000)/10 = 100000
> +	 */
> +	int tml_phy_lock = 100000 * tps4 / crtc_state->port_clock;

tml_phy_lock:

For 8b/10b: 
252 Symbol clocks * ( 10 / (Link Rate in Mbps) )

For 128b/132b:
396 Super Symbol clocks * ( 32/ (Link Rate in Mbps) )

In comment you should have:
* (10000 * 10000)/intel_dp_link_symbol_size()

> +	int lttpr_count = 0;
>  	int tcds, establishment_period;
>  
> -	tcds = (7 + DIV_ROUND_UP(6500, tml_phy_lock) + 1) *
> tml_phy_lock;
> -	establishment_period = (SILENCE_PERIOD_TIME + t1 + tcds);
> +	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP)) {
> +		tcds = (7 + DIV_ROUND_UP(6500, tml_phy_lock) + 1) *
> tml_phy_lock;
> +	} else {
> +		tcds = 7 * tml_phy_lock;
> +		lttpr_count = drm_dp_lttpr_count(intel_dp-
> >lttpr_common_caps);
> +	}
> +
> +	if (lttpr_count > 0) {
> +		/* tlw and tcs values are taken from DP2.1 spec */
> +		int tlw = 13000;
> +		int tcs = 10000;
> +		int tlfps_period = get_lfps_cycle_time(crtc_state);
> +		int tdcs = (SILENCE_PERIOD_TIME + t1 + tcs +
> +			    (lttpr_count - 1) * (tlw +
> tlfps_period));
> +		int tacds = 70000;
> +		int tds = 7 * tml_phy_lock;
> +
> +		/* tdrl is same as tcds*/
> +		establishment_period = tlw + tlfps_period + tdcs +
> tacds +
> +				       (lttpr_count - 1) * tds +
> tcds;
> +	} else {
> +		/* TODO: Add a check for data realign by DPCD
> 0x116[3] */
> +
> +		establishment_period = (SILENCE_PERIOD_TIME + t1 +
> tcds);
> +	}
>  
>  	return establishment_period;
>  }
> @@ -138,12 +181,16 @@ static int get_establishment_period(const
> struct intel_crtc_state *crtc_state)
>   * within the CDS period complete within the CDS period regardless
> of
>   * entry into the period
>   * tML_PHY_LOCK = TPS4 Length * ( 10 / (Link Rate in MHz) )

Here also switch 10 as Rate Div.

BR,
Jouni Högander

> - * TPS4 Length = 252 Symbols
> + * ML_PHY_LOCK Length (8b/10b): 252
> + * ML_PHY_LOCK Length (128b/132b): 396
> + * TPS4 Length = ML_PHY_LOCK * Rate Div
> + * Rate Div = 32 for 128b/132b and 10 for 8b/10b.
>   */
> -static int _lnl_compute_aux_less_wake_time(const struct
> intel_crtc_state *crtc_state)
> +static int _lnl_compute_aux_less_wake_time(struct intel_dp
> *intel_dp,
> +					   const struct
> intel_crtc_state *crtc_state)
>  {
>  	int tphy2_p2_to_p0 = get_tphy2_p2_to_p0(crtc_state);
> -	int establishment_period =
> get_establishment_period(crtc_state);
> +	int establishment_period =
> get_establishment_period(intel_dp, crtc_state);
>  
>  	return DIV_ROUND_UP(tphy2_p2_to_p0 +
> get_lfps_cycle_time(crtc_state) +
>  			    establishment_period, 1000);
> @@ -158,7 +205,7 @@ _lnl_compute_aux_less_alpm_params(struct intel_dp
> *intel_dp,
>  		lfps_half_cycle;
>  
>  	aux_less_wake_time =
> -		_lnl_compute_aux_less_wake_time(crtc_state);
> +		_lnl_compute_aux_less_wake_time(intel_dp,
> crtc_state);
>  	aux_less_wake_lines = intel_usecs_to_scanlines(&crtc_state-
> >hw.adjusted_mode,
>  						      
> aux_less_wake_time);
>  	silence_period = get_silence_period_symbols(crtc_state);


  reply	other threads:[~2026-09-29  5:34 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-15 18:04 [PATCH v6 00/18] Enable DP2.1 alpm Animesh Manna
2026-09-15 18:04 ` [PATCH v6 01/18] drm/i915/alpm: Add DPCD definition for DP2.1 ALPM capability Animesh Manna
2026-09-15 18:04 ` [PATCH v6 02/18] drm/i915/alpm: Move alpm sink capability readout into a separate function Animesh Manna
2026-09-28  9:58   ` Hogander, Jouni
2026-09-28 10:02     ` Hogander, Jouni
2026-09-30  8:04   ` Jani Nikula
2026-10-01 11:21     ` Manna, Animesh
2026-09-15 18:04 ` [PATCH v6 03/18] drm/i915/alpm: alpm_init() for DP2.1 Animesh Manna
2026-09-15 18:57   ` sashiko-bot
2026-09-28 10:44   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 04/18] drm/i915/alpm: Enable debugfs " Animesh Manna
2026-09-28 11:22   ` Hogander, Jouni
2026-10-01 11:32     ` Manna, Animesh
2026-10-08  4:40       ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 05/18] drm/i915/alpm: Refactor Auxless wake time calculation Animesh Manna
2026-09-28 11:49   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 06/18] drm/i915/alpm: Auxless wake time calculation for Xe3p Animesh Manna
2026-09-29  5:34   ` Hogander, Jouni [this message]
2026-09-29  5:38   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 07/18] drm/i915/alpm: Modify AUX_LESS_WAKE_TIME bitfield for xe3lpd Animesh Manna
2026-09-15 18:46   ` sashiko-bot
2026-09-29  6:20   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 08/18] drm/i915/alpm: table based establishment period Animesh Manna
2026-09-29  7:00   ` Hogander, Jouni
2026-09-29  7:31     ` Hogander, Jouni
2026-10-01 11:40     ` Manna, Animesh
2026-09-15 18:04 ` [PATCH v6 09/18] drm/i915/alpm: Half LFPS cycle calculation Animesh Manna
2026-09-15 18:53   ` sashiko-bot
2026-09-29  9:57   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 10/18] drm/i915/alpm: Modify LFPS cycle count for DP ALPM Animesh Manna
2026-10-01  6:22   ` Hogander, Jouni
2026-10-01 12:10     ` Hogander, Jouni
2026-10-01 12:11       ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 11/18] drm/i915/alpm: Modify LFPS_CYCLE_COUNT bitfield for xe3lpd Animesh Manna
2026-09-15 18:04 ` [PATCH v6 12/18] drm/i915/alpm: Program LTTPR count for DP 2.1 ALPM Animesh Manna
2026-09-15 18:56   ` sashiko-bot
2026-09-30  8:16   ` Jani Nikula
2026-10-02  4:41   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 13/18] drm/i915/alpm: Enable MAC Transmitting LFPS for LT PHY Animesh Manna
2026-10-02  5:04   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 14/18] drm/i915/alpm: Replace is_edp() with alpm_is_possible() Animesh Manna
2026-09-15 18:57   ` sashiko-bot
2026-10-02  6:41   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 15/18] drm/i915/alpm: Introduce has_alpm to decouple from pr/psr2/lobf Animesh Manna
2026-09-15 19:01   ` sashiko-bot
2026-10-02  8:37   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 16/18] drm/i915/alpm: Compute and program switch to active latency Animesh Manna
2026-09-15 18:55   ` sashiko-bot
2026-10-02  9:00   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 17/18] drm/i915/alpm: Program zero-based LFPS half cycle duration Animesh Manna
2026-09-15 19:02   ` sashiko-bot
2026-10-02  9:08   ` Hogander, Jouni
2026-09-15 18:04 ` [PATCH v6 18/18] drm/i915/alpm: Disallow ALPM for UHBR link rate Animesh Manna
2026-09-15 19:01   ` sashiko-bot
2026-10-02 11:33   ` Hogander, Jouni
2026-09-15 21:11 ` ✓ i915.CI.BAT: success for Enable DP2.1 alpm (rev6) Patchwork
2026-09-16  8:14 ` ✗ i915.CI.Full: failure " Patchwork

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=5cef6cb6f540b893730076330c1b527197ed3504.camel@intel.com \
    --to=jouni.hogander@intel.com \
    --cc=animesh.manna@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=jani.nikula@intel.com \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox