From: "Lisovskiy, Stanislav" <stanislav.lisovskiy@intel.com>
To: Matt Roper <matthew.d.roper@intel.com>
Cc: intel-gfx@lists.freedesktop.org,
Lucas De Marchi <lucas.demarchi@intel.com>
Subject: Re: [Intel-gfx] [PATCH v4 01/23] drm/i915/xelpd: Enhanced pipe underrun reporting
Date: Mon, 17 May 2021 09:52:59 +0300 [thread overview]
Message-ID: <20210517065259.GB17429@intel.com> (raw)
In-Reply-To: <20210515031035.2561658-2-matthew.d.roper@intel.com>
On Fri, May 14, 2021 at 08:10:13PM -0700, Matt Roper wrote:
> XE_LPD brings enhanced underrun recovery: the hardware can somewhat
> mitigate underruns by using an interpolated replacement pixel (soft
> underrun) or the previous pixel (hard underrun). Furthermore, underruns
> can now be caused downstream by the port, even if the pipe itself is
> operating properly. The interrupt register and PIPE_STATUS register
> give us extra bits to recognize hard/soft underruns and determine
> whether the underrun was caused by the port, so we'll use that
> information to print some more descriptive errors when underruns occur.
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
>
> v2:
> - Keep ICL's PIPE_STATUS defined separately from the old GMCH pipe
> status register. (Ville)
> - Only read/clear the PIPE_STATUS register on platforms with
> display ver >= 11. (Lucas)
> v3:
> - Actually enable+unmask all the new underrun interrupts, clear stale
> bits out from PIPE_STATUS before enabling the interrupts, report all
> FIFO underruns errors at once, rename a bunch of stuff to unconfuse
> vs. PIPESTAT. (Ville)
>
> Bspec: 50335
> Bspec: 50366
> Cc: Lucas De Marchi <lucas.demarchi@intel.com>
> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> ---
> .../drm/i915/display/intel_fifo_underrun.c | 57 +++++++++++++++++--
> drivers/gpu/drm/i915/i915_irq.c | 19 ++++++-
> drivers/gpu/drm/i915/i915_irq.h | 1 +
> drivers/gpu/drm/i915/i915_reg.h | 9 +++
> 4 files changed, 77 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_fifo_underrun.c b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c
> index 3315aa1d4d5a..eb841960840d 100644
> --- a/drivers/gpu/drm/i915/display/intel_fifo_underrun.c
> +++ b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c
> @@ -185,15 +185,34 @@ static void ivb_set_fifo_underrun_reporting(struct drm_device *dev,
> }
> }
>
> +static u32
> +icl_pipe_status_underrun_mask(struct drm_i915_private *dev_priv)
> +{
> + u32 mask = PIPE_STATUS_UNDERRUN;
> +
> + if (DISPLAY_VER(dev_priv) >= 13)
> + mask |= PIPE_STATUS_SOFT_UNDERRUN_XELPD |
> + PIPE_STATUS_HARD_UNDERRUN_XELPD |
> + PIPE_STATUS_PORT_UNDERRUN_XELPD;
> +
> + return mask;
> +}
> +
> static void bdw_set_fifo_underrun_reporting(struct drm_device *dev,
> enum pipe pipe, bool enable)
> {
> struct drm_i915_private *dev_priv = to_i915(dev);
> + u32 mask = gen8_de_pipe_underrun_mask(dev_priv);
>
> - if (enable)
> - bdw_enable_pipe_irq(dev_priv, pipe, GEN8_PIPE_FIFO_UNDERRUN);
> - else
> - bdw_disable_pipe_irq(dev_priv, pipe, GEN8_PIPE_FIFO_UNDERRUN);
> + if (enable) {
> + if (DISPLAY_VER(dev_priv) >= 11)
> + intel_de_write(dev_priv, ICL_PIPESTATUS(pipe),
> + icl_pipe_status_underrun_mask(dev_priv));
> +
> + bdw_enable_pipe_irq(dev_priv, pipe, mask);
> + } else {
> + bdw_disable_pipe_irq(dev_priv, pipe, mask);
> + }
> }
>
> static void ibx_set_fifo_underrun_reporting(struct drm_device *dev,
> @@ -373,6 +392,7 @@ void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
> enum pipe pipe)
> {
> struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe);
> + u32 underruns = 0;
>
> /* We may be called too early in init, thanks BIOS! */
> if (crtc == NULL)
> @@ -383,10 +403,35 @@ void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
> crtc->cpu_fifo_underrun_disabled)
> return;
>
> + /*
> + * Starting with display version 11, the PIPE_STAT register records
> + * whether an underrun has happened, and on XELPD+, it will also record
> + * whether the underrun was soft/hard and whether it was triggered by
> + * the downstream port logic. We should clear these bits (which use
> + * write-1-to-clear logic) too.
> + *
> + * Note that although the IIR gives us the same underrun and soft/hard
> + * information, PIPE_STAT is the only place we can find out whether
> + * the underrun was caused by the downstream port.
> + */
> + if (DISPLAY_VER(dev_priv) >= 11) {
> + underruns = intel_de_read(dev_priv, ICL_PIPESTATUS(pipe)) &
> + icl_pipe_status_underrun_mask(dev_priv);
> + intel_de_write(dev_priv, ICL_PIPESTATUS(pipe), underruns);
> + }
> +
> if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false)) {
> trace_intel_cpu_fifo_underrun(dev_priv, pipe);
> - drm_err(&dev_priv->drm, "CPU pipe %c FIFO underrun\n",
> - pipe_name(pipe));
> +
> + if (DISPLAY_VER(dev_priv) >= 11)
> + drm_err(&dev_priv->drm, "CPU pipe %c FIFO underrun: %s%s%s%s\n",
> + pipe_name(pipe),
> + underruns & PIPE_STATUS_SOFT_UNDERRUN_XELPD ? "soft," : "",
> + underruns & PIPE_STATUS_HARD_UNDERRUN_XELPD ? "hard," : "",
> + underruns & PIPE_STATUS_PORT_UNDERRUN_XELPD ? "port," : "",
> + underruns & PIPE_STATUS_UNDERRUN ? "transcoder," : "");
> + else
> + drm_err(&dev_priv->drm, "CPU pipe %c FIFO underrun\n", pipe_name(pipe));
> }
>
> intel_fbc_handle_fifo_underrun_irq(dev_priv);
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index d4611c643446..957d401186d2 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2425,6 +2425,17 @@ static u32 gen8_de_pipe_flip_done_mask(struct drm_i915_private *i915)
> return GEN8_PIPE_PRIMARY_FLIP_DONE;
> }
>
> +u32 gen8_de_pipe_underrun_mask(struct drm_i915_private *dev_priv)
> +{
> + u32 mask = GEN8_PIPE_FIFO_UNDERRUN;
> +
> + if (DISPLAY_VER(dev_priv) >= 13)
> + mask |= XELPD_PIPE_SOFT_UNDERRUN |
> + XELPD_PIPE_HARD_UNDERRUN;
> +
> + return mask;
> +}
> +
> static irqreturn_t
> gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
> {
> @@ -2536,7 +2547,7 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
> if (iir & GEN8_PIPE_CDCLK_CRC_DONE)
> hsw_pipe_crc_irq_handler(dev_priv, pipe);
>
> - if (iir & GEN8_PIPE_FIFO_UNDERRUN)
> + if (iir & gen8_de_pipe_underrun_mask(dev_priv))
> intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
>
> fault_errors = iir & gen8_de_pipe_fault_mask(dev_priv);
> @@ -3173,7 +3184,8 @@ void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
> u8 pipe_mask)
> {
> struct intel_uncore *uncore = &dev_priv->uncore;
> - u32 extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN |
> + u32 extra_ier = GEN8_PIPE_VBLANK |
> + gen8_de_pipe_underrun_mask(dev_priv) |
> gen8_de_pipe_flip_done_mask(dev_priv);
> enum pipe pipe;
>
> @@ -3757,7 +3769,8 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
> }
>
> de_pipe_enables = de_pipe_masked |
> - GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN |
> + GEN8_PIPE_VBLANK |
> + gen8_de_pipe_underrun_mask(dev_priv) |
> gen8_de_pipe_flip_done_mask(dev_priv);
>
> de_port_enables = de_port_masked;
> diff --git a/drivers/gpu/drm/i915/i915_irq.h b/drivers/gpu/drm/i915/i915_irq.h
> index 25f25cd95818..db34d5dbe402 100644
> --- a/drivers/gpu/drm/i915/i915_irq.h
> +++ b/drivers/gpu/drm/i915/i915_irq.h
> @@ -100,6 +100,7 @@ void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
> u8 pipe_mask);
> void gen8_irq_power_well_pre_disable(struct drm_i915_private *dev_priv,
> u8 pipe_mask);
> +u32 gen8_de_pipe_underrun_mask(struct drm_i915_private *dev_priv);
>
> bool intel_crtc_get_vblank_timestamp(struct drm_crtc *crtc, int *max_error,
> ktime_t *vblank_time, bool in_vblank_irq);
> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> index 089b5a59bed3..506a5aa0646b 100644
> --- a/drivers/gpu/drm/i915/i915_reg.h
> +++ b/drivers/gpu/drm/i915/i915_reg.h
> @@ -6158,6 +6158,13 @@ enum {
> #define SKL_BOTTOM_COLOR_CSC_ENABLE (1 << 30)
> #define SKL_BOTTOM_COLOR(pipe) _MMIO_PIPE2(pipe, _SKL_BOTTOM_COLOR_A)
>
> +#define _ICL_PIPE_A_STATUS 0x70058
> +#define ICL_PIPESTATUS(pipe) _MMIO_PIPE2(pipe, _ICL_PIPE_A_STATUS)
> +#define PIPE_STATUS_UNDERRUN REG_BIT(31)
> +#define PIPE_STATUS_SOFT_UNDERRUN_XELPD REG_BIT(28)
> +#define PIPE_STATUS_HARD_UNDERRUN_XELPD REG_BIT(27)
> +#define PIPE_STATUS_PORT_UNDERRUN_XELPD REG_BIT(26)
> +
> #define VLV_DPFLIPSTAT _MMIO(VLV_DISPLAY_BASE + 0x70028)
> #define PIPEB_LINE_COMPARE_INT_EN (1 << 29)
> #define PIPEB_HLINE_INT_EN (1 << 28)
> @@ -7814,6 +7821,8 @@ enum {
> #define GEN8_PIPE_FIFO_UNDERRUN (1 << 31)
> #define GEN8_PIPE_CDCLK_CRC_ERROR (1 << 29)
> #define GEN8_PIPE_CDCLK_CRC_DONE (1 << 28)
> +#define XELPD_PIPE_SOFT_UNDERRUN (1 << 22)
> +#define XELPD_PIPE_HARD_UNDERRUN (1 << 21)
> #define GEN8_PIPE_CURSOR_FAULT (1 << 10)
> #define GEN8_PIPE_SPRITE_FAULT (1 << 9)
> #define GEN8_PIPE_PRIMARY_FAULT (1 << 8)
> --
> 2.25.4
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2021-05-17 6:49 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-15 3:10 [Intel-gfx] [PATCH v4 00/23] Alder Lake-P Support Matt Roper
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 01/23] drm/i915/xelpd: Enhanced pipe underrun reporting Matt Roper
2021-05-17 6:52 ` Lisovskiy, Stanislav [this message]
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 02/23] drm/i915/xelpd: Support DP1.4 compression BPPs Matt Roper
2021-05-17 15:18 ` Jani Nikula
2021-05-18 6:33 ` Kulkarni, Vandita
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 03/23] drm/i915/xelpd: Calculate VDSC RC parameters Matt Roper
2021-05-18 18:06 ` Navare, Manasi
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 04/23] drm/i915/xelpd: Add rc_qp_table for rcparams calculation Matt Roper
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 05/23] drm/i915/xelpd: Add VRR guardband for VRR CTL Matt Roper
2021-05-24 13:40 ` Aditya Swarup
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 06/23] drm/i915/adl_p: Add dedicated SAGV watermarks Matt Roper
2021-05-17 6:49 ` Lisovskiy, Stanislav
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 07/23] drm/i915/adl_p: Setup ports/phys Matt Roper
2021-05-17 18:01 ` Imre Deak
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 08/23] drm/i915/adl_p: Handle TC cold Matt Roper
2021-05-17 14:53 ` Imre Deak
2021-05-17 23:15 ` Souza, Jose
2021-05-17 23:22 ` Souza, Jose
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 09/23] drm/i915/adl_p: Implement TC sequences Matt Roper
2021-05-17 15:12 ` Imre Deak
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 10/23] drm/i915/adl_p: Don't config MBUS and DBUF during display initialization Matt Roper
2021-05-18 11:58 ` Lisovskiy, Stanislav
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 11/23] drm/i915/adl_p: Add ddb allocation support Matt Roper
2021-05-18 12:22 ` Lisovskiy, Stanislav
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 12/23] drm/i915: Introduce MBUS relative dbuf offsets Matt Roper
2021-05-17 6:38 ` Lisovskiy, Stanislav
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 13/23] drm/i915/adl_p: MBUS programming Matt Roper
2023-07-17 10:32 ` Tvrtko Ursulin
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 14/23] drm/i915/adl_p: Tx escape clock with DSI Matt Roper
2021-05-17 7:36 ` Kulkarni, Vandita
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 15/23] drm/i915/display: Replace dc3co_enabled with dc3co_exitline on intel_psr struct Matt Roper
2021-05-17 6:39 ` Gupta, Anshuman
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 16/23] drm/i915/display: Remove a redundant function argument from intel_psr_enable_source() Matt Roper
2021-05-19 6:49 ` Anshuman Gupta
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 17/23] drm/i915/display: Add PSR interrupt error check function Matt Roper
2021-05-17 17:03 ` Souza, Jose
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 18/23] drm/i915/display: Introduce new intel_psr_pause/resume function Matt Roper
2021-05-17 16:58 ` Souza, Jose
2021-05-18 9:33 ` Mun, Gwan-gyeong
2021-05-18 11:06 ` Ville Syrjälä
2021-05-21 10:58 ` Mun, Gwan-gyeong
2021-05-21 21:52 ` Souza, Jose
2021-06-01 10:23 ` Mun, Gwan-gyeong
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 19/23] drm/i915/adl_p: Define and use ADL-P specific DP translation tables Matt Roper
2021-05-17 23:02 ` Clint Taylor
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 20/23] drm/i915/adl_p: Add PLL Support Matt Roper
2021-05-17 21:55 ` Clint Taylor
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 21/23] drm/i915/adl_p: Program DP/HDMI link rate to DDI_BUF_CTL Matt Roper
2021-05-17 17:01 ` Souza, Jose
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 22/23] drm/i915/adlp: Add PIPE_MISC2 programming Matt Roper
2021-05-17 21:46 ` Clint Taylor
2021-05-15 3:10 ` [Intel-gfx] [PATCH v4 23/23] drm/i915/adl_p: Update memory bandwidth parameters Matt Roper
2021-05-17 17:02 ` Souza, Jose
2021-05-15 4:51 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for Alder Lake-P Support (rev3) Patchwork
2021-05-15 4:53 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2021-05-15 5:22 ` [Intel-gfx] ✗ Fi.CI.BAT: failure " Patchwork
2021-05-17 23:41 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Alder Lake-P Support (rev4) 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=20210517065259.GB17429@intel.com \
--to=stanislav.lisovskiy@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=lucas.demarchi@intel.com \
--cc=matthew.d.roper@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