intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Matt Roper <matthew.d.roper@intel.com>
To: intel-gfx@lists.freedesktop.org
Subject: [Intel-gfx] [PATCH v4 17/18] drm/i915/dg2: Maintain backward-compatible nested batch behavior
Date: Thu, 29 Jul 2021 10:00:07 -0700	[thread overview]
Message-ID: <20210729170008.2836648-18-matthew.d.roper@intel.com> (raw)
In-Reply-To: <20210729170008.2836648-1-matthew.d.roper@intel.com>

For tgl+, the per-context setting of MI_MODE[12] determines whether
the bits of a nested MI_BATCH_BUFFER_START instruction should be
interpreted in the traditional manner or whether they should
instead use a new tgl+ meaning that breaks backward compatibility, but
allows nesting into 3rd-level batchbuffers.  For previous platforms,
the hardware default for this register bit is to maintain
backward-compatible behavior unless a context intentionally opts into
the new behavior; however Xe_HPG flips the hardware default behavior.

From a SW perspective, we want to maintain the backward-compatible
behavior for userspace, so we'll apply a fake workaround to set it back
to the legacy behavior on platforms where the hardware default is to
break compatibility.  At the moment there is no Linux userspace that
utilizes third-level batchbuffers, so this will avoid userspace from
needing to make any changes.  using the legacy meaning is the correct
thing to do.  If/when we have userspace consumers that want to utilize
third-level batch nesting, we can provide a context parameter to allow
them to opt-in.

Bspec: 45974, 45718
Cc: John Harrison <John.C.Harrison@Intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 drivers/gpu/drm/i915/gt/intel_workarounds.c | 39 +++++++++++++++++++--
 drivers/gpu/drm/i915/i915_reg.h             |  1 +
 2 files changed, 38 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c
index 6895b083523d..3e756b761526 100644
--- a/drivers/gpu/drm/i915/gt/intel_workarounds.c
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
@@ -644,6 +644,37 @@ static void dg1_ctx_workarounds_init(struct intel_engine_cs *engine,
 		     DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE);
 }
 
+static void fakewa_disable_nestedbb_mode(struct intel_engine_cs *engine,
+					 struct i915_wa_list *wal)
+{
+	/*
+	 * This is a "fake" workaround defined by software to ensure we
+	 * maintain reliable, backward-compatible behavior for userspace with
+	 * regards to how nested MI_BATCH_BUFFER_START commands are handled.
+	 *
+	 * The per-context setting of MI_MODE[12] determines whether the bits
+	 * of a nested MI_BATCH_BUFFER_START instruction should be interpreted
+	 * in the traditional manner or whether they should instead use a new
+	 * tgl+ meaning that breaks backward compatibility, but allows nesting
+	 * into 3rd-level batchbuffers.  When this new capability was first
+	 * added in TGL, it remained off by default unless a context
+	 * intentionally opted in to the new behavior.  However Xe_HPG now
+	 * flips this on by default and requires that we explicitly opt out if
+	 * we don't want the new behavior.
+	 *
+	 * From a SW perspective, we want to maintain the backward-compatible
+	 * behavior for userspace, so we'll apply a fake workaround to set it
+	 * back to the legacy behavior on platforms where the hardware default
+	 * is to break compatibility.  At the moment there is no Linux
+	 * userspace that utilizes third-level batchbuffers, so this will avoid
+	 * userspace from needing to make any changes.  using the legacy
+	 * meaning is the correct thing to do.  If/when we have userspace
+	 * consumers that want to utilize third-level batch nesting, we can
+	 * provide a context parameter to allow them to opt-in.
+	 */
+	wa_masked_dis(wal, RING_MI_MODE(engine->mmio_base), TGL_NESTED_BB_EN);
+}
+
 static void
 __intel_engine_init_ctx_wa(struct intel_engine_cs *engine,
 			   struct i915_wa_list *wal,
@@ -651,11 +682,15 @@ __intel_engine_init_ctx_wa(struct intel_engine_cs *engine,
 {
 	struct drm_i915_private *i915 = engine->i915;
 
+	wa_init_start(wal, name, engine->name);
+
+	/* Applies to all engines */
+	if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 55))
+		fakewa_disable_nestedbb_mode(engine, wal);
+
 	if (engine->class != RENDER_CLASS)
 		return;
 
-	wa_init_start(wal, name, engine->name);
-
 	if (IS_DG1(i915))
 		dg1_ctx_workarounds_init(engine, wal);
 	else if (GRAPHICS_VER(i915) == 12)
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index ed85f3ec1727..c192c1e096d7 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2893,6 +2893,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define MI_MODE		_MMIO(0x209c)
 # define VS_TIMER_DISPATCH				(1 << 6)
 # define MI_FLUSH_ENABLE				(1 << 12)
+# define TGL_NESTED_BB_EN				(1 << 12)
 # define ASYNC_FLIP_PERF_DISABLE			(1 << 14)
 # define MODE_IDLE					(1 << 9)
 # define STOP_RING					(1 << 8)
-- 
2.25.4

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2021-07-29 17:00 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-29 16:59 [Intel-gfx] [PATCH v4 00/18] Begin enabling Xe_HP SDV and DG2 platforms Matt Roper
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 01/18] drm/i915/xehp: handle new steering options Matt Roper
2021-08-04 19:53   ` Lucas De Marchi
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 02/18] drm/i915/xehpsdv: Define steering tables Matt Roper
2021-08-04 20:13   ` Lucas De Marchi
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 03/18] drm/i915/dg2: Add forcewake table Matt Roper
2021-08-04  0:15   ` Souza, Jose
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 04/18] drm/i915/dg2: Update LNCF steering ranges Matt Roper
2021-08-04 20:16   ` Lucas De Marchi
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 05/18] drm/i915/dg2: Add SQIDI steering Matt Roper
2021-08-04 20:22   ` Lucas De Marchi
2021-08-05 15:11     ` Matt Roper
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 06/18] drm/i915/xehp: Loop over all gslices for INSTDONE processing Matt Roper
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 07/18] drm/i915/dg2: Report INSTDONE_GEOM values in error state Matt Roper
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 08/18] drm/i915/xehp: Changes to ss/eu definitions Matt Roper
2021-08-04  0:17   ` Souza, Jose
2021-07-29 16:59 ` [Intel-gfx] [PATCH v4 09/18] drm/i915/xehpsdv: Add maximum sseu limits Matt Roper
2021-08-04 20:26   ` Lucas De Marchi
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 10/18] drm/i915/xehpsdv: Add compute DSS type Matt Roper
2021-08-04 20:36   ` Lucas De Marchi
2021-08-04 21:00     ` Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 11/18] drm/i915/dg2: DG2 uses the same sseu limits as XeHP SDV Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 12/18] drm/i915/xehpsdv: Define MOCS table for " Matt Roper
2021-07-29 17:31   ` Lucas De Marchi
2021-07-30  7:16     ` Siddiqui, Ayaz A
2021-07-30 17:01       ` Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 13/18] drm/i915/dg2: Define MOCS table for DG2 Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 14/18] drm/i915/xehpsdv: factor out function to read RP_STATE_CAP Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 15/18] drm/i915/xehpsdv: Read correct RP_STATE_CAP register Matt Roper
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 16/18] drm/i915/dg2: Add new LRI reg offsets Matt Roper
2021-07-29 17:00 ` Matt Roper [this message]
2021-07-29 17:00 ` [Intel-gfx] [PATCH v4 18/18] drm/i915/dg2: Configure PCON in DP pre-enable path Matt Roper
2021-07-29 20:59 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Begin enabling Xe_HP SDV and DG2 platforms (rev8) 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=20210729170008.2836648-18-matthew.d.roper@intel.com \
    --to=matthew.d.roper@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --subject='Re: [Intel-gfx] [PATCH v4 17/18] drm/i915/dg2: Maintain backward-compatible nested batch behavior' \
    /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

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).