All of lore.kernel.org
 help / color / mirror / Atom feed
From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: Intel-gfx@lists.freedesktop.org
Subject: [PATCH v2 2/4] drm/i915: Fix WaProgramMgsrForL3BankSpecificMmioReads
Date: Thu, 11 Jul 2019 16:59:39 +0100	[thread overview]
Message-ID: <20190711155939.18059-1-tvrtko.ursulin@linux.intel.com> (raw)
In-Reply-To: <20190709210620.15805-3-tvrtko.ursulin@linux.intel.com>

From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

Two issues in this code:

1.
fls() usage is incorrect causing off by one in subslice mask lookup,
which in other words means subslice mask of all zeroes is always used
(subslice mask of a slice which is not present, or even out of bounds
array access), rendering the checks in wa_init_mcr either futile or
random.

2.
Condition in WARN_ON is not correct. It is doing a bitwise and operation
between a positive (present subslices) and negative mask (disabled L3
banks).

This means that with corrected fls() usage the assert would always
incorrectly fail.

We can fix this by inverting the fuse bits in the check.

v2:
 * Simplify check for logic and redability.
 * Improve commentary explaining what is really happening ie. what the
   assert is really trying to check and why.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: fe864b76c2ab ("drm/i915: Implement WaProgramMgsrForL3BankSpecificMmioReads")
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> # v1
Cc: Michał Winiarski <michal.winiarski@intel.com>
---
 drivers/gpu/drm/i915/gt/intel_workarounds.c | 80 ++++++++++-----------
 1 file changed, 40 insertions(+), 40 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c
index 9e069286d3ce..80f1159e5cda 100644
--- a/drivers/gpu/drm/i915/gt/intel_workarounds.c
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
@@ -761,7 +761,27 @@ static void
 wa_init_mcr(struct drm_i915_private *i915, struct i915_wa_list *wal)
 {
 	const struct sseu_dev_info *sseu = &RUNTIME_INFO(i915)->sseu;
-	u32 mcr_slice_subslice_mask;
+	u32 mcr_mask, mcr;
+
+	/*
+	 * WaProgramMgsrForCorrectSliceSpecificMmioReads:cnl,icl
+	 * Before any MMIO read into slice/subslice specific registers, MCR
+	 * packet control register needs to be programmed to point to any
+	 * enabled s/ss pair. Otherwise, incorrect values will be returned.
+	 * This means each subsequent MMIO read will be forwarded to an
+	 * specific s/ss combination, but this is OK since these registers
+	 * are consistent across s/ss in almost all cases. In the rare
+	 * occasions, such as INSTDONE, where this value is dependent
+	 * on s/ss combo, the read should be done with read_subslice_reg.
+	 */
+	mcr = intel_calculate_mcr_s_ss_select(i915);
+
+	if (INTEL_GEN(i915) >= 11)
+		mcr_mask = GEN11_MCR_SLICE_MASK | GEN11_MCR_SUBSLICE_MASK;
+	else
+		mcr_mask = GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK;
+
+	wa_write_masked_or(wal, GEN8_MCR_SELECTOR, mcr_mask, mcr);
 
 	/*
 	 * WaProgramMgsrForL3BankSpecificMmioReads: cnl,icl
@@ -776,49 +796,29 @@ wa_init_mcr(struct drm_i915_private *i915, struct i915_wa_list *wal)
 	 * something more complex that requires checking the range of every
 	 * MMIO read).
 	 */
-	if (INTEL_GEN(i915) >= 10 &&
-	    is_power_of_2(sseu->slice_mask)) {
+	if (INTEL_GEN(i915) >= 10 && is_power_of_2(sseu->slice_mask)) {
 		/*
-		 * read FUSE3 for enabled L3 Bank IDs, if L3 Bank matches
-		 * enabled subslice, no need to redirect MCR packet
+		 * GEN8_MCR_SELECTOR contains dual-purpose bits which select
+		 * both to which subslice, or to which L3 bank, the respective
+		 * mmio reads will go.
+		 * Since we have selected one enabled subslice in
+		 * WaProgramMgsrForCorrectSliceSpecificMmioReads, we now
+		 * need to check if the L3 bank of the equal "instance" is also
+		 * enabled.
+		 * If that is not the case we could try to find a number which
+		 * works for both, or going even further, implement a dynamic
+		 * scheme where we switch at before every affected mmio read.
+		 * Fortunately neither seems to be needed at the moment for
+		 * current parts and current driver behaviour.
 		 */
-		u32 slice = fls(sseu->slice_mask);
-		u32 fuse3 =
-			intel_uncore_read(&i915->uncore, GEN10_MIRROR_FUSE3);
-		u8 ss_mask = sseu->subslice_mask[slice];
-
-		u8 enabled_mask = (ss_mask | ss_mask >>
-				   GEN10_L3BANK_PAIR_COUNT) & GEN10_L3BANK_MASK;
-		u8 disabled_mask = fuse3 & GEN10_L3BANK_MASK;
+		unsigned int mcr_ss = BIT((mcr >> 24) & 0x7);
+		unsigned int l3_fuse =
+			intel_uncore_read(&i915->uncore, GEN10_MIRROR_FUSE3) &
+			GEN10_L3BANK_MASK;
+		unsigned int l3_en = ~(l3_fuse << 4 | l3_fuse);
 
-		/*
-		 * Production silicon should have matched L3Bank and
-		 * subslice enabled
-		 */
-		WARN_ON((enabled_mask & disabled_mask) != enabled_mask);
+		WARN_ON(!(mcr_ss & l3_en));
 	}
-
-	if (INTEL_GEN(i915) >= 11)
-		mcr_slice_subslice_mask = GEN11_MCR_SLICE_MASK |
-					  GEN11_MCR_SUBSLICE_MASK;
-	else
-		mcr_slice_subslice_mask = GEN8_MCR_SLICE_MASK |
-					  GEN8_MCR_SUBSLICE_MASK;
-	/*
-	 * WaProgramMgsrForCorrectSliceSpecificMmioReads:cnl,icl
-	 * Before any MMIO read into slice/subslice specific registers, MCR
-	 * packet control register needs to be programmed to point to any
-	 * enabled s/ss pair. Otherwise, incorrect values will be returned.
-	 * This means each subsequent MMIO read will be forwarded to an
-	 * specific s/ss combination, but this is OK since these registers
-	 * are consistent across s/ss in almost all cases. In the rare
-	 * occasions, such as INSTDONE, where this value is dependent
-	 * on s/ss combo, the read should be done with read_subslice_reg.
-	 */
-	wa_write_masked_or(wal,
-			   GEN8_MCR_SELECTOR,
-			   mcr_slice_subslice_mask,
-			   intel_calculate_mcr_s_ss_select(i915));
 }
 
 static void
-- 
2.20.1

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

  parent reply	other threads:[~2019-07-11 15:59 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-09 21:06 [PATCH 0/4] MCR fixes Tvrtko Ursulin
2019-07-09 21:06 ` [PATCH 1/4] drm/i915: Fix GEN8_MCR_SELECTOR programming Tvrtko Ursulin
2019-07-09 21:09   ` Chris Wilson
2019-07-10  6:21     ` Tvrtko Ursulin
2019-07-09 21:06 ` [PATCH 2/4] drm/i915: Fix WaProgramMgsrForL3BankSpecificMmioReads Tvrtko Ursulin
2019-07-09 21:11   ` Chris Wilson
2019-07-11  9:15     ` Tvrtko Ursulin
2019-07-11 15:59   ` Tvrtko Ursulin [this message]
2019-07-11 23:51     ` [PATCH v2 " Summers, Stuart
2019-07-12  5:32       ` Tvrtko Ursulin
2019-07-09 21:06 ` [PATCH 3/4] drm/i915: Move intel_calculate_mcr_s_ss_select to intel_sseu.c Tvrtko Ursulin
2019-07-09 21:12   ` Chris Wilson
2019-07-09 21:06 ` [PATCH 4/4] drm/i915/icl: Verify engine workarounds in GEN8_L3SQCREG4 Tvrtko Ursulin
2019-07-09 21:12   ` Chris Wilson
2019-07-09 22:18 ` ✓ Fi.CI.BAT: success for MCR fixes Patchwork
2019-07-11  8:53 ` ✓ Fi.CI.IGT: " Patchwork
2019-07-11 19:05 ` ✓ Fi.CI.BAT: success for MCR fixes (rev2) Patchwork
2019-07-12 22:17 ` ✗ Fi.CI.IGT: 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=20190711155939.18059-1-tvrtko.ursulin@linux.intel.com \
    --to=tvrtko.ursulin@linux.intel.com \
    --cc=Intel-gfx@lists.freedesktop.org \
    /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.