All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric Anholt <eric@anholt.net>
To: dri-devel@lists.freedesktop.org
Cc: linux-kernel@vger.kernel.org,
	Paul Kocialkowski <paul.kocialkowski@bootlin.com>,
	Maxime Ripard <maxime.ripard@bootlin.com>,
	david.emett@broadcom.com, thomas.spurden@broadcom.com,
	Eric Anholt <eric@anholt.net>
Subject: [PATCH 4/4] drm/v3d: Fix and extend MMU error handling.
Date: Thu, 18 Apr 2019 17:10:14 -0700	[thread overview]
Message-ID: <20190419001014.23579-4-eric@anholt.net> (raw)
In-Reply-To: <20190419001014.23579-1-eric@anholt.net>

We were setting the wrong flags to enable PTI errors, so we were
seeing reads to invalid PTEs show up as write errors.  Also, we
weren't turning on the interrupts.  The AXI IDs we were dumping
included the outstanding write number and so they looked basically
random.  And the VIO_ADDR decoding was based on the MMU VA_WIDTH for
the first platform I worked on and was wrong on others.  In short,
this was a thorough mess from early HW enabling.

Tested on V3D 4.1 and 4.2 with intentional L2T, CLE, PTB, and TLB
faults.

Signed-off-by: Eric Anholt <eric@anholt.net>
---
 drivers/gpu/drm/v3d/v3d_drv.c  |  1 +
 drivers/gpu/drm/v3d/v3d_drv.h  |  2 ++
 drivers/gpu/drm/v3d/v3d_irq.c  | 31 +++++++++++++++++++++++++++----
 drivers/gpu/drm/v3d/v3d_mmu.c  |  7 +++++--
 drivers/gpu/drm/v3d/v3d_regs.h |  3 ++-
 5 files changed, 37 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/v3d/v3d_drv.c b/drivers/gpu/drm/v3d/v3d_drv.c
index 7ab36192e6bc..9ce2e4ef6c2a 100644
--- a/drivers/gpu/drm/v3d/v3d_drv.c
+++ b/drivers/gpu/drm/v3d/v3d_drv.c
@@ -494,6 +494,7 @@ static int v3d_platform_drm_probe(struct platform_device *pdev)
 	mmu_debug = V3D_READ(V3D_MMU_DEBUG_INFO);
 	dev->coherent_dma_mask =
 		DMA_BIT_MASK(30 + V3D_GET_FIELD(mmu_debug, V3D_MMU_PA_WIDTH));
+	v3d->va_width = 30 + V3D_GET_FIELD(mmu_debug, V3D_MMU_VA_WIDTH);
 
 	ident1 = V3D_READ(V3D_HUB_IDENT1);
 	v3d->ver = (V3D_GET_FIELD(ident1, V3D_HUB_IDENT1_TVER) * 10 +
diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h
index 6d31a6a5a08e..64682923018d 100644
--- a/drivers/gpu/drm/v3d/v3d_drv.h
+++ b/drivers/gpu/drm/v3d/v3d_drv.h
@@ -63,6 +63,8 @@ struct v3d_dev {
 	 */
 	void *mmu_scratch;
 	dma_addr_t mmu_scratch_paddr;
+	/* virtual address bits from V3D to the MMU. */
+	int va_width;
 
 	/* Number of V3D cores. */
 	u32 cores;
diff --git a/drivers/gpu/drm/v3d/v3d_irq.c b/drivers/gpu/drm/v3d/v3d_irq.c
index fac3c542860b..268d8a889ac5 100644
--- a/drivers/gpu/drm/v3d/v3d_irq.c
+++ b/drivers/gpu/drm/v3d/v3d_irq.c
@@ -162,10 +162,33 @@ v3d_hub_irq(int irq, void *arg)
 		      V3D_HUB_INT_MMU_PTI |
 		      V3D_HUB_INT_MMU_CAP)) {
 		u32 axi_id = V3D_READ(V3D_MMU_VIO_ID);
-		u64 vio_addr = (u64)V3D_READ(V3D_MMU_VIO_ADDR) << 8;
-
-		dev_err(v3d->dev, "MMU error from client %d at 0x%08llx%s%s%s\n",
-			axi_id, (long long)vio_addr,
+		u64 vio_addr = ((u64)V3D_READ(V3D_MMU_VIO_ADDR) <<
+				(v3d->va_width - 32));
+		static const char *const v3d41_axi_ids[] = {
+			"L2T",
+			"PTB",
+			"PSE",
+			"TLB",
+			"CLE",
+			"TFU",
+			"MMU",
+			"GMP",
+		};
+		const char *client = "?";
+
+		V3D_WRITE(V3D_MMU_CTL,
+			  V3D_READ(V3D_MMU_CTL) & (V3D_MMU_CTL_CAP_EXCEEDED |
+						   V3D_MMU_CTL_PT_INVALID |
+						   V3D_MMU_CTL_WRITE_VIOLATION));
+
+		if (v3d->ver >= 41) {
+			axi_id = axi_id >> 5;
+			if (axi_id < ARRAY_SIZE(v3d41_axi_ids))
+				client = v3d41_axi_ids[axi_id];
+		}
+
+		dev_err(v3d->dev, "MMU error from client %s (%d) at 0x%llx%s%s%s\n",
+			client, axi_id, (long long)vio_addr,
 			((intsts & V3D_HUB_INT_MMU_WRV) ?
 			 ", write violation" : ""),
 			((intsts & V3D_HUB_INT_MMU_PTI) ?
diff --git a/drivers/gpu/drm/v3d/v3d_mmu.c b/drivers/gpu/drm/v3d/v3d_mmu.c
index 7a21f1787ab1..395e81d97163 100644
--- a/drivers/gpu/drm/v3d/v3d_mmu.c
+++ b/drivers/gpu/drm/v3d/v3d_mmu.c
@@ -69,10 +69,13 @@ int v3d_mmu_set_page_table(struct v3d_dev *v3d)
 	V3D_WRITE(V3D_MMU_PT_PA_BASE, v3d->pt_paddr >> V3D_MMU_PAGE_SHIFT);
 	V3D_WRITE(V3D_MMU_CTL,
 		  V3D_MMU_CTL_ENABLE |
-		  V3D_MMU_CTL_PT_INVALID |
+		  V3D_MMU_CTL_PT_INVALID_ENABLE |
 		  V3D_MMU_CTL_PT_INVALID_ABORT |
+		  V3D_MMU_CTL_PT_INVALID_INT |
 		  V3D_MMU_CTL_WRITE_VIOLATION_ABORT |
-		  V3D_MMU_CTL_CAP_EXCEEDED_ABORT);
+		  V3D_MMU_CTL_WRITE_VIOLATION_INT |
+		  V3D_MMU_CTL_CAP_EXCEEDED_ABORT |
+		  V3D_MMU_CTL_CAP_EXCEEDED_INT);
 	V3D_WRITE(V3D_MMU_ILLEGAL_ADDR,
 		  (v3d->mmu_scratch_paddr >> V3D_MMU_PAGE_SHIFT) |
 		  V3D_MMU_ILLEGAL_ADDR_ENABLE);
diff --git a/drivers/gpu/drm/v3d/v3d_regs.h b/drivers/gpu/drm/v3d/v3d_regs.h
index eda1e289976f..9bcb57781d31 100644
--- a/drivers/gpu/drm/v3d/v3d_regs.h
+++ b/drivers/gpu/drm/v3d/v3d_regs.h
@@ -152,7 +152,8 @@
 # define V3D_MMU_CTL_PT_INVALID_ABORT                  BIT(19)
 # define V3D_MMU_CTL_PT_INVALID_INT                    BIT(18)
 # define V3D_MMU_CTL_PT_INVALID_EXCEPTION              BIT(17)
-# define V3D_MMU_CTL_WRITE_VIOLATION                   BIT(16)
+# define V3D_MMU_CTL_PT_INVALID_ENABLE                 BIT(16)
+# define V3D_MMU_CTL_WRITE_VIOLATION                   BIT(12)
 # define V3D_MMU_CTL_WRITE_VIOLATION_ABORT             BIT(11)
 # define V3D_MMU_CTL_WRITE_VIOLATION_INT               BIT(10)
 # define V3D_MMU_CTL_WRITE_VIOLATION_EXCEPTION         BIT(9)
-- 
2.20.1


  parent reply	other threads:[~2019-04-19  0:10 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-19  0:10 [PATCH 1/4] drm/v3d: Fix debugfs reads of MMU regs Eric Anholt
2019-04-19  0:10 ` Eric Anholt
2019-04-19  0:10 ` [PATCH 2/4] drm/v3d: Set the correct DMA mask according to the MMU's limits Eric Anholt
2019-04-19  8:47   ` Paul Kocialkowski
2019-04-19  8:47     ` Paul Kocialkowski
2019-04-19  0:10 ` [PATCH 3/4] drm/v3d: Dump V3D error debug registers in debugfs, and one at reset Eric Anholt
2019-04-19  8:48   ` Paul Kocialkowski
2019-04-19  8:48     ` Paul Kocialkowski
2019-04-19  0:10 ` Eric Anholt [this message]
2019-04-19  8:52   ` [PATCH 4/4] drm/v3d: Fix and extend MMU error handling Paul Kocialkowski
2019-04-19  8:52     ` Paul Kocialkowski
2019-04-19  8:47 ` [PATCH 1/4] drm/v3d: Fix debugfs reads of MMU regs Paul Kocialkowski
2019-04-19  8:47   ` Paul Kocialkowski

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=20190419001014.23579-4-eric@anholt.net \
    --to=eric@anholt.net \
    --cc=david.emett@broadcom.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maxime.ripard@bootlin.com \
    --cc=paul.kocialkowski@bootlin.com \
    --cc=thomas.spurden@broadcom.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.