[PATCH 085/222] drm/radeon: stop page faults from hanging the system (v2)

From: Herton Ronaldo Krzesinski
Date: Wed Jan 16 2013 - 11:02:41 EST


3.5.7.3 -stable review patch. If anyone has any objections, please let me know.

------------------

From: =?UTF-8?q?Christian=20K=C3=B6nig?= <deathsimple@xxxxxxxxxxx>

commit ae133a1129790ec288b429b5f08ab4701633844a upstream.

Redirect invalid memory accesses to the default page
instead of locking up the memory controller. Also
enable the invalid memory access interrupts and
start spamming system log with it.

v2 (agd5f): fix up against 2 level PT changes

Signed-off-by: Christian KÃnig <deathsimple@xxxxxxxxxxx>
Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx>
[ herton: adjust context ]
Signed-off-by: Herton Ronaldo Krzesinski <herton.krzesinski@xxxxxxxxxxxxx>
---
drivers/gpu/drm/radeon/evergreen.c | 10 ++++++++++
drivers/gpu/drm/radeon/evergreend.h | 3 +++
drivers/gpu/drm/radeon/ni.c | 16 +++++++++++++---
drivers/gpu/drm/radeon/nid.h | 11 +++++++++++
drivers/gpu/drm/radeon/si.c | 25 +++++++++++++++++++++++--
drivers/gpu/drm/radeon/sid.h | 14 ++++++++++++++
6 files changed, 74 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 9c61b96..fb868675 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2913,6 +2913,16 @@ restart_ih:
break;
}
break;
+ case 146:
+ case 147:
+ dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
+ dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n",
+ RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+ dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+ RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+ /* reset addr and status */
+ WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
+ break;
case 176: /* CP_INT in ring buffer */
case 177: /* CP_INT in IB1 */
case 178: /* CP_INT in IB2 */
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index bfdbafe..d2f0528 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -647,6 +647,7 @@
#define PAGE_TABLE_DEPTH(x) (((x) & 3) << 1)
#define RANGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 4)
#define VM_CONTEXT1_CNTL 0x1414
+#define VM_CONTEXT1_CNTL2 0x1434
#define VM_CONTEXT0_PAGE_TABLE_BASE_ADDR 0x153C
#define VM_CONTEXT0_PAGE_TABLE_END_ADDR 0x157C
#define VM_CONTEXT0_PAGE_TABLE_START_ADDR 0x155C
@@ -668,6 +669,8 @@
#define CACHE_UPDATE_MODE(x) ((x) << 6)
#define VM_L2_STATUS 0x140C
#define L2_BUSY (1 << 0)
+#define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x14FC
+#define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x14DC

#define WAIT_UNTIL 0x8040

diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 2732a70..6a0664d 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -780,10 +780,20 @@ int cayman_pcie_gart_enable(struct radeon_device *rdev)
/* enable context1-7 */
WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
(u32)(rdev->dummy_page.addr >> 12));
- WREG32(VM_CONTEXT1_CNTL2, 0);
- WREG32(VM_CONTEXT1_CNTL, 0);
+ WREG32(VM_CONTEXT1_CNTL2, 4);
WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
- RANGE_PROTECTION_FAULT_ENABLE_DEFAULT);
+ RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ RANGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+ DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+ PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ PDE0_PROTECTION_FAULT_ENABLE_DEFAULT |
+ VALID_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ VALID_PROTECTION_FAULT_ENABLE_DEFAULT |
+ READ_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ READ_PROTECTION_FAULT_ENABLE_DEFAULT |
+ WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ WRITE_PROTECTION_FAULT_ENABLE_DEFAULT);

cayman_pcie_gart_tlb_flush(rdev);
DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index a0b9806..5e76c2f5 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -80,7 +80,18 @@
#define VM_CONTEXT0_CNTL 0x1410
#define ENABLE_CONTEXT (1 << 0)
#define PAGE_TABLE_DEPTH(x) (((x) & 3) << 1)
+#define RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 3)
#define RANGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 4)
+#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 6)
+#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 7)
+#define PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 9)
+#define PDE0_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 10)
+#define VALID_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 12)
+#define VALID_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 13)
+#define READ_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 15)
+#define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16)
+#define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18)
+#define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19)
#define VM_CONTEXT1_CNTL 0x1414
#define VM_CONTEXT0_CNTL2 0x1430
#define VM_CONTEXT1_CNTL2 0x1434
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index ec8abab..29e42de 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2391,9 +2391,20 @@ int si_pcie_gart_enable(struct radeon_device *rdev)
/* enable context1-15 */
WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
(u32)(rdev->dummy_page.addr >> 12));
- WREG32(VM_CONTEXT1_CNTL2, 0);
+ WREG32(VM_CONTEXT1_CNTL2, 4);
WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
- RANGE_PROTECTION_FAULT_ENABLE_DEFAULT);
+ RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ RANGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+ DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+ PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ PDE0_PROTECTION_FAULT_ENABLE_DEFAULT |
+ VALID_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ VALID_PROTECTION_FAULT_ENABLE_DEFAULT |
+ READ_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ READ_PROTECTION_FAULT_ENABLE_DEFAULT |
+ WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+ WRITE_PROTECTION_FAULT_ENABLE_DEFAULT);

si_pcie_gart_tlb_flush(rdev);
DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
@@ -3604,6 +3615,16 @@ restart_ih:
break;
}
break;
+ case 146:
+ case 147:
+ dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
+ dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n",
+ RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+ dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+ RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+ /* reset addr and status */
+ WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
+ break;
case 176: /* RINGID0 CP_INT */
radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
break;
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 810a83e..06b17e6 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -91,7 +91,18 @@
#define VM_CONTEXT0_CNTL 0x1410
#define ENABLE_CONTEXT (1 << 0)
#define PAGE_TABLE_DEPTH(x) (((x) & 3) << 1)
+#define RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 3)
#define RANGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 4)
+#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 6)
+#define DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 7)
+#define PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 9)
+#define PDE0_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 10)
+#define VALID_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 12)
+#define VALID_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 13)
+#define READ_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 15)
+#define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16)
+#define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18)
+#define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19)
#define VM_CONTEXT1_CNTL 0x1414
#define VM_CONTEXT0_CNTL2 0x1430
#define VM_CONTEXT1_CNTL2 0x1434
@@ -104,6 +115,9 @@
#define VM_CONTEXT14_PAGE_TABLE_BASE_ADDR 0x1450
#define VM_CONTEXT15_PAGE_TABLE_BASE_ADDR 0x1454

+#define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x14FC
+#define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x14DC
+
#define VM_INVALIDATE_REQUEST 0x1478
#define VM_INVALIDATE_RESPONSE 0x147c

--
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/