Skip to contentSkip to conversation

FAILED: patch "[PATCH] iommu/vt-d: Fix PASID directory pointer coherency" failed to apply to 5.4-stable tree

PATCH [email protected] 1 participant 1 message Mar 7, 2023

Subsystems (3)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y
git checkout FETCH_HEAD
git cherry-pick -x 194b3348bdbb7db65375c72f3f774aee4cc6614e
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '[email protected]' --subject-prefix 'PATCH 5.4.y' HEAD^..

Possible dependencies:

194b3348bdbb ("iommu/vt-d: Fix PASID directory pointer coherency") 803766cbf85f ("iommu/vt-d: Fix lockdep splat in intel_pasid_get_entry()")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 194b3348bdbb7db65375c72f3f774aee4cc6614e Mon Sep 17 00:00:00 2001
From: Jacob Pan <[email protected]>
Date: Thu, 16 Feb 2023 21:08:15 +0800
Subject: [PATCH] iommu/vt-d: Fix PASID directory pointer coherency

On platforms that do not support IOMMU Extended capability bit 0 Page-walk Coherency, CPU caches are not snooped when IOMMU is accessing any translation structures. IOMMU access goes only directly to memory. Intel IOMMU code was missing a flush for the PASID table directory that resulted in the unrecoverable fault as shown below.

This patch adds clflush calls whenever allocating and updating a PASID table directory to ensure cache coherency.

On the reverse direction, there's no need to clflush the PASID directory pointer when we deactivate a context entry in that IOMMU hardware will not see the old PASID directory pointer after we clear the context entry. PASID directory entries are also never freed once allocated.

DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Read NO_PASID] Request device [00:0d.2] fault addr 0x1026a4000

       [fault reason 0x51] SM: Present bit in Directory Entry is clear

DMAR: Dump dmar1 table entries for IOVA 0x1026a4000
DMAR: scalable mode root entry: hi 0x0000000102448001, low 0x0000000101b3e001
DMAR: context entry: hi 0x0000000000000000, low 0x0000000101b4d401
DMAR: pasid dir entry: 0x0000000101b4e001
DMAR: pasid table entry[0]: 0x0000000000000109
DMAR: pasid table entry[1]: 0x0000000000000001
DMAR: pasid table entry[2]: 0x0000000000000000
DMAR: pasid table entry[3]: 0x0000000000000000
DMAR: pasid table entry[4]: 0x0000000000000000
DMAR: pasid table entry[5]: 0x0000000000000000
DMAR: pasid table entry[6]: 0x0000000000000000
DMAR: pasid table entry[7]: 0x0000000000000000
DMAR: PTE not present at level 4

drivers/iommu/intel/pasid.c +7 -0
index ec964ac7d797..9d2f05cf6164 100644
@@ -128,6 +128,9 @@ int intel_pasid_alloc_table(struct device *dev)
 	pasid_table->max_pasid = 1 << (order + PAGE_SHIFT + 3);
 	info->pasid_table = pasid_table;
 
+	if (!ecap_coherent(info->iommu->ecap))
+		clflush_cache_range(pasid_table->table, size);
+
 	return 0;
 }
 
@@ -215,6 +218,10 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid)
 			free_pgtable_page(entries);
 			goto retry;
 		}
+		if (!ecap_coherent(info->iommu->ecap)) {
+			clflush_cache_range(entries, VTD_PAGE_SIZE);
+			clflush_cache_range(&dir[dir_index].val, sizeof(*dir));
+		}
 	}
 
 	return &entries[index];