Patch "iommu/arm-smmu: really fix page table locking" has been added to the 3.13-stable tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



This is a note to let you know that I've just added the patch titled

    iommu/arm-smmu: really fix page table locking

to the 3.13-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     iommu-arm-smmu-really-fix-page-table-locking.patch
and it can be found in the queue-3.13 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.


>From c9d09e2748eaa55cac2af274574baa6368189bc1 Mon Sep 17 00:00:00 2001
From: Will Deacon <will.deacon@xxxxxxx>
Date: Tue, 4 Feb 2014 22:12:42 +0000
Subject: iommu/arm-smmu: really fix page table locking

From: Will Deacon <will.deacon@xxxxxxx>

commit c9d09e2748eaa55cac2af274574baa6368189bc1 upstream.

Commit a44a9791e778 ("iommu/arm-smmu: use mutex instead of spinlock for
locking page tables") replaced the page table spinlock with a mutex, to
allow blocking allocations to satisfy lazy mapping requests.

Unfortunately, it turns out that IOMMU mappings are created from atomic
context (e.g. spinlock held during a dma_map), so this change doesn't
really help us in practice.

This patch is a partial revert of the offending commit, bringing back
the original spinlock but replacing our page table allocations for any
levels below the pgd (which is allocated during domain init) with
GFP_ATOMIC instead of GFP_KERNEL.

Reported-by: Andreas Herrmann <andreas.herrmann@xxxxxxxxxxx>
Signed-off-by: Will Deacon <will.deacon@xxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 drivers/iommu/arm-smmu.c |   20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

--- a/drivers/iommu/arm-smmu.c
+++ b/drivers/iommu/arm-smmu.c
@@ -392,7 +392,7 @@ struct arm_smmu_domain {
 	struct arm_smmu_cfg		root_cfg;
 	phys_addr_t			output_mask;
 
-	struct mutex			lock;
+	spinlock_t			lock;
 };
 
 static DEFINE_SPINLOCK(arm_smmu_devices_lock);
@@ -900,7 +900,7 @@ static int arm_smmu_domain_init(struct i
 		goto out_free_domain;
 	smmu_domain->root_cfg.pgd = pgd;
 
-	mutex_init(&smmu_domain->lock);
+	spin_lock_init(&smmu_domain->lock);
 	domain->priv = smmu_domain;
 	return 0;
 
@@ -1137,7 +1137,7 @@ static int arm_smmu_attach_dev(struct io
 	 * Sanity check the domain. We don't currently support domains
 	 * that cross between different SMMU chains.
 	 */
-	mutex_lock(&smmu_domain->lock);
+	spin_lock(&smmu_domain->lock);
 	if (!smmu_domain->leaf_smmu) {
 		/* Now that we have a master, we can finalise the domain */
 		ret = arm_smmu_init_domain_context(domain, dev);
@@ -1152,7 +1152,7 @@ static int arm_smmu_attach_dev(struct io
 			dev_name(device_smmu->dev));
 		goto err_unlock;
 	}
-	mutex_unlock(&smmu_domain->lock);
+	spin_unlock(&smmu_domain->lock);
 
 	/* Looks ok, so add the device to the domain */
 	master = find_smmu_master(smmu_domain->leaf_smmu, dev->of_node);
@@ -1162,7 +1162,7 @@ static int arm_smmu_attach_dev(struct io
 	return arm_smmu_domain_add_master(smmu_domain, master);
 
 err_unlock:
-	mutex_unlock(&smmu_domain->lock);
+	spin_unlock(&smmu_domain->lock);
 	return ret;
 }
 
@@ -1209,7 +1209,7 @@ static int arm_smmu_alloc_init_pte(struc
 
 	if (pmd_none(*pmd)) {
 		/* Allocate a new set of tables */
-		pgtable_t table = alloc_page(PGALLOC_GFP);
+		pgtable_t table = alloc_page(GFP_ATOMIC|__GFP_ZERO);
 		if (!table)
 			return -ENOMEM;
 
@@ -1314,7 +1314,7 @@ static int arm_smmu_alloc_init_pmd(struc
 
 #ifndef __PAGETABLE_PMD_FOLDED
 	if (pud_none(*pud)) {
-		pmd = pmd_alloc_one(NULL, addr);
+		pmd = (pmd_t *)get_zeroed_page(GFP_ATOMIC);
 		if (!pmd)
 			return -ENOMEM;
 
@@ -1346,7 +1346,7 @@ static int arm_smmu_alloc_init_pud(struc
 
 #ifndef __PAGETABLE_PUD_FOLDED
 	if (pgd_none(*pgd)) {
-		pud = pud_alloc_one(NULL, addr);
+		pud = (pud_t *)get_zeroed_page(GFP_ATOMIC);
 		if (!pud)
 			return -ENOMEM;
 
@@ -1400,7 +1400,7 @@ static int arm_smmu_handle_mapping(struc
 	if (paddr & ~output_mask)
 		return -ERANGE;
 
-	mutex_lock(&smmu_domain->lock);
+	spin_lock(&smmu_domain->lock);
 	pgd += pgd_index(iova);
 	end = iova + size;
 	do {
@@ -1416,7 +1416,7 @@ static int arm_smmu_handle_mapping(struc
 	} while (pgd++, iova != end);
 
 out_unlock:
-	mutex_unlock(&smmu_domain->lock);
+	spin_unlock(&smmu_domain->lock);
 
 	/* Ensure new page tables are visible to the hardware walker */
 	if (smmu->features & ARM_SMMU_FEAT_COHERENT_WALK)


Patches currently in stable-queue which might be from will.deacon@xxxxxxx are

queue-3.13/perf-fix-hotplug-splat.patch
queue-3.13/iommu-arm-smmu-really-fix-page-table-locking.patch
queue-3.13/arm-7953-1-mm-ensure-tlb-invalidation-is-complete-before-enabling-mmu.patch
queue-3.13/iommu-arm-smmu-set-cbarn.bpshcfg-to-nsh-for-s1-s2-bypass-contexts.patch
queue-3.13/iommu-arm-smmu-fix-table-flushing-during-initial-allocations.patch
queue-3.13/iommu-arm-smmu-fix-pud-pmd-entry-fill-sequence.patch
queue-3.13/arm-7955-1-spinlock-ensure-we-have-a-compiler-barrier-before-sev.patch
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]