iommu/iova: Fix alloc iova overflows issue
Subsystems (4)
PATCH v2 1/1 iommu/iova: Fix alloc iova overflows issue
From: Yunfei Wang <[email protected]>
In __alloc_and_insert_iova_range, there is an issue that retry_pfn overflows. The value of iovad->anchor.pfn_hi is ~0UL, then when iovad->cached_node is iovad->anchor, curr_iova->pfn_hi + 1 will overflow. As a result, if the retry logic is executed, low_pfn is updated to 0, and then new_pfn < low_pfn returns false to make the allocation successful.
This issue occurs in the following two situations:
1. The first iova size exceeds the domain size. When initializing iova domain, iovad->cached_node is assigned as iovad->anchor. For example, the iova domain size is 10M, start_pfn is 0x1_F000_0000, and the iova size allocated for the first time is 11M. The following is the log information, new->pfn_lo is smaller than iovad->cached_node.
Example log as follows:
[ 223.798112][T1705487] sh: [name:iova&]__alloc_and_insert_iova_range start_pfn:0x1f0000,retry_pfn:0x0,size:0xb00,limit_pfn:0x1f0a00 [ 223.799590][T1705487] sh: [name:iova&]__alloc_and_insert_iova_range
success start_pfn:0x1f0000,new->pfn_lo:0x1efe00,new->pfn_hi:0x1f08ff
2. The node with the largest iova->pfn_lo value in the iova domain is deleted, iovad->cached_node will be updated to iovad->anchor, and then the alloc iova size exceeds the maximum iova size that can be allocated in the domain.
After judging that retry_pfn is less than limit_pfn, call retry_pfn+1 to fix the overflow issue.
---
v2: Update patch
1. Cc [email protected]
This patch needs to be merged stable branch, add [email protected] in mail list.
2. Refer robin's suggestion to update patch.
--- drivers/iommu/iova.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iommu/iova.c +2 -2
@@ -197,7 +197,7 @@ static int __alloc_and_insert_iova_range(struct iova_domain *iovad, curr =__get_cached_rbnode (iovad, limit_pfn); curr_iova =to_iova (curr); - retry_pfn = curr_iova->pfn_hi +1 ; + retry_pfn = curr_iova->pfn_hi; retry: do {
@@ -211,7 +211,7 @@ static int __alloc_and_insert_iova_range(struct iova_domain *iovad,if (high_pfn < size || new_pfn < low_pfn) {if (low_pfn == iovad->start_pfn && retry_pfn < limit_pfn) { high_pfn = limit_pfn; - low_pfn = retry_pfn; + low_pfn = retry_pfn +1 ; curr =iova_find_limit (iovad, limit_pfn); curr_iova =to_iova (curr); goto retry;
Signature · 1 line
2.18.0
Quoted 28 lines · On 2023-01-11 06:38, [email protected] wrote:
From: Yunfei Wang <[email protected]>
In
__alloc_and_insert_iova_range, there is an issue thatretry_pfnoverflows. The value ofiovad->anchor.pfn_hiis ~0UL, then wheniovad->cached_nodeisiovad->anchor,curr_iova->pfn_hi+ 1 will overflow. As a result, if the retry logic is executed,low_pfnis updated to 0, and thennew_pfn<low_pfnreturns false to make the allocation successful.This issue occurs in the following two situations:
1. The first iova size exceeds the domain size. When initializing iova domain,iovad->cached_nodeis assigned asiovad->anchor. For example, the iova domain size is 10M,start_pfnis 0x1_F000_0000, and the iova size allocated for the first time is 11M. The following is the log information,new->pfn_lois smaller thaniovad->cached_node.Example log as follows:
[ 223.798112][T1705487] sh: [name:iova&]__alloc_and_insert_iova_range start_pfn:0x1f0000,retry_pfn:0x0,size:0xb00,limit_pfn:0x1f0a00 [ 223.799590][T1705487] sh: [name:iova&]__alloc_and_insert_iova_rangesuccess start_pfn:0x1f0000,new->pfn_lo:0x1efe00,new->pfn_hi:0x1f08ff
2. The node with the largest
iova->pfn_lovalue in the iova domain is deleted,iovad->cached_nodewill be updated toiovad->anchor, and then the alloc iova size exceeds the maximum iova size that can be allocated in the domain.
Quoted 6 lines
After judging that
retry_pfnis less thanlimit_pfn, call retry_pfn+1 to fix the overflow issue.Signed-off-by jianjiao zeng <[email protected]>Signed-off-by Yunfei Wang <[email protected]>Cc <[email protected]> # 5.15.*
Quoted 33 lines
---
v2: Update patch
1. Cc [email protected]
This patch needs to be merged stable branch, add [email protected] in mail list.
2. Refer robin's suggestion to update patch.--- drivers/iommu/iova.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/iommu/iova.c +2
-2@@ -197,7 +197,7 @@ static int __alloc_and_insert_iova_range(struct iova_domain *iovad, curr =__get_cached_rbnode (iovad, limit_pfn); curr_iova =to_iova (curr); - retry_pfn = curr_iova->pfn_hi +1 ; + retry_pfn = curr_iova->pfn_hi; retry: do {@@ -211,7 +211,7 @@ static int __alloc_and_insert_iova_range(struct iova_domain *iovad,if (high_pfn < size || new_pfn < low_pfn) {if (low_pfn == iovad->start_pfn && retry_pfn < limit_pfn) { high_pfn = limit_pfn; - low_pfn = retry_pfn; + low_pfn = retry_pfn +1 ; curr =iova_find_limit (iovad, limit_pfn); curr_iova =to_iova (curr); goto retry;
Joerg Roedel
maintainerQuoted 5 lines · On Wed, Jan 11, 2023 at 02:38:00PM +0800, [email protected] wrote:
v2: Update patch
1. Cc [email protected]
This patch needs to be merged stable branch, add [email protected] in mail list.
2. Refer robin's suggestion to update patch.
Applied for 6.2, thanks.