Skip to contentSkip to conversation

xsk: fix number of pinned pages/umem size discrepancy

PATCH Applied Björn Töpel 4 participants 4 messages Sep 10, 2020 – Sep 15, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xsk: fix number of pinned pages/umem size discrepancy

From: Björn Töpel <[email protected]>

For AF_XDP sockets, there was a discrepancy between the number of of pinned pages and the size of the umem region.

The size of the umem region is used to validate the AF_XDP descriptor addresses. The logic that pinned the pages covered by the region only took whole pages into consideration, creating a mismatch between the size and pinned pages. A user could then pass AF_XDP addresses outside the range of pinned pages, but still within the size of the region, crashing the kernel.

This change correctly calculates the number of pages to be pinned. Further, the size check for the aligned mode is simplified. Now the code simply checks if the size is divisible by the chunk size.

Fixes bbff2f321a86 ("xsk: new descriptor addressing scheme")
---
 net/xdp/xdp_umem.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)
net/xdp/xdp_umem.c +8 -9
index e97db37354e4..b010bfde0149 100644
@@ -303,10 +303,10 @@ static int xdp_umem_account_pages(struct xdp_umem *umem)
 
 static int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
 {
+	u32 npgs_rem, chunk_size = mr->chunk_size, headroom = mr->headroom;
 	bool unaligned_chunks = mr->flags & XDP_UMEM_UNALIGNED_CHUNK_FLAG;
-	u32 chunk_size = mr->chunk_size, headroom = mr->headroom;
 	u64 npgs, addr = mr->addr, size = mr->len;
-	unsigned int chunks, chunks_per_page;
+	unsigned int chunks, chunks_rem;
 	int err;
 
 	if (chunk_size < XDP_UMEM_MIN_CHUNK_SIZE || chunk_size > PAGE_SIZE) {
@@ -336,19 +336,18 @@ static int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
 	if ((addr + size) < addr)
 		return -EINVAL;
 
-	npgs = size >> PAGE_SHIFT;
+	npgs = div_u64_rem(size, PAGE_SIZE, &npgs_rem);
+	if (npgs_rem)
+		npgs++;
 	if (npgs > U32_MAX)
 		return -EINVAL;
 
-	chunks = (unsigned int)div_u64(size, chunk_size);
+	chunks = (unsigned int)div_u64_rem(size, chunk_size, &chunks_rem);
 	if (chunks == 0)
 		return -EINVAL;
 
-	if (!unaligned_chunks) {
-		chunks_per_page = PAGE_SIZE / chunk_size;
-		if (chunks < chunks_per_page || chunks % chunks_per_page)
-			return -EINVAL;
-	}
+	if (!unaligned_chunks && chunks_rem)
+		return -EINVAL;
 
 	if (headroom >= chunk_size - XDP_PACKET_HEADROOM)
 		return -EINVAL;

base-commit: 746f534a4809e07f427f7d13d10f3a6a9641e5c3

Signature · 1 line
2.25.1
Tested-by on v1 1/1
#1 lore raw
Quoted 13 lines

From: Björn Töpel <[email protected]>

For AF_XDP sockets, there was a discrepancy between the number of of pinned pages and the size of the umem region.

The size of the umem region is used to validate the AF_XDP descriptor addresses. The logic that pinned the pages covered by the region only took whole pages into consideration, creating a mismatch between the size and pinned pages. A user could then pass AF_XDP addresses outside the range of pinned pages, but still within the size of the region, crashing the kernel.

Quoted 8 lines

This change correctly calculates the number of pages to be pinned. Further, the size check for the aligned mode is simplified. Now the code simply checks if the size is divisible by the chunk size.

Fixes bbff2f321a86 ("xsk: new descriptor addressing scheme")

Thanks for the patch Björn.

Quoted 59 lines
---
 net/xdp/xdp_umem.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)
net/xdp/xdp_umem.c +2 -2
index e97db37354e4..b010bfde0149 100644
@@ -303,10 +303,10 @@ static int xdp_umem_account_pages(struct
 xdp_umem *umem)
 
 static int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg
 *mr)
 {
+	u32 npgs_rem, chunk_size = mr->chunk_size, headroom = mr-
 >headroom;
 	bool unaligned_chunks = mr->flags &
 XDP_UMEM_UNALIGNED_CHUNK_FLAG;
-	u32 chunk_size = mr->chunk_size, headroom = mr->headroom;
 	u64 npgs, addr = mr->addr, size = mr->len;
-	unsigned int chunks, chunks_per_page;
+	unsigned int chunks, chunks_rem;
 	int err;
 
 	if (chunk_size < XDP_UMEM_MIN_CHUNK_SIZE || chunk_size >

PAGE_SIZE) {

net/xdp/xdp_umem.c +6 -7
@@ -336,19 +336,18 @@ static int xdp_umem_reg(struct xdp_umem
 *umem, struct xdp_umem_reg *mr)
 	if ((addr + size) < addr)
 		return -EINVAL;
 
-	npgs = size >> PAGE_SHIFT;
+	npgs = div_u64_rem(size, PAGE_SIZE, &npgs_rem);
+	if (npgs_rem)
+		npgs++;
 	if (npgs > U32_MAX)
 		return -EINVAL;
 
-	chunks = (unsigned int)div_u64(size, chunk_size);
+	chunks = (unsigned int)div_u64_rem(size, chunk_size,
 &chunks_rem);
 	if (chunks == 0)
 		return -EINVAL;
 
-	if (!unaligned_chunks) {
-		chunks_per_page = PAGE_SIZE / chunk_size;
-		if (chunks < chunks_per_page || chunks %
 chunks_per_page)
-			return -EINVAL;
-	}
+	if (!unaligned_chunks && chunks_rem)
+		return -EINVAL;
net/xdp/xdp_umem.c +0 -0
 	if (headroom >= chunk_size - XDP_PACKET_HEADROOM)
 		return -EINVAL;

base-commit: 746f534a4809e07f427f7d13d10f3a6a9641e5c3 --
2.25.1

reviewer Acked-by
#2 lore raw
Quoted 19 lines · Ciara
Quoted 19 lines

From: Björn Töpel <[email protected]>

For AF_XDP sockets, there was a discrepancy between the number of of pinned pages and the size of the umem region.

The size of the umem region is used to validate the AF_XDP descriptor addresses. The logic that pinned the pages covered by the region only took whole pages into consideration, creating a mismatch between the size and pinned pages. A user could then pass AF_XDP addresses outside the range of pinned pages, but still within the size of the region, crashing the kernel.

This change correctly calculates the number of pages to be pinned. Further, the size check for the aligned mode is simplified. Now the code simply checks if the size is divisible by the chunk size.

Quoted 7 lines
Quoted 3 lines
Fixes bbff2f321a86 ("xsk: new descriptor addressing scheme")

Thanks for the patch Björn.

#3 lore raw
In reply to Björn Töpel
Quoted 13 lines · Björn Töpel

From: Björn Töpel <[email protected]>

For AF_XDP sockets, there was a discrepancy between the number of of pinned pages and the size of the umem region.

The size of the umem region is used to validate the AF_XDP descriptor addresses. The logic that pinned the pages covered by the region only took whole pages into consideration, creating a mismatch between the size and pinned pages. A user could then pass AF_XDP addresses outside the range of pinned pages, but still within the size of the region, crashing the kernel.

Quoted 8 lines

This change correctly calculates the number of pages to be pinned. Further, the size check for the aligned mode is simplified. Now the code simply checks if the size is divisible by the chunk size.

Fixes bbff2f321a86 ("xsk: new descriptor addressing scheme")

Applied. Thanks