[PATCH bpf-next] bpf: Remove unnecessary ring buffer size check

[PATCH bpf-next] bpf: Remove unnecessary ring buffer size check

[Hou Tao]

From: Hou Tao <houtao1@huawei.com>

The theoretical maximum size of ring buffer is about 64GB, but now the
size of ring buffer is specified by max_entries in bpf_attr and its
maximum value is (4GB - 1), and it won't be possible for overflow.

So just remove the unnecessary size check in ringbuf_map_alloc() but
keep the comments for possible extension in future.

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/bpf/9c636a63-1f3d-442d-9223-96c2dccb9469@moroto.mountain
Signed-off-by: Hou Tao <houtao1@huawei.com>
---
 kernel/bpf/ringbuf.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

diff --git a/kernel/bpf/ringbuf.c b/kernel/bpf/ringbuf.c
index 875ac9b698d9..f045fde632e5 100644
--- a/kernel/bpf/ringbuf.c
+++ b/kernel/bpf/ringbuf.c
@@ -23,15 +23,6 @@
 
 #define RINGBUF_MAX_RECORD_SZ (UINT_MAX/4)
 
-/* Maximum size of ring buffer area is limited by 32-bit page offset within
- * record header, counted in pages. Reserve 8 bits for extensibility, and take
- * into account few extra pages for consumer/producer pages and
- * non-mmap()'able parts. This gives 64GB limit, which seems plenty for single
- * ring buffer.
- */
-#define RINGBUF_MAX_DATA_SZ \
-	(((1ULL << 24) - RINGBUF_POS_PAGES - RINGBUF_PGOFF) * PAGE_SIZE)
-
 struct bpf_ringbuf {
 	wait_queue_head_t waitq;
 	struct irq_work work;
@@ -161,6 +152,17 @@ static void bpf_ringbuf_notify(struct irq_work *work)
 	wake_up_all(&rb->waitq);
 }
 
+/* Maximum size of ring buffer area is limited by 32-bit page offset within
+ * record header, counted in pages. Reserve 8 bits for extensibility, and
+ * take into account few extra pages for consumer/producer pages and
+ * non-mmap()'able parts, the current maximum size would be:
+ *
+ *     (((1ULL << 24) - RINGBUF_POS_PAGES - RINGBUF_PGOFF) * PAGE_SIZE)
+ *
+ * This gives 64GB limit, which seems plenty for single ring buffer. Now
+ * considering that the maximum value of data_sz is (4GB - 1), there
+ * will be no overflow, so just note the size limit in the comments.
+ */
 static struct bpf_ringbuf *bpf_ringbuf_alloc(size_t data_sz, int numa_node)
 {
 	struct bpf_ringbuf *rb;
@@ -193,12 +195,6 @@ static struct bpf_map *ringbuf_map_alloc(union bpf_attr *attr)
 	    !PAGE_ALIGNED(attr->max_entries))
 		return ERR_PTR(-EINVAL);
 
-#ifdef CONFIG_64BIT
-	/* on 32-bit arch, it's impossible to overflow record's hdr->pgoff */
-	if (attr->max_entries > RINGBUF_MAX_DATA_SZ)
-		return ERR_PTR(-E2BIG);
-#endif
-
 	rb_map = bpf_map_area_alloc(sizeof(*rb_map), NUMA_NO_NODE);
 	if (!rb_map)
 		return ERR_PTR(-ENOMEM);
-- 
2.29.2

Re: [PATCH bpf-next] bpf: Remove unnecessary ring buffer size check

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to bpf/bpf-next.git (master)
by Daniel Borkmann <daniel@iogearbox.net>:

On Tue,  4 Jul 2023 15:40:14 +0800 you wrote:
> From: Hou Tao <houtao1@huawei.com>
> 
> The theoretical maximum size of ring buffer is about 64GB, but now the
> size of ring buffer is specified by max_entries in bpf_attr and its
> maximum value is (4GB - 1), and it won't be possible for overflow.
> 
> So just remove the unnecessary size check in ringbuf_map_alloc() but
> keep the comments for possible extension in future.
> 
> [...]

Here is the summary with links:
  - [bpf-next] bpf: Remove unnecessary ring buffer size check
    https://git.kernel.org/bpf/bpf-next/c/cf6eeb8f9dac

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html