Skip to contentSkip to conversation

bpf: propagate __user annotations properly

PATCH Applied Lukas Bulwahn 3 participants 4 messages Dec 7, 2020 – Dec 8, 2020

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

PATCH 1/1 bpf: propagate __user annotations properly

__htab_map_lookup_and_delete_batch() stores a user pointer in the local variable ubatch and uses that in copy_{from,to}_user(), but ubatch misses a __user annotation.

So, sparse warns in the various assignments and uses of ubatch:

kernel/bpf/hashtab.c:1415:24: warning: incorrect type in initializer
(different address spaces)
kernel/bpf/hashtab.c:1415:24: expected void *ubatch
kernel/bpf/hashtab.c:1415:24: got void [noderef] __user *

kernel/bpf/hashtab.c:1444:46: warning: incorrect type in argument 2
(different address spaces)
kernel/bpf/hashtab.c:1444:46: expected void const [noderef] __user *from
kernel/bpf/hashtab.c:1444:46: got void *ubatch

kernel/bpf/hashtab.c:1608:16: warning: incorrect type in assignment
(different address spaces)
kernel/bpf/hashtab.c:1608:16: expected void *ubatch
kernel/bpf/hashtab.c:1608:16: got void [noderef] __user *

kernel/bpf/hashtab.c:1609:26: warning: incorrect type in argument 1
(different address spaces)
kernel/bpf/hashtab.c:1609:26: expected void [noderef] __user *to
kernel/bpf/hashtab.c:1609:26: got void *ubatch

Add the __user annotation to repair this chain of propagating __user annotations in __htab_map_lookup_and_delete_batch().

---
applies cleanly on current master (v5.10-rc7) and next-20201204

BPF maintainers, please pick this minor non-urgent clean-up patch.

 kernel/bpf/hashtab.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/bpf/hashtab.c +1 -1
index fe7a0733a63a..76c791def033 100644
@@ -1412,7 +1412,7 @@ __htab_map_lookup_and_delete_batch(struct bpf_map *map,
 	void *keys = NULL, *values = NULL, *value, *dst_key, *dst_val;
 	void __user *uvalues = u64_to_user_ptr(attr->batch.values);
 	void __user *ukeys = u64_to_user_ptr(attr->batch.keys);
-	void *ubatch = u64_to_user_ptr(attr->batch.in_batch);
+	void __user *ubatch = u64_to_user_ptr(attr->batch.in_batch);
 	u32 batch, max_count, size, bucket_size;
 	struct htab_elem *node_to_free = NULL;
 	u64 elem_map_flags, map_flags;
Signature · 1 line
2.17.1
Acked-by on v1 1/1
#1 lore raw
Quoted 21 lines · Lukas Bulwahn

__htab_map_lookup_and_delete_batch() stores a user pointer in the local variable ubatch and uses that in copy_{from,to}_user(), but ubatch misses a __user annotation.

So, sparse warns in the various assignments and uses of ubatch:

kernel/bpf/hashtab.c:1415:24: warning: incorrect type in initializer
(different address spaces)
kernel/bpf/hashtab.c:1415:24: expected void *ubatch
kernel/bpf/hashtab.c:1415:24: got void [noderef] __user *

kernel/bpf/hashtab.c:1444:46: warning: incorrect type in argument 2
(different address spaces)
kernel/bpf/hashtab.c:1444:46: expected void const [noderef] __user *from
kernel/bpf/hashtab.c:1444:46: got void *ubatch

kernel/bpf/hashtab.c:1608:16: warning: incorrect type in assignment
(different address spaces)
kernel/bpf/hashtab.c:1608:16: expected void *ubatch
kernel/bpf/hashtab.c:1608:16: got void [noderef] __user *

Quoted 7 lines

kernel/bpf/hashtab.c:1609:26: warning: incorrect type in argument 1
(different address spaces)
kernel/bpf/hashtab.c:1609:26: expected void [noderef] __user *to
kernel/bpf/hashtab.c:1609:26: got void *ubatch

Add the __user annotation to repair this chain of propagating __user annotations in __htab_map_lookup_and_delete_batch().

Add fix tag?

Fixes 057996380a42 ("bpf: Add batch ops to all htab bpf map")
Quoted 2 lines

Thanks for the fix. LGTM. I guess either bpf or bpf-next tree is fine since this is not a correctness issue.

#2 lore raw
Quoted 30 lines · Yonghong Song
Quoted 26 lines · Lukas Bulwahn

__htab_map_lookup_and_delete_batch() stores a user pointer in the local variable ubatch and uses that in copy_{from,to}_user(), but ubatch misses a __user annotation.

So, sparse warns in the various assignments and uses of ubatch:

kernel/bpf/hashtab.c:1415:24: warning: incorrect type in initializer
(different address spaces)
kernel/bpf/hashtab.c:1415:24: expected void *ubatch
kernel/bpf/hashtab.c:1415:24: got void [noderef] __user *

kernel/bpf/hashtab.c:1444:46: warning: incorrect type in argument 2
(different address spaces)
kernel/bpf/hashtab.c:1444:46: expected void const [noderef] __user *from
kernel/bpf/hashtab.c:1444:46: got void *ubatch

kernel/bpf/hashtab.c:1608:16: warning: incorrect type in assignment
(different address spaces)
kernel/bpf/hashtab.c:1608:16: expected void *ubatch
kernel/bpf/hashtab.c:1608:16: got void [noderef] __user *

kernel/bpf/hashtab.c:1609:26: warning: incorrect type in argument 1
(different address spaces)
kernel/bpf/hashtab.c:1609:26: expected void [noderef] __user *to
kernel/bpf/hashtab.c:1609:26: got void *ubatch

Quoted 7 lines
Quoted 2 lines

Add the __user annotation to repair this chain of propagating __user annotations in __htab_map_lookup_and_delete_batch().

Add fix tag?

Fixes 057996380a42 ("bpf: Add batch ops to all htab bpf map")

Fixes tag can be added by the maintainers when they pick it, but I personally am not a fan of adding a Fixes tag for such a minor fix here.

It is purely a syntactic change and change for the sparse semantic parser, but it really does not need to be backported and nothing observable in the binary was broken.

That is my rationale for not adding a Fixes: tag here. It is your final call.

Quoted 6 lines
Quoted 2 lines

Thanks for the fix. LGTM. I guess either bpf or bpf-next tree is fine since this is not a correctness issue.

Agree, and it is no functional change, nor a change in the object code. So risks of regressions are very, very low (zero).

Thanks for the review,

Lukas

#3 lore raw
In reply to Lukas Bulwahn
This patch was applied to bpf/bpf-next.git (refs/heads/master):

Hello:

This patch was applied to bpf/bpf-next.git (refs/heads/master):

Quoted 4 lines · On Mon, 7 Dec 2020 13:37:20 +0100 you wrote:

__htab_map_lookup_and_delete_batch() stores a user pointer in the local variable ubatch and uses that in copy_{from,to}_user(), but ubatch misses a __user annotation.

Quoted 8 lines

So, sparse warns in the various assignments and uses of ubatch:

kernel/bpf/hashtab.c:1415:24: warning: incorrect type in initializer
(different address spaces)
kernel/bpf/hashtab.c:1415:24: expected void *ubatch
kernel/bpf/hashtab.c:1415:24: got void [noderef] __user *

[...]

Here is the summary with links:
- bpf: propagate __user annotations properly

    https://git.kernel.org/bpf/bpf-next/c/2f4b03195fe8

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