On 5/27/25 9:54 AM, Jerome Marchand wrote:
The second argument of bpf_sysctl_get_name() helper is a pointer to a
buffer that is being written to. However that isn't specify in the
prototype.
Until commit 37cce22dbd51a ("bpf: verifier: Refactor helper access
type tracking"), all helper accesses were considered as a possible
write access by the verifier, so no big harm was done. However, since
then, the verifier might make wrong asssumption about the content of
that address which might lead it to make faulty optimizations (such as
removing code that was wrongly labeled dead). This is what happens in
Could you give more detailed example about the above statement?
the verifier might make wrong asssumption about the content of
that address which might lead it to make faulty optimizations (such as
removing code that was wrongly labeled dead)
This patch actually may cause a behavior change.
Without this patch, typically the whole buffer will be initialized
to 0 and then the helper itself will copy bytes until seeing a '\0'.
With this patch, bpf prog does not need to initialize the buffer.
Inside the helper, the copied bytes may not cover the whole buffer.
test_sysctl selftest to the tests related to sysctl_get_name.
Correctly mark the second argument of bpf_sysctl_get_name() as
ARG_PTR_TO_UNINIT_MEM.
Signed-off-by: Jerome Marchand <jmarchan@xxxxxxxxxx>
---
kernel/bpf/cgroup.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 84f58f3d028a3..09c02a592d24a 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -2104,7 +2104,7 @@ static const struct bpf_func_proto bpf_sysctl_get_name_proto = {
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
- .arg2_type = ARG_PTR_TO_MEM,
+ .arg2_type = ARG_PTR_TO_UNINIT_MEM,
.arg3_type = ARG_CONST_SIZE,
.arg4_type = ARG_ANYTHING,
};
