[PATCH bpf 1/2] bpf, cgroup: Assign cgroup in cgroup_sk_alloc when called from interrupt

[Daniel Borkmann <daniel@xxxxxxxxxxxxx>]

If cgroup_sk_alloc() is called from interrupt context, then just assign the
root cgroup to skcd->cgroup. Prior to commit 8520e224f547 ("bpf, cgroups:
Fix cgroup v2 fallback on v1/v2 mixed mode") we would just return, and later
on in sock_cgroup_ptr(), we were NULL-testing the cgroup in fast-path. Rather
than re-adding the NULL-test to the fast-path we can just assign it once from
cgroup_sk_alloc() given v1/v2 handling has been simplified.

Fixes: 8520e224f547 ("bpf, cgroups: Fix cgroup v2 fallback on v1/v2 mixed mode")
Reported-by: syzbot+df709157a4ecaf192b03@xxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
Tested-by: syzbot+df709157a4ecaf192b03@xxxxxxxxxxxxxxxxxxxxxxxxx
Cc: Tejun Heo <tj@xxxxxxxxxx>
Cc: Stanislav Fomichev <sdf@xxxxxxxxxx>
---
 kernel/cgroup/cgroup.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 8afa8690d288..570b0c97392a 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -6574,22 +6574,29 @@ int cgroup_parse_float(const char *input, unsigned dec_shift, s64 *v)
 
 void cgroup_sk_alloc(struct sock_cgroup_data *skcd)
 {
-	/* Don't associate the sock with unrelated interrupted task's cgroup. */
-	if (in_interrupt())
-		return;
+	struct cgroup *cgroup;
 
 	rcu_read_lock();
+	/* Don't associate the sock with unrelated interrupted task's cgroup. */
+	if (in_interrupt()) {
+		cgroup = &cgrp_dfl_root.cgrp;
+		cgroup_get(cgroup);
+		goto out;
+	}
+
 	while (true) {
 		struct css_set *cset;
 
 		cset = task_css_set(current);
 		if (likely(cgroup_tryget(cset->dfl_cgrp))) {
-			skcd->cgroup = cset->dfl_cgrp;
-			cgroup_bpf_get(cset->dfl_cgrp);
+			cgroup = cset->dfl_cgrp;
 			break;
 		}
 		cpu_relax();
 	}
+out:
+	skcd->cgroup = cgroup;
+	cgroup_bpf_get(cgroup);
 	rcu_read_unlock();
 }
 
-- 
2.21.0