bpf: Fix missing migration protection in __rhtab_map_lookup_and_delete_batch()
PATCH v2 1/1 bpf: Fix missing migration protection in __rhtab_map_lookup_and_delete_batch()
bpf_mem_cache_free_rcu() uses this_cpu_ptr() which requires migration to be disabled. All callers of rhtab_delete_elem() disable migration except __rhtab_map_lookup_and_delete_batch(), which calls it under rcu_read_lock() only.
On CONFIG_PREEMPT_RCU, rcu_read_lock() does not disable preemption or migration, so the task can migrate between CPUs during the delete loop, causing this_cpu_ptr() to trigger:
BUG: using smp_processor_id() in preemptible [00000000] code
Fix by wrapping the delete loop in migrate_disable()/migrate_enable() in __rhtab_map_lookup_and_delete_batch(), matching the migration protection that the other callers already provide.
---
Changes in v2:
- Fix root cause in __rhtab_map_lookup_and_delete_batch() with migrate_disable()/migrate_enable() instead of stretching bpf_disable_instrumentation() in rhtab_delete_elem().
- Fix Fixes: tag to 818e00848227.
- Fix commit message: migration disabled, not preemption.
Tested with syzkaller:
- Before the fix: the reported warning was reproduced reliably.
- After the fix: the warning was no longer reproducible.
kernel/bpf/hashtab.c | 2 ++ 1 file changed, 2 insertions(+)
kernel/bpf/hashtab.c +2 -0
@@ -3369,8 +3369,10 @@ static int __rhtab_map_lookup_and_delete_batch(struct bpf_map *map, }if (do_delete) { +migrate_disable ();for (i =0 ; i < total; i++)rhtab_delete_elem (rhtab, del_elems[i],NULL ,0 ); +migrate_enable (); }rcu_read_unlock ();
--
2.55.0
Quoted 15 lines · Ömer Mete Kaya
bpf_mem_cache_free_rcu()usesthis_cpu_ptr()which requires migration to be disabled. All callers ofrhtab_delete_elem()disable migration except__rhtab_map_lookup_and_delete_batch(), which calls it underrcu_read_lock()only.On
CONFIG_PREEMPT_RCU,rcu_read_lock()does not disable preemption or migration, so the task can migrate between CPUs during the delete loop, causingthis_cpu_ptr()to trigger:BUG: using smp_processor_id() in preemptible [00000000] codeFix by wrapping the delete loop in migrate_disable()/migrate_enable() in
__rhtab_map_lookup_and_delete_batch(), matching the migration protection that the other callers already provide.
Quoted 6 lines
Fixes 818e00848227 ("bpf: Implement iteration ops for resizable hashtab")Reported-by [email protected]Suggested-by Alexei Starovoitov <[email protected]>Signed-off-by Ömer Mete Kaya <[email protected]>---
Quoted 33 lines
Changes in v2:
- Fix root cause in__rhtab_map_lookup_and_delete_batch()with migrate_disable()/migrate_enable() instead of stretchingbpf_disable_instrumentation()inrhtab_delete_elem().
- Fix Fixes: tag to818e00848227.
- Fix commit message: migration disabled, not preemption.Tested with syzkaller:
- Before the fix: the reported warning was reproduced reliably.
- After the fix: the warning was no longer reproducible.kernel/bpf/hashtab.c | 2 ++ 1 file changed, 2 insertions(+)kernel/bpf/hashtab.c +2
-1@@ -3369,8 +3369,10 @@ static int __rhtab_map_lookup_and_delete_batch(struct bpf_map *map, }if (do_delete) { +migrate_disable ();for (i =0 ; i < total; i++)rhtab_delete_elem (rhtab, del_elems[i],NULL ,0 ); +migrate_enable (); }rcu_read_unlock (); --2.55.0
This patch was applied to bpf/bpf.git (master) by Alexei Starovoitov <[email protected]>:
Hello:
This patch was applied to bpf/bpf.git (master)
by Alexei Starovoitov <[email protected]>:
Quoted 10 lines · On Tue, 29 Sep 2026 11:10:34 +0300 you wrote:
bpf_mem_cache_free_rcu()usesthis_cpu_ptr()which requires migration to be disabled. All callers ofrhtab_delete_elem()disable migration except__rhtab_map_lookup_and_delete_batch(), which calls it underrcu_read_lock()only.On
CONFIG_PREEMPT_RCU,rcu_read_lock()does not disable preemption or migration, so the task can migrate between CPUs during the delete loop, causingthis_cpu_ptr()to trigger:[...]
Here is the summary with links:
- [bpf,v2] bpf: Fix missing migration protection in __rhtab_map_lookup_and_delete_batch()
https://git.kernel.org/bpf/bpf/c/de020dc8049b
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html