* [PATCH net-next] tcx: Fix splat during dev unregister
@ 2023-07-28 21:47 Daniel Borkmann
2023-07-30 9:31 ` Leon Romanovsky
2023-07-31 21:50 ` patchwork-bot+netdevbpf
0 siblings, 2 replies; 3+ messages in thread
From: Daniel Borkmann @ 2023-07-28 21:47 UTC (permalink / raw)
To: kuba
Cc: bpf, netdev, Martin KaFai Lau, syzbot+376a289e86a0fd02b9ba,
Leon Romanovsky, Daniel Borkmann
From: Martin KaFai Lau <martin.lau@kernel.org>
During unregister_netdevice_many_notify(), the ordering of our concerned
function calls is like this:
unregister_netdevice_many_notify
dev_shutdown
qdisc_put
clsact_destroy
tcx_uninstall
The syzbot reproducer triggered a case that the qdisc refcnt is not
zero during dev_shutdown().
tcx_uninstall() will then WARN_ON_ONCE(tcx_entry(entry)->miniq_active)
because the miniq is still active and the entry should not be freed.
The latter assumed that qdisc destruction happens before tcx teardown.
This fix is to avoid tcx_uninstall() doing tcx_entry_free() when the
miniq is still alive and let the clsact_destroy() do the free later, so
that we do not assume any specific ordering for either of them.
If still active, tcx_uninstall() does clear the entry when flushing out
the prog/link. clsact_destroy() will then notice the "!tcx_entry_is_active()"
and then does the tcx_entry_free() eventually.
Fixes: e420bed02507 ("bpf: Add fd-based tcx multi-prog infra with link support")
Reported-by: syzbot+376a289e86a0fd02b9ba@syzkaller.appspotmail.com
Reported-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Co-developed-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: syzbot+376a289e86a0fd02b9ba@syzkaller.appspotmail.com
---
[ Sending directly to net-next given the issue was reported there by Leon. ]
include/linux/bpf_mprog.h | 16 ++++++++++++++++
kernel/bpf/tcx.c | 12 ++++++++----
2 files changed, 24 insertions(+), 4 deletions(-)
diff --git a/include/linux/bpf_mprog.h b/include/linux/bpf_mprog.h
index 2b429488f840..929225f7b095 100644
--- a/include/linux/bpf_mprog.h
+++ b/include/linux/bpf_mprog.h
@@ -256,6 +256,22 @@ static inline void bpf_mprog_entry_copy(struct bpf_mprog_entry *dst,
memcpy(dst->fp_items, src->fp_items, sizeof(src->fp_items));
}
+static inline void bpf_mprog_entry_clear(struct bpf_mprog_entry *dst)
+{
+ memset(dst->fp_items, 0, sizeof(dst->fp_items));
+}
+
+static inline void bpf_mprog_clear_all(struct bpf_mprog_entry *entry,
+ struct bpf_mprog_entry **entry_new)
+{
+ struct bpf_mprog_entry *peer;
+
+ peer = bpf_mprog_peer(entry);
+ bpf_mprog_entry_clear(peer);
+ peer->parent->count = 0;
+ *entry_new = peer;
+}
+
static inline void bpf_mprog_entry_grow(struct bpf_mprog_entry *entry, int idx)
{
int total = bpf_mprog_total(entry);
diff --git a/kernel/bpf/tcx.c b/kernel/bpf/tcx.c
index 69a272712b29..13f0b5dc8262 100644
--- a/kernel/bpf/tcx.c
+++ b/kernel/bpf/tcx.c
@@ -94,15 +94,19 @@ int tcx_prog_detach(const union bpf_attr *attr, struct bpf_prog *prog)
void tcx_uninstall(struct net_device *dev, bool ingress)
{
+ struct bpf_mprog_entry *entry, *entry_new = NULL;
struct bpf_tuple tuple = {};
- struct bpf_mprog_entry *entry;
struct bpf_mprog_fp *fp;
struct bpf_mprog_cp *cp;
+ bool active;
entry = tcx_entry_fetch(dev, ingress);
if (!entry)
return;
- tcx_entry_update(dev, NULL, ingress);
+ active = tcx_entry(entry)->miniq_active;
+ if (active)
+ bpf_mprog_clear_all(entry, &entry_new);
+ tcx_entry_update(dev, entry_new, ingress);
tcx_entry_sync();
bpf_mprog_foreach_tuple(entry, fp, cp, tuple) {
if (tuple.link)
@@ -111,8 +115,8 @@ void tcx_uninstall(struct net_device *dev, bool ingress)
bpf_prog_put(tuple.prog);
tcx_skeys_dec(ingress);
}
- WARN_ON_ONCE(tcx_entry(entry)->miniq_active);
- tcx_entry_free(entry);
+ if (!active)
+ tcx_entry_free(entry);
}
int tcx_prog_query(const union bpf_attr *attr, union bpf_attr __user *uattr)
--
2.21.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH net-next] tcx: Fix splat during dev unregister
2023-07-28 21:47 [PATCH net-next] tcx: Fix splat during dev unregister Daniel Borkmann
@ 2023-07-30 9:31 ` Leon Romanovsky
2023-07-31 21:50 ` patchwork-bot+netdevbpf
1 sibling, 0 replies; 3+ messages in thread
From: Leon Romanovsky @ 2023-07-30 9:31 UTC (permalink / raw)
To: Daniel Borkmann
Cc: kuba, bpf, netdev, Martin KaFai Lau, syzbot+376a289e86a0fd02b9ba
On Fri, Jul 28, 2023 at 11:47:17PM +0200, Daniel Borkmann wrote:
> From: Martin KaFai Lau <martin.lau@kernel.org>
>
> During unregister_netdevice_many_notify(), the ordering of our concerned
> function calls is like this:
>
> unregister_netdevice_many_notify
> dev_shutdown
> qdisc_put
> clsact_destroy
> tcx_uninstall
>
> The syzbot reproducer triggered a case that the qdisc refcnt is not
> zero during dev_shutdown().
>
> tcx_uninstall() will then WARN_ON_ONCE(tcx_entry(entry)->miniq_active)
> because the miniq is still active and the entry should not be freed.
> The latter assumed that qdisc destruction happens before tcx teardown.
>
> This fix is to avoid tcx_uninstall() doing tcx_entry_free() when the
> miniq is still alive and let the clsact_destroy() do the free later, so
> that we do not assume any specific ordering for either of them.
>
> If still active, tcx_uninstall() does clear the entry when flushing out
> the prog/link. clsact_destroy() will then notice the "!tcx_entry_is_active()"
> and then does the tcx_entry_free() eventually.
>
> Fixes: e420bed02507 ("bpf: Add fd-based tcx multi-prog infra with link support")
> Reported-by: syzbot+376a289e86a0fd02b9ba@syzkaller.appspotmail.com
> Reported-by: Leon Romanovsky <leonro@nvidia.com>
> Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
> Co-developed-by: Daniel Borkmann <daniel@iogearbox.net>
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Tested-by: syzbot+376a289e86a0fd02b9ba@syzkaller.appspotmail.com
> ---
> [ Sending directly to net-next given the issue was reported there by Leon. ]
>
> include/linux/bpf_mprog.h | 16 ++++++++++++++++
> kernel/bpf/tcx.c | 12 ++++++++----
> 2 files changed, 24 insertions(+), 4 deletions(-)
>
Thanks,
Tested-by: Leon Romanovsky <leonro@nvidia.com>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH net-next] tcx: Fix splat during dev unregister
2023-07-28 21:47 [PATCH net-next] tcx: Fix splat during dev unregister Daniel Borkmann
2023-07-30 9:31 ` Leon Romanovsky
@ 2023-07-31 21:50 ` patchwork-bot+netdevbpf
1 sibling, 0 replies; 3+ messages in thread
From: patchwork-bot+netdevbpf @ 2023-07-31 21:50 UTC (permalink / raw)
To: Daniel Borkmann
Cc: kuba, bpf, netdev, martin.lau, syzbot+376a289e86a0fd02b9ba,
leonro
Hello:
This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:
On Fri, 28 Jul 2023 23:47:17 +0200 you wrote:
> From: Martin KaFai Lau <martin.lau@kernel.org>
>
> During unregister_netdevice_many_notify(), the ordering of our concerned
> function calls is like this:
>
> unregister_netdevice_many_notify
> dev_shutdown
> qdisc_put
> clsact_destroy
> tcx_uninstall
>
> [...]
Here is the summary with links:
- [net-next] tcx: Fix splat during dev unregister
https://git.kernel.org/netdev/net-next/c/079082c60aff
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-07-31 21:50 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-07-28 21:47 [PATCH net-next] tcx: Fix splat during dev unregister Daniel Borkmann
2023-07-30 9:31 ` Leon Romanovsky
2023-07-31 21:50 ` patchwork-bot+netdevbpf
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).