The commit 74e7e1efdad4 (xen/netback: don't call kfree_skb() with interrupts disabled) fixes deadlock in Linux netback driver. This seems to be a good candidate for the stable trees. This patch didn't apply cleanly in 5.15 kernel due to difference in function prototypes in drivers/net/xen-netback/common.h.
Juergen Gross (1): xen/netback: don't call kfree_skb() with interrupts disabled
drivers/net/xen-netback/common.h | 2 +- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/rx.c | 8 +++++--- 3 files changed, 10 insertions(+), 6 deletions(-)
From: Juergen Gross jgross@suse.com
commit 74e7e1efdad45580cc3839f2a155174cf158f9b5 upstream.
It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So remove kfree_skb() from the spin_lock_irqsave() section and use the already existing "drop" label in xenvif_start_xmit() for dropping the SKB. At the same time replace the dev_kfree_skb() call there with a call of dev_kfree_skb_any(), as xenvif_start_xmit() can be called with disabled interrupts.
This is XSA-424 / CVE-2022-42328 / CVE-2022-42329.
Fixes: be81992f9086 ("xen/netback: don't queue unlimited number of packages") Reported-by: Yang Yingliang yangyingliang@huawei.com Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Jan Beulich jbeulich@suse.com Signed-off-by: Juergen Gross jgross@suse.com Signed-off-by: Meena Shanmugam meenashanmugam@google.com --- drivers/net/xen-netback/common.h | 2 +- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/rx.c | 8 +++++--- 3 files changed, 10 insertions(+), 6 deletions(-)
diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h index d9dea4829c86..bdb3139c7162 100644 --- a/drivers/net/xen-netback/common.h +++ b/drivers/net/xen-netback/common.h @@ -395,7 +395,7 @@ irqreturn_t xenvif_ctrl_irq_fn(int irq, void *data);
bool xenvif_have_rx_work(struct xenvif_queue *queue, bool test_kthread); void xenvif_rx_action(struct xenvif_queue *queue); -void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb); +bool xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb);
void xenvif_carrier_on(struct xenvif *vif);
diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c index c58996c1e230..6a35772fde7a 100644 --- a/drivers/net/xen-netback/interface.c +++ b/drivers/net/xen-netback/interface.c @@ -269,14 +269,16 @@ xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev) if (vif->hash.alg == XEN_NETIF_CTRL_HASH_ALGORITHM_NONE) skb_clear_hash(skb);
- xenvif_rx_queue_tail(queue, skb); + if (!xenvif_rx_queue_tail(queue, skb)) + goto drop; + xenvif_kick_thread(queue);
return NETDEV_TX_OK;
drop: vif->dev->stats.tx_dropped++; - dev_kfree_skb(skb); + dev_kfree_skb_any(skb); return NETDEV_TX_OK; }
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c index a0335407be42..c2671eb6ad93 100644 --- a/drivers/net/xen-netback/rx.c +++ b/drivers/net/xen-netback/rx.c @@ -82,9 +82,10 @@ static bool xenvif_rx_ring_slots_available(struct xenvif_queue *queue) return false; }
-void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb) +bool xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb) { unsigned long flags; + bool ret = true;
spin_lock_irqsave(&queue->rx_queue.lock, flags);
@@ -92,8 +93,7 @@ void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb) struct net_device *dev = queue->vif->dev;
netif_tx_stop_queue(netdev_get_tx_queue(dev, queue->id)); - kfree_skb(skb); - queue->vif->dev->stats.rx_dropped++; + ret = false; } else { if (skb_queue_empty(&queue->rx_queue)) xenvif_update_needed_slots(queue, skb); @@ -104,6 +104,8 @@ void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb) }
spin_unlock_irqrestore(&queue->rx_queue.lock, flags); + + return ret; }
static struct sk_buff *xenvif_rx_dequeue(struct xenvif_queue *queue)
On 13.12.22 22:53, Meena Shanmugam wrote:
The commit 74e7e1efdad4 (xen/netback: don't call kfree_skb() with interrupts disabled) fixes deadlock in Linux netback driver. This seems to be a good candidate for the stable trees. This patch didn't apply cleanly in 5.15 kernel due to difference in function prototypes in drivers/net/xen-netback/common.h.
Juergen Gross (1): xen/netback: don't call kfree_skb() with interrupts disabled
drivers/net/xen-netback/common.h | 2 +- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/rx.c | 8 +++++--- 3 files changed, 10 insertions(+), 6 deletions(-)
It has already been picked.
Juergen
On Tue, Dec 13, 2022 at 09:53:38PM +0000, Meena Shanmugam wrote:
The commit 74e7e1efdad4 (xen/netback: don't call kfree_skb() with interrupts disabled) fixes deadlock in Linux netback driver. This seems to be a good candidate for the stable trees. This patch didn't apply cleanly in 5.15 kernel due to difference in function prototypes in drivers/net/xen-netback/common.h.
Juergen Gross (1): xen/netback: don't call kfree_skb() with interrupts disabled
drivers/net/xen-netback/common.h | 2 +- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/rx.c | 8 +++++--- 3 files changed, 10 insertions(+), 6 deletions(-)
-- 2.39.0.rc1.256.g54fd8350bd-goog
Can you just test the latest stable -rc releases that were announced a few days ago instead? It has this commit in it, right?
thanks,
greg k-h
On Wed, Dec 14, 2022 at 12:53 AM Greg KH gregkh@linuxfoundation.org wrote:
On Tue, Dec 13, 2022 at 09:53:38PM +0000, Meena Shanmugam wrote:
The commit 74e7e1efdad4 (xen/netback: don't call kfree_skb() with interrupts disabled) fixes deadlock in Linux netback driver. This seems to be a good candidate for the stable trees. This patch didn't apply cleanly in 5.15 kernel due to difference in function prototypes in drivers/net/xen-netback/common.h.
Juergen Gross (1): xen/netback: don't call kfree_skb() with interrupts disabled
drivers/net/xen-netback/common.h | 2 +- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/rx.c | 8 +++++--- 3 files changed, 10 insertions(+), 6 deletions(-)
-- 2.39.0.rc1.256.g54fd8350bd-goog
Can you just test the latest stable -rc releases that were announced a few days ago instead? It has this commit in it, right?
thanks,
greg k-h
Sorry, I was testing using 5.15.82 and I didn't realize that it was already queued for 5.15.83.
Thanks, Meena
linux-stable-mirror@lists.linaro.org
-
Greg KH -
Juergen Gross -
Meena Shanmugam