| Message ID | 20241115103016.86461-1-graf@amazon.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | vsock/virtio: Remove queued_replies pushback logic | expand |
On Fri, Nov 15, 2024 at 10:30:16AM +0000, Alexander Graf wrote: >Ever since the introduction of the virtio vsock driver, it included >pushback logic that blocks it from taking any new RX packets until the >TX queue backlog becomes shallower than the virtqueue size. > >This logic works fine when you connect a user space application on the >hypervisor with a virtio-vsock target, because the guest will stop >receiving data until the host pulled all outstanding data from the VM. So, why not skipping this only when talking with a sibling VM? > >With Nitro Enclaves however, we connect 2 VMs directly via vsock: > > Parent Enclave > > RX -------- TX > TX -------- RX > >This means we now have 2 virtio-vsock backends that both have the pushback >logic. If the parent's TX queue runs full at the same time as the >Enclave's, both virtio-vsock drivers fall into the pushback path and >no longer accept RX traffic. However, that RX traffic is TX traffic on >the other side which blocks that driver from making any forward >progress. We're not in a deadlock. > >To resolve this, let's remove that pushback logic altogether and rely on >higher levels (like credits) to ensure we do not consume unbounded >memory. I spoke quickly with Stefan who has been following the development from the beginning and actually pointed out that there might be problems with the control packets, since credits only covers data packets, so it doesn't seem like a good idea remove this mechanism completely. > >Fixes: 0ea9e1d3a9e3 ("VSOCK: Introduce virtio_transport.ko") I'm not sure we should add this Fixes tag, this seems very risky backporting on stable branches IMHO. If we cannot find a better mechanism to replace this with something that works both guest <-> host and guest <-> guest, I would prefer to do this just for guest <-> guest communication. Because removing this completely seems too risky for me, at least without a proof that control packets are fine. Thanks, Stefano >Signed-off-by: Alexander Graf <graf@amazon.com> >--- > net/vmw_vsock/virtio_transport.c | 51 ++------------------------------ > 1 file changed, 2 insertions(+), 49 deletions(-) > >diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c >index 64a07acfef12..53e79779886c 100644 >--- a/net/vmw_vsock/virtio_transport.c >+++ b/net/vmw_vsock/virtio_transport.c >@@ -44,8 +44,6 @@ struct virtio_vsock { > struct work_struct send_pkt_work; > struct sk_buff_head send_pkt_queue; > >- atomic_t queued_replies; >- > /* The following fields are protected by rx_lock. vqs[VSOCK_VQ_RX] > * must be accessed with rx_lock held. > */ >@@ -171,17 +169,6 @@ virtio_transport_send_pkt_work(struct work_struct *work) > > virtio_transport_deliver_tap_pkt(skb); > >- if (reply) { >- struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX]; >- int val; >- >- val = atomic_dec_return(&vsock->queued_replies); >- >- /* Do we now have resources to resume rx processing? */ >- if (val + 1 == virtqueue_get_vring_size(rx_vq)) >- restart_rx = true; >- } >- > added = true; > } > >@@ -218,9 +205,6 @@ virtio_transport_send_pkt(struct sk_buff *skb) > goto out_rcu; > } > >- if (virtio_vsock_skb_reply(skb)) >- atomic_inc(&vsock->queued_replies); >- > virtio_vsock_skb_queue_tail(&vsock->send_pkt_queue, skb); > queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work); > >@@ -233,7 +217,7 @@ static int > virtio_transport_cancel_pkt(struct vsock_sock *vsk) > { > struct virtio_vsock *vsock; >- int cnt = 0, ret; >+ int ret; > > rcu_read_lock(); > vsock = rcu_dereference(the_virtio_vsock); >@@ -242,17 +226,7 @@ virtio_transport_cancel_pkt(struct vsock_sock *vsk) > goto out_rcu; > } > >- cnt = virtio_transport_purge_skbs(vsk, &vsock->send_pkt_queue); >- >- if (cnt) { >- struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX]; >- int new_cnt; >- >- new_cnt = atomic_sub_return(cnt, &vsock->queued_replies); >- if (new_cnt + cnt >= virtqueue_get_vring_size(rx_vq) && >- new_cnt < virtqueue_get_vring_size(rx_vq)) >- queue_work(virtio_vsock_workqueue, &vsock->rx_work); >- } >+ virtio_transport_purge_skbs(vsk, &vsock->send_pkt_queue); > > ret = 0; > >@@ -323,18 +297,6 @@ static void virtio_transport_tx_work(struct work_struct *work) > queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work); > } > >-/* Is there space left for replies to rx packets? */ >-static bool virtio_transport_more_replies(struct virtio_vsock *vsock) >-{ >- struct virtqueue *vq = vsock->vqs[VSOCK_VQ_RX]; >- int val; >- >- smp_rmb(); /* paired with atomic_inc() and atomic_dec_return() */ >- val = atomic_read(&vsock->queued_replies); >- >- return val < virtqueue_get_vring_size(vq); >-} >- > /* event_lock must be held */ > static int virtio_vsock_event_fill_one(struct virtio_vsock *vsock, > struct virtio_vsock_event *event) >@@ -581,14 +543,6 @@ static void virtio_transport_rx_work(struct work_struct *work) > struct sk_buff *skb; > unsigned int len; > >- if (!virtio_transport_more_replies(vsock)) { >- /* Stop rx until the device processes already >- * pending replies. Leave rx virtqueue >- * callbacks disabled. >- */ >- goto out; >- } >- > skb = virtqueue_get_buf(vq, &len); > if (!skb) > break; >@@ -735,7 +689,6 @@ static int virtio_vsock_probe(struct virtio_device *vdev) > > vsock->rx_buf_nr = 0; > vsock->rx_buf_max_nr = 0; >- atomic_set(&vsock->queued_replies, 0); > > mutex_init(&vsock->tx_lock); > mutex_init(&vsock->rx_lock); >-- >2.40.1 > > > > >Amazon Web Services Development Center Germany GmbH >Krausenstr. 38 >10117 Berlin >Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss >Eingetragen am Amtsgericht Charlottenburg unter HRB 257764 B >Sitz: Berlin >Ust-ID: DE 365 538 597 > >
Hi Stefano, On 15.11.24 12:59, Stefano Garzarella wrote: > > On Fri, Nov 15, 2024 at 10:30:16AM +0000, Alexander Graf wrote: >> Ever since the introduction of the virtio vsock driver, it included >> pushback logic that blocks it from taking any new RX packets until the >> TX queue backlog becomes shallower than the virtqueue size. >> >> This logic works fine when you connect a user space application on the >> hypervisor with a virtio-vsock target, because the guest will stop >> receiving data until the host pulled all outstanding data from the VM. > > So, why not skipping this only when talking with a sibling VM? I don't think there is a way to know, is there? > >> >> With Nitro Enclaves however, we connect 2 VMs directly via vsock: >> >> Parent Enclave >> >> RX -------- TX >> TX -------- RX >> >> This means we now have 2 virtio-vsock backends that both have the >> pushback >> logic. If the parent's TX queue runs full at the same time as the >> Enclave's, both virtio-vsock drivers fall into the pushback path and >> no longer accept RX traffic. However, that RX traffic is TX traffic on >> the other side which blocks that driver from making any forward >> progress. We're not in a deadlock. >> >> To resolve this, let's remove that pushback logic altogether and rely on >> higher levels (like credits) to ensure we do not consume unbounded >> memory. > > I spoke quickly with Stefan who has been following the development from > the beginning and actually pointed out that there might be problems > with the control packets, since credits only covers data packets, so > it doesn't seem like a good idea remove this mechanism completely. Can you help me understand which situations the current mechanism really helps with, so we can look at alternatives? > >> >> Fixes: 0ea9e1d3a9e3 ("VSOCK: Introduce virtio_transport.ko") > > I'm not sure we should add this Fixes tag, this seems very risky > backporting on stable branches IMHO. Which situations do you believe it will genuinely break anything in? As it stands today, if you run upstream parent and enclave and hammer them with vsock traffic, you get into a deadlock. Even without the flow control, you will never hit a deadlock. But you may get a brown-out like situation while Linux is flushing its buffers. Ideally we want to have actual flow control to mitigate the problem altogether. But I'm not quite sure how and where. Just blocking all receiving traffic causes problems. > If we cannot find a better mechanism to replace this with something > that works both guest <-> host and guest <-> guest, I would prefer > to do this just for guest <-> guest communication. > Because removing this completely seems too risky for me, at least > without a proof that control packets are fine. So your concern is that control packets would not receive pushback, so we would allow unbounded traffic to get queued up? Can you suggest options to help with that? Alex Amazon Web Services Development Center Germany GmbH Krausenstr. 38 10117 Berlin Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss Eingetragen am Amtsgericht Charlottenburg unter HRB 257764 B Sitz: Berlin Ust-ID: DE 365 538 597
diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c index 64a07acfef12..53e79779886c 100644 --- a/net/vmw_vsock/virtio_transport.c +++ b/net/vmw_vsock/virtio_transport.c @@ -44,8 +44,6 @@ struct virtio_vsock { struct work_struct send_pkt_work; struct sk_buff_head send_pkt_queue; - atomic_t queued_replies; - /* The following fields are protected by rx_lock. vqs[VSOCK_VQ_RX] * must be accessed with rx_lock held. */ @@ -171,17 +169,6 @@ virtio_transport_send_pkt_work(struct work_struct *work) virtio_transport_deliver_tap_pkt(skb); - if (reply) { - struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX]; - int val; - - val = atomic_dec_return(&vsock->queued_replies); - - /* Do we now have resources to resume rx processing? */ - if (val + 1 == virtqueue_get_vring_size(rx_vq)) - restart_rx = true; - } - added = true; } @@ -218,9 +205,6 @@ virtio_transport_send_pkt(struct sk_buff *skb) goto out_rcu; } - if (virtio_vsock_skb_reply(skb)) - atomic_inc(&vsock->queued_replies); - virtio_vsock_skb_queue_tail(&vsock->send_pkt_queue, skb); queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work); @@ -233,7 +217,7 @@ static int virtio_transport_cancel_pkt(struct vsock_sock *vsk) { struct virtio_vsock *vsock; - int cnt = 0, ret; + int ret; rcu_read_lock(); vsock = rcu_dereference(the_virtio_vsock); @@ -242,17 +226,7 @@ virtio_transport_cancel_pkt(struct vsock_sock *vsk) goto out_rcu; } - cnt = virtio_transport_purge_skbs(vsk, &vsock->send_pkt_queue); - - if (cnt) { - struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX]; - int new_cnt; - - new_cnt = atomic_sub_return(cnt, &vsock->queued_replies); - if (new_cnt + cnt >= virtqueue_get_vring_size(rx_vq) && - new_cnt < virtqueue_get_vring_size(rx_vq)) - queue_work(virtio_vsock_workqueue, &vsock->rx_work); - } + virtio_transport_purge_skbs(vsk, &vsock->send_pkt_queue); ret = 0; @@ -323,18 +297,6 @@ static void virtio_transport_tx_work(struct work_struct *work) queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work); } -/* Is there space left for replies to rx packets? */ -static bool virtio_transport_more_replies(struct virtio_vsock *vsock) -{ - struct virtqueue *vq = vsock->vqs[VSOCK_VQ_RX]; - int val; - - smp_rmb(); /* paired with atomic_inc() and atomic_dec_return() */ - val = atomic_read(&vsock->queued_replies); - - return val < virtqueue_get_vring_size(vq); -} - /* event_lock must be held */ static int virtio_vsock_event_fill_one(struct virtio_vsock *vsock, struct virtio_vsock_event *event) @@ -581,14 +543,6 @@ static void virtio_transport_rx_work(struct work_struct *work) struct sk_buff *skb; unsigned int len; - if (!virtio_transport_more_replies(vsock)) { - /* Stop rx until the device processes already - * pending replies. Leave rx virtqueue - * callbacks disabled. - */ - goto out; - } - skb = virtqueue_get_buf(vq, &len); if (!skb) break; @@ -735,7 +689,6 @@ static int virtio_vsock_probe(struct virtio_device *vdev) vsock->rx_buf_nr = 0; vsock->rx_buf_max_nr = 0; - atomic_set(&vsock->queued_replies, 0); mutex_init(&vsock->tx_lock); mutex_init(&vsock->rx_lock);
Ever since the introduction of the virtio vsock driver, it included pushback logic that blocks it from taking any new RX packets until the TX queue backlog becomes shallower than the virtqueue size. This logic works fine when you connect a user space application on the hypervisor with a virtio-vsock target, because the guest will stop receiving data until the host pulled all outstanding data from the VM. With Nitro Enclaves however, we connect 2 VMs directly via vsock: Parent Enclave RX -------- TX TX -------- RX This means we now have 2 virtio-vsock backends that both have the pushback logic. If the parent's TX queue runs full at the same time as the Enclave's, both virtio-vsock drivers fall into the pushback path and no longer accept RX traffic. However, that RX traffic is TX traffic on the other side which blocks that driver from making any forward progress. We're not in a deadlock. To resolve this, let's remove that pushback logic altogether and rely on higher levels (like credits) to ensure we do not consume unbounded memory. Fixes: 0ea9e1d3a9e3 ("VSOCK: Introduce virtio_transport.ko") Signed-off-by: Alexander Graf <graf@amazon.com> --- net/vmw_vsock/virtio_transport.c | 51 ++------------------------------ 1 file changed, 2 insertions(+), 49 deletions(-)