search for: buf_alloc

...gt; avoid the starvation of others sockets. > > The guests re-fill the "free" queue when the available buffers are > > less than half. > > > > Each peer have these variables in the per-socket state: > > /* local vars */ > > buf_alloc /* max bytes usable by this socket > > [exposed to the other peer] */ > > fwd_cnt /* increased when RX packet is consumed by the > > user space [exposed to the other peer] */ > > tx_cnt...

...gt; avoid the starvation of others sockets. > > The guests re-fill the "free" queue when the available buffers are > > less than half. > > > > Each peer have these variables in the per-socket state: > > /* local vars */ > > buf_alloc /* max bytes usable by this socket > > [exposed to the other peer] */ > > fwd_cnt /* increased when RX packet is consumed by the > > user space [exposed to the other peer] */ > > tx_cnt...

fwd_cnt is written with rx_lock, so we should read it using the same spinlock also if we are in the TX path. Move also buf_alloc under rx_lock and add a missing locking when we modify it. Signed-off-by: Stefano Garzarella <sgarzare at redhat.com> --- include/linux/virtio_vsock.h | 2 +- net/vmw_vsock/virtio_transport_common.c | 6 ++++-- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/includ...

...te: Fri, 31 May 2019 15:39:51 +0200 > @@ -434,7 +434,9 @@ void virtio_transport_set_buffer_size(struct vsock_sock *vsk, u64 val) > if (val > vvs->buf_size_max) > vvs->buf_size_max = val; > vvs->buf_size = val; > + spin_lock_bh(&vvs->rx_lock); > vvs->buf_alloc = val; > + spin_unlock_bh(&vvs->rx_lock); This locking doesn't do anything other than to strongly order the buf_size store to occur before the buf_alloc one. If you need a memory barrier, use one.

On Wed, Jul 17, 2019 at 01:30:28PM +0200, Stefano Garzarella wrote: > fwd_cnt and last_fwd_cnt are protected by rx_lock, so we should use > the same spinlock also if we are in the TX path. > > Move also buf_alloc under the same lock. > > Signed-off-by: Stefano Garzarella <sgarzare at redhat.com> Wait a second is this a bugfix? If it's used under the wrong lock won't values get corrupted? Won't traffic then stall or more data get to sent than credits? > --- > include/linux/v...

...f others sockets. > > > The guests re-fill the "free" queue when the available buffers are > > > less than half. > > > > > > Each peer have these variables in the per-socket state: > > > /* local vars */ > > > buf_alloc /* max bytes usable by this socket > > > [exposed to the other peer] */ > > > fwd_cnt /* increased when RX packet is consumed by the > > > user space [exposed to the other peer] */ > > >...

...f others sockets. > > > The guests re-fill the "free" queue when the available buffers are > > > less than half. > > > > > > Each peer have these variables in the per-socket state: > > > /* local vars */ > > > buf_alloc /* max bytes usable by this socket > > > [exposed to the other peer] */ > > > fwd_cnt /* increased when RX packet is consumed by the > > > user space [exposed to the other peer] */ > > >...

...2 files changed, 12 insertions(+), 3 deletions(-) diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h index e223e2632edd..6d7a22cc20bf 100644 --- a/include/linux/virtio_vsock.h +++ b/include/linux/virtio_vsock.h @@ -37,6 +37,7 @@ struct virtio_vsock_sock { u32 tx_cnt; u32 buf_alloc; u32 peer_fwd_cnt; + u32 last_fwd_cnt; u32 peer_buf_alloc; /* Protected by rx_lock */ diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c index 602715fc9a75..f32301d823f5 100644 --- a/net/vmw_vsock/virtio_transport_common.c +++ b/net/vmw_vsock/virt...

...2 of this series I discovered an huge use of memory, so I added patch 1 to mitigate this issue. I put it in this series in order to better track the performance trends. v4: - rebased all patches on current master (conflicts is Patch 4) - Patch 1: added Stefan's R-b - Patch 3: removed lock when buf_alloc is written [David]; moved this patch after "vsock/virtio: reduce credit update messages" to make it clearer - Patch 4: vhost_exceeds_weight() is recently introduced, so I've solved some conflicts v3: https://patchwork.kernel.org/cover/10970145 v2: ht...

...2 of this series I discovered an huge use of memory, so I added patch 1 to mitigate this issue. I put it in this series in order to better track the performance trends. v4: - rebased all patches on current master (conflicts is Patch 4) - Patch 1: added Stefan's R-b - Patch 3: removed lock when buf_alloc is written [David]; moved this patch after "vsock/virtio: reduce credit update messages" to make it clearer - Patch 4: vhost_exceeds_weight() is recently introduced, so I've solved some conflicts v3: https://patchwork.kernel.org/cover/10970145 v2: ht...

...trace is of a simple request response (which both fit in one buffer in each direction), the lines without an "?X:" prefix are my annotations/guesses as to what is going on: TX: =>SRC:00000002.00010002 DST:00000003.00000948 TX: LEN:00000000 TYPE:0001 OP:1=REQUEST TX: FLAGS:00000000 BUF_ALLOC:00008000 FWD_CNT:00000000 TX: <=SRC:00000003.00000948 DST:00000002.00010002 TX: LEN:00000000 TYPE:0001 OP:2=RESPONSE TX: FLAGS:00000000 BUF_ALLOC:00040000 FWD_CNT:00000000 REQUEST + RESPONSE == Channel open successfully RX: =>SRC:00000002.00010002 DST:00000003.00000948 RX: LEN:0000005...

...virtio_vsock_sock *vvs, > void virtio_transport_inc_tx_pkt(struct virtio_vsock_sock *vvs, struct virtio_vsock_pkt *pkt) > { > spin_lock_bh(&vvs->tx_lock); > + vvs->last_fwd_cnt = vvs->fwd_cnt; > pkt->hdr.fwd_cnt = cpu_to_le32(vvs->fwd_cnt); > pkt->hdr.buf_alloc = cpu_to_le32(vvs->buf_alloc); > spin_unlock_bh(&vvs->tx_lock); > @@ -261,6 +262,7 @@ virtio_transport_stream_do_dequeue(struct vsock_sock *vsk, > struct virtio_vsock_sock *vvs = vsk->trans; > struct virtio_vsock_pkt *pkt; > size_t bytes, total = 0; > + u32 fr...

...virtio_vsock_sock *vvs, > void virtio_transport_inc_tx_pkt(struct virtio_vsock_sock *vvs, struct virtio_vsock_pkt *pkt) > { > spin_lock_bh(&vvs->tx_lock); > + vvs->last_fwd_cnt = vvs->fwd_cnt; > pkt->hdr.fwd_cnt = cpu_to_le32(vvs->fwd_cnt); > pkt->hdr.buf_alloc = cpu_to_le32(vvs->buf_alloc); > spin_unlock_bh(&vvs->tx_lock); > @@ -261,6 +262,7 @@ virtio_transport_stream_do_dequeue(struct vsock_sock *vsk, > struct virtio_vsock_sock *vvs = vsk->trans; > struct virtio_vsock_pkt *pkt; > size_t bytes, total = 0; > + u32 fr...

...m is per-socket, I think to > avoid the starvation of others sockets. > The guests re-fill the "free" queue when the available buffers are > less than half. > > Each peer have these variables in the per-socket state: > /* local vars */ > buf_alloc /* max bytes usable by this socket > [exposed to the other peer] */ > fwd_cnt /* increased when RX packet is consumed by the > user space [exposed to the other peer] */ > tx_cnt /* increased whe...

fwd_cnt and last_fwd_cnt are protected by rx_lock, so we should use the same spinlock also if we are in the TX path. Move also buf_alloc under the same lock. Signed-off-by: Stefano Garzarella <sgarzare at redhat.com> --- include/linux/virtio_vsock.h | 2 +- net/vmw_vsock/virtio_transport_common.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/include/linux/virtio_vsock.h b/include/linux/v...

...t; avoid the starvation of others sockets. > > The guests re-fill the "free" queue when the available buffers are > > less than half. > > > > Each peer have these variables in the per-socket state: > > /* local vars */ > > buf_alloc /* max bytes usable by this socket > > [exposed to the other peer] */ > > fwd_cnt /* increased when RX packet is consumed by the > > user space [exposed to the other peer] */ > > tx_cnt...

On Mon, Apr 11, 2016 at 03:54:08PM +0300, Michael S. Tsirkin wrote: > On Mon, Apr 11, 2016 at 11:45:48AM +0100, Stefan Hajnoczi wrote: > > On Fri, Apr 08, 2016 at 04:35:05PM +0100, Ian Campbell wrote: > > > On Fri, 2016-04-01 at 15:23 +0100, Stefan Hajnoczi wrote: > > > > This series is based on Michael Tsirkin's vhost branch (v4.5-rc6). > > > > >

On Mon, Apr 11, 2016 at 03:54:08PM +0300, Michael S. Tsirkin wrote: > On Mon, Apr 11, 2016 at 11:45:48AM +0100, Stefan Hajnoczi wrote: > > On Fri, Apr 08, 2016 at 04:35:05PM +0100, Ian Campbell wrote: > > > On Fri, 2016-04-01 at 15:23 +0100, Stefan Hajnoczi wrote: > > > > This series is based on Michael Tsirkin's vhost branch (v4.5-rc6). > > > > >

...148.30 148.87 143.89 490.037 256K 148.90 152.77 149.37 528.606 512K 140.68 153.65 152.64 516.622 [1] https://github.com/stefano-garzarella/iperf/ Stefano Garzarella (5): vsock/virtio: limit the memory used per-socket vsock/virtio: fix locking for fwd_cnt and buf_alloc vsock/virtio: reduce credit update messages vhost/vsock: split packets to send using multiple buffers vsock/virtio: change the maximum packet size allowed drivers/vhost/vsock.c | 53 ++++++++++--- include/linux/virtio_vsock.h | 4 +- net/vmw_vsock/virtio_tran...

This series makes the credit mechanism implemented in the virtio-vsock devices more robust. Patch 1 sends an update to the remote peer when the buf_alloc change. Patch 2 prevents a malicious peer (especially the guest) can consume all the memory of the other peer, discarding packets when the credit available is not respected. Stefano Garzarella (2): vsock/virtio: send a credit update when buffer size is changed vsock/virtio: discard packets if...