Changes in vsock.c

e6ab4500 - vsock/virtio: add SIOCOUTQ support for all virtio based transports

Tue, 30 Jul 2024 19:43:07 +0000

vsock/virtio: add SIOCOUTQ support for all virtio based transportsIntroduce support for virtio_transport_unsent_bytesioctl for virtio_transport, vhost_vsock and vsock_loopback.For all transports the unsent bytes counter is incrementedin virtio_transport_get_credit.In virtio_transport (G2H) and in vhost-vsock (H2G) the counteris decremented when the skbuff is consumed. In vsock_loopback thesame skbuff is passed from the transmitter to the receiver, sothe counter is decremented before queuing the skbuff to thereceiver.Signed-off-by: Luigi Leonardi Reviewed-by: Stefano Garzarella Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

1e1fdcbd - vhost/vsock: always initialize seqpacket_allow

Mon, 22 Apr 2024 14:03:13 +0000

vhost/vsock: always initialize seqpacket_allowThere are two issues around seqpacket_allow:1. seqpacket_allow is not initialized when socket is created. Thus if features are never set, it will be read uninitialized.2. if VIRTIO_VSOCK_F_SEQPACKET is set and then cleared, then seqpacket_allow will not be cleared appropriately (existing apps I know about don't usually do this but it's legal and there's no way to be sure no one relies on this).To fix: - initialize seqpacket_allow after allocation - set it unconditionally in set_featuresReported-by: syzbot+6c21aeb59d0e82eb2782@syzkaller.appspotmail.comReported-by: Jeongjun Park Fixes: ced7b713711f ("vhost/vsock: support SEQPACKET for transport").Tested-by: Arseny Krasnov Cc: David S. Miller Cc: Stefan Hajnoczi Message-ID: <20240422100010-mutt-send-email-mst@kernel.org>Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang Reviewed-by: Stefano Garzarella Reviewed-by: Eugenio Pérez Acked-by: Jakub Kicinski List of files: /linux-6.15/drivers/vhost/vsock.c

0fe17989 - virtio/vsock: send credit update during setting SO_RCVLOWAT

Thu, 14 Dec 2023 12:52:29 +0000

virtio/vsock: send credit update during setting SO_RCVLOWATSend credit update message when SO_RCVLOWAT is updated and it is biggerthan number of bytes in rx queue. It is needed, because 'poll()' willwait until number of bytes in rx queue will be not smaller thanO_RCVLOWAT, so kick sender to send more data. Otherwise mutual hungupfor tx/rx is possible: sender waits for free space and receiver iswaiting data in 'poll()'.Rename 'set_rcvlowat' callback to 'notify_set_rcvlowat' and set'sk->sk_rcvlowat' only in one place (i.e. 'vsock_set_rcvlowat'), so thetransport doesn't need to do it.Fixes: b89d882dc9fc ("vsock/virtio: reduce credit update messages")Signed-off-by: Arseniy Krasnov Reviewed-by: Stefano Garzarella Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

3719c48d - vhost/vsock: support MSG_ZEROCOPY for transport

Tue, 10 Oct 2023 19:15:17 +0000

vhost/vsock: support MSG_ZEROCOPY for transportAdd 'msgzerocopy_allow()' callback for vhost transport.Signed-off-by: Arseniy Krasnov Reviewed-by: Stefano Garzarella Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

0df7cd3c - vsock/virtio/vhost: read data from non-linear skb

Sat, 16 Sep 2023 13:09:15 +0000

vsock/virtio/vhost: read data from non-linear skbThis is preparation patch for MSG_ZEROCOPY support. It adds handling ofnon-linear skbs by replacing direct calls of 'memcpy_to_msg()' with'skb_copy_datagram_iter()'. Main advantage of the second one is that itcan handle paged part of the skb by using 'kmap()' on each page, but ifthere are no pages in the skb, it behaves like simple copying to ioviterator. This patch also adds new field to the control block of skb -this value shows current offset in the skb to read next portion of data(it doesn't matter linear it or not). Idea behind this field is that'skb_copy_datagram_iter()' handles both types of skb internally - itjust needs an offset from which to copy data from the given skb. Thisoffset is incremented on each read from skb. This approach allows tosimplify handling of both linear and non-linear skbs, because forlinear skb we need to call 'skb_pull()' after reading data from it,while in non-linear case we need to update 'data_len'.Signed-off-by: Arseniy Krasnov Reviewed-by: Stefano Garzarella Acked-by: Michael S. Tsirkin Signed-off-by: Paolo Abeni List of files: /linux-6.15/drivers/vhost/vsock.c

9e09d0ec - vhost_sock: convert to vhost_vq_work_queue

Mon, 26 Jun 2023 23:22:58 +0000

vhost_sock: convert to vhost_vq_work_queueConvert from vhost_work_queue to vhost_vq_work_queue, so we can dropvhost_work_queue.Signed-off-by: Mike Christie Message-Id: <20230626232307.97930-9-michael.christie@oracle.com>Signed-off-by: Michael S. Tsirkin List of files: /linux-6.15/drivers/vhost/vsock.c

634f1a71 - vsock: support sockmap

Mon, 27 Mar 2023 19:11:51 +0000

vsock: support sockmapThis patch adds sockmap support for vsock sockets. It is intended to beusable by all transports, but only the virtio and loopback transportsare implemented.SOCK_STREAM, SOCK_DGRAM, and SOCK_SEQPACKET are all supported.Signed-off-by: Bobby Eshleman Acked-by: Michael S. Tsirkin Reviewed-by: Stefano Garzarella Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

759aba1e - vhost: remove unused paramete

Tue, 10 Jan 2023 02:44:45 +0000

vhost: remove unused paramete"enabled" is defined in vhost_init_device_iotlb,but it is never used. Let's remove it.Signed-off-by: Liming Wu Message-Id: <20230110024445.303-1-liming.wu@jaguarmicro.com>Signed-off-by: Michael S. Tsirkin Reviewed-by: Stefano Garzarella Reviewed-by: Eugenio Pérez List of files: /linux-6.15/drivers/vhost/vsock.c

71dc9ec9 - virtio/vsock: replace virtio_vsock_pkt with sk_buff

Fri, 13 Jan 2023 22:21:37 +0000

virtio/vsock: replace virtio_vsock_pkt with sk_buffThis commit changes virtio/vsock to use sk_buff instead ofvirtio_vsock_pkt. Beyond better conforming to other net code, usingsk_buff allows vsock to use sk_buff-dependent features in the future(such as sockmap) and improves throughput.This patch introduces the following performance changes:Tool: UperfEnv: Phys Host + L1 GuestPayload: 64kThreads: 16Test Runs: 10Type: SOCK_STREAMBefore: commit b7bfaa761d760 ("Linux 6.2-rc3")Before------g2h: 16.77Gb/sh2g: 10.56Gb/sAfter-----g2h: 21.04Gb/sh2g: 10.76Gb/sSigned-off-by: Bobby Eshleman Reviewed-by: Stefano Garzarella Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

7a4efe18 - vhost/vsock: Fix error handling in vhost_vsock_init()

Tue, 08 Nov 2022 10:17:05 +0000

vhost/vsock: Fix error handling in vhost_vsock_init()A problem about modprobe vhost_vsock failed is triggered with thefollowing log given:modprobe: ERROR: could not insert 'vhost_vsock': Device or resource busyThe reason is that vhost_vsock_init() returns misc_register() directlywithout checking its return value, if misc_register() failed, it returnswithout calling vsock_core_unregister() on vhost_transport, resulting thevhost_vsock can never be installed later.A simple call graph is shown as below: vhost_vsock_init() vsock_core_register() # register vhost_transport misc_register() device_create_with_groups() device_create_groups_vargs() dev = kzalloc(...) # OOM happened # return without unregister vhost_transportFix by calling vsock_core_unregister() when misc_register() returns error.Fixes: 433fc58e6bf2 ("VSOCK: Introduce vhost_vsock.ko")Signed-off-by: Yuan Can Message-Id: <20221108101705.45981-1-yuancan@huawei.com>Signed-off-by: Michael S. Tsirkin Reviewed-by: Stefano Garzarella Acked-by: Jason Wang List of files: /linux-6.15/drivers/vhost/vsock.c

de4eda9d - use less confusing names for iov_iter direction initializers

Fri, 16 Sep 2022 00:25:47 +0000

use less confusing names for iov_iter direction initializersREAD/WRITE proved to be actively confusing - the meanings are"data destination, as used with read(2)" and "data source, asused with write(2)", but people keep interpreting those as"we read data from it" and "we write data to it", i.e. exactlythe wrong way.Call them ITER_DEST and ITER_SOURCE - at least that is harderto misinterpret...Signed-off-by: Al Viro List of files: /linux-6.15/drivers/vhost/vsock.c

0e3f7293 - vhost/vsock: Use kvmalloc/kvfree for larger packets.

Wed, 28 Sep 2022 06:45:38 +0000

vhost/vsock: Use kvmalloc/kvfree for larger packets.When copying a large file over sftp over vsock, data size is usually 32kB,and kmalloc seems to fail to try to allocate 32 32kB regions. vhost-5837: page allocation failure: order:4, mode:0x24040c0 Call Trace: [] dump_stack+0x97/0xdb [] warn_alloc_failed+0x10f/0x138 [] ? __alloc_pages_direct_compact+0x38/0xc8 [] __alloc_pages_nodemask+0x84c/0x90d [] alloc_kmem_pages+0x17/0x19 [] kmalloc_order_trace+0x2b/0xdb [] __kmalloc+0x177/0x1f7 [] ? copy_from_iter+0x8d/0x31d [] vhost_vsock_handle_tx_kick+0x1fa/0x301 [vhost_vsock] [] vhost_worker+0xf7/0x157 [vhost] [] kthread+0xfd/0x105 [] ? vhost_dev_set_owner+0x22e/0x22e [vhost] [] ? flush_kthread_worker+0xf3/0xf3 [] ret_from_fork+0x4e/0x80 [] ? flush_kthread_worker+0xf3/0xf3Work around by doing kvmalloc instead.Fixes: 433fc58e6bf2 ("VSOCK: Introduce vhost_vsock.ko")Signed-off-by: Junichi Uekawa Reviewed-by: Stefano Garzarella Acked-by: Michael S. Tsirkin Link: https://lore.kernel.org/r/20220928064538.667678-1-uekawa@chromium.orgSigned-off-by: Jakub Kicinski List of files: /linux-6.15/drivers/vhost/vsock.c

b2ffa407 - vhost: rename vhost_work_dev_flush

Tue, 17 May 2022 18:08:50 +0000

vhost: rename vhost_work_dev_flushThis patch renames vhost_work_dev_flush to just vhost_dev_flush torelfect that it flushes everything on the device and that driversdon't know/care that polls are based on vhost_works. Drivers justflush the entire device and polls, and works for vhost-scsimanagement TMFs and IO net virtqueues, etc all are flushed.Signed-off-by: Mike Christie Acked-by: Jason Wang Reviewed-by: Stefano Garzarella Message-Id: <20220517180850.198915-9-michael.christie@oracle.com>Signed-off-by: Michael S. Tsirkin List of files: /linux-6.15/drivers/vhost/vsock.c

15538ba5 - vhost_vsock: simplify vhost_vsock_flush()

Tue, 17 May 2022 18:08:47 +0000

vhost_vsock: simplify vhost_vsock_flush()vhost_vsock_flush() calls vhost_work_dev_flush(vsock->vqs[i].poll.dev)before vhost_work_dev_flush(&vsock->dev). This seems pointlessas vsock->vqs[i].poll.dev is the same as &vsock->dev and several flushesin a row doesn't do anything useful, one is just enough.Signed-off-by: Andrey Ryabinin Reviewed-by: Stefano Garzarella Signed-off-by: Mike Christie Acked-by: Jason Wang Message-Id: <20220517180850.198915-6-michael.christie@oracle.com>Signed-off-by: Michael S. Tsirkin List of files: /linux-6.15/drivers/vhost/vsock.c

6fcf224c - vhost: get rid of vhost_poll_flush() wrapper

Tue, 17 May 2022 18:08:43 +0000

vhost: get rid of vhost_poll_flush() wrappervhost_poll_flush() is a simple wrapper around vhost_work_dev_flush().It gives wrong impression that we are doing some work over vhost_poll,while in fact it flushes vhost_poll->dev.It only complicate understanding of the code and leads to mistakeslike flushing the same vhost_dev several times in a row.Just remove vhost_poll_flush() and call vhost_work_dev_flush() directly.Signed-off-by: Andrey Ryabinin [merge vhost_poll_flush removal from Stefano Garzarella]Signed-off-by: Mike Christie Reviewed-by: Chaitanya Kulkarni Acked-by: Jason Wang Reviewed-by: Stefano Garzarella Message-Id: <20220517180850.198915-2-michael.christie@oracle.com>Signed-off-by: Michael S. Tsirkin List of files: /linux-6.15/drivers/vhost/vsock.c

8e6ed963 - vsock: each transport cycles only on its own sockets

Fri, 11 Mar 2022 02:00:16 +0000

vsock: each transport cycles only on its own socketsWhen iterating over sockets using vsock_for_each_connected_socket, makesure that a transport filters out sockets that don't belong to thetransport.There actually was an issue caused by this; in a nested VMconfiguration, destroying the nested VM (which often involves theclosing of /dev/vhost-vsock if there was h2g connections to the nestedVM) kills not only the h2g connections, but also all existing g2hconnections to the (outmost) host which are totally unrelated.Tested: Executed the following steps on Cuttlefish (Android running on aVM) [1]: (1) Enter into an `adb shell` session - to have a g2hconnection inside the VM, (2) open and then close /dev/vhost-vsock by`exec 3< /dev/vhost-vsock && exec 3<&-`, (3) observe that the adbsession is not reset.[1] https://android.googlesource.com/device/google/cuttlefish/Fixes: c0cfa2d8a788 ("vsock: add multi-transports support")Reviewed-by: Stefano Garzarella Acked-by: Michael S. Tsirkin Signed-off-by: Jiyong Park Link: https://lore.kernel.org/r/20220311020017.1509316-1-jiyong@google.comSigned-off-by: Jakub Kicinski List of files: /linux-6.15/drivers/vhost/vsock.c

a58da53f - vhost/vsock: don't check owner in vhost_vsock_stop() while releasing

Tue, 22 Feb 2022 09:47:42 +0000

vhost/vsock: don't check owner in vhost_vsock_stop() while releasingvhost_vsock_stop() calls vhost_dev_check_owner() to check the deviceownership. It expects current->mm to be valid.vhost_vsock_stop() is also called by vhost_vsock_dev_release() whenthe user has not done close(), so when we are in do_exit(). In thiscase current->mm is invalid and we're releasing the device, so weshould clean it anyway.Let's check the owner only when vhost_vsock_stop() is calledby an ioctl.When invoked from release we can not fail so we don't check returncode of vhost_vsock_stop(). We need to stop vsock even if it's notthe owner.Fixes: 433fc58e6bf2 ("VSOCK: Introduce vhost_vsock.ko")Cc: stable@vger.kernel.orgReported-by: syzbot+1e3ea63db39f2b4440e0@syzkaller.appspotmail.comReported-and-tested-by: syzbot+3140b17cb44a7b174008@syzkaller.appspotmail.comSigned-off-by: Stefano Garzarella Acked-by: Jason Wang Signed-off-by: David S. Miller List of files: /linux-6.15/drivers/vhost/vsock.c

11708ff9 - vhost/vsock: cleanup removing `len` variable

Mon, 22 Nov 2021 16:35:25 +0000

vhost/vsock: cleanup removing `len` variableWe can increment `total_len` directly and remove `len` since itis no longer used for vhost_add_used().Signed-off-by: Stefano Garzarella Link: https://lore.kernel.org/r/20211122163525.294024-3-sgarzare@redhat.comSigned-off-by: Michael S. Tsirkin Reviewed-by: Stefan Hajnoczi List of files: /linux-6.15/drivers/vhost/vsock.c

49d8c5ff - vhost/vsock: fix incorrect used length reported to the guest

Mon, 22 Nov 2021 16:35:24 +0000

vhost/vsock: fix incorrect used length reported to the guestThe "used length" reported by calling vhost_add_used() must be thenumber of bytes written by the device (using "in" buffers).In vhost_vsock_handle_tx_kick() the device only reads the guestbuffers (they are all "out" buffers), without writing anything,so we must pass 0 as "used length" to comply virtio spec.Fixes: 433fc58e6bf2 ("VSOCK: Introduce vhost_vsock.ko")Cc: stable@vger.kernel.orgReported-by: Halil Pasic Suggested-by: Jason Wang Signed-off-by: Stefano Garzarella Link: https://lore.kernel.org/r/20211122163525.294024-2-sgarzare@redhat.comSigned-off-by: Michael S. Tsirkin Reviewed-by: Stefan Hajnoczi Reviewed-by: Halil Pasic List of files: /linux-6.15/drivers/vhost/vsock.c

1af7e555 - vhost/vsock: support MSG_EOR bit processing

Fri, 03 Sep 2021 12:32:35 +0000

vhost/vsock: support MSG_EOR bit processing'MSG_EOR' handling has similar logic as 'MSG_EOM' - if bit presentin packet's header, reset it to 0. Then restore it back if packetprocessing wasn't completed. Instead of bool variable for eachflag, bit mask variable was added: it has logical OR of 'MSG_EOR'and 'MSG_EOM' if needed, to restore flags, this variable is ORedwith flags field of packet.Signed-off-by: Arseny Krasnov Link: https://lore.kernel.org/r/20210903123238.3273526-1-arseny.krasnov@kaspersky.comSigned-off-by: Michael S. Tsirkin Reviewed-by: Stefano Garzarella List of files: /linux-6.15/drivers/vhost/vsock.c