packages/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
qemu/hw-net-virtio-net-fix-qemu-set-used-ring-flag-even-v.patch
Jiabo Feng 7bb77ea74f QEMU update to version 6.2.0-103: 
- Introduce the SM4 cipher algorithms (OSCCA GB/T 32907-2016).
- intel_iommu: Add missed sanity check for 256-bit invalidation queue
- linux-user: use 'max' instead of 'qemu32' / 'qemu64' by default
- chardev/baum: Use definitions to avoid dynamic stack allocation
- ui/console: Get tab completion working again in the SDL monitor vc
- s390x/tcg: Fix opcode for lzrf
- virtiofsd: use g_date_time_get_microsecond to get subsecond
- ui/curses: Avoid dynamic stack allocation
- target/m68k: always call gen_exit_tb() after writes to SR
- target/m68k: Perform writback before modifying SR
- target/m68k: Fix MACSR to CCR
- target/m68k: Implement atomic test-and-set
- block/nvme: nvme_process_completion() fix bound for cid
- hw/pci-host: pnv_phb{3, 4}: Fix heap out-of-bound access failure
- target/ppc: Zero second doubleword of VSR registers for FPR insns
- target/ppc: Set OV32 when OV is set
- target/ppc: Zero second doubleword for VSX madd instructions
- target/ppc: Set result to QNaN for DENBCD when VXCVI occurs
- hw/pci: Add parenthesis to PCI_BUILD_BDF macro
- intel_iommu: Send IQE event when setting reserved bit in IQT_TAIL
- acpi: cpuhp: fix guest-visible maximum access size to the legacy reg block
- acpi: ged: Add macro for acpi sleep control register
- hw/pci-bridge: Add a Kconfig switch for the normal PCI bridge
- ui/vnc: fix handling of VNC_FEATURE_XVP
- s390/sclp: fix SCLP facility map
- docs/tools/qemu-img.rst: fix typo (sumarizes)
- chardev/char: fix qemu_chr_is_busy() check
- edu: fix DMA range upper bound check
- platform-bus: fix refcount leak
- hw/net/virtio-net: fix qemu set used ring flag even vhost started
- hw/net/can/sja1000: fix bug for single acceptance filter and standard frame
- tests/avocado: fix typo in replay_linux
- util/userfaultfd: Remove unused uffd_poll_events
- hw/core/ptimer: fix timer zero period condition for freq > 1GHz
- hcd-ohci: Drop ohci_service_iso_td() if ed->head & OHCI_DPTR_MASK is zero
- tests/unit/test-vmstate: Avoid dynamic stack allocation
- hw/usb/hcd-ohci: Use definition to avoid dynamic stack allocation
- hw/i386/multiboot: Avoid dynamic stack allocation
- hw/ppc/spapr: Fix code style problems reported by checkpatch
- chardev/baum: Replace magic values by X_MAX / Y_MAX definitions
- hw/intc/xics: Avoid dynamic stack allocation
- hw/net/e1000e_core: Use definition to avoid dynamic stack allocation
- intel_iommu: Fix invalidation descriptor type field
- configs: Fix typo in the sh4-softmmu devices config file

Signed-off-by: Jiabo Feng <fengjiabo1@huawei.com>
(cherry picked from commit 9813ed21ec2499c50cb58ac5fb114a1641708eb2)
2024-11-30 11:32:24 +08:00

74 lines
4.1 KiB
Diff
Raw Blame History

From 8784e43349fe54c50ecd9df804bb610c2b7f6f3c Mon Sep 17 00:00:00 2001
From: Zhang Jiao <zhangjiao2_yewu@cmss.chinamobile.com>
Date: Thu, 17 Oct 2024 12:14:27 +0800
Subject: [PATCH] hw/net/virtio-net: fix qemu set used ring flag even vhost
started
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
cheery-pick from 4c54f5bc8e1d38f15cc35b6a6932d8fbe219c692
When vhost-user or vhost-kernel is handling virtio net datapath,
QEMU should not touch used ring.
But with vhost-user socket reconnect scenario, in a very rare case
(has pending kick event). VRING_USED_F_NO_NOTIFY is set by QEMU in
following code path:
#0 virtio_queue_split_set_notification (vq=0x7ff5f4c920a8, enable=0) at ../hw/virtio/virtio.c:511
#1 0x0000559d6dbf033b in virtio_queue_set_notification (vq=0x7ff5f4c920a8, enable=0) at ../hw/virtio/virtio.c:576
#2 0x0000559d6dbbbdbc in virtio_net_handle_tx_bh (vdev=0x559d703a6aa0, vq=0x7ff5f4c920a8) at ../hw/net/virtio-net.c:2801
#3 0x0000559d6dbf4791 in virtio_queue_notify_vq (vq=0x7ff5f4c920a8) at ../hw/virtio/virtio.c:2248
#4 0x0000559d6dbf79da in virtio_queue_host_notifier_read (n=0x7ff5f4c9211c) at ../hw/virtio/virtio.c:3525
#5 0x0000559d6d9a5814 in virtio_bus_cleanup_host_notifier (bus=0x559d703a6a20, n=1) at ../hw/virtio/virtio-bus.c:321
#6 0x0000559d6dbf83c9 in virtio_device_stop_ioeventfd_impl (vdev=0x559d703a6aa0) at ../hw/virtio/virtio.c:3774
#7 0x0000559d6d9a55c8 in virtio_bus_stop_ioeventfd (bus=0x559d703a6a20) at ../hw/virtio/virtio-bus.c:259
#8 0x0000559d6d9a53e8 in virtio_bus_grab_ioeventfd (bus=0x559d703a6a20) at ../hw/virtio/virtio-bus.c:199
#9 0x0000559d6dbf841c in virtio_device_grab_ioeventfd (vdev=0x559d703a6aa0) at ../hw/virtio/virtio.c:3783
#10 0x0000559d6d9bde18 in vhost_dev_enable_notifiers (hdev=0x559d707edd70, vdev=0x559d703a6aa0) at ../hw/virtio/vhost.c:1592
#11 0x0000559d6d89a0b8 in vhost_net_start_one (net=0x559d707edd70, dev=0x559d703a6aa0) at ../hw/net/vhost_net.c:266
#12 0x0000559d6d89a6df in vhost_net_start (dev=0x559d703a6aa0, ncs=0x559d7048d890, data_queue_pairs=31, cvq=0) at ../hw/net/vhost_net.c:412
#13 0x0000559d6dbb5b89 in virtio_net_vhost_status (n=0x559d703a6aa0, status=15 '\017') at ../hw/net/virtio-net.c:311
#14 0x0000559d6dbb5e34 in virtio_net_set_status (vdev=0x559d703a6aa0, status=15 '\017') at ../hw/net/virtio-net.c:392
#15 0x0000559d6dbb60d8 in virtio_net_set_link_status (nc=0x559d7048d890) at ../hw/net/virtio-net.c:455
#16 0x0000559d6da64863 in qmp_set_link (name=0x559d6f0b83d0 "hostnet1", up=true, errp=0x7ffdd76569f0) at ../net/net.c:1459
#17 0x0000559d6da7226e in net_vhost_user_event (opaque=0x559d6f0b83d0, event=CHR_EVENT_OPENED) at ../net/vhost-user.c:301
#18 0x0000559d6ddc7f63 in chr_be_event (s=0x559d6f2ffea0, event=CHR_EVENT_OPENED) at ../chardev/char.c:62
#19 0x0000559d6ddc7fdc in qemu_chr_be_event (s=0x559d6f2ffea0, event=CHR_EVENT_OPENED) at ../chardev/char.c:82
This issue causes guest kernel stop kicking device and traffic stop.
Add vhost_started check in virtio_net_handle_tx_bh to fix this wrong
VRING_USED_F_NO_NOTIFY set.
Signed-off-by: Yajun Wu <yajunw@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-ID: <20240402045109.97729-1-yajunw@nvidia.com>
[PMD: Use unlikely()]
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Zhang Jiao <zhangjiao2_yewu@cmss.chinamobile.com>
---
hw/net/virtio-net.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
index cef7fc0aa0..7c87aa28ef 100644
--- a/hw/net/virtio-net.c
+++ b/hw/net/virtio-net.c
@@ -2709,6 +2709,10 @@ static void virtio_net_handle_tx_bh(VirtIODevice *vdev, VirtQueue *vq)
VirtIONet *n = VIRTIO_NET(vdev);
VirtIONetQueue *q = &n->vqs[vq2q(virtio_get_queue_index(vq))];
+ if (unlikely(n->vhost_started)) {
+ return;
+ }
+
if (unlikely((n->status & VIRTIO_NET_S_LINK_UP) == 0)) {
virtio_net_drop_tx_queue_data(vdev, vq);
return;
--
2.41.0.windows.1
Reference in New Issue View Git Blame Copy Permalink