All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
	Greg Kurz <groug@kaod.org>, Yanghang Liu <yanghliu@redhat.com>,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Maxime Coquelin <maxime.coquelin@redhat.com>
Subject: [PULL 54/56] Revert "vhost-user: Monitor slave channel in vhost_user_read()"
Date: Mon, 30 Jan 2023 15:21:39 -0500	[thread overview]
Message-ID: <20230130201810.11518-55-mst@redhat.com> (raw)
In-Reply-To: <20230130201810.11518-1-mst@redhat.com>

From: Greg Kurz <groug@kaod.org>

This reverts commit db8a3772e300c1a656331a92da0785d81667dc81.

Motivation : this is breaking vhost-user with DPDK as reported in [0].

Received unexpected msg type. Expected 22 received 40
Fail to update device iotlb
Received unexpected msg type. Expected 40 received 22
Received unexpected msg type. Expected 22 received 11
Fail to update device iotlb
Received unexpected msg type. Expected 11 received 22
vhost VQ 1 ring restore failed: -71: Protocol error (71)
Received unexpected msg type. Expected 22 received 11
Fail to update device iotlb
Received unexpected msg type. Expected 11 received 22
vhost VQ 0 ring restore failed: -71: Protocol error (71)
unable to start vhost net: 71: falling back on userspace virtio

The failing sequence that leads to the first error is :
- QEMU sends a VHOST_USER_GET_STATUS (40) request to DPDK on the master
  socket
- QEMU starts a nested event loop in order to wait for the
  VHOST_USER_GET_STATUS response and to be able to process messages from
  the slave channel
- DPDK sends a couple of legitimate IOTLB miss messages on the slave
  channel
- QEMU processes each IOTLB request and sends VHOST_USER_IOTLB_MSG (22)
  updates on the master socket
- QEMU assumes to receive a response for the latest VHOST_USER_IOTLB_MSG
  but it gets the response for the VHOST_USER_GET_STATUS instead

The subsequent errors have the same root cause : the nested event loop
breaks the order by design. It lures QEMU to expect responses to the
latest message sent on the master socket to arrive first.

Since this was only needed for DAX enablement which is still not merged
upstream, just drop the code for now. A working solution will have to
be merged later on. Likely protect the master socket with a mutex
and service the slave channel with a separate thread, as discussed with
Maxime in the mail thread below.

[0] https://lore.kernel.org/qemu-devel/43145ede-89dc-280e-b953-6a2b436de395@redhat.com/

Reported-by: Yanghang Liu <yanghliu@redhat.com>
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2155173
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20230119172424.478268-2-groug@kaod.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>
---
 hw/virtio/vhost-user.c | 35 +++--------------------------------
 1 file changed, 3 insertions(+), 32 deletions(-)

diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c
index eca9e104ba..ad8883113b 100644
--- a/hw/virtio/vhost-user.c
+++ b/hw/virtio/vhost-user.c
@@ -356,35 +356,6 @@ end:
     return G_SOURCE_REMOVE;
 }
 
-static gboolean slave_read(QIOChannel *ioc, GIOCondition condition,
-                           gpointer opaque);
-
-/*
- * This updates the read handler to use a new event loop context.
- * Event sources are removed from the previous context : this ensures
- * that events detected in the previous context are purged. They will
- * be re-detected and processed in the new context.
- */
-static void slave_update_read_handler(struct vhost_dev *dev,
-                                      GMainContext *ctxt)
-{
-    struct vhost_user *u = dev->opaque;
-
-    if (!u->slave_ioc) {
-        return;
-    }
-
-    if (u->slave_src) {
-        g_source_destroy(u->slave_src);
-        g_source_unref(u->slave_src);
-    }
-
-    u->slave_src = qio_channel_add_watch_source(u->slave_ioc,
-                                                G_IO_IN | G_IO_HUP,
-                                                slave_read, dev, NULL,
-                                                ctxt);
-}
-
 static int vhost_user_read(struct vhost_dev *dev, VhostUserMsg *msg)
 {
     struct vhost_user *u = dev->opaque;
@@ -406,7 +377,6 @@ static int vhost_user_read(struct vhost_dev *dev, VhostUserMsg *msg)
      * be prepared for re-entrancy. So we create a new one and switch chr
      * to use it.
      */
-    slave_update_read_handler(dev, ctxt);
     qemu_chr_be_update_read_handlers(chr->chr, ctxt);
     qemu_chr_fe_add_watch(chr, G_IO_IN | G_IO_HUP, vhost_user_read_cb, &data);
 
@@ -418,7 +388,6 @@ static int vhost_user_read(struct vhost_dev *dev, VhostUserMsg *msg)
      * context that have been processed by the nested loop are purged.
      */
     qemu_chr_be_update_read_handlers(chr->chr, prev_ctxt);
-    slave_update_read_handler(dev, NULL);
 
     g_main_loop_unref(loop);
     g_main_context_unref(ctxt);
@@ -1809,7 +1778,9 @@ static int vhost_setup_slave_channel(struct vhost_dev *dev)
         return -ECONNREFUSED;
     }
     u->slave_ioc = ioc;
-    slave_update_read_handler(dev, NULL);
+    u->slave_src = qio_channel_add_watch_source(u->slave_ioc,
+                                                G_IO_IN | G_IO_HUP,
+                                                slave_read, dev, NULL, NULL);
 
     if (reply_supported) {
         msg.hdr.flags |= VHOST_USER_NEED_REPLY_MASK;
-- 
MST



  parent reply	other threads:[~2023-01-30 20:31 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-30 20:18 [PULL 00/56] virtio,pc,pci: features, cleanups, fixes Michael S. Tsirkin
2023-01-30 20:18 ` [PULL 01/56] shpc: disallow unplug when power indicator is blinking Michael S. Tsirkin
2023-01-30 20:18 ` [PULL 02/56] hw/i386/acpi-build: Remove unused attributes Michael S. Tsirkin
2023-01-30 20:18 ` [PULL 03/56] hw/isa/isa-bus: Turn isa_build_aml() into qbus_build_aml() Michael S. Tsirkin
2023-01-30 20:18 ` [PULL 04/56] hw/acpi/piix4: No need to #include "hw/southbridge/piix.h" Michael S. Tsirkin
2023-01-30 20:18 ` [PULL 05/56] hw/acpi/acpi_dev_interface: Remove unused parameter from AcpiDeviceIfClass::madt_cpu Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 06/56] vhost-user: Correct a reference of TARGET_AARCH64 Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 07/56] hw/pci-host: Use register definitions from PCI standard Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 08/56] virtio-rng-pci: fix migration compat for vectors Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 09/56] intel-iommu: Document iova_tree Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 10/56] x86: don't let decompressed kernel image clobber setup_data Michael S. Tsirkin
2023-01-30 20:19   ` Michael S. Tsirkin
2023-01-30 20:19     ` Michael S. Tsirkin
2023-01-31 19:39     ` Jason A. Donenfeld
2023-01-31 21:27       ` Michael S. Tsirkin
2023-01-31 20:54     ` H. Peter Anvin
2023-01-31 21:22       ` Jason A. Donenfeld
2023-02-01  5:40         ` H. Peter Anvin
2023-01-31 23:32     ` Jason A. Donenfeld
2023-01-30 20:19 ` [PULL 12/56] tests: acpi: cleanup arguments to make them more readable Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 13/56] tests: acpi: whitelist DSDT blobs for tests that use pci-bridges Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 14/56] tests: acpi: extend pcihp with nested bridges Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 15/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 16/56] tests: acpi: cleanup use_uefi argument usage Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 17/56] pci_bridge: remove whitespace Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 18/56] x86: acpi: pcihp: clean up duplicate bridge_in_acpi assignment Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 19/56] pci: acpi hotplug: rename x-native-hotplug to x-do-not-expose-native-hotplug-cap Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 20/56] pcihp: piix4: do not call acpi_pcihp_reset() when ACPI PCI hotplug is disabled Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 21/56] pci: acpihp: assign BSEL only to coldplugged bridges Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 22/56] x86: pcihp: fix invalid AML PCNT calls to hotplugged bridges Michael S. Tsirkin
2023-01-30 20:19 ` [PULL 23/56] tests: boot_sector_test: avoid crashing if status is not available yet Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 24/56] tests: acpi: extend bridge tests with hotplugged bridges Michael S. Tsirkin
2023-01-30 20:20   ` Michael S. Tsirkin
2023-01-30 20:20     ` Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 11/56] tests: qtest: print device_add error before failing test Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 26/56] tests: acpi: add reboot cycle to bridge test Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 27/56] tests: acpi: whitelist DSDT before refactoring acpi based PCI hotplug machinery Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 28/56] pcihp: drop pcihp_bridge_en dependency when composing PCNT method Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 29/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 30/56] tests: acpi: whitelist DSDT before refactoring acpi based PCI hotplug machinery Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 31/56] pcihp: compose PCNT callchain right before its user _GPE._E01 Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 32/56] pcihp: do not put empty PCNT in DSDT Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 25/56] tests: boot_sector_test(): make it multi-shot Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 33/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 34/56] whitelist DSDT before adding endpoint devices to bridge testcases Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 35/56] tests: acpi: add endpoint devices to bridges Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 36/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 37/56] x86: pcihp: acpi: prepare slot ignore rule to work with self describing bridges Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 38/56] pci: acpi: wire up AcpiDevAmlIf interface to generic bridge Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 39/56] pcihp: make bridge describe itself using AcpiDevAmlIfClass:build_dev_aml Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 40/56] pci: make sure pci_bus_is_express() won't error out with "discards ‘const’ qualifier" Michael S. Tsirkin
2023-01-30 20:20 ` [PULL 41/56] pcihp: isolate rule whether slot should be described in DSDT Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 42/56] tests: acpi: whitelist DSDT before decoupling PCI hotplug code from basic slots description Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 43/56] pcihp: acpi: decouple hotplug and generic " Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 44/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 45/56] tests: acpi: whitelist DSDT blobs before removing dynamic _DSM on coldplugged bridges Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 46/56] pcihp: acpi: ignore coldplugged bridges when composing hotpluggable slots Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 47/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 48/56] tests: acpi: whitelist DSDT before moving non-hotpluggble slots description from hotplug path Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 49/56] pcihp: generate populated non-hotpluggble slot descriptions on non-hotplug path Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 50/56] tests: acpi: update expected blobs Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 51/56] vhost-user: Skip unnecessary duplicated VHOST_USER_ADD/REM_MEM_REG requests Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 52/56] hw: Use TYPE_PCI_BUS definition where appropriate Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 53/56] tests/qtest/bios-tables-test: Make the test less verbose by default Michael S. Tsirkin
2023-01-30 20:21 ` Michael S. Tsirkin [this message]
2023-01-30 20:21 ` [PULL 55/56] Revert "vhost-user: Introduce nested event loop in vhost_user_read()" Michael S. Tsirkin
2023-01-30 20:21 ` [PULL 56/56] docs/pcie.txt: Replace ioh3420 with pcie-root-port Michael S. Tsirkin
2023-02-02 13:42 ` [PULL 00/56] virtio,pc,pci: features, cleanups, fixes Peter Maydell

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230130201810.11518-55-mst@redhat.com \
    --to=mst@redhat.com \
    --cc=groug@kaod.org \
    --cc=maxime.coquelin@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=yanghliu@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.