From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
David Hildenbrand <david@redhat.com>,
yuanminghao <yuanmh12@chinatelecom.cn>,
Igor Mammedov <imammedo@redhat.com>,
Stefano Garzarella <sgarzare@redhat.com>
Subject: [PULL 07/97] vhost: Fix used memslot tracking when destroying a vhost device
Date: Mon, 14 Jul 2025 19:06:31 -0400 [thread overview]
Message-ID: <9f749129e2629b19f424df106c92c5a5647e396c.1752534227.git.mst@redhat.com> (raw)
In-Reply-To: <cover.1752534227.git.mst@redhat.com>
From: David Hildenbrand <david@redhat.com>
When we unplug a vhost device, we end up calling vhost_dev_cleanup()
where we do a memory_listener_unregister().
This memory_listener_unregister() call will end up disconnecting the
listener from the address space through listener_del_address_space().
In that process, we effectively communicate the removal of all memory
regions from that listener, resulting in region_del() + commit()
callbacks getting triggered.
So in case of vhost, we end up calling vhost_commit() with no remaining
memory slots (0).
In vhost_commit() we end up overwriting the global variables
used_memslots / used_shared_memslots, used for detecting the number
of free memslots. With used_memslots / used_shared_memslots set to 0
by vhost_commit() during device removal, we'll later assume that the
other vhost devices still have plenty of memslots left when calling
vhost_get_free_memslots().
Let's fix it by simply removing the global variables and depending
only on the actual per-device count.
Easy to reproduce by adding two vhost-user devices to a VM and then
hot-unplugging one of them.
While at it, detect unexpected underflows in vhost_get_free_memslots()
and issue a warning.
Reported-by: yuanminghao <yuanmh12@chinatelecom.cn>
Link: https://lore.kernel.org/qemu-devel/20241121060755.164310-1-yuanmh12@chinatelecom.cn/
Fixes: 2ce68e4cf5be ("vhost: add vhost_has_free_slot() interface")
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20250603111336.1858888-1-david@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/vhost.c | 37 ++++++++++---------------------------
1 file changed, 10 insertions(+), 27 deletions(-)
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index fc43853704..c87861b31f 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -47,12 +47,6 @@ static struct vhost_log *vhost_log[VHOST_BACKEND_TYPE_MAX];
static struct vhost_log *vhost_log_shm[VHOST_BACKEND_TYPE_MAX];
static QLIST_HEAD(, vhost_dev) vhost_log_devs[VHOST_BACKEND_TYPE_MAX];
-/* Memslots used by backends that support private memslots (without an fd). */
-static unsigned int used_memslots;
-
-/* Memslots used by backends that only support shared memslots (with an fd). */
-static unsigned int used_shared_memslots;
-
static QLIST_HEAD(, vhost_dev) vhost_devices =
QLIST_HEAD_INITIALIZER(vhost_devices);
@@ -74,15 +68,15 @@ unsigned int vhost_get_free_memslots(void)
QLIST_FOREACH(hdev, &vhost_devices, entry) {
unsigned int r = hdev->vhost_ops->vhost_backend_memslots_limit(hdev);
- unsigned int cur_free;
+ unsigned int cur_free = r - hdev->mem->nregions;
- if (hdev->vhost_ops->vhost_backend_no_private_memslots &&
- hdev->vhost_ops->vhost_backend_no_private_memslots(hdev)) {
- cur_free = r - used_shared_memslots;
+ if (unlikely(r < hdev->mem->nregions)) {
+ warn_report_once("used (%u) vhost backend memory slots exceed"
+ " the device limit (%u).", hdev->mem->nregions, r);
+ free = 0;
} else {
- cur_free = r - used_memslots;
+ free = MIN(free, cur_free);
}
- free = MIN(free, cur_free);
}
return free;
}
@@ -666,13 +660,6 @@ static void vhost_commit(MemoryListener *listener)
dev->mem = g_realloc(dev->mem, regions_size);
dev->mem->nregions = dev->n_mem_sections;
- if (dev->vhost_ops->vhost_backend_no_private_memslots &&
- dev->vhost_ops->vhost_backend_no_private_memslots(dev)) {
- used_shared_memslots = dev->mem->nregions;
- } else {
- used_memslots = dev->mem->nregions;
- }
-
for (i = 0; i < dev->n_mem_sections; i++) {
struct vhost_memory_region *cur_vmr = dev->mem->regions + i;
struct MemoryRegionSection *mrs = dev->mem_sections + i;
@@ -1619,15 +1606,11 @@ int vhost_dev_init(struct vhost_dev *hdev, void *opaque,
QLIST_INSERT_HEAD(&vhost_devices, hdev, entry);
/*
- * The listener we registered properly updated the corresponding counter.
- * So we can trust that these values are accurate.
+ * The listener we registered properly setup the number of required
+ * memslots in vhost_commit().
*/
- if (hdev->vhost_ops->vhost_backend_no_private_memslots &&
- hdev->vhost_ops->vhost_backend_no_private_memslots(hdev)) {
- used = used_shared_memslots;
- } else {
- used = used_memslots;
- }
+ used = hdev->mem->nregions;
+
/*
* We assume that all reserved memslots actually require a real memslot
* in our vhost backend. This might not be true, for example, if the
--
MST
next prev parent reply other threads:[~2025-07-14 23:21 UTC|newest]
Thread overview: 103+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-14 23:06 [PULL 00/97] virtio,pci,pc: features, fixes, tests Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 01/97] qdev-properties: Add DEFINE_PROP_ON_OFF_AUTO_BIT64() Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 02/97] net/vhost-vdpa: Report hashing capability Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 03/97] virtio-net: Move virtio_net_get_features() down Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 04/97] virtio-net: Retrieve peer hashing capability Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 05/97] net/vhost-vdpa: Remove dummy SetSteeringEBPF Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 06/97] virtio-net: Add hash type options Michael S. Tsirkin
2025-07-14 23:06 ` Michael S. Tsirkin [this message]
2025-07-14 23:06 ` [PULL 08/97] softmmu/runstate: add a way to detect force shutdowns Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 09/97] vhost: add a helper for force stopping a device Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 10/97] vhost-user-blk: add an option to skip GET_VRING_BASE for force shutdown Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 11/97] tests/acpi: Add empty ACPI data files for LoongArch Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 12/97] tests/qtest/bios-tables-test: Add basic testing " Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 13/97] rebuild-expected-aml.sh: Add support " Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 14/97] tests/acpi: Fill acpi table data " Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 15/97] tests/acpi: Remove stale allowed tables Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 16/97] hw/acpi: Fix GPtrArray memory leak in crs_range_merge Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 17/97] amd_iommu: Fix Miscellaneous Information Register 0 encoding Michael S. Tsirkin
2025-07-14 23:06 ` [PULL 18/97] amd_iommu: Fix Device ID decoding for INVALIDATE_IOTLB_PAGES command Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 19/97] amd_iommu: Update bitmasks representing DTE reserved fields Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 20/97] amd_iommu: Fix masks for various IOMMU MMIO Registers Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 21/97] amd_iommu: Fix mask to retrieve Interrupt Table Root Pointer from DTE Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 22/97] amd_iommu: Fix the calculation for Device Table size Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 23/97] amd_iommu: Remove duplicated definitions Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 24/97] amd_iommu: Fix truncation of oldval in amdvi_writeq Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 25/97] acpi: Add machine option to disable SPCR table Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 26/97] tests/qtest/bios-tables-test: Add test for disabling SPCR on AArch64 Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 27/97] tests/qtest/bios-tables-test: Add test for disabling SPCR on RISC-V Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 28/97] rust: bindings: allow any number of params Michael S. Tsirkin
2025-07-14 23:30 ` Manos Pitsidianakis
2025-07-15 6:17 ` Michael S. Tsirkin
2025-07-15 6:56 ` [PULL v2 " Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 29/97] pci: Add a memory attribute for pre-translated DMA operations Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 30/97] memory: Add permissions in IOMMUAccessFlags Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 31/97] memory: Allow to store the PASID in IOMMUTLBEntry Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 32/97] intel_iommu: Fill the PASID field when creating an IOMMUTLBEntry Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 33/97] intel_iommu: Declare supported PASID size Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 34/97] intel_iommu: Implement vtd_get_iotlb_info from PCIIOMMUOps Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 35/97] intel_iommu: Implement the PCIIOMMUOps callbacks related to invalidations of device-IOTLB Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 36/97] intel_iommu: Return page walk level even when the translation fails Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 37/97] intel_iommu: Set address mask when a translation fails and adjust W permission Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 38/97] intel_iommu: Add support for ATS Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 39/97] target/qmp: Use target_cpu_type() Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 40/97] qemu/target-info: Factor target_arch() out Michael S. Tsirkin
2025-07-14 23:07 ` [PULL 41/97] qemu/target-info: Add %target_arch field to TargetInfo Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 42/97] qemu/target-info: Add target_endian_mode() Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 43/97] qemu: Convert target_words_bigendian() to TargetInfo API Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 44/97] gdbstub/helpers: Replace TARGET_BIG_ENDIAN -> target_big_endian() Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 45/97] qemu: Declare all load/store helper in 'qemu/bswap.h' Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 46/97] hw/virtio: Build various files once Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 47/97] hw/i386/acpi-build: Make aml_pci_device_dsm() static Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 48/97] hw/acpi: Rename and move build_x86_acpi_pci_hotplug to pcihp Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 49/97] hw/pci-host/gpex-acpi: Add native_pci_hotplug arg to acpi_dsdt_add_pci_osc Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 50/97] tests/qtest/bios-tables-test: Prepare for changes in the DSDT table Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 51/97] hw/pci-host/gpex-acpi: Split host bridge OSC and DSM generation Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 52/97] hw/acpi/ged: Add a acpi-pci-hotplug-with-bridge-support property Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 53/97] hw/pci-host/gpex-acpi: Use GED acpi pcihp property Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 54/97] hw/i386/acpi-build: Turn build_q35_osc_method into a generic method Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 55/97] hw/pci-host/gpex-acpi: Use build_pci_host_bridge_osc_method Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 56/97] tests/qtest/bios-tables-test: Update DSDT blobs after GPEX _OSC change Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 57/97] hw/i386/acpi-build: Introduce build_append_pcihp_resources() helper Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 58/97] hw/acpi/pcihp: Add an AmlRegionSpace arg to build_acpi_pci_hotplug Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 59/97] hw/i386/acpi-build: Move build_append_notification_callback to pcihp Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 60/97] hw/i386/acpi-build: Move build_append_pci_bus_devices/pcihp_slots " Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 61/97] hw/i386/acpi-build: Use AcpiPciHpState::root in acpi_set_pci_info Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 62/97] hw/i386/acpi-build: Move aml_pci_edsm to a generic place Michael S. Tsirkin
2025-07-14 23:08 ` [PULL 63/97] qtest/bios-tables-test: Prepare for fixing the aarch64 viot test Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 64/97] qtest/bios-tables-test: Add a variant to " Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 65/97] qtest/bios-tables-test: Generate DSDT.viot Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 66/97] tests/qtest/bios-tables-test: Prepare for changes in the arm virt DSDT table Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 67/97] hw/arm/virt-acpi-build: Let non hotplug ports support static acpi-index Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 68/97] tests/qtest/bios-tables-test: Update ARM DSDT reference blobs Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 69/97] hw/arm/virt-acpi-build: Modify the DSDT ACPI table to enable ACPI PCI hotplug Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 70/97] hw/acpi/ged: Add a bus link property Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 71/97] hw/arm/virt: Pass the bus on the ged creation Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 72/97] hw/acpi/ged: Call pcihp plug callbacks in hotplug handler implementation Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 73/97] hw/acpi/pcihp: Remove root arg in acpi_pcihp_init Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 74/97] hw/acpi/ged: Prepare the device to react to PCI hotplug events Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 75/97] hw/acpi/ged: Support migration of AcpiPciHpState Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 76/97] hw/core/sysbus: Introduce sysbus_mmio_map_name() helper Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 77/97] hw/arm/virt: Minor code reshuffling in create_acpi_ged Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 78/97] hw/arm/virt: Let virt support pci hotplug/unplug GED event Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 79/97] tests/qtest/bios-tables-test: Prepare for addition of acpi pci hp tests Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 80/97] tests/qtest/bios-tables-test: Add aarch64 ACPI PCI hotplug test Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 81/97] qtest/bios-tables-test: Generate reference blob for DSDT.hpoffacpiindex Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 82/97] qtest/bios-tables-test: Generate reference blob for DSDT.acpipcihp Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 83/97] tests: virt: Allow changes to PPTT test table Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 84/97] hw/acpi/aml-build: Set identical implementation flag for PPTT processor nodes Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 85/97] hw/acpi/aml-build: Build a root node in the PPTT table Michael S. Tsirkin
2025-07-14 23:09 ` [PULL 86/97] tests: virt: Update expected ACPI tables for virt test Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 87/97] hw/cxl: fix DC extent capacity tracking Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 88/97] hw/cxl: mailbox-utils: 0x5600 - FMAPI Get DCD Info Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 89/97] hw/mem: cxl_type3: Add dsmas_flags to CXLDCRegion struct Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 90/97] hw/cxl: mailbox-utils: 0x5601 - FMAPI Get Host Region Config Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 91/97] hw/cxl: Move definition for dynamic_capacity_uuid and enum for DC event types to header Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 92/97] hw/mem: cxl_type3: Add DC Region bitmap lock Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 93/97] hw/cxl: mailbox-utils: 0x5602 - FMAPI Set DC Region Config Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 94/97] hw/cxl: mailbox-utils: 0x5603 - FMAPI Get DC Region Extent Lists Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 95/97] hw/cxl: Create helper function to create DC Event Records from extents Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 96/97] hw/cxl: mailbox-utils: 0x5604 - FMAPI Initiate DC Add Michael S. Tsirkin
2025-07-14 23:10 ` [PULL 97/97] hw/cxl: mailbox-utils: 0x5605 - FMAPI Initiate DC Release Michael S. Tsirkin
2025-07-15 6:57 ` [PULL 00/97] virtio,pci,pc: features, fixes, tests Michael S. Tsirkin
2025-07-16 12:39 ` Stefan Hajnoczi
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=9f749129e2629b19f424df106c92c5a5647e396c.1752534227.git.mst@redhat.com \
--to=mst@redhat.com \
--cc=david@redhat.com \
--cc=imammedo@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=sgarzare@redhat.com \
--cc=yuanmh12@chinatelecom.cn \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).