qemu-kvm/kvm-scsi-only-access-SCSIDevice-requests-from-one-thread.patch

308 lines
9.6 KiB
Diff
Raw Normal View History

* Wed Jan 24 2024 Miroslav Rezanina <mrezanin@redhat.com> - 8.2.0-3 - kvm-hw-arm-virt-Add-properties-to-disable-high-memory-re.patch [RHEL-19738] - kvm-vfio-Introduce-base-object-for-VFIOContainer-and-tar.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Introduce-a-empty-VFIOIOMMUOps.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Switch-to-dma_map-unmap-API.patch [RHEL-19302 RHEL-21057] - kvm-vfio-common-Introduce-vfio_container_init-destroy-he.patch [RHEL-19302 RHEL-21057] - kvm-vfio-common-Move-giommu_list-in-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-space-field-to-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Switch-to-IOMMU-BE-set_dirty_page_tra.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-per-container-device-list-in-bas.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Convert-functions-to-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-pgsizes-and-dma_max_mappings-to-.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-vrdl_list-to-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-listener-to-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-dirty_pgsizes-and-max_dirty_bitm.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Move-iova_ranges-to-base-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Implement-attach-detach_device.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Introduce-spapr-backend-and-target-interf.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-switch-to-spapr-IOMMU-BE-add-del_section_.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Move-prereg_listener-into-spapr-container.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Move-hostwin_list-into-spapr-container.patch [RHEL-19302 RHEL-21057] - kvm-backends-iommufd-Introduce-the-iommufd-object.patch [RHEL-19302 RHEL-21057] - kvm-util-char_dev-Add-open_cdev.patch [RHEL-19302 RHEL-21057] - kvm-vfio-common-return-early-if-space-isn-t-empty.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Implement-the-iommufd-backend.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Relax-assert-check-for-iommufd-backend.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Add-support-for-iova_ranges-and-pgsizes.patch [RHEL-19302 RHEL-21057] - kvm-vfio-pci-Extract-out-a-helper-vfio_pci_get_pci_hot_r.patch [RHEL-19302 RHEL-21057] - kvm-vfio-pci-Introduce-a-vfio-pci-hot-reset-interface.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Enable-pci-hot-reset-through-iommufd-cd.patch [RHEL-19302 RHEL-21057] - kvm-vfio-pci-Allow-the-selection-of-a-given-iommu-backen.patch [RHEL-19302 RHEL-21057] - kvm-vfio-pci-Make-vfio-cdev-pre-openable-by-passing-a-fi.patch [RHEL-19302 RHEL-21057] - kvm-vfio-platform-Allow-the-selection-of-a-given-iommu-b.patch [RHEL-19302 RHEL-21057] - kvm-vfio-platform-Make-vfio-cdev-pre-openable-by-passing.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ap-Allow-the-selection-of-a-given-iommu-backend.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ap-Make-vfio-cdev-pre-openable-by-passing-a-fil.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ccw-Allow-the-selection-of-a-given-iommu-backen.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ccw-Make-vfio-cdev-pre-openable-by-passing-a-fi.patch [RHEL-19302 RHEL-21057] - kvm-vfio-Make-VFIOContainerBase-poiner-parameter-const-i.patch [RHEL-19302 RHEL-21057] - kvm-hw-arm-Activate-IOMMUFD-for-virt-machines.patch [RHEL-19302 RHEL-21057] - kvm-kconfig-Activate-IOMMUFD-for-s390x-machines.patch [RHEL-19302 RHEL-21057] - kvm-hw-i386-Activate-IOMMUFD-for-q35-machines.patch [RHEL-19302 RHEL-21057] - kvm-vfio-pci-Move-VFIODevice-initializations-in-vfio_ins.patch [RHEL-19302 RHEL-21057] - kvm-vfio-platform-Move-VFIODevice-initializations-in-vfi.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ap-Move-VFIODevice-initializations-in-vfio_ap_i.patch [RHEL-19302 RHEL-21057] - kvm-vfio-ccw-Move-VFIODevice-initializations-in-vfio_ccw.patch [RHEL-19302 RHEL-21057] - kvm-vfio-Introduce-a-helper-function-to-initialize-VFIOD.patch [RHEL-19302 RHEL-21057] - kvm-docs-devel-Add-VFIO-iommufd-backend-documentation.patch [RHEL-19302 RHEL-21057] - kvm-hw-ppc-Kconfig-Imply-VFIO_PCI.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Extend-VFIOIOMMUOps-with-a-release-handle.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Introduce-vfio_legacy_setup-for-furth.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Initialize-VFIOIOMMUOps-under-vfio_in.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Introduce-a-VFIOIOMMU-QOM-interface.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Introduce-a-VFIOIOMMU-legacy-QOM-inte.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Intoduce-a-new-VFIOIOMMUClass-setup-h.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Introduce-a-sPAPR-VFIOIOMMU-QOM-interface.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Introduce-a-VFIOIOMMU-iommufd-QOM-inter.patch [RHEL-19302 RHEL-21057] - kvm-vfio-spapr-Only-compile-sPAPR-IOMMU-support-when-nee.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Remove-CONFIG_IOMMUFD-usage.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Replace-basename-with-g_path_get_base.patch [RHEL-19302 RHEL-21057] - kvm-hw-vfio-fix-iteration-over-global-VFIODevice-list.patch [RHEL-19302 RHEL-21057] - kvm-vfio-iommufd-Remove-the-use-of-stat-to-check-file-ex.patch [RHEL-19302 RHEL-21057] - kvm-vfio-container-Rename-vfio_init_container-to-vfio_se.patch [RHEL-19302 RHEL-21057] - kvm-vfio-migration-Add-helper-function-to-set-state-or-r.patch [RHEL-19302 RHEL-21057] - kvm-backends-iommufd-Remove-check-on-number-of-backend-u.patch [RHEL-19302 RHEL-21057] - kvm-backends-iommufd-Remove-mutex.patch [RHEL-19302 RHEL-21057] - kvm-Compile-IOMMUFD-object-on-aarch64.patch [RHEL-19302 RHEL-21057] - kvm-Compile-IOMMUFD-on-s390x.patch [RHEL-19302 RHEL-21057] - kvm-Compile-IOMMUFD-on-x86_64.patch [RHEL-19302 RHEL-21057] - kvm-target-s390x-kvm-pv-Provide-some-more-useful-informa.patch [RHEL-18212] - kvm-nbd-server-avoid-per-NBDRequest-nbd_client_get-put.patch [RHEL-15965] - kvm-nbd-server-only-traverse-NBDExport-clients-from-main.patch [RHEL-15965] - kvm-nbd-server-introduce-NBDClient-lock-to-protect-field.patch [RHEL-15965] - kvm-block-file-posix-set-up-Linux-AIO-and-io_uring-in-th.patch [RHEL-15965] - kvm-virtio-blk-add-lock-to-protect-s-rq.patch [RHEL-15965] - kvm-virtio-blk-don-t-lock-AioContext-in-the-completion-c.patch [RHEL-15965] - kvm-virtio-blk-don-t-lock-AioContext-in-the-submission-c.patch [RHEL-15965] - kvm-scsi-only-access-SCSIDevice-requests-from-one-thread.patch [RHEL-15965] - kvm-virtio-scsi-don-t-lock-AioContext-around-virtio_queu.patch [RHEL-15965] - kvm-scsi-don-t-lock-AioContext-in-I-O-code-path.patch [RHEL-15965] - kvm-dma-helpers-don-t-lock-AioContext-in-dma_blk_cb.patch [RHEL-15965] - kvm-virtio-scsi-replace-AioContext-lock-with-tmf_bh_lock.patch [RHEL-15965] - kvm-scsi-assert-that-callbacks-run-in-the-correct-AioCon.patch [RHEL-15965] - kvm-tests-remove-aio_context_acquire-tests.patch [RHEL-15965] - kvm-aio-make-aio_context_acquire-aio_context_release-a-n.patch [RHEL-15965] - kvm-graph-lock-remove-AioContext-locking.patch [RHEL-15965] - kvm-block-remove-AioContext-locking.patch [RHEL-15965] - kvm-block-remove-bdrv_co_lock.patch [RHEL-15965] - kvm-scsi-remove-AioContext-locking.patch [RHEL-15965] - kvm-aio-wait-draw-equivalence-between-AIO_WAIT_WHILE-and.patch [RHEL-15965] - kvm-aio-remove-aio_context_acquire-aio_context_release-A.patch [RHEL-15965] - kvm-docs-remove-AioContext-lock-from-IOThread-docs.patch [RHEL-15965] - kvm-scsi-remove-outdated-AioContext-lock-comment.patch [RHEL-15965] - kvm-job-remove-outdated-AioContext-locking-comments.patch [RHEL-15965] - kvm-block-remove-outdated-AioContext-locking-comments.patch [RHEL-15965] - kvm-block-coroutine-wrapper-use-qemu_get_current_aio_con.patch [RHEL-15965] - kvm-s390x-pci-avoid-double-enable-disable-of-aif.patch [RHEL-21169] - kvm-s390x-pci-refresh-fh-before-disabling-aif.patch [RHEL-21169] - kvm-s390x-pci-drive-ISM-reset-from-subsystem-reset.patch [RHEL-21169] - kvm-include-ui-rect.h-fix-qemu_rect_init-mis-assignment.patch [RHEL-21570] - kvm-virtio-gpu-block-migration-of-VMs-with-blob-true.patch [RHEL-7565] - kvm-spec-Enable-zstd.patch [RHEL-7361] - Resolves: RHEL-19738 (Enable properties allowing to disable high memory regions) - Resolves: RHEL-19302 (NVIDIA:Grace-Hopper Backport QEMU IOMMUFD Backend) - Resolves: RHEL-21057 (Request backport of 9353b6da430f90e47f352dbf6dc31120c8914da6) - Resolves: RHEL-18212 ([RHEL9][Secure-execution][s390x] The error message is not clear when boot up a SE guest with wrong encryption) - Resolves: RHEL-15965 ( [qemu-kvm] Remove AioContext lock (no response with QMP command block_resize)) - Resolves: RHEL-21169 ([s390x] VM fails to start with ISM passed through QEMU 8.2) - Resolves: RHEL-21570 (Critical performance degradation for input devices in virtio vnc session) - Resolves: RHEL-7565 (qemu crashed when migrate guest with blob resources enabled) - Resolves: RHEL-7361 ([qemu-kvm] Enable zstd support for qcow2 files)
2024-01-24 09:26:42 +00:00
From 42dd1357310bd1a68d6cacaa53cd5b1d1b02880d Mon Sep 17 00:00:00 2001
From: Stefan Hajnoczi <stefanha@redhat.com>
Date: Mon, 4 Dec 2023 11:42:56 -0500
Subject: [PATCH 077/101] scsi: only access SCSIDevice->requests from one
thread
RH-Author: Kevin Wolf <kwolf@redhat.com>
RH-MergeRequest: 214: Remove AioContext lock
RH-Jira: RHEL-15965
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
RH-Commit: [8/26] 9df662e82a63e93d184b5763bebbe7e43bc9dabe (kmwolf/centos-qemu-kvm)
Stop depending on the AioContext lock and instead access
SCSIDevice->requests from only one thread at a time:
- When the VM is running only the BlockBackend's AioContext may access
the requests list.
- When the VM is stopped only the main loop may access the requests
list.
These constraints protect the requests list without the need for locking
in the I/O code path.
Note that multiple IOThreads are not supported yet because the code
assumes all SCSIRequests are executed from a single AioContext. Leave
that as future work.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-ID: <20231204164259.1515217-2-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
hw/scsi/scsi-bus.c | 181 ++++++++++++++++++++++++++++-------------
include/hw/scsi/scsi.h | 7 +-
2 files changed, 131 insertions(+), 57 deletions(-)
diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
index fc4b77fdb0..b649cdf555 100644
--- a/hw/scsi/scsi-bus.c
+++ b/hw/scsi/scsi-bus.c
@@ -85,6 +85,89 @@ SCSIDevice *scsi_device_get(SCSIBus *bus, int channel, int id, int lun)
return d;
}
+/*
+ * Invoke @fn() for each enqueued request in device @s. Must be called from the
+ * main loop thread while the guest is stopped. This is only suitable for
+ * vmstate ->put(), use scsi_device_for_each_req_async() for other cases.
+ */
+static void scsi_device_for_each_req_sync(SCSIDevice *s,
+ void (*fn)(SCSIRequest *, void *),
+ void *opaque)
+{
+ SCSIRequest *req;
+ SCSIRequest *next_req;
+
+ assert(!runstate_is_running());
+ assert(qemu_in_main_thread());
+
+ QTAILQ_FOREACH_SAFE(req, &s->requests, next, next_req) {
+ fn(req, opaque);
+ }
+}
+
+typedef struct {
+ SCSIDevice *s;
+ void (*fn)(SCSIRequest *, void *);
+ void *fn_opaque;
+} SCSIDeviceForEachReqAsyncData;
+
+static void scsi_device_for_each_req_async_bh(void *opaque)
+{
+ g_autofree SCSIDeviceForEachReqAsyncData *data = opaque;
+ SCSIDevice *s = data->s;
+ AioContext *ctx;
+ SCSIRequest *req;
+ SCSIRequest *next;
+
+ /*
+ * If the AioContext changed before this BH was called then reschedule into
+ * the new AioContext before accessing ->requests. This can happen when
+ * scsi_device_for_each_req_async() is called and then the AioContext is
+ * changed before BHs are run.
+ */
+ ctx = blk_get_aio_context(s->conf.blk);
+ if (ctx != qemu_get_current_aio_context()) {
+ aio_bh_schedule_oneshot(ctx, scsi_device_for_each_req_async_bh,
+ g_steal_pointer(&data));
+ return;
+ }
+
+ QTAILQ_FOREACH_SAFE(req, &s->requests, next, next) {
+ data->fn(req, data->fn_opaque);
+ }
+
+ /* Drop the reference taken by scsi_device_for_each_req_async() */
+ object_unref(OBJECT(s));
+}
+
+/*
+ * Schedule @fn() to be invoked for each enqueued request in device @s. @fn()
+ * runs in the AioContext that is executing the request.
+ */
+static void scsi_device_for_each_req_async(SCSIDevice *s,
+ void (*fn)(SCSIRequest *, void *),
+ void *opaque)
+{
+ assert(qemu_in_main_thread());
+
+ SCSIDeviceForEachReqAsyncData *data =
+ g_new(SCSIDeviceForEachReqAsyncData, 1);
+
+ data->s = s;
+ data->fn = fn;
+ data->fn_opaque = opaque;
+
+ /*
+ * Hold a reference to the SCSIDevice until
+ * scsi_device_for_each_req_async_bh() finishes.
+ */
+ object_ref(OBJECT(s));
+
+ aio_bh_schedule_oneshot(blk_get_aio_context(s->conf.blk),
+ scsi_device_for_each_req_async_bh,
+ data);
+}
+
static void scsi_device_realize(SCSIDevice *s, Error **errp)
{
SCSIDeviceClass *sc = SCSI_DEVICE_GET_CLASS(s);
@@ -144,20 +227,18 @@ void scsi_bus_init_named(SCSIBus *bus, size_t bus_size, DeviceState *host,
qbus_set_bus_hotplug_handler(BUS(bus));
}
-static void scsi_dma_restart_bh(void *opaque)
+void scsi_req_retry(SCSIRequest *req)
{
- SCSIDevice *s = opaque;
- SCSIRequest *req, *next;
-
- qemu_bh_delete(s->bh);
- s->bh = NULL;
+ req->retry = true;
+}
- aio_context_acquire(blk_get_aio_context(s->conf.blk));
- QTAILQ_FOREACH_SAFE(req, &s->requests, next, next) {
- scsi_req_ref(req);
- if (req->retry) {
- req->retry = false;
- switch (req->cmd.mode) {
+/* Called in the AioContext that is executing the request */
+static void scsi_dma_restart_req(SCSIRequest *req, void *opaque)
+{
+ scsi_req_ref(req);
+ if (req->retry) {
+ req->retry = false;
+ switch (req->cmd.mode) {
case SCSI_XFER_FROM_DEV:
case SCSI_XFER_TO_DEV:
scsi_req_continue(req);
@@ -166,37 +247,22 @@ static void scsi_dma_restart_bh(void *opaque)
scsi_req_dequeue(req);
scsi_req_enqueue(req);
break;
- }
}
- scsi_req_unref(req);
}
- aio_context_release(blk_get_aio_context(s->conf.blk));
- /* Drop the reference that was acquired in scsi_dma_restart_cb */
- object_unref(OBJECT(s));
-}
-
-void scsi_req_retry(SCSIRequest *req)
-{
- /* No need to save a reference, because scsi_dma_restart_bh just
- * looks at the request list. */
- req->retry = true;
+ scsi_req_unref(req);
}
static void scsi_dma_restart_cb(void *opaque, bool running, RunState state)
{
SCSIDevice *s = opaque;
+ assert(qemu_in_main_thread());
+
if (!running) {
return;
}
- if (!s->bh) {
- AioContext *ctx = blk_get_aio_context(s->conf.blk);
- /* The reference is dropped in scsi_dma_restart_bh.*/
- object_ref(OBJECT(s));
- s->bh = aio_bh_new_guarded(ctx, scsi_dma_restart_bh, s,
- &DEVICE(s)->mem_reentrancy_guard);
- qemu_bh_schedule(s->bh);
- }
+
+ scsi_device_for_each_req_async(s, scsi_dma_restart_req, NULL);
}
static bool scsi_bus_is_address_free(SCSIBus *bus,
@@ -1657,15 +1723,16 @@ void scsi_device_set_ua(SCSIDevice *sdev, SCSISense sense)
}
}
+static void scsi_device_purge_one_req(SCSIRequest *req, void *opaque)
+{
+ scsi_req_cancel_async(req, NULL);
+}
+
void scsi_device_purge_requests(SCSIDevice *sdev, SCSISense sense)
{
- SCSIRequest *req;
+ scsi_device_for_each_req_async(sdev, scsi_device_purge_one_req, NULL);
aio_context_acquire(blk_get_aio_context(sdev->conf.blk));
- while (!QTAILQ_EMPTY(&sdev->requests)) {
- req = QTAILQ_FIRST(&sdev->requests);
- scsi_req_cancel_async(req, NULL);
- }
blk_drain(sdev->conf.blk);
aio_context_release(blk_get_aio_context(sdev->conf.blk));
scsi_device_set_ua(sdev, sense);
@@ -1737,31 +1804,33 @@ static char *scsibus_get_fw_dev_path(DeviceState *dev)
/* SCSI request list. For simplicity, pv points to the whole device */
+static void put_scsi_req(SCSIRequest *req, void *opaque)
+{
+ QEMUFile *f = opaque;
+
+ assert(!req->io_canceled);
+ assert(req->status == -1 && req->host_status == -1);
+ assert(req->enqueued);
+
+ qemu_put_sbyte(f, req->retry ? 1 : 2);
+ qemu_put_buffer(f, req->cmd.buf, sizeof(req->cmd.buf));
+ qemu_put_be32s(f, &req->tag);
+ qemu_put_be32s(f, &req->lun);
+ if (req->bus->info->save_request) {
+ req->bus->info->save_request(f, req);
+ }
+ if (req->ops->save_request) {
+ req->ops->save_request(f, req);
+ }
+}
+
static int put_scsi_requests(QEMUFile *f, void *pv, size_t size,
const VMStateField *field, JSONWriter *vmdesc)
{
SCSIDevice *s = pv;
- SCSIBus *bus = DO_UPCAST(SCSIBus, qbus, s->qdev.parent_bus);
- SCSIRequest *req;
- QTAILQ_FOREACH(req, &s->requests, next) {
- assert(!req->io_canceled);
- assert(req->status == -1 && req->host_status == -1);
- assert(req->enqueued);
-
- qemu_put_sbyte(f, req->retry ? 1 : 2);
- qemu_put_buffer(f, req->cmd.buf, sizeof(req->cmd.buf));
- qemu_put_be32s(f, &req->tag);
- qemu_put_be32s(f, &req->lun);
- if (bus->info->save_request) {
- bus->info->save_request(f, req);
- }
- if (req->ops->save_request) {
- req->ops->save_request(f, req);
- }
- }
+ scsi_device_for_each_req_sync(s, put_scsi_req, f);
qemu_put_sbyte(f, 0);
-
return 0;
}
diff --git a/include/hw/scsi/scsi.h b/include/hw/scsi/scsi.h
index 3692ca82f3..10c4e8288d 100644
--- a/include/hw/scsi/scsi.h
+++ b/include/hw/scsi/scsi.h
@@ -69,14 +69,19 @@ struct SCSIDevice
{
DeviceState qdev;
VMChangeStateEntry *vmsentry;
- QEMUBH *bh;
uint32_t id;
BlockConf conf;
SCSISense unit_attention;
bool sense_is_ua;
uint8_t sense[SCSI_SENSE_BUF_SIZE];
uint32_t sense_len;
+
+ /*
+ * The requests list is only accessed from the AioContext that executes
+ * requests or from the main loop when IOThread processing is stopped.
+ */
QTAILQ_HEAD(, SCSIRequest) requests;
+
uint32_t channel;
uint32_t lun;
int blocksize;
--
2.39.3