qemu-kvm/kvm-block-file-posix-set-up-Linux-AIO-and-io_uring-in-th.patch
Miroslav Rezanina 9ef7cdf7ca * 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 04:26:42 -05:00

218 lines
8.1 KiB
Diff

From 25cce5df341861e8ba8ec57722558e2dee3ce56a Mon Sep 17 00:00:00 2001
From: Stefan Hajnoczi <stefanha@redhat.com>
Date: Thu, 14 Sep 2023 10:00:58 -0400
Subject: [PATCH 073/101] block/file-posix: set up Linux AIO and io_uring in
the current 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: [4/26] 74c7daf805daefe706378308c3afeb28d861164b (kmwolf/centos-qemu-kvm)
The file-posix block driver currently only sets up Linux AIO and
io_uring in the BDS's AioContext. In the multi-queue block layer we must
be able to submit I/O requests in AioContexts that do not have Linux AIO
and io_uring set up yet since any thread can call into the block driver.
Set up Linux AIO and io_uring for the current AioContext during request
submission. We lose the ability to return an error from
.bdrv_file_open() when Linux AIO and io_uring setup fails (e.g. due to
resource limits). Instead the user only gets warnings and we fall back
to aio=threads. This is still better than a fatal error after startup.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230914140101.1065008-2-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block/file-posix.c | 103 ++++++++++++++++++++++-----------------------
1 file changed, 51 insertions(+), 52 deletions(-)
diff --git a/block/file-posix.c b/block/file-posix.c
index b862406c71..35684f7e21 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -712,17 +712,11 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
#ifdef CONFIG_LINUX_AIO
/* Currently Linux does AIO only for files opened with O_DIRECT */
- if (s->use_linux_aio) {
- if (!(s->open_flags & O_DIRECT)) {
- error_setg(errp, "aio=native was specified, but it requires "
- "cache.direct=on, which was not specified.");
- ret = -EINVAL;
- goto fail;
- }
- if (!aio_setup_linux_aio(bdrv_get_aio_context(bs), errp)) {
- error_prepend(errp, "Unable to use native AIO: ");
- goto fail;
- }
+ if (s->use_linux_aio && !(s->open_flags & O_DIRECT)) {
+ error_setg(errp, "aio=native was specified, but it requires "
+ "cache.direct=on, which was not specified.");
+ ret = -EINVAL;
+ goto fail;
}
#else
if (s->use_linux_aio) {
@@ -733,14 +727,7 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
}
#endif /* !defined(CONFIG_LINUX_AIO) */
-#ifdef CONFIG_LINUX_IO_URING
- if (s->use_linux_io_uring) {
- if (!aio_setup_linux_io_uring(bdrv_get_aio_context(bs), errp)) {
- error_prepend(errp, "Unable to use io_uring: ");
- goto fail;
- }
- }
-#else
+#ifndef CONFIG_LINUX_IO_URING
if (s->use_linux_io_uring) {
error_setg(errp, "aio=io_uring was specified, but is not supported "
"in this build.");
@@ -2444,6 +2431,48 @@ static bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov)
return true;
}
+#ifdef CONFIG_LINUX_IO_URING
+static inline bool raw_check_linux_io_uring(BDRVRawState *s)
+{
+ Error *local_err = NULL;
+ AioContext *ctx;
+
+ if (!s->use_linux_io_uring) {
+ return false;
+ }
+
+ ctx = qemu_get_current_aio_context();
+ if (unlikely(!aio_setup_linux_io_uring(ctx, &local_err))) {
+ error_reportf_err(local_err, "Unable to use linux io_uring, "
+ "falling back to thread pool: ");
+ s->use_linux_io_uring = false;
+ return false;
+ }
+ return true;
+}
+#endif
+
+#ifdef CONFIG_LINUX_AIO
+static inline bool raw_check_linux_aio(BDRVRawState *s)
+{
+ Error *local_err = NULL;
+ AioContext *ctx;
+
+ if (!s->use_linux_aio) {
+ return false;
+ }
+
+ ctx = qemu_get_current_aio_context();
+ if (unlikely(!aio_setup_linux_aio(ctx, &local_err))) {
+ error_reportf_err(local_err, "Unable to use Linux AIO, "
+ "falling back to thread pool: ");
+ s->use_linux_aio = false;
+ return false;
+ }
+ return true;
+}
+#endif
+
static int coroutine_fn raw_co_prw(BlockDriverState *bs, int64_t *offset_ptr,
uint64_t bytes, QEMUIOVector *qiov, int type)
{
@@ -2474,13 +2503,13 @@ static int coroutine_fn raw_co_prw(BlockDriverState *bs, int64_t *offset_ptr,
if (s->needs_alignment && !bdrv_qiov_is_aligned(bs, qiov)) {
type |= QEMU_AIO_MISALIGNED;
#ifdef CONFIG_LINUX_IO_URING
- } else if (s->use_linux_io_uring) {
+ } else if (raw_check_linux_io_uring(s)) {
assert(qiov->size == bytes);
ret = luring_co_submit(bs, s->fd, offset, qiov, type);
goto out;
#endif
#ifdef CONFIG_LINUX_AIO
- } else if (s->use_linux_aio) {
+ } else if (raw_check_linux_aio(s)) {
assert(qiov->size == bytes);
ret = laio_co_submit(s->fd, offset, qiov, type,
s->aio_max_batch);
@@ -2567,39 +2596,13 @@ static int coroutine_fn raw_co_flush_to_disk(BlockDriverState *bs)
};
#ifdef CONFIG_LINUX_IO_URING
- if (s->use_linux_io_uring) {
+ if (raw_check_linux_io_uring(s)) {
return luring_co_submit(bs, s->fd, 0, NULL, QEMU_AIO_FLUSH);
}
#endif
return raw_thread_pool_submit(handle_aiocb_flush, &acb);
}
-static void raw_aio_attach_aio_context(BlockDriverState *bs,
- AioContext *new_context)
-{
- BDRVRawState __attribute__((unused)) *s = bs->opaque;
-#ifdef CONFIG_LINUX_AIO
- if (s->use_linux_aio) {
- Error *local_err = NULL;
- if (!aio_setup_linux_aio(new_context, &local_err)) {
- error_reportf_err(local_err, "Unable to use native AIO, "
- "falling back to thread pool: ");
- s->use_linux_aio = false;
- }
- }
-#endif
-#ifdef CONFIG_LINUX_IO_URING
- if (s->use_linux_io_uring) {
- Error *local_err = NULL;
- if (!aio_setup_linux_io_uring(new_context, &local_err)) {
- error_reportf_err(local_err, "Unable to use linux io_uring, "
- "falling back to thread pool: ");
- s->use_linux_io_uring = false;
- }
- }
-#endif
-}
-
static void raw_close(BlockDriverState *bs)
{
BDRVRawState *s = bs->opaque;
@@ -3896,7 +3899,6 @@ BlockDriver bdrv_file = {
.bdrv_co_copy_range_from = raw_co_copy_range_from,
.bdrv_co_copy_range_to = raw_co_copy_range_to,
.bdrv_refresh_limits = raw_refresh_limits,
- .bdrv_attach_aio_context = raw_aio_attach_aio_context,
.bdrv_co_truncate = raw_co_truncate,
.bdrv_co_getlength = raw_co_getlength,
@@ -4266,7 +4268,6 @@ static BlockDriver bdrv_host_device = {
.bdrv_co_copy_range_from = raw_co_copy_range_from,
.bdrv_co_copy_range_to = raw_co_copy_range_to,
.bdrv_refresh_limits = raw_refresh_limits,
- .bdrv_attach_aio_context = raw_aio_attach_aio_context,
.bdrv_co_truncate = raw_co_truncate,
.bdrv_co_getlength = raw_co_getlength,
@@ -4402,7 +4403,6 @@ static BlockDriver bdrv_host_cdrom = {
.bdrv_co_pwritev = raw_co_pwritev,
.bdrv_co_flush_to_disk = raw_co_flush_to_disk,
.bdrv_refresh_limits = cdrom_refresh_limits,
- .bdrv_attach_aio_context = raw_aio_attach_aio_context,
.bdrv_co_truncate = raw_co_truncate,
.bdrv_co_getlength = raw_co_getlength,
@@ -4528,7 +4528,6 @@ static BlockDriver bdrv_host_cdrom = {
.bdrv_co_pwritev = raw_co_pwritev,
.bdrv_co_flush_to_disk = raw_co_flush_to_disk,
.bdrv_refresh_limits = cdrom_refresh_limits,
- .bdrv_attach_aio_context = raw_aio_attach_aio_context,
.bdrv_co_truncate = raw_co_truncate,
.bdrv_co_getlength = raw_co_getlength,
--
2.39.3