qemu-kvm/kvm-aio-warn-about-iohandler_ctx-special-casing.patch

65 lines
2.5 KiB
Diff
Raw Normal View History

* Mon Jul 01 2024 Miroslav Rezanina <mrezanin@redhat.com> - 9.0.0-3 - kvm-qtest-x86-numa-test-do-not-use-the-obsolete-pentium-.patch [RHEL-28972] - kvm-tests-qtest-libqtest-add-qtest_has_cpu_model-api.patch [RHEL-28972] - kvm-tests-qtest-x86-check-for-availability-of-older-cpu-.patch [RHEL-28972] - kvm-target-cpu-models-x86-Remove-the-existing-deprecated.patch [RHEL-28972] - kvm-x86-cpu-deprecate-cpu-models-that-do-not-support-x86.patch [RHEL-28971] - kvm-virtio-gpu-fix-v2-migration.patch [RHEL-36329] - kvm-rhel-9.4.0-machine-type-compat-for-virtio-gpu-migrat.patch [RHEL-36329] - kvm-s390x-remove-deprecated-rhel-machine-types.patch [RHEL-39898] - kvm-s390x-select-correct-components-for-no-board-build.patch [RHEL-39898] - kvm-target-s390x-Add-a-CONFIG-switch-to-disable-legacy-C.patch [RHEL-39898] - kvm-target-s390x-cpu_models-Disable-everything-up-to-the.patch [RHEL-39898] - kvm-target-s390x-Revert-the-old-s390x-CPU-model-disablem.patch [RHEL-39898] - kvm-Revert-monitor-use-aio_co_reschedule_self.patch [RHEL-43409 RHEL-43410] - kvm-aio-warn-about-iohandler_ctx-special-casing.patch [RHEL-43409 RHEL-43410] - Resolves: RHEL-28972 (x86: Remove the existing deprecated CPU models on RHEL10) - Resolves: RHEL-28971 (Consider deprecating CPU models like "Nehalem" / "IvyBridge" on RHEL 10) - Resolves: RHEL-36329 ([RHEL10.0.beta][stable_guest_abi]Failed to migrate VM with (qemu) qemu-kvm: Missing section footer for 0000:00:01.0/virtio-gpu qemu-kvm: load of migration failed: Invalid argument) - Resolves: RHEL-39898 (s390: Remove the legacy CPU models on RHEL10) - Resolves: RHEL-43409 (aio=io_uring: Assertion failure `luringcb->co->ctx == s->aio_context' with block_resize) - Resolves: RHEL-43410 (aio=native: Assertion failure `laiocb->co->ctx == laiocb->ctx->aio_context' with block_resize)
2024-07-01 09:14:15 +00:00
From 6c8da957fd534b3546354a8b8252c01cf9ee3511 Mon Sep 17 00:00:00 2001
From: Stefan Hajnoczi <stefanha@redhat.com>
Date: Mon, 6 May 2024 15:06:22 -0400
Subject: [PATCH 14/14] aio: warn about iohandler_ctx special casing
RH-Author: Kevin Wolf <kwolf@redhat.com>
RH-MergeRequest: 253: Revert "monitor: use aio_co_reschedule_self()"
RH-Jira: RHEL-43409 RHEL-43410
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Acked-by: Hanna Czenczek <hreitz@redhat.com>
RH-Commit: [2/2] 895231553731f09f51275c1abbf50c3440fe977f (kmwolf/centos-qemu-kvm)
The main loop has two AioContexts: qemu_aio_context and iohandler_ctx.
The main loop runs them both, but nested aio_poll() calls on
qemu_aio_context exclude iohandler_ctx.
Which one should qemu_get_current_aio_context() return when called from
the main loop? Document that it's always qemu_aio_context.
This has subtle effects on functions that use
qemu_get_current_aio_context(). For example, aio_co_reschedule_self()
does not work when moving from iohandler_ctx to qemu_aio_context because
qemu_get_current_aio_context() does not differentiate these two
AioContexts.
Document this in order to reduce the chance of future bugs.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240506190622.56095-3-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
(cherry picked from commit e669e800fc9ef8806af5c5578249ab758a4f8a5a)
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
include/block/aio.h | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/include/block/aio.h b/include/block/aio.h
index 8378553eb9..4ee81936ed 100644
--- a/include/block/aio.h
+++ b/include/block/aio.h
@@ -629,6 +629,9 @@ void aio_co_schedule(AioContext *ctx, Coroutine *co);
*
* Move the currently running coroutine to new_ctx. If the coroutine is already
* running in new_ctx, do nothing.
+ *
+ * Note that this function cannot reschedule from iohandler_ctx to
+ * qemu_aio_context.
*/
void coroutine_fn aio_co_reschedule_self(AioContext *new_ctx);
@@ -661,6 +664,9 @@ void aio_co_enter(AioContext *ctx, Coroutine *co);
* If called from an IOThread this will be the IOThread's AioContext. If
* called from the main thread or with the "big QEMU lock" taken it
* will be the main loop AioContext.
+ *
+ * Note that the return value is never the main loop's iohandler_ctx and the
+ * return value is the main loop AioContext instead.
*/
AioContext *qemu_get_current_aio_context(void);
--
2.39.3