It is pretty rare for aio_notify to actually set the EventNotifier. It can happen with worker threads such as thread-pool.c's, but otherwise it should never be set thanks to the ctx->notify_me optimization. The previous patch, unfortunately, added an unconditional call to event_notifier_test_and_clear; now add a userspace fast path that avoids the call. Note that it is not possible to do the same with event_notifier_set; it would break, as proved (again) by the included formal model. This patch survived over 3000 reboots on aarch64 KVM. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> Tested-by: Richard W.M. Jones <rjones@redhat.com> Message-id: 1437487673-23740-7-git-send-email-pbonzini@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> |
||
---|---|---|
.. | ||
qmp | ||
specs | ||
aio_notify.promela | ||
aio_notify_accept.promela | ||
aio_notify_bug.promela | ||
atomics.txt | ||
bitmaps.md | ||
blkdebug.txt | ||
blkverify.txt | ||
bootindex.txt | ||
ccid.txt | ||
ich9-ehci-uhci.cfg | ||
image-fuzzer.txt | ||
libcacard.txt | ||
live-block-ops.txt | ||
memory-hotplug.txt | ||
memory.txt | ||
migration.txt | ||
multi-thread-compression.txt | ||
multiple-iothreads.txt | ||
multiseat.txt | ||
pci_expander_bridge.txt | ||
q35-chipset.cfg | ||
qapi-code-gen.txt | ||
qdev-device-use.txt | ||
qemupciserial.inf | ||
rcu.txt | ||
rdma.txt | ||
spice-port-fqdn.txt | ||
tracing.txt | ||
usb-storage.txt | ||
usb2.txt | ||
virtio-balloon-stats.txt | ||
vnc-ledstate-Pseudo-encoding.txt | ||
writing-qmp-commands.txt | ||
xbzrle.txt | ||
xen-save-devices-state.txt |