ed6e216171
laio_init() can fail for a couple of reasons, which will lead to a NULL pointer dereference in laio_attach_aio_context(). To solve this, add a aio_setup_linux_aio() function which is called early in raw_open_common. If this fails, propagate the error up. The signature of aio_get_linux_aio() was not modified, because it seems preferable to return the actual errno from the possible failing initialization calls. Additionally, when the AioContext changes, we need to associate a LinuxAioState with the new AioContext. Use the bdrv_attach_aio_context callback and call the new aio_setup_linux_aio(), which will allocate a new AioContext if needed, and return errors on failures. If it fails for any reason, fallback to threaded AIO with an error message, as the device is already in-use by the guest. Add an assert that aio_get_linux_aio() cannot return NULL. Signed-off-by: Nishanth Aravamudan <naravamudan@digitalocean.com> Message-id: 20180622193700.6523-1-naravamudan@digitalocean.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> |
||
---|---|---|
.. | ||
arch-query-cpu-def.c | ||
arch-query-cpu-model-baseline.c | ||
arch-query-cpu-model-comparison.c | ||
arch-query-cpu-model-expansion.c | ||
bdrv-next-monitor-owned.c | ||
blk-commit-all.c | ||
blockdev-close-all-bdrv-states.c | ||
change-state-handler.c | ||
clock-warp.c | ||
cpu-get-clock.c | ||
cpu-get-icount.c | ||
dump.c | ||
error-printf.c | ||
fd-register.c | ||
fdset.c | ||
gdbstub.c | ||
get-vm-name.c | ||
iothread-lock.c | ||
iothread.c | ||
is-daemonized.c | ||
linux-aio.c | ||
machine-init-done.c | ||
Makefile.objs | ||
migr-blocker.c | ||
monitor.c | ||
notify-event.c | ||
pc_madt_cpu_entry.c | ||
pci-host-piix.c | ||
qmp_memory_device.c | ||
qtest.c | ||
ram-block.c | ||
replay.c | ||
runstate-check.c | ||
set-fd-handler.c | ||
slirp.c | ||
sysbus.c | ||
target-get-monitor-def.c | ||
target-monitor-defs.c | ||
tpm.c | ||
trace-control.c | ||
uuid.c | ||
vm-stop.c | ||
vmgenid.c | ||
vmstate.c | ||
xen-common.c | ||
xen-hvm.c |