When deleting an fd event there is a chance the object doesn't get
deleted, but only ->deleted set positive and deleted somewhere later.
Now, if we create a handler for the fd again before the actual
deletion occurs, we end up writing data into an object that has
->deleted set, which is obviously wrong.
I see two ways to fix this:
1. Don't return ->deleted objects in the search
2. Unset ->deleted in the search
This patch implements 1. which feels safer to do. It fixes AIO issues
I've seen with curl, as libcurl unsets fd event listeners pretty
frequently.
Signed-off-by: Alexander Graf <alex@csgraf.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This was spotted by valgrind.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@5470 c046a42c-6fe2-441c-8c8c-71466251a162
This patch refactors the AIO layer to allow multiple AIO implementations. It's
only possible because of the recent signalfd() patch.
Right now, the AIO infrastructure is pretty specific to the block raw backend.
For other block devices to implement AIO, the qemu_aio_wait function must
support registration. This patch introduces a new function,
qemu_aio_set_fd_handler, which can be used to register a file descriptor to be
called back. qemu_aio_wait() now polls a set of file descriptors registered
with this function until one becomes readable or writable.
This patch should allow the implementation of alternative AIO backends (via a
thread pool or linux-aio) and AIO backends in non-traditional block devices
(like NBD).
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@5297 c046a42c-6fe2-441c-8c8c-71466251a162
