Images with disk size 0 may be used for
VM snapshots, but not to save normal block data.
It is possible to create such images using
qemu-img, but opening them later fails.
So even "qemu-img info image.qcow2" is not
possible for an image created with
"qemu-img create -f qcow2 image.qcow2 0".
This is fixed here.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The context parameter in paio_submit isn't used anyway, so there is no reason
why block drivers should need to remember it. This also avoids passing a Linux
AIO context to paio_submit (which doesn't do any harm as long as the parameter
is unused, but it is highly confusing).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
It was merely a workaround and the real fix is done now.
This reverts commit ef845c3bf4.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
We'll leave some AIO completions unhandled when we can't call the callback.
qemu_aio_process_queue() is used later to run any callbacks that are left and
can be run then.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When using Linux AIO raw still falls back to POSIX AIO sometimes, so we should
initialize it.
Not initializing it happens to work if POSIX AIO is used by another drive, or
if the format is not specified (probing the format uses POSIX AIO) or by pure
luck (e.g. it doesn't seem to happen any more with qcow2 since we have re-added
synchronous qcow2 functions).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
In case of failure, we haven't increased the refcount for the newly allocated
cluster yet. Therefore we must not free the cluster or its refcount will become
negative (and endless recursion is possible).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When the synchronous read and write functions were dropped, they were replaced
by generic emulation functions. Unfortunately, these emulation functions don't
provide the same semantics as the original functions did.
The original bdrv_read would mean that we read some data synchronously and that
we won't be interrupted during this read. The latter assumption is no longer
true with the emulation function which needs to use qemu_aio_poll and therefore
allows the callback of any other concurrent AIO request to be run during the
read. Which in turn means that (meta)data read earlier could have changed and
be invalid now. qcow2 is not prepared to work in this way and it's just scary
how many places there are where other requests could run.
I'm not sure yet where exactly it breaks, but you'll see breakage with virtio
on qcow2 with a backing file. Providing synchronous functions again fixes the
problem for me.
Patchworks-ID: 35437
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Today host_devices have a create function, so they also need a create_options
field to prevent qemu-img from complaining.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch increases the maximum qcow2 cluster size to 2 MB. Starting with 128k
clusters, L2 tables span 2 GB or more of virtual disk space, causing 32 bit
truncation and wraparound of signed integers. Therefore some variables need to
use a larger data type.
While being at reviewing data types, change some integers that are used for
array indices to unsigned. In some places they were checked against some upper
limit but not for negative values. This could avoid potential segfaults with
corrupted qcow2 images.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
If available, the Universally Unique Identifier library
is used by the vdi block driver.
Other parts of QEMU (vl.c) could also use it.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
In the very least, a change like this requires discussion on the list.
The naming convention is goofy and it causes a massive merge problem. Something
like this _must_ be presented on the list first so people can provide input
and cope with it.
This reverts commit 99a0949b72.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Put space between = and & when taking a pointer,
to avoid confusion with old-style "&=".
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Problem: Our file sys-queue.h is a copy of the BSD file, but there are
some additions and it's not entirely compatible. Because of that, there have
been conflicts with system headers on BSD systems. Some hacks have been
introduced in the commits 15cc923584,
f40d753718,
96555a96d7 and
3990d09adf but the fixes were fragile.
Solution: Avoid the conflict entirely by renaming the functions and the
file. Revert the previous hacks.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Instead stalling the VCPU while serving a cache flush try to do it
asynchronously. Use our good old helper thread pool to issue an
asynchronous fdatasync for raw-posix. Note that while Linux AIO
implements a fdatasync operation it is not useful for us because
it isn't actually implement in asynchronous fashion.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
If we are flushing the caches for our image files we only care about the
data (including the metadata required for accessing it) but not things
like timestamp updates. So try to use fdatasync instead of fsync to
implement the flush operations.
Unfortunately many operating systems still do not support fdatasync,
so we add a qemu_fdatasync wrapper that uses fdatasync if available
as per the _POSIX_SYNCHRONIZED_IO feature macro or fsync otherwise.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When two AIO requests write to the same cluster, and this cluster is
unallocated, currently both requests allocate a new cluster and the second one
merges the first one when it is completed. This means an cluster allocation, a
read and a cluster deallocation which cause some overhead. If we simply let the
second request wait until the first one is done, we improve overall performance
with AIO requests (specifially, qcow2/virtio combinations).
This patch maintains a list of in-flight requests that have allocated new
clusters. A second request touching the same cluster is limited so that it
either doesn't touch the allocation of the first request (so it can have a
non-overlapping allocation) or it waits for the first request to complete.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The wrong version of the preallocation patch has been applied, so this is the
remaining diff.
We can't use truncate to grow the image file to the right size because we don't
know if metadata has been written after the last data cluster. In this case
truncate would shrink the file and destroy its metadata. Write a zero sector at
the end of the virtual disk instead to ensure that the file is big enough.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The company which made Virtual PC was Connectix.
They use the magic string "conectix" in their disk images.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch fixes linker errors when building QEMU without Linux AIO support.
It is based on suggestions from malc and Kevin Wolf.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Now that do have a nicer interface to work against we can add Linux native
AIO support. It's an extremly thing layer just setting up an iocb for
the io_submit system call in the submission path, and registering an
eventfd with the qemu poll handler to do complete the iocbs directly
from there.
This started out based on Anthony's earlier AIO patch, but after
estimated 42,000 rewrites and just as many build system changes
there's not much left of it.
To enable native kernel aio use the aio=native sub-command on the
drive command line. I have also added an option to qemu-io to
test the aio support without needing a guest.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Currently the raw-posix.c code contains a lot of knowledge about the
asynchronous I/O scheme that is mostly implemented in posix-aio-compat.c.
All this code does not really belong here and is getting a bit in the
way of implementing native AIO on Linux.
So instead move all the guts of the AIO implementation into
posix-aio-compat.c (which might need a better name, btw).
There's now a very small interface between the AIO providers and raw-posix.c:
- an init routine is called from raw_open_common to return an AIO context
for this drive. An AIO implementation may either re-use one context
for all drives, or use a different one for each as the Linux native
AIO support will do.
- an submit routine is called from the aio_reav/writev methods to submit
an AIO request
There are no indirect calls involved in this interface as we need to
decide which one to call manually. We will only call the Linux AIO native
init function if we were requested to by vl.c, and we will only call
the native submit function if we are asked to and the request is properly
aligned. That's also the reason why the alignment check actually does
the inverse move and now goes into raw-posix.c.
The old posix-aio-compat.h headers is removed now that most of it's
content is private to posix-aio-compat.c, and instead we add a new
block/raw-posix-aio.h headers is created containing only the tiny interface
between raw-posix.c and the AIO implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This introduces a qemu-img create option for qcow2 which allows the metadata to
be preallocated, i.e. clusters are reserved in the refcount table and L1/L2
tables, but no data is written to them. Metadata is quite small, so this
happens in almost no time.
Especially with qcow2 on virtio this helps to gain a bit of performance during
the initial writes. However, as soon as create a snapshot, we're back to the
normal slow speed, obviously. So this isn't the real fix, but kind of a cheat
while we're still having trouble with qcow2 on virtio.
Note that the option is disabled by default and needs to be specified
explicitly using qemu-img create -f qcow2 -o preallocation=metadata.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
* The code for option '-static' was wrong, so image creation
always created static images.
* Static images created with qemu-img did not set header entry
blocks_allocated.
* The size of the block map must be rounded to the next multiple
of SECTOR_SIZE, otherwise the block map is only read partially
for block map sizes which are not a multiple of SECTOR_SIZE.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
These errors come up when compiling with gcc-4.3.3 and some older headers:
/scratch/froydnj/qemu.git/block/vpc.c: In function 'vpc_create':
/scratch/froydnj/qemu.git/block/vpc.c:514: error: value computed is not used
/scratch/froydnj/qemu.git/block/vpc.c:516: error: value computed is not used
/scratch/froydnj/qemu.git/block/vpc.c:517: error: value computed is not used
/scratch/froydnj/qemu.git/block/vpc.c:566: error: value computed is not used
Use memcpy to copy the strings instead of strncpy.
Signed-off-by: Nathan Froyd <froydnj@codesourcery.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
As requested by Anthony make pthreads mandatory. This means we will always
have AIO available on posix hosts, and it will also allow enabling the I/O
thread unconditionally once it's ready.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This is a new block driver written from scratch
to support the VDI format in QEMU.
VDI is the native format used by Innotek / SUN VirtualBox.
Latest changes:
* stripped down version
(code for synchronous operations and experimental code removed)
* don't open VDI snapshot images (with uuid_link or uuid_parent)
* modified vdi_aio_cancel
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Message-Id:
Instead of storing the backing file in its own BlockDriverState, VMDK uses the
BlockDriverState of the raw image file it opened. This is wrong and breaks
functions that access the backing file or protocols. This fix replaces all
occurrences of s->hd->backing_* with bs->backing_*.
This fixes qemu-iotests failure in 020 (Commit changes to backing file).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
I used the following command to enable debugging:
perl -p -i -e 's/^\/\/#define DEBUG/#define DEBUG/g' * */* */*/*
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
In qemu-iotests, some large images are created using qemu-img.
Without checks for errors, qemu-img will just create an
empty image, and later read / write tests will fail.
With the patch, failures during image creation are detected
and reported.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The VM state offset is a concept internal to the image format. Replace
the old bdrv_{get,put}_buffer method that require an index into the
image file that is constructed from the VM state offset and an offset
into the vmstate with the bdrv_{load,save}_vmstate that just take an
offset into the VM state.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Contrary to what one could expect, the size of L1 tables is not cluster
aligned. So as we're writing whole sectors now instead of single entries,
we need to ensure that the L1 table in memory is large enough; otherwise
write would access memory after the end of the L1 table.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The performance of qcow2 has improved meanwhile, so we don't need to
special-case it any more. Switch the default to write-through caching
like all other block drivers.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
The really time consuming part of snapshotting is to adjust the reference count
of all clusters. Currently after each adjusted cluster the refcount block is
written to disk.
Don't write each single byte immediately to disk but cache all writes to the
refcount block and write them out once we're done with the block.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When using O_DIRECT, qcow2 snapshots didn't work any more for me. In the
process of creating the snapshot, qcow2 tries to pwrite some new information
(e.g. new L1 table) which will often end up being after the old end of the
image file. Now pwrite tries to align things and reads the old contents of the
file, read returns 0 because there is nothing to read after the end of file and
pwrite is stuck in an endless loop.
This patch allows to pread beyond the end of an image file. Whenever the
given offset is after the end of the image file, the read succeeds and fills
the buffer with zeros.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Problem: It is impossible to feed filenames with the character colon because
qemu interprets such names as a protocol. For example filename scsi:0, is
interpreted as a protocol by name "scsi".
This patch allows user to espace colon characters. For example the above
filename can now be expressed either as 'scsi\:0' or as file:scsi:0
anything following the "file:" tag is interpreted verbatin. However if "file:"
tag is omitted then any colon characters in the string must be escaped using
backslash.
Here are couple of examples:
scsi\:0\:abc is a local file scsi:0:abc
http\://myweb is a local file by name http://myweb
file:scsi:0:abc is a local file scsi:0:abc
file:http://myweb is a local file by name http://myweb
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When updating the refcount blocks in update_refcount(), write complete sectors
instead of updating single entries.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When updating the L2 tables in alloc_cluster_link_l2(), write complete
sectors instead of updating single entries.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
When modifying the L1 table, l2_allocate() needs to write complete sectors
instead of single entries. The L1 table is already in memory, reading it from
disk in the block layer to align the request is wasted performance.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
