Account the total latency for read/write/flush requests. This allows
management tools to average it based on a snapshot of the nr ops
counters and allow checking for SLAs or provide statistics.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Decouple the I/O accounting from bdrv_aio_readv/writev/flush and
make the hardware models call directly into the accounting helpers.
This means:
- we do not count internal requests from image formats in addition
to guest originating I/O
- we do not double count I/O ops if the device model handles it
chunk wise
- we only account I/O once it actuall is done
- can extent I/O accounting to synchronous or coroutine I/O easily
- implement I/O latency tracking easily (see the next patch)
I've conveted the existing device model callers to the new model,
device models that are using synchronous I/O and weren't accounted
before haven't been updated yet. Also scsi hasn't been converted
to the end-to-end accounting as I want to defer that after the pending
scsi layer overhaul.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add new block driver callbacks bdrv_co_readv/writev, which work on a
QEMUIOVector like bdrv_aio_*, but don't need a callback. The function may only
be called inside a coroutine, so a block driver implementing this interface can
yield instead of blocking during I/O.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Callees always return 0, except for FreeBSD's cdrom_eject(), which
returns -ENOTSUP when the device is in a terminally wedged state.
The only caller is bdrv_eject(), and it maps -ENOTSUP to 0 since
commit 4be9762a.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The only caller is bdrv_set_locked(), and it ignores the value.
Callees always return 0, except for FreeBSD's cdrom_set_locked(),
which returns -ENOTSUP when the device is in a terminally wedged
state.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
qemu-img.c wants to count allocated file size of image. Previously it
counts a single bs->file by 'stat' or Window API. As VMDK introduces
multiple file support, the operation becomes format specific with
platform specific meanwhile.
The functions are moved to block/raw-{posix,win32}.c and qemu-img.c calls
bdrv_get_allocated_file_size to count the bs. And also added VMDK code
to count his own extents.
Signed-off-by: Fam Zheng <famcool@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add create option 'format', with enums:
monolithicSparse
monolithicFlat
twoGbMaxExtentSparse
twoGbMaxExtentFlat
Each creates a subformat image file. The default is monolithicSparse.
Signed-off-by: Fam Zheng <famcool@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
No users of bdrv_get_type_hint() left. bdrv_set_type_hint() can make
the media removable by side effect. Make that explicit.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Certain operations such as drive_del or resize cannot be performed
while external users (eg. block migration) reference the block device.
Add a flag to indicate that.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Extend the change_cb callback with a reason argument, and use it
to tell drivers about size changes.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This patch introduces the qed on-disk layout and implements image
creation. Later patches add read/write and other functionality.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add a new bdrv_discard method to free blocks in a mapping image, and a new
drive property to set the granularity for these discard. If no discard
granularity support is set discard support is disabled.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If -6 or -e is specified, an error message is printed and we exit. It
does not print help() to avoid the error message getting lost in the
noise.
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If bootindex is specified on command line a string that describes device
in firmware readable way is added into sorted list. Later this list will
be passed into firmware to control boot order.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This changes bdrv_flush to return 0 on success and -errno in case of failure.
It's a requirement for implementing proper error handle in users of bdrv_flush.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
In order to backup snapshots, created from QCOW2 iamge, we want to copy snapshots out of QCOW2 disk to a seperate storage.
The following patch adds a new option in "qemu-img": qemu-img convert -f qcow2 -O qcow2 -s snapshot_name src_img bck_img.
Right now, it only supports to copy the full snapshot, delta snapshot is on the way.
Changes from V1: all the comments from Kevin are addressed:
Add read-only checking
Fix coding style
Change the name from bdrv_snapshot_load to bdrv_snapshot_load_tmp
Signed-off-by: Disheng Su <edison@cloud.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This reverts commit 79368c81bf.
Conflicts:
block.c
I haven't been able to come up with a solution yet for the corruption caused by
unaligned requests from the IDE disk so revert until a solution can be written.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
bdrv_eject() gets called when a device model opens or closes the tray.
If the block driver implements method bdrv_eject(), that method gets
called. Drivers host_cdrom implements it, and it opens and closes the
physical tray, and nothing else. When a device model opens, then
closes the tray, media changes only if the user actively changes the
physical media while the tray is open. This is matches how physical
hardware behaves.
If the block driver doesn't implement method bdrv_eject(), we do
something quite different: opening the tray severs the connection to
the image by calling bdrv_close(), and closing the tray does nothing.
When the device model opens, then closes the tray, media is gone,
unless the user actively inserts another one while the tray is open,
with a suitable change command in the monitor. This isn't how
physical hardware behaves. Rather inconvenient when programs
"helpfully" eject media to give you a chance to change it. The way
bdrv_eject() behaves here turns that chance into a must, which is not
what these programs or their users expect.
Change the default action not to call bdrv_close(). Instead, note the
tray status in new BlockDriverState member tray_open. Use it in
bdrv_is_inserted().
Arguably, the device models should keep track of tray status
themselves. But this is less invasive.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Assuming that any image on a block device is not properly zero-initialized is
actually wrong: Only raw images have this problem. Any other image format
shouldn't care about it, they initialize everything properly themselves.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Currently we set them to 512 bytes unless manually specified. Unforuntaly
some brain-dead partitioning tools create unaligned partitions if they
get low enough optiomal I/O size values, so don't report any at all
unless explicitly set.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
CVE-2008-2004 described a vulnerability in QEMU whereas a malicious user could
trick the block probing code into accessing arbitrary files in a guest. To
mitigate this, we added an explicit format parameter to -drive which disabling
block probing.
Fast forward to today, and the vast majority of users do not use this parameter.
libvirt does not use this by default nor does virt-manager.
Most users want block probing so we should try to make it safer.
This patch adds some logic to the raw device which attempts to detect a write
operation to the beginning of a raw device. If the first 4 bytes happen to
match an image file that has a backing file that we support, it scrubs the
signature to all zeros. If a user specifies an explicit format parameter, this
behavior is disabled.
I contend that while a legitimate guest could write such a signature to the
header, we would behave incorrectly anyway upon the next invocation of QEMU.
This simply changes the incorrect behavior to not involve a security
vulnerability.
I've tested this pretty extensively both in the positive and negative case. I'm
not 100% confident in the block layer's ability to deal with zero sized writes
particularly with respect to the aio functions so some additional eyes would be
appreciated.
Even in the case of a single sector write, we have to make sure to invoked the
completion from a bottom half so just removing the zero sized write is not an
option.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This distinguishes between harmless leaks and real corruption. Hopefully users
better understand what qemu-img check wants to tell them.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For instance, -device scsi-disk,drive=foo -device scsi-disk,drive=foo
happily creates two SCSI disks connected to the same block device.
It's all downhill from there.
Device usb-storage deliberately attaches twice to the same blockdev,
which fails with the fix in place. Detach before the second attach
there.
Also catch attempt to delete while a guest device model is attached.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Make the property point to BlockDriverState, cutting out the DriveInfo
middleman. This prepares the ground for block devices that don't have
a DriveInfo.
Currently all user-defined ones have a DriveInfo, because the only way
to define one is -drive & friends (they go through drive_init()).
DriveInfo is closely tied to -drive, and like -drive, it mixes
information about host and guest part of the block device. I'm
working towards a new way to define block devices, with clean
host/guest separation, and I need to get DriveInfo out of the way for
that.
Fortunately, the device models are perfectly happy with
BlockDriverState, except for two places: ide_drive_initfn() and
scsi_disk_initfn() need to check the DriveInfo for a serial number set
with legacy -drive serial=... Use drive_get_by_blockdev() there.
Device model code should now use DriveInfo only when explicitly
dealing with drives defined the old way, i.e. without -device.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Both SCSI and virtio expect the physical block size relative to the
logical block size. So get the factor first before calculating the
log2.
Reported-by: Mike Cao <bcao@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
That's where they belong semantically (block device host part), even
though the actions are actually executed by guest device code.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds the wr_highest_sector blockstat which implements what is generally
known as the high watermark. It is the highest offset of a sector written to
the respective BlockDriverState since it has been opened.
The query-blockstat QMP command is extended to add this value to the result,
and also to add the statistics of the underlying protocol in a new "parent"
field. Note that to get the "high watermark" of a qcow2 image, you need to look
into the wr_highest_sector field of the parent (which can be a file, a
host_device, ...). The wr_highest_sector of the qcow2 BlockDriverState itself
is the highest offset on the _virtual_ disk that the guest has written to.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Format drivers shouldn't need to bother with things like file names, but rather
just get an open BlockDriverState for the underlying protocol. This patch
introduces this behaviour for bdrv_open implementation. For protocols which
need to access the filename to open their file/device/connection/... a new
callback bdrv_file_open is introduced which doesn't get an underlying file
opened.
For now, also some of the more obscure formats use bdrv_file_open because they
open() the file themselves instead of using the block.c functions. They need to
be fixed in later patches.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The bdrv_first linked list of BlockDriverStates is currently extern so
that block migration can iterate the list. However, since there is
already a bdrv_iterate() function there is no need to expose bdrv_first.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Block drivers can trigger a blkdebug event whenever they reach a place where it
could be useful to inject an error for testing/debugging purposes.
Rules are read from a blkdebug config file and describe which action is taken
when an event is triggered. For now this is only injecting an error (with a few
options) or changing the state (which is an integer). Rules can be declared to
be active only in a specific state; this way later rules can distiguish on
which path we came to trigger their event.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add a logical block size attribute as various guest side tools only
increase the filesystem sector size based on it, not the advisory
physical block size.
For scsi we already have support for a different logical block size
in place for CDROMs that we can built upon. Only my recent block
device characteristics VPD page needs some fixups. Note that we
leave the logial block size for CDROMs hardcoded as the 2k value
is expected for it in general.
For virtio-blk we already have a feature flag claiming to support
a variable logical block size that was added for the s390 kuli
hypervisor. Interestingly it does not actually change the units
in which the protocol works, which is still fixed at 512 bytes,
but only communicates a different minimum I/O granularity. So
all we need to do in virtio is to add a trap for unaligned I/O
and round down the device size to the next multiple of the logical
block size.
IDE does not support any other logical block size than 512 bytes.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Open backing file read-only where possible
Upgrade backing file to read-write during commit, back to read-only after commit
If upgrade fail, back to read-only. If also fail, "disconnect" the drive.
Signed-off-by: Naphtali Sprei <nsprei@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Add three new qdev properties to export block topology information to
the guest. This is needed to get optimal I/O alignment for RAID arrays
or SSDs.
The options are:
- physical_block_size to specify the physical block size of the device,
this is going to increase from 512 bytes to 4096 kilobytes for many
modern storage devices
- min_io_size to specify the minimal I/O size without performance impact,
this is typically set to the RAID chunk size for arrays.
- opt_io_size to specify the optimal sustained I/O size, this is
typically the RAID stripe width for arrays.
I decided to not auto-probe these values from blkid which might easily
be possible as I don't know how to deal with these issues on migration.
Note that we specificly only set the physical_block_size, and not the
logial one which is the unit all I/O is described in. The reason for
that is that IDE does not support increasing the logical block size and
at last for now I want to stick to one meachnisms in queue and allow
for easy switching of transports for a given backing image which would
not be possible if scsi and virtio use real 4k sectors, while ide only
uses the physical block exponent.
To make this more common for the different block drivers introduce a
new BlockConf structure holding all common block properties and a
DEFINE_BLOCK_PROPERTIES macro to add them all together, mirroring
what is done for network drivers. Also switch over all block drivers
to use it, except for the floppy driver which has weird driveA/driveB
properties and probably won't require any advanced block options ever.
Example usage for a virtio device with 4k physical block size and
8k optimal I/O size:
-drive file=scratch.img,media=disk,cache=none,id=scratch \
-device virtio-blk-pci,drive=scratch,physical_block_size=4096,opt_io_size=8192
aliguori: updated patch to take into account BLOCK events
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This will manage dirty counter for each device and will allow to get the
dirty counter from above.
Signed-off-by: Liran Schour <lirans@il.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Introduce the functions needed to change the backing file of an image. The
function is implemented for qcow2.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
I haven't heard yet of anyone using qemu-img to copy an image to a real floppy,
but it's a valid use case.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This switches the dirty bitmap to a true bitmap, reducing its footprint
(specifically in caches). It moreover fixes off-by-one bugs in
set_dirty_bitmap (nb_sectors+1 were marked) and bdrv_get_dirty (limit
check allowed one sector behind end of drive). And is drops redundant
dirty_tracking field from BlockDriverState.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
To support live migration without shared storage we need to be able to trace
writes to disk while migrating. This Patch expose dirty block tracking per
device to be polled from upper layer.
Changes from v4:
- Register dirty tracking for each block device.
- Minor coding style issues.
- Block.c will now manage a dirty bitmap per device once
bdrv_set_dirty_tracking() is called. Bitmap is polled by the upper
layer (block-migration.c).
Signed-off-by: Liran Schour <lirans@il.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.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>
Add a enable_write_cache flag in the block driver state, and use it to
decide if we claim to have a volatile write cache that needs controlled
flushing from the guest. The flag is off if cache=writethrough is
defined because O_DSYNC guarantees that every write goes to stable
storage, and it is on for cache=none and cache=writeback.
Both scsi-disk and ide now use the new flage, changing from their
defaults of always off (ide) or always on (scsi-disk).
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
One performance problem of qcow2 during the initial image growth are
sequential writes that are not cluster aligned. In this case, when a first
requests requires to allocate a new cluster but writes only to the first
couple of sectors in that cluster, the rest of the cluster is zeroed - just
to be overwritten by the following second request that fills up the cluster.
Let's try to merge sequential write requests to the same cluster, so we can
avoid to write the zero padding to the disk in the first place.
As a nice side effect, also other formats take advantage of dealing with less
and larger requests.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
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 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>
Add a bdrv_probe_device method to all BlockDriver instances implementing
host devices to move matching of host device types into the actual drivers.
For now we keep exacly the old matching behaviour based on the devices names,
although we really should have better detetion methods based on device
information in the future.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Now that we have a separate aio pool structure we can remove those
aio pool details from BlockDriver.
Every driver supporting AIO now needs to declare a static AIOPool
with the aiocb size and the cancellation method. This cleans up the
current code considerably and will make it cleaner and more obvious
to support two different aio implementations behind a single
BlockDriver.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Add an option to specify the cluster size of a newly created qcow2 image.
Default is 4k which is the same value that was hard-coded before.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
