Add a simple test case for qemu-iotests that covers read/write
with encrypted qcow2 files.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The qemu-io tool does not check if the image is encrypted so
historically would silently corrupt the sectors by writing
plain text data into them instead of cipher text. The earlier
commit turns this mistake into a fatal abort, so check for
encryption and prompt for key when required.
This enables us to add unit tests to ensure we don't break
the ability of qemu-img to convert existing encrypted qcow2
files into a non-encrypted format.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The qemu_read_password() method looks for \r to terminate the
reading of the a password. This is what will be seen when
reading the password from a TTY. When scripting though, it is
useful to be able to send the password via a pipe, in which
case we must look for \n to terminate password input.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The qemu-img.c file has a read_password() method impl that is
used to prompt for passwords on the console, with impls for
POSIX and Windows. This will be needed by qemu-io.c too, so
move it into the QEMU osdep/oslib files where it can be shared
without code duplication
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When a qcow[2] file is opened, if the header reports an
encryption method, this is used to set the 'crypt_method_header'
field on the BDRVQcow[2]State struct, and the 'encrypted' flag
in the BDRVState struct.
When doing I/O operations, the 'crypt_method' field on the
BDRVQcow[2]State struct is checked to determine if encryption
needs to be applied.
The crypt_method_header value is copied into crypt_method when
the bdrv_set_key() method is called.
The QEMU code which opens a block device is expected to always
do a check
if (bdrv_is_encrypted(bs)) {
bdrv_set_key(bs, ....key...);
}
If code forgets to do this, then 'crypt_method' is never set
and so when I/O is performed, QEMU writes plain text data
into a sector which is expected to contain cipher text, or
when reading, will return cipher text instead of plain
text.
Change the qcow[2] code to consult bs->encrypted when deciding
whether encryption is required, and assert(s->crypt_method)
to protect against cases where the caller forgets to set the
encryption key.
Also put an assert in the set_key methods to protect against
the case where the caller sets an encryption key on a block
device that does not have encryption
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Adding "-d" option. The output goes to "tee" so it appears in your
console. Also, raise the verbosity of unnitest runner.
When testing a topic branch, it's possible that a bug introduced by a
code change makes the python test case hang, with debug output, it is
much easier to locate the problem.
This can also be helpful if you want to watch the progress of a python
test, it offers you a way to sense the speed of each test case method
you're writing.
Note: because there is no easy way to get *both* the verbose output and
the output expected by ./check comparison, the case would always fail
with an "output mismatch". The sole purpose of using this option is
giving developers a quick way to debug when things go wrong.
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The image is contributed by Richard W.M. Jones.
Cc: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Bogus image may have a large total_sectors that will overflow the
multiplication. For cleanness, fix the return code so the error message
will be meaningful.
Reported-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
POSIX says getopt() returns -1 on completion. While Linux happens
to define EOF as -1, this definition is not required by POSIX, and
there may be platforms where checking for EOF instead of -1 would
lead to an infinite loop.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Fix pointer declaration to make it consistent with the rest of the
code.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This function never receives an invalid table pointer, so we can make
it void and remove all the error checking code.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The current cache algorithm traverses the array starting always from
the beginning, so the average number of comparisons needed to perform
a lookup is proportional to the size of the array.
By using a hash of the offset as the starting point, lookups are
faster and independent from the array size.
The hash is computed using the cluster number of the table, multiplied
by 4 to make it perform better when there are collisions.
In my tests, using a cache with 2048 entries, this reduces the average
number of comparisons per lookup from 430 to 2.5.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
A cache miss means that the whole array was traversed and the entry
we were looking for was not found, so there's no need to traverse it
again in order to select an entry to replace.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The current algorithm to evict entries from the cache gives always
preference to those in the lowest positions. As the size of the cache
increases, the chances of the later elements of being removed decrease
exponentially.
In a scenario with random I/O and lots of cache misses, entries in
positions 8 and higher are rarely (if ever) evicted. This can be seen
even with the default cache size, but with larger caches the problem
becomes more obvious.
Using an LRU algorithm makes the chances of being removed from the
cache independent from the position.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Since all tables are now stored together, it is possible to obtain
the position of a particular table directly from its address, so the
operation becomes O(1).
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The qcow2 L2/refcount cache contains one separate table for each cache
entry. Doing one allocation per table adds unnecessary overhead and it
also requires us to store the address of each table separately.
Since the size of the cache is constant during its lifetime, it's
better to have an array that contains all the tables using one single
allocation.
In my tests measuring freshly created caches with sizes 128MB (L2) and
32MB (refcount) this uses around 10MB of RAM less.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Richard Jones caught this bug with afl fuzzer.
In fact, that's the only possible value to overflow (extent->l1_size =
0x20000000) l1_size:
l1_size = extent->l1_size * sizeof(long) => 0x80000000;
g_try_malloc returns NULL because l1_size is interpreted as negative
during type casting from 'int' to 'gsize', which yields a enormous
value. Hence, by coincidence, we get a "not too bad" behavior:
qemu-img: Could not open '/tmp/afl6.img': Could not open
'/tmp/afl6.img': Cannot allocate memory
Values larger than 0x20000000 will be refused by the validation in
vmdk_add_extent.
Values smaller than 0x20000000 will not overflow l1_size.
Cc: qemu-stable@nongnu.org
Reported-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Tested-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This fixes the bug introduced by commit c6ac36e (vmdk: Optimize cluster
allocation).
Sometimes, write_len could be larger than cluster size, because it
contains both data and marker. We must advance next_cluster_sector in
this case, otherwise the image gets corrupted.
Cc: qemu-stable@nongnu.org
Reported-by: Antoni Villalonga <qemu-list@friki.cat>
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The SCSI emulation in the Linux NVMe driver really wants to know
if a device has a volatile write cache. Given that qemu has moved
away from a model where we report the backing store WCE bit to
one where the WCE bit is supposed to be part of the migratable
guest-visible state we always return 1 here.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Before a freed cluster can be reused, pending discards for this cluster
must be processed.
The original assumption was that this was not a problem because discards
are only cached during discard/write zeroes operations, which are
synchronous so that no concurrent write requests can cause cluster
allocations.
However, the discard/write zeroes operation itself can allocate a new L2
table (and it has to in order to put zero flags there), so make sure we
can cope with the situation.
This fixes https://bugs.launchpad.net/bugs/1349972.
Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
This instruction was introduced by the new Aurix platform.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Those instruction were introduced in the new Aurix platform.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Those instruction were introduced in the new Aurix platform.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
The aurix platform contains of several different cpu models and uses
the 1.6.1 ISA. This patch changes the generic aurix model to the more
specific tc27x cpu model and sets specific features.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
This reverts commit 32ad48abd7.
Unfortunately the SSE2 code here fails to compile on some versions
of gcc:
target-alpha/int_helper.c:77:24: error: invalid operands to binary >=
(have '__vector(16) unsigned char' and '__vector(16) unsigned char')
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
A bit of Boolean algebra (and common sense) tells us that the
second "if" here is looking for blocks that are not allocated.
This is the opposite of the "if" that sets BDRV_BLOCK_ALLOCATED,
and thus it can use an "else".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 1431599702-10431-1-git-send-email-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Test zero write in byte range 512~1024 for 4k alignment.
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 1431522721-3266-4-git-send-email-famz@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
For zero write, callers pass in NULL qiov (qemu-io "write -z" or
scsi-disk "write same").
Commit fc3959e466 fixed bdrv_co_write_zeroes which is the common case
for this bug, but it still exists in bdrv_aio_write_zeroes. A simpler
fix would be in bdrv_co_do_pwritev which is the NULL dereference point
and covers both cases.
So don't access it in bdrv_co_do_pwritev in this case, use three aligned
writes.
[Initialize ret to 0 in bdrv_co_do_zero_pwritev() to avoid uninitialized
variable warning with gcc 4.9.2.
--Stefan]
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-id: 1431522721-3266-3-git-send-email-famz@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
This reverts commit fc3959e466.
The core write code already handles the case, so remove this
duplication.
Because commit 61007b316 moved the touched code from block.c to
block/io.c, the change is manually reverted.
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 1431522721-3266-2-git-send-email-famz@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The following sequence
int fd = open(argv[1], O_RDWR | O_CREAT | O_DIRECT, 0644);
for (i = 0; i < 100000; i++)
write(fd, buf, 4096);
performs 5% better if buf is aligned to 4096 bytes.
The difference is quite reliable.
On the other hand we do not want at the moment to enforce bounce
buffering if guest request is aligned to 512 bytes.
The patch changes default bounce buffer optimal alignment to
MAX(page size, 4k). 4k is chosen as maximal known sector size on real
HDD.
The justification of the performance improve is quite interesting.
From the kernel point of view each request to the disk was split
by two. This could be seen by blktrace like this:
9,0 11 1 0.000000000 11151 Q WS 312737792 + 1023 [qemu-img]
9,0 11 2 0.000007938 11151 Q WS 312738815 + 8 [qemu-img]
9,0 11 3 0.000030735 11151 Q WS 312738823 + 1016 [qemu-img]
9,0 11 4 0.000032482 11151 Q WS 312739839 + 8 [qemu-img]
9,0 11 5 0.000041379 11151 Q WS 312739847 + 1016 [qemu-img]
9,0 11 6 0.000042818 11151 Q WS 312740863 + 8 [qemu-img]
9,0 11 7 0.000051236 11151 Q WS 312740871 + 1017 [qemu-img]
9,0 5 1 0.169071519 11151 Q WS 312741888 + 1023 [qemu-img]
After the patch the pattern becomes normal:
9,0 6 1 0.000000000 12422 Q WS 314834944 + 1024 [qemu-img]
9,0 6 2 0.000038527 12422 Q WS 314835968 + 1024 [qemu-img]
9,0 6 3 0.000072849 12422 Q WS 314836992 + 1024 [qemu-img]
9,0 6 4 0.000106276 12422 Q WS 314838016 + 1024 [qemu-img]
and the amount of requests sent to disk (could be calculated counting
number of lines in the output of blktrace) is reduced about 2 times.
Both qemu-img and qemu-io are affected while qemu-kvm is not. The guest
does his job well and real requests comes properly aligned (to page).
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 1431441056-26198-3-git-send-email-den@openvz.org
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The patch introduces new concept: minimal memory alignment for bounce
buffers. Original so called "optimal" value is actually minimal required
value for aligment. It should be used for validation that the IOVec
is properly aligned and bounce buffer is not required.
Though, from the performance point of view, it would be better if
bounce buffer or IOVec allocated by QEMU will be aligned stricter.
The patch does not change any alignment value yet.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 1431441056-26198-2-git-send-email-den@openvz.org
CC: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
This is the behavior in the operating system, for example Linux's
blkdev_write_iter has the following:
if (bdev_read_only(I_BDEV(bd_inode)))
return -EPERM;
This does not apply to opening a device for read/write, when the
device only supports read-only operation. In this case any of
EACCES, EPERM or EROFS is acceptable depending on why writing is
not possible.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1431013548-22492-1-git-send-email-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Test if ccache is interfering with semantic analysis of macros,
disable its habit of trying to compile already pre-processed
versions of code if so. ccache attempts to save time by compiling
pre-processed versions of code, but this disturbs clang's static
analysis enough to produce false positives.
ccache allows us to disable this feature, opting instead to
compile the original version instead of its preprocessed version.
This makes ccache much slower for cache misses, but at least it
becomes usable with QEMU/clang.
This workaround only activates for users using ccache AND clang,
and only if their configuration is observed to be producing warnings.
You may need to clear your ccache for builds started without -Werror,
as those may continue to produce warnings from the cache.
Thanks to Peter Eisentraut for his writeup on the issue:
http://peter.eisentraut.org/blog/2014/12/01/ccache-and-clang-part-3/
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1427324259-1481-5-git-send-email-jsnow@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The glib headers use GCC attributes. Unfortunately the __GNUC__ and
__GNUC_MINOR__ version macros are also defined by clang, but clang
doesn't support the same attributes as GCC.
clang 3.5.0 does not support the __alloc_size__ attribute:
c047507a9a
The following warning is produced:
gstrfuncs.h:257:44: warning: unknown attribute '__alloc_size__' ignored [-Wunknown-attributes]
G_GNUC_MALLOC G_GNUC_ALLOC_SIZE(2);
gmacros.h:67:45: note: expanded from macro 'G_GNUC_ALLOC_SIZE'
#define G_GNUC_ALLOC_SIZE(x) __attribute__((__alloc_size__(x)))
This patch checks whether glib headers cause warnings and disables
-Wunknown-attributes if it is able to.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1427324259-1481-4-git-send-email-jsnow@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Factor out the function that checks if a compiler
flag is supported or not.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1427324259-1481-3-git-send-email-jsnow@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
gcc 4.9.2 treats -nopie as an error:
cc: error: unrecognized command line option ‘-nopie’
clang 3.5.0 treats -nopie as a warning:
clang: warning: argument unused during compilation: '-nopie'
The causes ./configure to fail with clang:
ERROR: configure test passed without -Werror but failed with -Werror.
Make the -nopie test use -Werror so that compile_prog works for both gcc
and clang.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1427324259-1481-2-git-send-email-jsnow@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Try to perform IO for the biggest continuous block possible.
All blocks abscent in the image are accounted in the same type
and preallocation is made for all of them at once.
The performance for sequential write is increased from 200 Mb/sec to
235 Mb/sec on my SSD HDD.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Roman Kagan <rkagan@parallels.com>
Signed-off-by: Roman Kagan <rkagan@parallels.com>
Message-id: 1430207220-24458-28-git-send-email-den@openvz.org
CC: Kevin Wolf <kwolf@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Plain image expansion spends a lot of time to update image file size.
This seriously affects the performance. The following simple test
qemu_img create -f parallels -o cluster_size=64k ./1.hds 64G
qemu_io -n -c "write -P 0x11 0 1024M" ./1.hds
could be improved if the format driver will pre-allocate some space
in the image file with a reasonable chunk.
This patch preallocates 128 Mb using bdrv_write_zeroes, which should
normally use fallocate() call inside. Fallback to older truncate()
could be used as a fallback using image open options thanks to the
previous patch.
The benefit is around 15%.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Roman Karan <rkagan@parallels.com>
Signed-off-by: Roman Kagan <rkagan@parallels.com>
Message-id: 1430207220-24458-27-git-send-email-den@openvz.org
CC: Kevin Wolf <kwolf@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
This is preparational commit for tweaks in Parallels image expansion.
The idea is that enlarge via truncate by one data block is slow. It
would be much better to use fallocate via bdrv_write_zeroes and
expand by some significant amount at once.
Original idea with sequential file writing to the end of the file without
fallocate/truncate would be slower than this approach if the image is
expanded with several operations:
- each image expanding means file metadata update, i.e. filesystem
journal write. Truncate/write to newly truncated space update file
metadata twice thus truncate removal helps. With fallocate call
inside bdrv_write_zeroes file metadata is updated only once and
this should happen infrequently thus this approach is the best one
for the image expansion
- tail writes are ordered, i.e. the guest IO queue could not be sent
immediately to the host introducing additional IO delays
This patch just adds proper parameters into BDRVParallelsState and
performs options parsing in parallels_open.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Roman Kagan <rkagan@parallels.com>
Signed-off-by: Roman Kagan <rkagan@parallels.com>
Message-id: 1430207220-24458-26-git-send-email-den@openvz.org
CC: Roman Kagan <rkagan@parallels.com>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The idea is that we do not need to immediately sync BAT to the image as
from the guest point of view there is a possibility that IO is lost
even in the physical controller until flush command was finished.
bdrv_co_flush_to_os is exactly the right place for this purpose.
Technically the patch uses loaded BAT data as a cache and performs
actual on-disk metadata updates in parallels_co_flush_to_os callback.
This patch speed ups
qemu-img create -f parallels -o cluster_size=64k ./1.hds 64G
qemu-io -f parallels -c "write -P 0x11 0 1024k" 1.hds
writing from 50-60 Mb/sec to 80-90 Mb/sec on rotational media and
from 160 Mb/sec to 190 Mb/sec on SSD disk.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Roman Kagan <rkagan@parallels.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Roman Kagan <rkagan@parallels.com>
Message-id: 1430207220-24458-25-git-send-email-den@openvz.org
CC: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Markus Armbruster