The QCowL2Meta structure is used to store information about a part of
a write request that touches clusters that need changes in their L2
entries. This happens with newly-allocated clusters or subclusters.
This structure has changed a bit since it was first created and its
current documentation is not quite up-to-date.
A write request can span a region consisting of a combination of
clusters of different types, and qcow2_alloc_host_offset() can
repeatedly call handle_copied() and handle_alloc() to add more
clusters to the mix as long as they all are contiguous on the image
file.
Because of this a write request has a list of QCowL2Meta structures,
one for each part of the request that needs changes in the L2
metadata.
Each one of them spans nb_clusters and has two copy-on-write regions
located immediately before and after the middle region touched by that
part of the write request. Even when those regions themselves are
empty their offsets must be correct because they are used to know the
location of the middle region.
This was not always the case but it is not a problem anymore
because the only two places where QCowL2Meta structures are created
(calculate_l2_meta() and qcow2_co_truncate()) ensure that the
copy-on-write regions are correctly defined, and so do assertions like
the ones in perform_cow().
The conditional initialization of the 'written_to' variable is
therefore unnecessary and is removed by this patch.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20201007161323.4667-1-berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Since commit c8bb23cbdb when a write
request results in a new allocation QEMU first tries to see if the
rest of the cluster outside the written area contains only zeroes.
In that case, instead of doing a normal copy-on-write operation and
writing explicit zero buffers to disk, the code zeroes the whole
cluster efficiently using pwrite_zeroes() with BDRV_REQ_NO_FALLBACK.
This improves performance very significantly but it only happens when
we are writing to an area that was completely unallocated before. Zero
clusters (QCOW2_CLUSTER_ZERO_*) are treated like normal clusters and
are therefore slower to allocate.
This happens because the code uses bdrv_is_allocated_above() rather
bdrv_block_status_above(). The former is not as accurate for this
purpose but it is faster. However in the case of qcow2 the underlying
call does already report zero clusters just fine so there is no reason
why we cannot use that information.
After testing 4KB writes on an image that only contains zero clusters
this patch results in almost five times more IOPS.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <6d77cab968c501c44d6e1089b9bc91b04170b49e.1603731354.git.berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_co_block_status_above has several design problems with handling
short backing files:
1. With want_zeros=true, it may return ret with BDRV_BLOCK_ZERO but
without BDRV_BLOCK_ALLOCATED flag, when actually short backing file
which produces these after-EOF zeros is inside requested backing
sequence.
2. With want_zero=false, it may return pnum=0 prior to actual EOF,
because of EOF of short backing file.
Fix these things, making logic about short backing files clearer.
With fixed bdrv_block_status_above we also have to improve is_zero in
qcow2 code, otherwise iotest 154 will fail, because with this patch we
stop to merge zeros of different types (produced by fully unallocated
in the whole backing chain regions vs produced by short backing files).
Note also, that this patch leaves for another day the general problem
around block-status: misuse of BDRV_BLOCK_ALLOCATED as is-fs-allocated
vs go-to-backing.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20200924194003.22080-2-vsementsov@virtuozzo.com
[Fix s/comes/come/ as suggested by Eric Blake
--Stefan]
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
qcow2_alloc_cluster_offset() takes an (unaligned) guest offset and
returns the (aligned) offset of the corresponding cluster in the qcow2
image.
In practice none of the callers need to know where the cluster starts
so this patch makes the function calculate and return the final host
offset directly. The function is also renamed accordingly.
See 388e581615 for a similar change to qcow2_get_cluster_offset().
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <9bfef50ec9200d752413be4fc2aeb22a28378817.1599833007.git.berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
This function preallocates metadata structures and then extends the
image to its new size, but that new size calculation is wrong because
it doesn't take into account that the host_offset variable is always
cluster-aligned.
This problem can be reproduced with preallocation=metadata when the
original size is not cluster-aligned but the new size is. In this case
the final image size will be shorter than expected.
qemu-img create -f qcow2 img.qcow2 31k
qemu-img resize --preallocation=metadata img.qcow2 128k
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <adeb8b059917b141d5f5b3bd2a016262d3052c79.1599833007.git.berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
[mreitz: Mark compat=0.10 unsupported for iotest 125]
Signed-off-by: Max Reitz <mreitz@redhat.com>
This function checks the current status of a (sub)cluster in order to
see if an unaligned 'write zeroes' request can be done efficiently by
simply updating the L2 metadata and without having to write actual
zeroes to disk.
If the situation does not allow using the fast path then the function
returns -ENOTSUP and the caller falls back to writing zeroes.
If can happen however that the aforementioned check returns an actual
error code so in this case we should pass it to the caller.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20200909123739.719-1-berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
If qcow2_alloc_cluster_offset() or qcow2_alloc_cluster_link_l2() fail
then this function simply returns the error code, potentially leaking
the QCowL2Meta structure and leaving stale items in s->cluster_allocs.
A second problem is that this function calls qcow2_free_any_clusters()
on failure but passing a host cluster offset instead of an L2 entry.
Luckily for normal uncompressed clusters a raw offset also works like
a valid L2 entry so it works just the same, but we should be using
qcow2_free_clusters() instead.
This patch fixes both problems by using qcow2_handle_l2meta().
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <cd3a6b9abd43f9c0b60be413d760f0cacc67eb66.1599573989.git.berto@igalia.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
In the past, when a new cluster was allocated the l2meta structure was
a variable in the stack so it was necessary to have a way to tell
whether it had been initialized and contained valid data or not. The
nb_clusters field was used for this purpose. Since commit f50f88b9fe
this is no longer the case, l2meta (nowadays a pointer to a list) is
only allocated when needed and nb_clusters is guaranteed to be > 0 so
this check is unnecessary.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <ab0b67c29c7ba26e598db35f12aa5ab5982539c1.1599150873.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
This patch replaces instances of sizeof(uint64_t) in the qcow2 driver
with macros that indicate what those sizes are actually referring to.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20200828110828.13833-1-berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Traditional qcow2 images don't allow preallocation if a backing file
is set. This is because once a cluster is allocated there is no way to
tell that its data should be read from the backing file.
Extended L2 entries have individual allocation bits for each
subcluster, and therefore it is perfectly possible to have an
allocated cluster with all its subclusters unallocated.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <6d5b0f38e7dc5f2f31d8cab1cb92044e9909aece.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Now that the implementation of subclusters is complete we can finally
add the necessary options to create and read images with this feature,
which we call "extended L2 entries".
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <6476caaa73216bd05b7bb2d504a20415e1665176.1594396418.git.berto@igalia.com>
[mreitz: %s/5\.1/5.2/; fixed 302's and 303's reference output]
Signed-off-by: Max Reitz <mreitz@redhat.com>
This field allows us to indicate that the L2 metadata update does not
come from a write request with actual data but from a preallocation
request.
For traditional images this does not make any difference, but for
images with extended L2 entries this means that the clusters are
allocated normally in the L2 table but individual subclusters are
marked as unallocated.
This will allow preallocating images that have a backing file.
There is one special case: when we resize an existing image we can
also request that the new clusters are preallocated. If the image
already had a backing file then we have to hide any possible stale
data and zero out the new clusters (see commit 955c7d6687 for more
details).
In this case the subclusters cannot be left as unallocated so the L2
bitmap must be updated.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <960d4c444a4f5a870e2b47e5da322a73cd9a2f5a.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Extended L2 entries are bigger than normal L2 entries so this has an
impact on the amount of metadata needed for a qcow2 file.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <7efae2efd5e36b42d2570743a12576d68ce53685.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
This works now at the subcluster level and pwrite_zeroes_alignment is
updated accordingly.
qcow2_cluster_zeroize() is turned into qcow2_subcluster_zeroize() with
the following changes:
- The request can now be subcluster-aligned.
- The cluster-aligned body of the request is still zeroized using
zero_in_l2_slice() as before.
- The subcluster-aligned head and tail of the request are zeroized
with the new zero_l2_subclusters() function.
There is just one thing to take into account for a possible future
improvement: compressed clusters cannot be partially zeroized so
zero_l2_subclusters() on the head or the tail can return -ENOTSUP.
This makes the caller repeat the *complete* request and write actual
zeroes to disk. This is sub-optimal because
1) if the head area was compressed we would still be able to use
the fast path for the body and possibly the tail.
2) if the tail area was compressed we are writing zeroes to the
head and the body areas, which are already zeroized.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <17e05e2ee7e12f10dcf012da81e83ebe27eb3bef.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
The bdrv_co_pwrite_zeroes() call here fills complete clusters with
zeroes, but it can happen that some subclusters are not part of the
write request or the copy-on-write. This patch makes sure that only
the affected subclusters are overwritten.
A potential improvement would be to also fill with zeroes the other
subclusters if we can guarantee that we are not overwriting existing
data. However this would waste more disk space, so we should first
evaluate if it's really worth doing.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <b3dc97e8e2240ddb5191a4f930e8fc9653f94621.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
When dealing with subcluster types there is a new value called
QCOW2_SUBCLUSTER_UNALLOCATED_ALLOC that has no equivalent in
QCow2ClusterType.
This patch handles that value in all places where subcluster types
are processed.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <bf09e2e2439a468a901bb96ace411eed9ee50295.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
In order to support extended L2 entries some functions of the qcow2
driver need to start dealing with subclusters instead of clusters.
qcow2_get_host_offset() is modified to return the subcluster type
instead of the cluster type, and all callers are updated to replace
all values of QCow2ClusterType with their QCow2SubclusterType
equivalents.
This patch only changes the data types, there are no semantic changes.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <f6c29737c295f32cbee74c903c30b01820363b34.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
This function returns an integer that can be either an error code or a
cluster type (a value from the QCow2ClusterType enum).
We are going to start using subcluster types instead of cluster types
in some functions so it's better to use the exact data types instead
of integers for clarity and in order to detect errors more easily.
This patch makes qcow2_get_host_offset() return 0 on success and
puts the returned cluster type in a separate parameter. There are no
semantic changes.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <396b6eab1859a271551dcd7dcba77f8934aa3c3f.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
qcow2 images with subclusters have 128-bit L2 entries. The first 64
bits contain the same information as traditional images and the last
64 bits form a bitmap with the status of each individual subcluster.
Because of that we cannot assume that L2 entries are sizeof(uint64_t)
anymore. This function returns the proper value for the image.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <d34d578bd0380e739e2dde3e8dd6187d3d249fa9.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
This patch adds the following new fields to BDRVQcow2State:
- subclusters_per_cluster: Number of subclusters in a cluster
- subcluster_size: The size of each subcluster, in bytes
- subcluster_bits: No. of bits so 1 << subcluster_bits = subcluster_size
Images without subclusters are treated as if they had exactly one
subcluster per cluster (i.e. subcluster_size = cluster_size).
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <55bfeac86b092fa2c9d182a95cbeb479ff7eca4f.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
qcow2_get_cluster_offset() takes an (unaligned) guest offset and
returns the (aligned) offset of the corresponding cluster in the qcow2
image.
In practice none of the callers need to know where the cluster starts
so this patch makes the function calculate and return the final host
offset directly. The function is also renamed accordingly.
There is a pre-existing exception with compressed clusters: in this
case the function returns the complete cluster descriptor (containing
the offset and size of the compressed data). This does not change with
this patch but it is now documented.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <ffae6cdc5ca8950e8280ac0f696dcc376cb07095.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
The file_cluster_offset field of Qcow2AioTask stores a cluster-aligned
host offset. In practice this is not very useful because all users(*)
of this structure need the final host offset into the cluster, which
they calculate using
host_offset = file_cluster_offset + offset_into_cluster(s, offset)
There is no reason why Qcow2AioTask cannot store host_offset directly
and that is what this patch does.
(*) compressed clusters are the exception: in this case what
file_cluster_offset was storing was the full compressed cluster
descriptor (offset + size). This does not change with this patch
but it is documented now.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <07c4b15c644dcf06c9459f98846ac1c4ea96e26f.1594396418.git.berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Make the capitalization of the hexadecimal numbers consistent for the
QCOW2 header extension constants in docs/interop/qcow2.txt.
Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <1594973699-781898-2-git-send-email-andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
For now, this is a mechanical addition; all callers pass false. But
the next patch will use it to improve 'qemu-img rebase -u' when
selecting a backing file with no format.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Message-Id: <20200706203954.341758-10-eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The use of 'qemu-img amend' to change qcow2 backing files is not
tested very well. In particular, our implementation has a bug where
if a new backing file is provided without a format, then the prior
format is blindly reused, even if this results in data corruption, but
this is not caught by iotests.
There are also situations where amending other options needs access to
the original backing file (for example, on a downgrade to a v2 image,
knowing whether a v3 zero cluster must be allocated or may be left
unallocated depends on knowing whether the backing file already reads
as zero), but the command line does not have a nice way to tell us
both the backing file to use for opening the image as well as the
backing file to install after the operation is complete.
Even if we do allow changing the backing file, it is redundant with
the existing ability to change backing files via 'qemu-img rebase -u'.
It is time to deprecate this support (leaving the existing behavior
intact, even if it is buggy), and at a point in the future, require
the use of only 'qemu-img rebase' for adjusting backing chain
relations, saving 'qemu-img amend' for changes unrelated to the
backing chain.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200706203954.341758-8-eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Convert
visit_type_FOO(v, ..., &ptr, &err);
...
if (err) {
...
}
to
visit_type_FOO(v, ..., &ptr, errp);
...
if (!ptr) {
...
}
for functions that set @ptr to non-null / null on success / error.
Eliminate error_propagate() that are now unnecessary. Delete @err
that are now unused.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200707160613.848843-40-armbru@redhat.com>
When all we do with an Error we receive into a local variable is
propagating to somewhere else, we can just as well receive it there
right away. The previous commit did that with a Coccinelle script I
consider fairly trustworthy. This commit uses the same script with
the matching of return taken out, i.e. we convert
if (!foo(..., &err)) {
...
error_propagate(errp, err);
...
}
to
if (!foo(..., errp)) {
...
...
}
This is unsound: @err could still be read between afterwards. I don't
know how to express "no read of @err without an intervening write" in
Coccinelle. Instead, I manually double-checked for uses of @err.
Suboptimal line breaks tweaked manually. qdev_realize() simplified
further to placate scripts/checkpatch.pl.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200707160613.848843-36-armbru@redhat.com>
Currently this field only set by qed and qcow2. But in fact, all
backing-supporting formats (parallels, qcow, qcow2, qed, vmdk) share
these semantics: on unallocated blocks, if there is no backing file they
just memset the buffer with zeroes.
So, document this behavior for .supports_backing and drop
.unallocated_blocks_are_zero
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200528094405.145708-10-vsementsov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Currently the implementation only supports amending the encryption
options, unlike the qemu-img version
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200608094030.670121-14-mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Now that we have all the infrastructure in place,
wire it in the qcow2 driver and expose this to the user.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200608094030.670121-9-mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Some qcow2 create options can't be used for amend.
Remove them from the qcow2 create options and add generic logic to detect
such options in qemu-img
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[mreitz: Dropped some iotests reference output hunks that became
unnecessary thanks to
"iotests: Make _filter_img_create more active"]
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200625125548.870061-12-mreitz@redhat.com>
Some options are only useful for creation
(or hard to be amended, like cluster size for qcow2), while some other
options are only useful for amend, like upcoming keyslot management
options for luks
Since currently only qcow2 supports amend, move all its options
to a common macro and then include it in each action option list.
In future it might be useful to remove some options which are
not supported anyway from amend list, which currently
cause an error message if amended.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200608094030.670121-5-mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
'force' option will be used for some unsafe amend operations.
This includes things like erasing last keyslot in luks based formats
(which destroys the data, unless the master key is backed up
by external means), but that _might_ be desired result.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200608094030.670121-4-mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
When resizing an image with qcow2_co_truncate() using the falloc or
full preallocation modes the code assumes that both the old and new
sizes are cluster-aligned.
There are two problems with this:
1) The calculation of how many clusters are involved does not always
get the right result.
Example: creating a 60KB image and resizing it (with
preallocation=full) to 80KB won't allocate the second cluster.
2) No copy-on-write is performed, so in the previous example if
there is a backing file then the first 60KB of the first cluster
won't be filled with data from the backing file.
This patch fixes both issues.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20200617140036.20311-1-berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
It's useful to know how much space can be occupied by qcow2 persistent
bitmaps, even though such metadata is unrelated to the guest-visible
data. Report this value as an additional QMP field, present when
measuring an existing image and output format that both support
bitmaps. Update iotest 178 and 190 to updated output, as well as new
coverage in 190 demonstrating non-zero values made possible with the
recently-added qemu-img bitmap command (see 3b51ab4b).
The new 'bitmaps size:' field is displayed automatically as part of
'qemu-img measure' any time it is present in QMP (that is, any time
both the source image being measured and destination format support
bitmaps, even if the measurement is 0 because there are no bitmaps
present). If the field is absent, it means that no bitmaps can be
copied (source, destination, or both lack bitmaps, including when
measuring based on size rather than on a source image). This behavior
is compatible with an upcoming patch adding 'qemu-img convert
--bitmaps': that command will fail in the same situations where this
patch omits the field.
The addition of a new field demonstrates why we should always
zero-initialize qapi C structs; while the qcow2 driver still fully
populates all fields, the raw and crypto drivers had to be tweaked to
avoid uninitialized data.
Consideration was also given towards having a 'qemu-img measure
--bitmaps' which errors out when bitmaps are not possible, and
otherwise sums the bitmaps into the existing allocation totals rather
than displaying as a separate field, as a potential convenience
factor. But this was ultimately decided to be more complexity than
necessary when the QMP interface was sufficient enough with bitmaps
remaining a separate field.
See also: https://bugzilla.redhat.com/1779904
Reported-by: Nir Soffer <nsoffer@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200521192137.1120211-3-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Upcoming patches will enhance bitmap support in qemu-img, but in doing
so, it turns out to be nice to suppress output when persistent bitmaps
make no sense (such as on a qcow2 v2 image). Add a hook to make this
easier to query.
This patch adds a new callback .bdrv_supports_persistent_dirty_bitmap,
rather than trying to shoehorn the answer in via existing callbacks.
In particular, while it might have been possible to overload
.bdrv_co_can_store_new_dirty_bitmap to special-case a NULL input to
answer whether any persistent bitmaps are supported, that is at odds
with whether a particular bitmap can be stored (for example, even on
an image that supports persistent bitmaps but has currently filled up
the maximum number of bitmaps, attempts to store another one should
fail); and the new functionality doesn't require coroutine safety.
Similarly, we could have added one more piece of information to
.bdrv_get_info, but then again, most callers to that function tend to
already discard extraneous information, and making it a catch-all
rather than a series of dedicated scalar queries hasn't really
simplified life.
In the future, when we improve the ability to look up bitmaps through
a filter, we will probably also want to teach the block layer to
automatically let filters pass this request on through.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200513011648.166876-4-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
bdrv_default_perms() can decide which permission profile to use based on
the BdrvChildRole, so block drivers do not need to select it explicitly.
The blkverify driver now no longer shares the WRITE permission for the
image to verify. We thus have to adjust two places in
test-block-iothread not to take it. (Note that in theory, blkverify
should behave like quorum in this regard and share neither WRITE nor
RESIZE for both of its children. In practice, it does not really
matter, because blkverify is used only for debugging, so we might as
well keep its permissions rather liberal.)
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200513110544.176672-30-mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Commonly, they need to pass the BDRV_CHILD_IMAGE set as the
BdrvChildRole; but there are exceptions for drivers with external data
files (qcow2 and vmdk).
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200513110544.176672-26-mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For now, it is always set to 0. Later patches in this series will
ensure that all callers pass an appropriate combination of flags.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200513110544.176672-6-mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We want to unify child_format and child_file at some point. One of the
important things that set format drivers apart from other drivers is
that they do not expect other format nodes under them (except in the
backing chain), i.e. we must not probe formats inside of formats. That
means we need something on which to distinguish format drivers from
others, and hence this flag.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20200513110544.176672-3-mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
zstd significantly reduces cluster compression time.
It provides better compression performance maintaining
the same level of the compression ratio in comparison with
zlib, which, at the moment, is the only compression
method available.
The performance test results:
Test compresses and decompresses qemu qcow2 image with just
installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2G
The test was conducted with brd disk to reduce the influence
of disk subsystem to the test results.
The results is given in seconds.
compress cmd:
time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
src.img [zlib|zstd]_compressed.img
decompress cmd
time ./qemu-img convert -O qcow2
[zlib|zstd]_compressed.img uncompressed.img
compression decompression
zlib zstd zlib zstd
------------------------------------------------------------
real 65.5 16.3 (-75 %) 1.9 1.6 (-16 %)
user 65.0 15.8 5.3 2.5
sys 3.3 0.2 2.0 2.0
Both ZLIB and ZSTD gave the same compression ratio: 1.57
compressed image size in both cases: 1.4G
Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
QAPI part:
Acked-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200507082521.29210-4-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
The patch adds some preparation parts for incompatible compression type
feature to qcow2 allowing the use different compression methods for
image clusters (de)compressing.
It is implied that the compression type is set on the image creation and
can be changed only later by image conversion, thus compression type
defines the only compression algorithm used for the image, and thus,
for all image clusters.
The goal of the feature is to add support of other compression methods
to qcow2. For example, ZSTD which is more effective on compression than ZLIB.
The default compression is ZLIB. Images created with ZLIB compression type
are backward compatible with older qemu versions.
Adding of the compression type breaks a number of tests because now the
compression type is reported on image creation and there are some changes
in the qcow2 header in size and offsets.
The tests are fixed in the following ways:
* filter out compression_type for many tests
* fix header size, feature table size and backing file offset
affected tests: 031, 036, 061, 080
header_size +=8: 1 byte compression type
7 bytes padding
feature_table += 48: incompatible feature compression type
backing_file_offset += 56 (8 + 48 -> header_change + feature_table_change)
* add "compression type" for test output matching when it isn't filtered
affected tests: 049, 060, 061, 065, 082, 085, 144, 182, 185, 198, 206,
242, 255, 274, 280
Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
QAPI part:
Acked-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200507082521.29210-2-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Now that there are no clients of bdrv_has_zero_init_truncate, none of
the drivers need to worry about providing it.
What's more, this eliminates a source of some confusion: a literal
reading of the documentation as written in ceaca56f and implemented in
commit 1dcaf527 claims that a driver which returns 0 for
bdrv_has_zero_init_truncate() must not return 1 for
bdrv_has_zero_init(); this condition was violated for parallels, qcow,
and sometimes for vdi, although in practice it did not matter since
those drivers also lacked .bdrv_co_truncate.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200428202905.770727-10-eblake@redhat.com>
Acked-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Calling bdrv_getlength() to get the pre-truncate file size will not
really work on block devices, because they have always the same length,
and trying to write beyond it will fail with a rather cryptic error
message.
Instead, we should use qcow2_get_last_cluster() and bdrv_getlength()
only as a fallback.
Before this patch:
$ truncate -s 1G test.img
$ sudo losetup -f --show test.img
/dev/loop0
$ sudo qemu-img create -f qcow2 -o preallocation=full /dev/loop0 64M
Formatting '/dev/loop0', fmt=qcow2 size=67108864 cluster_size=65536
preallocation=full lazy_refcounts=off refcount_bits=16
qemu-img: /dev/loop0: Could not resize image: Failed to resize refcount
structures: No space left on device
With this patch:
$ sudo qemu-img create -f qcow2 -o preallocation=full /dev/loop0 64M
Formatting '/dev/loop0', fmt=qcow2 size=67108864 cluster_size=65536
preallocation=full lazy_refcounts=off refcount_bits=16
qemu-img: /dev/loop0: Could not resize image: Failed to resize
underlying file: Preallocation mode 'full' unsupported for this
non-regular file
So as you can see, it still fails, but now the problem is missing
support on the block device level, so we at least get a better error
message.
Note that we cannot preallocate block devices on truncate by design,
because we do not know what area to preallocate. Their length is always
the same, the truncate operation does not change it.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200505141801.1096763-1-mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
After commit f01643fb8b when an image is
extended and BDRV_REQ_ZERO_WRITE is set then the new clusters are
zeroized.
The code however does not detect correctly situations when the old and
the new end of the image are within the same cluster. The problem can
be reproduced with these steps:
qemu-img create -f qcow2 backing.qcow2 1M
qemu-img create -f qcow2 -F qcow2 -b backing.qcow2 top.qcow2
qemu-img resize --shrink top.qcow2 520k
qemu-img resize top.qcow2 567k
In the last step offset - zero_start causes an integer wraparound.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20200504155217.10325-1-berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Our comment did not actually match the code. Rewrite the comment to
be less sensitive to any future changes to qcow2-bitmap.c that might
implement scenarios that we currently reject.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200428192648.749066-4-eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
We originally refused to allow resize of images with internal
snapshots because the v2 image format did not require the tracking of
snapshot size, making it impossible to safely revert to a snapshot
with a different size than the current view of the image. But the
snapshot size tracking was rectified in v3, and our recent fixes to
qemu-img amend (see 0a85af35) guarantee that we always have a valid
snapshot size. Thus, we no longer need to artificially limit image
resizes, but it does become one more thing that would prevent a
downgrade back to v2. And now that we support different-sized
snapshots, it's also easy to fix reverting to a snapshot to apply the
new size.
Upgrade iotest 61 to cover this (we previously had NO coverage of
refusal to resize while snapshots exist). Note that the amend process
can fail but still have effects: in particular, since we break things
into upgrade, resize, downgrade, a failure during resize does not roll
back changes made during upgrade, nor does failure in downgrade roll
back a resize. But this situation is pre-existing even without this
patch; and without journaling, the best we could do is minimize the
chance of partial failure by collecting all changes prior to doing any
writes - which adds a lot of complexity but could still fail with EIO.
On the other hand, we are careful that even if we have partial
modification but then fail, the image is left viable (that is, we are
careful to sequence things so that after each successful cluster
write, there may be transient leaked clusters but no corrupt
metadata). And complicating the code to make it more transaction-like
is not worth the effort: a user can always request multiple 'qemu-img
amend' changing one thing each, if they need finer-grained control
over detecting the first failure than what they get by letting qemu
decide how to sequence multiple changes.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200428192648.749066-3-eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
There are several callers that need to create a new block backend from
an existing BDS; make the task slightly easier with a common helper
routine.
Suggested-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200424190903.522087-2-eblake@redhat.com>
[mreitz: Set @ret only in error paths, see
https://lists.nongnu.org/archive/html/qemu-block/2020-04/msg01216.html]
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200428192648.749066-2-eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
The BDRV_REQ_ZERO_WRITE is currently implemented in a way that first the
image is possibly preallocated and then the zero flag is added to all
clusters. This means that a copy-on-write operation may be needed when
writing to these clusters, despite having used preallocation, negating
one of the major benefits of preallocation.
Instead, try to forward the BDRV_REQ_ZERO_WRITE to the protocol driver,
and if the protocol driver can ensure that the new area reads as zeros,
we can skip setting the zero flag in the qcow2 layer.
Unfortunately, the same approach doesn't work for metadata
preallocation, so we'll still set the zero flag there.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200424142701.67053-1-kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Alberto Garcia
Vladimir Sementsov-Ogievskiy
Andrey Shinkevich
Eric Blake
Markus Armbruster
Maxim Levitsky
Max Reitz
Denis Plotnikov
Kevin Wolf