If we know we've already locked the bytes, don't do it again; similarly
don't unlock a byte if we haven't locked it. This doesn't change the
behavior, but fixes a corner case explained below.
Libvirt had an error handling bug that an image can get its (ownership,
file mode, SELinux) permissions ch
v5: Address Max's comments (Thanks for reviewing):
- Clean up after test done.
- Add rev-by to patch 1 and 2.
v4: Fix test on systems without OFD. [Patchew]
The first patch reduces chances of QEMU crash in unusual (but not unlikely)
cases especially when used by Libvirt (see commit messag
The lock_fd field is not strictly necessary because transferring locked
bytes from old fd to the new one shouldn't fail anyway. This spares the
user one fd per image.
Signed-off-by: Fam Zheng
Reviewed-by: Max Reitz
---
block/file-posix.c | 37 +
1 file change
Signed-off-by: Fam Zheng
---
tests/Makefile.include | 2 +
tests/test-image-locking.c | 157 +
2 files changed, 159 insertions(+)
create mode 100644 tests/test-image-locking.c
diff --git a/tests/Makefile.include b/tests/Makefile.include
index 7a3059bf6c
On Tue, Oct 09, 2018 at 21:35:20 +0200, Kevin Wolf wrote:
> Peter, would this provide what libvirt urgently needs for backing files
> vs. the commit block job?
This looks fine for us with one exception. I'd prefer if the curl driver
implemented this too so that we don't have to hardcode the readon
This doesn't have any practical effect at the moment because the
values of BDRV_SECTOR_SIZE, QCRYPTO_BLOCK_LUKS_SECTOR_SIZE and
QCRYPTO_BLOCK_QCOW_SECTOR_SIZE are all the same (512 bytes), but
future encryption methods could have different requirements.
Signed-off-by: Alberto Garcia
---
block/qc
On Thu, Oct 11, 2018 at 01:58:02PM +0300, Alberto Garcia wrote:
> This doesn't have any practical effect at the moment because the
> values of BDRV_SECTOR_SIZE, QCRYPTO_BLOCK_LUKS_SECTOR_SIZE and
> QCRYPTO_BLOCK_QCOW_SECTOR_SIZE are all the same (512 bytes), but
> future encryption methods could ha
Am 10.10.2018 um 23:31 hat Philippe Mathieu-Daudé geschrieben:
> Both Python 2.7 and 3 support the same io.StringIO to
> handle unicode strings.
>
> Use the common form to use indistinctly Python 2.7 or 3.
>
> http://python-future.org/compatible_idioms.html#stringio
>
> This fixes running tests
On 11/10/2018 13:33, Kevin Wolf wrote:
> 045 1s ...[13:31:47] [13:31:47] [failed, exit status 1] - output
> mismatch (see 045.out.bad)
> --- /home/kwolf/source/qemu/tests/qemu-iotests/045.out 2017-01-24
> 14:49:48.0 +0100
> +++ /home/kwolf/source/qemu/tests/qemu-iotests/045.out.b
On Sat 06 Oct 2018 01:40:16 AM CEST, Max Reitz wrote:
> Currently, BlockDriver.bdrv_refresh_filename() is supposed to both
> refresh the filename (BDS.exact_filename) and set BDS.full_open_options.
> Now that we have generic code in the central bdrv_refresh_filename() for
> creating BDS.full_open_o
Paolo Bonzini writes:
> On 11/10/2018 13:33, Kevin Wolf wrote:
>> 045 1s ...[13:31:47] [13:31:47] [failed, exit status 1] - output
>> mismatch (see 045.out.bad)
>> --- /home/kwolf/source/qemu/tests/qemu-iotests/045.out 2017-01-24
>> 14:49:48.0 +0100
>> +++ /home/kwolf/source/qe
On Sat 06 Oct 2018 01:40:23 AM CEST, Max Reitz wrote:
> Signed-off-by: Max Reitz
Reviewed-by: Alberto Garcia
Berto
On 11/10/2018 16:01, Markus Armbruster wrote:
> Paolo Bonzini writes:
>
>> On 11/10/2018 13:33, Kevin Wolf wrote:
>>> 045 1s ...[13:31:47] [13:31:47] [failed, exit status 1] - output
>>> mismatch (see 045.out.bad)
>>> --- /home/kwolf/source/qemu/tests/qemu-iotests/045.out 2017-01-24
>>
Am 11.10.2018 um 16:01 hat Markus Armbruster geschrieben:
> Paolo Bonzini writes:
>
> > On 11/10/2018 13:33, Kevin Wolf wrote:
> >> 045 1s ...[13:31:47] [13:31:47] [failed, exit status 1] - output
> >> mismatch (see 045.out.bad)
> >> --- /home/kwolf/source/qemu/tests/qemu-iotests/045.out
qcow2_inc_refcounts_imrt() (through realloc_refcount_array()) can eat
an unpredictable amount of memory on corrupted table entries, which are
referencing regions far beyond the end of file.
Prevent this, by skipping such regions from further processing.
Signed-off-by: Vladimir Sementsov-Ogievskiy
Separate offset and size of compressed cluster.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2-refcount.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index afaa1a1409..23b105b43b 100644
--- a/block
Misaligned entries will lead to fatal qcow2 driver corruption on read
or write to corresponding offset, so there is no sense to take them
into account.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2-refcount.c | 25 +
1 file changed, 13 insertions(+), 12 dele
Increase corruptions_fixed only after successful fix.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2-refcount.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index 3c539f02e5..b453d87a3f 100644
--- a/block
Hi all!
v3:
01: s/l2_fixed_entries/l2_dirty/
02: add comment, s/>/>=
03: s/BDRV_SECTOR_SIZE/512
04: add Max's r-b
05: new
06-08: mostly rewritten
v2:
02, 06: check bdrv_getlength error return code
v1:
We've faced the following problem: after host fs corruption, vm images
becomes invalid. And w
Rewrite corrupted L2 table entry, which reference space out of
underlying file.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2-refcount.c | 66 +-
1 file changed, 65 insertions(+), 1 deletion(-)
diff --git a/block/qcow2-refcount.c b/block/qco
Split entry repairing to separate function, to be reused later in
check_refcounts_l2, prepare the whole pipeline for adding more
corruption types which may be repaired.
Note: entry in in-memory l2 table (local variable in
check_refcounts_l2) is not updated after this patch.
Signed-off-by: Vladimi
Reduce number of structures ignored in overlap check: when checking
active table ignore active tables, when checking inactive table ignore
inactive ones.
Signed-off-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Max Reitz
---
block/qcow2-refcount.c | 16 +---
1 file changed, 9 insert
Zero out corrupted L1 table entry, which reference L2 table out of
underlying file.
Zero L1 table entry means that "the L2 table and all clusters described
by this L2 table are unallocated."
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2-refcount.c | 26 ++
1
ping
17.09.2018 17:57, Vladimir Sementsov-Ogievskiy wrote:
> Hi all.
>
> 1. bdrv_dirty_iter_next_area don't use hbitmap_next_zero and uses
> inefficient loop instead. Let's improve it.
>
> 2. bdrv_dirty_iter_next_area don't handle unaligned offset and
> max_offset correctly. I'm not sure that it i
ping
19.09.2018 15:43, Vladimir Sementsov-Ogievskiy wrote:
> Hi all!
>
> We finally implemented fleecing scheme, create a test case for it (222)
> and even made it safe (f8d59dfb40).
>
> However, fleecing is used inside block-replication long ago, and to make
> the scheme safe it uses direct inter
Am 11.10.2018 um 12:58 hat Alberto Garcia geschrieben:
> This doesn't have any practical effect at the moment because the
> values of BDRV_SECTOR_SIZE, QCRYPTO_BLOCK_LUKS_SECTOR_SIZE and
> QCRYPTO_BLOCK_QCOW_SECTOR_SIZE are all the same (512 bytes), but
> future encryption methods could have differ
On Thu, Oct 11, 2018 at 04:56:04PM +0200, Kevin Wolf wrote:
> Am 11.10.2018 um 16:01 hat Markus Armbruster geschrieben:
[...]
> > I think we need to invest more resources in Python 2/3 compatiblity, or
> > else we'll miss our hard deadline of January 1, 2020.
>
> Did we decide until when we want t
27 matches
Mail list logo
