If btrfs_get_path_rootid fails, prints an error message but continues.
Replace with a helper that returns an error.
Signed-off-by: David Sterba <dsterba@suse.com>
Since new "receive --dump" has better output and structure, it's time
to remove old and function-weak send-test tool.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Automated build checks for the devel branch and integration to coverity,
using travis.org. Roman's initial .yml version has been adjusted for
btrfs-progs needs.
Author: Roman Lebedev <lebedev.ri@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
While most guys are using ctags and cscope with vim, new completion tool
like vim-clang_completion is gaining its popularity, due to its compiler
level accuracy simpleness to use.
Since ctags and cscope are already in .gitignore, I see no reason to
reject .clang_complete.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add test case to check the basic function of --clear-space-cache.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Kernel clear_cache mount option will only rebuild free space cache if
the used space of that chunk has changed.
So it won't ensure any corrupted free space cache get cleared.
So add a new option "--clear-space-cache v1|v2" to btrfsck, to
completely wipe out free space cache.
So kernel won't complain again.
Reported-by: Ivan P <chrnosphered@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ adjusted error messages, doc wording changes ]
Signed-off-by: David Sterba <dsterba@suse.com>
While performing an fsck, an assertion failure occurs because of reusing path in a loop.
ctree.c:1112: btrfs_search_slot: Warning: assertion `p->nodes[0] != NULL` failed, value 0
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 854437ca(btrfs-progs: extent-tree: avoid allocating tree block
that crosses stripe boundary) introduces check for logical bytenr not
crossing stripe boundary.
However that check is not completely correct.
It only checks if the logical bytenr and length agaist absolute logical
offset.
That's to say, it only check if a tree block lies in 64K logical stripe.
But in fact, it's possible a block group starts at bytenr unaligned with
64K, just like the following case.
Then btrfsck will give false alert.
0 32K 64K 96K 128K 160K ...
|--------------- Block group A ---------------------
|<-----TB 32K------>|
|/Scrub stripe unit/|
| WRONG UNIT |
In that case, TB(tree block) at bytenr 32K in fact fits into the kernel
scrub stripe unit.
But doesn't fit into the pure logical 64K stripe.
Fix check_crossing_stripes() to compare bytenr to block group start, not
to absolute logical bytenr.
Reported-by: Jussi Kansanen <jussi.kansanen@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The build fails if ioctl.h is included without kerncompat.h which
defines BUILD_ASSERT, as reported. The proposed patches add kerncompat.h
but I don't want to pull that header via ioctl.h, so the fix will
provide an empty macro instead.
Reported-by: Eric Sandeen <sandeen@redhat.com>
Reported-by: Sergei Trofimovich <slyfox@gentoo.org>
Reported-by: Mikhail Pukhlikov <cynede@gentoo.org>
Signed-off-by: David Sterba <dsterba@suse.com>
Move btrfs_extref_hash() from inode-item.c to hash.h,
so that the function can be called elsewhere.
Signed-off-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Some unnecessary codes are deleted.
- the setting of subvol is duplicated
- read only check was already done by the previous loop
Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
It seems like bad idea to use a library name (lblkid) within generic
function name. The currently used scanning library is implementation
detail and this detail should be hidden for rest of the code.
Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
The FIEMAP_EXTENT_SHARED fiemap flag was introduced in 2.6.33. If the
headers do not provide the definition, the build will fail. The support
of the fiemap sharing depends on the running kernel. There are still
systems with 2.6.32 kernel headers but running newer versions.
To support such environment, don't fail build, provide own defintion of
the structure and detect if there's an old kernel in use in the relevant
command (btrfs fi du).
Reported-by: Abhay Sachan <lkp.abhay@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Remove various BUG_ON in raid56 write routine, including:
1) Memory allocation error
Old codes allocates memory when code needs new memory in a loop, and
catch the error using BUG_ON().
New codes allocates memory in a allocation loop first, if any failure
is caught, freeing already allocated memories then throw -ENOMEM.
2) Write error
Change BUG_ON() to correct return value.
3) Validation check
Same as write error.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This allows us to put raid5 codes into that file other than creating a
new raid5.c.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For image with tree reloc tree, if its height is over level 2, the root
node's backref will point to itself. It's valid for kernel, but quota
verify code can't handle it and cause a infinite call, overflowing the
stack.
Add minimal image to reproduce the bug, as regression test.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ minor cleanups in test.sh ]
Signed-off-by: David Sterba <dsterba@suse.com>
Fixed by commit 7c646c538e btrfs-progs: qgroup: Fix regression leads
to corrupted qgroup status.
Add minimal test image for that fix.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ minor cleanups in test.sh ]
Signed-off-by: David Sterba <dsterba@suse.com>
For tree reloc tree whose level is >= 2, the root node's parent will
point to itself.
In this case it will make btrfsck overflow its stack and cause segfault.
While for tree reloc tree, it doesn't affect qgroup and kernel can
handle it well.
So add tree reloc tree check for qgroup-verify.c and fix the bug.
Test case will follow soon after I make a minimal image for it.
Current xz ziped image is still over 10M for a 512M fs.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In v4.8-rc1, gcc 5.3.1 gives following warning. Fixed it.
[CC] btrfs-image.o
btrfs-image.c: In function 'flush_pending':
btrfs-image.c:708:17: warning: 'start' may be used uninitialized in this function [-Wmaybe-uninitialized]
header->bytenr = cpu_to_le64(start);
^
btrfs-image.c:927:6: note: 'start' was declared here
u64 start;
^
Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add a new test case to check if btrfs-convert copies common inode flags
like append(a), immutable(i).
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ minor fixes ]
Signed-off-by: David Sterba <dsterba@suse.com>
Before this patch, btrfs-convert never copy ext* inode flags to
corresponding btrfs inode flags.
This makes common flags like APPEND/SYNC/SYNCDIR/IMMUTABLE not copied to
btrfs inode.
This patch introduces ext2_convert_inode_flags() function to handle the
convert, so btrfs-convert can copy as many inode flags as possible.
Reported-by: Lakshmipathi.G <lakshmipathi.g@gmail.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Before this patch, only 3 inode flags have readable string: NODATACOW,
NODATASUM, READONLY.
This patch will output all readable strings for remaining inode flags,
making debug-tree tool more handy.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ minor adjustments ]
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs-progs header lacks quite a lot inode flags.
Copy them from kernel for later use.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add another target that'll use default compiler options, eg. if we don't
know the bitness of the architecture or gcc does not recognize the
-m32/-m64 options.
Signed-off-by: David Sterba <dsterba@suse.com>
Enhance the test to verify ioctl uniqueness, compare the defined values
against the hardcoded expected values, and take care of the 32bit/64bit
compatibility workarounds.
Signed-off-by: David Sterba <dsterba@suse.com>
The ioctl value of SEND will be different on 32bit userspace and 64bit
kernel due to different pointer type width, that unfortunatelly made it
into the structure definition.
To maintain backward compatibility, we must do it in the 64bit->32bit
way, because we don't have the kernel side workardound like
SET_RECEIVED_SUBVOL has. Changing value of SEND would then break
existing users of the raw ioctl.
The compatibility structure and ioctl should not be used, exists for
documentation, and testing.
Signed-off-by: David Sterba <dsterba@suse.com>
The ioctl value of SET_RECEIVED_SUBVOL will be different on 32bit
userspace and 64bit kernel. This is transparently handled on the kernel
side. We now define the same structure so we can verify the ioctl value.
The defined structure and ioctl should not be used. It is exists only
for testing purposes, documenting the mess and as a warning for the
future.
Signed-off-by: David Sterba <dsterba@suse.com>
Temporarily make the build checks optional. The structure sizes could
change on arches due to alignment requirements or padding inserted into
the structures. We need more extensive tests to make sure we'd not break
ioctls.
Signed-off-by: David Sterba <dsterba@suse.com>
If we disable backtrace, btrfs-progs can't be compiled since we don't
have warning_trace defined.
Fix by move it out of #ifndef BTRFS_DISABLE_BACKTRACE block.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
[ move warning_trace to the right place ]
Signed-off-by: David Sterba <dsterba@suse.com>
All structures that become part of ioctl definition must never change
size. Add the build time checks for that.
Signed-off-by: David Sterba <dsterba@suse.com>
David Sterba