This pulls in the kernel's uapi/btrfs_tree.h, which now has all of the
on-disk definitions. Include this into ctree.h, and then yank out all
the duplicate code from ctree.h.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We want to keep this file locally as we want to be uptodate with
upstream, so we can build btrfs-progs regardless of which kernel is
currently installed. Sync this with the upstream version and put it in
kernel-shared/uapi to maintain some semblance of where this file comes
from.
There are some changes that need to be synced back to kernel. A local
definition of static_assert is used to avoid compilation problems on gcc
(< 9) due to mandatory 2nd parameter.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now that this is unused by these helpers and only used by the repair
related code we can remove this argument from the main helpers.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs check wants to be able to record corrupted extents if it finds any
bad blocks. This has been done directly inside of the
btrfs_check_leaf/btrfs_check_node helpers, but these are going to be
sync'ed from the kernel in the future. Add another helper and move the
corrupt block handling into this helper and keep it inside of the check
code.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This can be pulled out of the extent buffer that is passed in, drop the
fs_info argument from the function.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now that the libbtrfs stuff has it's own local copy of ctree.h and
ioctl.h, let's rename these qgroup struct members to match the kernel
names, this way it'll make it easier to sync the kernel code into
btrfs-progs.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
These helpers all do variations on the same thing, so add a helper to
just do the printf part, and a macro to handle the special prefix and
postfix, and then make the helpers just use the macro and new helper.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have different helpers for warning_on and error_on(), which take the
condition and then do the printf. However we can just check the
condition in the macro and call the normal warning or error helper, so
clean this usage up and delete the unneeded message helpers.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Everybody who calls print_trace wraps it around this check, move the
check instead to print_trace and remove the check from all the callers.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
While syncing messages.[ch] I had to back out the ASSERT() code in
kerncompat.h, which means we now rely on the kernel code for ASSERT().
In order to maintain some semblance of separation introduce UASSERT()
and use that in all the purely userspace code.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now that we're properly separated with libbtrfs/kerncompat.h and
include/kerncompat.h, go ahead and add the __init definition back so we
can have it available for the kernel synced files.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
While running make test-convert as a normal user I ran into this problem
where we do sudo find <blah> into a mktemp file that's created as the
normal user. This results in find getting a EPERM while trying to mess
with that temp file. Fix this by using $SUDO_HELPER for all the
tempfile manipulations so that root is the owner of everything, which
allows the convert tests to run as a normal user.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We're keeping a libbtrfs compatible kerncompat.h around to make it
easier to modify the rest of btrfs-progs. Unfortunately we also use
some of kernel-lib in libbtrfs, and those also include kerncompat.h.
Those getting included first means we'll pull include/kerncompat.h
instead of libbtrfs/kerncompat.h, which will mess things up.
Fix this by making sure we include our local copy of kerncompat.h first
before we include any other header in btrfs-progs.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The new test case would create a fs without free space tree, then
populate it, convert to free-space-tree feature, and make sure
everything is fine.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
From the very beginning of free-space-tree feature, we allow mount
option "space_cache=v2" to convert the filesystem to the new feature.
But this is not the proper practice for new features (no matter if it's
incompat or compat_ro).
This is already making the clear_cache/space_cache mount option more
complex.
Thus this patch introduces the proper way to enable free-space-tree, and
I hope one day we can deprecate the "space_cache=" mount option.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently check/clear-cache.c still uses a lot of global variables like
gfs_info and g_task_ctx, which are only implemented in check/main.c.
Since we have separated clear-cache code into its own c and header files,
we should not utilize those global variables.
Or it would lead to compiling errors for any tools which only has
check/clear-cache.o linked:
/usr/bin/ld: check/clear-cache.o: in function `clear_free_space_cache':
/home/adam/btrfs/btrfs-progs/check/clear-cache.c:46: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:56: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:67: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:73: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:84: undefined reference to `gfs_info'
/usr/bin/ld: check/clear-cache.o:/home/adam/btrfs/btrfs-progs/check/clear-cache.c:85: more undefined references to `gfs_info' follow
/usr/bin/ld: check/clear-cache.o: in function `check_space_cache':
/home/adam/btrfs/btrfs-progs/check/clear-cache.c:357: undefined reference to `g_task_ctx'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:357: undefined reference to `g_task_ctx'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:358: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:365: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:373: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:374: undefined reference to `gfs_info'
/usr/bin/ld: /home/adam/btrfs/btrfs-progs/check/clear-cache.c:382: undefined reference to `gfs_info'
/usr/bin/ld: check/clear-cache.o:/home/adam/btrfs/btrfs-progs/check/clear-cache.c:383: more undefined references to `gfs_info' follow
collect2: error: ld returned 1 exit status
This provides the basis for later clear cache usage out of check realm.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
I have a setup where null_blk is not a module but is built-in, so to check
if the kernel supports null_blk, use 'modinfo -n'.
Also fix a comment.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
-R option is deprecated since commit 4dbe66ca2f ("btrfs-progs: mkfs: make
-R|--runtime-features option deprecated"), migrate the test case to
follow the change.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When compiling the documents, we have the following warning:
~/btrfs-progs/Documentation/dev/On-disk-format.rst:369: WARNING: Bullet list ends without a blank line; unexpected unindent.
It's caused by a mismatched indent.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add helper scripts to build static binaries manually or clean old runs.
Requires the 'gh' command line tool and write access the git repository.
Signed-off-by: David Sterba <dsterba@suse.com>
Add new CI image target based on 32bit Alpine with musl. It's a copy of
musl and requires docker's buildx extension to properly select the
platform.
Signed-off-by: David Sterba <dsterba@suse.com>
Calculate checksums for static binaries that are published as release
artifacts. Do that when they're built in the workflow so the checksum
appears in the build log and also provide the artifacts of the checksums
so they can be published along the releases.
Signed-off-by: David Sterba <dsterba@suse.com>
Rename the options so they more accurately reflect what the command is
actually doing. The feature is enabled/disabled in the end but it's not
a simple on/off like for others, the conversion takes time.
Signed-off-by: David Sterba <dsterba@suse.com>
There has been no update since 2014 and it's likely broken due to many
new features added to btrfs since then.
Signed-off-by: David Sterba <dsterba@suse.com>
- "Wipe" in storage terms is often understood as some kind of secure
deletion. Use "remove" instead in order to indicate that the space
cache is fully removed (and not just cleared and then e.g.
automatically rebuild).
- The --clear-space-cache option for btrfs check actually clears the
whole space cache, just as documented. Thus move any documentation
about the clear_cache mount option not doing so for v1 to that.
Instead, refer to the mount option.
- Also note that when clear_cache is used with v1, the free space cache
for block groups that are modified gets always cleared, but rebuilt
only if nospace_cache is not used.
Pull-request: #571
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Signed-off-by: David Sterba <dsterba@suse.com>
Clarify active vs. inactive swapfile regarding snapshotting and further minor
improvements.
Issue: #608
Pull-request: #618
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently "btrfs filesystem du" auto-completes for directories only,
but it can also be used against files to determine shared vs exclusive
extents.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
The combination of --force and --repair was mistakenly documented as not
working on a mounted filesystem. Update help text, documentation and the
warning itself.
Issue: #503
Signed-off-by: David Sterba <dsterba@suse.com>
Running 'btrfs inspect tree-stats' on a mounted filesystem works though
it reads directly from block devices. This can lead to inconsistent
data, warnings, errors or or a crash. More checks could be added but at
least explain things in more detail.
Issue: #520
Signed-off-by: David Sterba <dsterba@suse.com>
Attempting to create a snapshot of subvolume with an active swapfile
prints the errno message corresponding to ETXTBSY but this is confusing
so change it to be more descriptove to:
ERROR: cannot snapshot '/hibernate': source subvolume contains an active swapfile (Text file busy)
Issue: #607
Signed-off-by: David Sterba <dsterba@suse.com>
The *64 interfaces, such as fstat64, off64_t, etc, are legacy interfaces
created at a time when 64-bit file support was still new. They are
generally exposed when defining a macro named _LARGEFILE64_SOURCE, as
e.g. the glibc docs[0] say.
The modern way to utilise largefile support, is to continue to use the
regular interfaces (off_t, fstat, ..), and define _FILE_OFFSET_BITS=64.
We already use the autoconf macro AC_SYS_LARGEFILE[1] which arranges this
and sets this macro for us. Therefore, we can utilise the non-64 names
without fear of breaking on 32-bit systems.
This fixes the build against musl libc, ever since musl dropped the
*64 compat from interfaces by default[2] just for _GNU_SOURCE, unless
_LARGEFILE64_SOURCE is defined. However, there are plans for a future
removal of the whole *64 header API, and that workaround (adding another
define) might cease to exist.
So, rename all *64 API use to the regular non-suffixed names. For
consistency, rename the internal functions that were *64 named
(lstat64_path, ..) too.
This should have no regressions on any platform.
[0]: https://www.gnu.org/software/libc/manual/html_node/Feature-Test-Macros.html#index-_005fLARGEFILE64_005fSOURCE
[1]: https://www.gnu.org/software/autoconf/manual/autoconf-2.67/html_node/System-Services.html
[2]: 25e6fee27f
Pull-request: #615
Signed-off-by: psykose <alice@ayaya.dev>
Signed-off-by: David Sterba <dsterba@suse.com>
Executing the script inside the directories as './test.sh' is not
supposed to work but could happen accidentally. With an exit after
attempting to source the we can fix that.
Signed-off-by: David Sterba <dsterba@suse.com>
We've changed mkfs defaults in 5.15 so it's not necessary to test the
features separately with convert. Instead use only defaults and add
other features that can be selected independently.
Signed-off-by: David Sterba <dsterba@suse.com>
[BUG]
There is an internal bug report that, after mkfs.btrfs there is a chance
that no /dev/disk/by-uuid/<uuid> symlink is not created at all.
[CAUSE]
That uuid symlink is created by udev, which listens to inotify
IN_CLOSE_WRITE events from all block devices.
After such IN_CLOSE_WRITE event is triggered, udev would *disable*
inotify for that block device, and do a blkid scan on it.
After the blkid scan is done, re-enables the inotify listening.
This means normally mkfs tools should open the fd, do all the writes,
and close the fd after everything is done.
But unfortunately for mkfs.btrfs, it's not the case, we have a lot of
phases separated by different close() calls:
open_ctree() would open fds of each involved device
and close them at close_ctree()
Only after close_ctree() we have a valid superblock -\
|
|<------- A -------->|<--------- B --------->|<------- C ------->|
| |
| `- open a new fd for make_btrfs()
| and close it before open_ctree()
| The device contains invalid sb.
|
`- open a new fd for each device, then call
btrfs_prepare_device(), then close the fd.
The device would contain no valid superblock.
If at the close() of phase A udev event is triggered, while doing udev
scan we go into phase C (but before the new valid super blocks written),
udev would only see no superblock or invalid superblock.
Then phase C finished, udev resumes its inotify listening, but at this
time mkfs is finished, while udev only sees the premature data from
phase A, and misses the IN_CLOSE_WRITE events from phase C.
[FIX]
Instead of opening and closing a new fd for each device, re-use the fd
opened during prepare_one_device(), and close all the fds until
close_ctree() is called.
By this, although we may still have race between close_ctree() and
explicit close() calls, at least udev can always see the properly
written super blocks.
To compensate the change, some extra cleanups are made:
- Do not touch @device_count
Which makes later prepare_ctx iteration much easier.
- Remove top-level @fd variable
Instead go with prepare_ctx[i].fd.
- Do not open with O_RDWR in test_dev_for_mkfs()
as test_dev_for_mkfs() would close the fd, if we go O_RDWR, it can
cause the udev race.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
With previous btrfstune support to convert to block-group-tree, it has
implemented most of the infrastructure for bi-directional conversion.
This patch will implement the remaining conversion support to go back to
extent tree.
The modification includes:
- New convert_to_extent_tree() function in btrfstune.c
It's almost the same as convert_to_bg_tree(), but with small changes:
* No need to set extra features like NO_HOLES/FST.
* Need to delete the block group tree when everything finished.
- Update btrfs_delete_and_free_root() to handle non-global roots
Currently the function can only accepts global roots (extent/csum/free
space trees)
If we pass a non-global root into the function, we will screw up
global_roots_tree and crash.
Since we're going to use btrfs_delete_and_free_root() to free block
group tree which is not a global tree, this is needed.
- New handling for half converted fs in get_last_converted_bg()
There are two cases need to be handled:
* The bg tree is already empty
We need to grab the first bg in extent tree.
Or at conversion function we will fail at grabbing the first bg.
* The bg tree is not empty
Then we need to grab the last bg in extent tree.
- Extra root switching in involved functions. This involves:
* read_converting_block_groups()
* insert_block_group_item()
* update_block_group_item()
We just need to update our target root according to the current
compat_ro and super flags.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The free tier of travis-ci has ended on 05/2021, remove the scripts.
Nowadays zstd has made it to all distros so we don't need to build
it from sources, reiserfs support is fading out so we don't need to
build it either. The integration with coverity scan needs to be done
again. The rest has been moved to new github actions CI.
Signed-off-by: David Sterba <dsterba@suse.com>
