During the nocow writeback path, we currently iterate the rbtree of block
groups twice: once for checking if the target block group is RO with the
call to btrfs_extent_readonly()), and once again for getting a nocow
reference on the block group with a call to btrfs_inc_nocow_writers().
Since btrfs_inc_nocow_writers() already returns false when the target
block group is RO, remove the call to btrfs_extent_readonly(). Not only
we avoid searching the blocks group rbtree twice, it also helps reduce
contention on the lock that protects it (specially since it is a spin
lock and not a read-write lock). That may make a noticeable difference
on very large filesystems, with thousands of allocated block groups.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
During allocation the allocator will try to allocate an extent using
cluster policy. Once the current cluster is exhausted it will remove the
entry under btrfs_free_cluster::lock and subsequently acquire
btrfs_free_space_ctl::tree_lock to dispose of the already-deleted entry
and adjust btrfs_free_space_ctl::total_bitmap. This poses a problem
because there exists a race condition between removing the entry under
one lock and doing the necessary accounting holding a different lock
since extent freeing only uses the 2nd lock. This can result in the
following situation:
T1: T2:
btrfs_alloc_from_cluster insert_into_bitmap <holds tree_lock>
if (entry->bytes == 0) if (block_group && !list_empty(&block_group->cluster_list)) {
rb_erase(entry)
spin_unlock(&cluster->lock);
(total_bitmaps is still 4) spin_lock(&cluster->lock);
<doesn't find entry in cluster->root>
spin_lock(&ctl->tree_lock); <goes to new_bitmap label, adds
<blocked since T2 holds tree_lock> <a new entry and calls add_new_bitmap>
recalculate_thresholds <crashes,
due to total_bitmaps
becoming 5 and triggering
an ASSERT>
To fix this ensure that once depleted, the cluster entry is deleted when
both cluster lock and tree locks are held in the allocator (T1), this
ensures that even if there is a race with a concurrent
insert_into_bitmap call it will correctly find the entry in the cluster
and add the new space to it.
CC: <stable@vger.kernel.org> # 4.4+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently check_compressed_csum() completely relies on sectorsize ==
PAGE_SIZE to do checksum verification for compressed extents.
To make it subpage compatible, this patch will:
- Do extra calculation for the csum range
Since we have multiple sectors inside a page, we need to only hash
the range we want, not the full page anymore.
- Do sector-by-sector hash inside the page
With this patch and previous conversion on
btrfs_submit_compressed_read(), now we can read subpage compressed
extents properly, and do proper csum verification.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For compressed read, we always submit page read using page size. This
doesn't work well with subpage, as for subpage one page can contain
several sectors. Such submission will read range out of what we want,
and cause problems.
Thankfully to make it subpage compatible, we only need to change how the
last page of the compressed extent is read.
Instead of always adding a full page to the compressed read bio, if we're
at the last page, calculate the size using compressed length, so that we
only add part of the range into the compressed read bio.
Since we are here, also change the PAGE_SIZE used in
lookup_extent_mapping() to sectorsize.
This modification won't cause any functional change, as
lookup_extent_mapping() can handle the case where the search range is
larger than found extent range.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When a qstripe is required an extra page is allocated and mapped. There
were 3 problems:
1) There is no corresponding call of kunmap() for the qstripe page.
2) There is no reason to map the qstripe page more than once if the
number of bits set in rbio->dbitmap is greater than one.
3) There is no reason to map the parity page and unmap it each time
through the loop.
The page memory can continue to be reused with a single mapping on each
iteration by raid6_call.gen_syndrome() without remapping. So map the
page for the duration of the loop.
Similarly, improve the algorithm by mapping the parity page just 1 time.
Fixes: 5a6ac9eacb ("Btrfs, raid56: support parity scrub on raid56")
CC: stable@vger.kernel.org # 4.4.x: c17af96554: btrfs: raid56: simplify tracking of Q stripe presence
CC: stable@vger.kernel.org # 4.4.x
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There's an I/O error on fsync() in a detached loop device
if it has been previously attached.
The issue is write cache is enabled in the attach path in
loop_configure() but it isn't disabled in the detach path;
thus it remains enabled in the block device regardless of
whether it is attached or not.
Now fsync() can get an I/O request that will just be failed
later in loop_queue_rq() as device's state is not 'Lo_bound'.
So, disable write cache in the detach path.
Do so based on the queue flag, not the loop device flag for
read-only (used to enable) as the queue flag can be changed
via sysfs even on read-only loop devices (e.g., losetup -r.)
Test-case:
# DEV=/dev/loop7
# IMG=/tmp/image
# truncate --size 1M $IMG
# losetup $DEV $IMG
# losetup -d $DEV
Before:
# strace -e fsync parted -s $DEV print 2>&1 | grep fsync
fsync(3) = -1 EIO (Input/output error)
Warning: Error fsyncing/closing /dev/loop7: Input/output error
[ 982.529929] blk_update_request: I/O error, dev loop7, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0
After:
# strace -e fsync parted -s $DEV print 2>&1 | grep fsync
fsync(3) = 0
Co-developed-by: Eric Desrochers <eric.desrochers@canonical.com>
Signed-off-by: Eric Desrochers <eric.desrochers@canonical.com>
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Tested-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If duty cycle is first set to a value that is sufficiently high to
enable the output (e.g. 10000 ns) but then lowered to a value that
is quantized to zero (e.g. 1000 ns), the output is disabled as the
device cannot drive a constant zero (as expected).
However if the device is later re-initialized due to watchdog bite,
the output is re-enabled at the next-to-last duty cycle (10000 ns).
This is because the iqs620_pwm->out_en flag unconditionally tracks
state->enabled instead of what was actually written to the device.
To solve this problem, use one state variable that encodes all 257
states of the output (duty_scale) with 0 representing tri-state, 1
representing the minimum available duty cycle and 256 representing
100% duty cycle.
Signed-off-by: Jeff LaBundy <jeff@labundy.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
If state->duty_cycle is 0x100000000000000, the previous calculation of
duty_scale overflows and yields a duty cycle ratio of 0% instead of
100%. Fix this by clamping the requested duty cycle to the maximal
possible duty cycle first. This way it is possible to use a native
integer division instead of a (depending on the architecture) more
expensive 64bit division.
With this change in place duty_scale cannot be bigger than 256 which
allows to simplify the calculation of duty_val.
Fixes: 6f0841a819 ("pwm: Add support for Azoteq IQS620A PWM generator")
Tested-by: Jeff LaBundy <jeff@labundy.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
QUEUE_FLAG_POLL flag will be cleared when turning off 'io_poll', while
at that moment there may be IOs stuck in hw queue uncompleted. The
following polling routine won't help reap these IOs, since blk_poll()
will return immediately because of cleared QUEUE_FLAG_POLL flag. Thus
these IOs will hang until they finnaly time out. The hang out can be
observed by 'fio --engine=io_uring iodepth=1', while turning off
'io_poll' at the same time.
To fix this, freeze and flush the request queue first when turning off
'io_poll'.
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The commit f3bdc62fd8 ("blktrace: Provide event for request merging")
added the comment for blk_rq_merge() tracepoint. Remove the duplicate
word from the tracepoint documentation.
Fixes: f3bdc62fd8 ("blktrace: Provide event for request merging")
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The commit 881245dcff ("Add DocBook documentation for the block tracepoints.")
added the comment for blk_rq_issue() tracepoint. Remove the duplicate
word from the tracepoint documentation.
Fixes: 881245dcff ("Add DocBook documentation for the block tracepoints.")
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The last parameter for the function blk_fill_rwbs() was added in
5782138e47 ("tracing/events: convert block trace points to
TRACE_EVENT()") in order to signal read request and use of that parameter
was replaced with using switch case REQ_OP_READ with
1b9a9ab78b ("blktrace: use op accessors"), but the parameter was never
removed.
Remove the unused parameter and adjust the respective call sites.
Fixes: 1b9a9ab78b ("blktrace: use op accessors")
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit a1ce35fa49 ("block: remove dead elevator code") removed the last
callers of blk_pm_requeue_request(), blk_pm_add_request() and
blk_pm_put_request(). Hence remove the definitions of these functions.
Removing these functions removes all users of the struct request nr_pending
member. Hence also remove 'nr_pending'. Note: 'nr_pending' is no longer
used since commit 7cedffec8e ("block: Make blk_get_request() block for
non-PM requests while suspended").
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When CONFIG_EPOLL is not set/enabled, sys_oabi-compat.c has build
errors. Fix these by surrounding them with ifdef CONFIG_EPOLL/endif
and providing stubs for the "EPOLL is not set" case.
../arch/arm/kernel/sys_oabi-compat.c: In function 'sys_oabi_epoll_ctl':
../arch/arm/kernel/sys_oabi-compat.c:257:6: error: implicit declaration of function 'ep_op_has_event' [-Werror=implicit-function-declaration]
257 | if (ep_op_has_event(op) &&
| ^~~~~~~~~~~~~~~
../arch/arm/kernel/sys_oabi-compat.c:264:9: error: implicit declaration of function 'do_epoll_ctl'; did you mean 'sys_epoll_ctl'? [-Werror=implicit-function-declaration]
264 | return do_epoll_ctl(epfd, op, fd, &kernel, false);
| ^~~~~~~~~~~~
Fixes: c281634c86 ("ARM: compat: remove KERNEL_DS usage in sys_oabi_epoll_ctl()")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com> # from an lkp .config file
Cc: linux-arm-kernel@lists.infradead.org
Cc: Nicolas Pitre <nico@fluxnic.net>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: patches@armlinux.org.uk
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Pull clockevent fixes from Daniel Lezcano
- Fix harmless warning with the ixp4xx when the TIMER_OF option is not
selected (Arnd Bergmann)
- Make sure channel clock supply is enabled on sh_cmt (Geert Uytterhoeven)
- Fix compilation error when DEBUG is defined with the mxs_timer (Tom Rix)
Link: https://lore.kernel.org/r/ae3bcda6-5180-639d-6246-d2dfd271c3e7@linaro.org
Pull tpm updates from Jarkko Sakkinen:
"New features:
- Cr50 I2C TPM driver
- sysfs exports of PCR registers in TPM 2.0 chips
Bug fixes:
- bug fixes for tpm_tis driver, which had a racy wait for hardware
state change to be ready to send a command to the TPM chip. The bug
has existed already since 2006, but has only made itself known in
recent past. This is the same as the "last time" :-)
- Otherwise there's bunch of fixes for not as alarming regressions. I
think the list is about the same as last time, except I added fixes
for some disjoint bugs in trusted keys that I found some time ago"
* tag 'tpmdd-next-v5.12-rc1-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd:
KEYS: trusted: Reserve TPM for seal and unseal operations
KEYS: trusted: Fix migratable=1 failing
KEYS: trusted: Fix incorrect handling of tpm_get_random()
tpm/ppi: Constify static struct attribute_group
ABI: add sysfs description for tpm exports of PCR registers
tpm: add sysfs exports for all banks of PCR registers
keys: Update comment for restrict_link_by_key_or_keyring_chain
tpm: Remove tpm_dev_wq_lock
char: tpm: add i2c driver for cr50
tpm: Fix fall-through warnings for Clang
tpm_tis: Clean up locality release
tpm_tis: Fix check_locality for correct locality acquisition
Pull seccomp updates from Kees Cook:
"Two small seccomp updates.
This contains a fix for a build failure that went unnoticed for many
years, and a memory barrier correction:
- Fix a non-FILTER build failure for some architectures (Paul
Cercueil)
- Improve performance with correct memory barrier (wanghongzhe)"
* tag 'seccomp-v5.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
seccomp: Improve performace by optimizing rmb()
seccomp: Add missing return in non-void function
Pull smack updates from Casey Schaufler:
"Bounds checking for writes to smackfs interfaces"
* tag 'Smack-for-v5.12' of git://github.com/cschaufler/smack-next:
smackfs: restrict bytes count in smackfs write functions
Pull IMA updates from Mimi Zohar:
"New is IMA support for measuring kernel critical data, as per usual
based on policy. The first example measures the in memory SELinux
policy. The second example measures the kernel version.
In addition are four bug fixes to address memory leaks and a missing
'static' function declaration"
* tag 'integrity-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity:
integrity: Make function integrity_add_key() static
ima: Free IMA measurement buffer after kexec syscall
ima: Free IMA measurement buffer on error
IMA: Measure kernel version in early boot
selinux: include a consumer of the new IMA critical data hook
IMA: define a builtin critical data measurement policy
IMA: extend critical data hook to limit the measurement based on a label
IMA: limit critical data measurement based on a label
IMA: add policy rule to measure critical data
IMA: define a hook to measure kernel integrity critical data
IMA: add support to measure buffer data hash
IMA: generalize keyring specific measurement constructs
evm: Fix memleak in init_desc
Pull audit updates from Paul Moore:
"Three very trivial patches for audit this time"
* tag 'audit-pr-20210215' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit:
audit: Make audit_filter_syscall() return void
audit: Remove leftover reference to the audit_tasklet
kernel/audit: convert comma to semicolon
syzbot reported a warning which could cause shift-out-of-bounds issue.
Call Trace:
__dump_stack lib/dump_stack.c:79 [inline]
dump_stack+0x183/0x22e lib/dump_stack.c:120
ubsan_epilogue lib/ubsan.c:148 [inline]
__ubsan_handle_shift_out_of_bounds+0x432/0x4d0 lib/ubsan.c:395
exfat_read_boot_sector fs/exfat/super.c:471 [inline]
__exfat_fill_super fs/exfat/super.c:556 [inline]
exfat_fill_super+0x2acb/0x2d00 fs/exfat/super.c:624
get_tree_bdev+0x406/0x630 fs/super.c:1291
vfs_get_tree+0x86/0x270 fs/super.c:1496
do_new_mount fs/namespace.c:2881 [inline]
path_mount+0x1937/0x2c50 fs/namespace.c:3211
do_mount fs/namespace.c:3224 [inline]
__do_sys_mount fs/namespace.c:3432 [inline]
__se_sys_mount+0x2f9/0x3b0 fs/namespace.c:3409
do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
entry_SYSCALL_64_after_hwframe+0x44/0xa9
exfat specification describe sect_per_clus_bits field of boot sector
could be at most 25 - sect_size_bits and at least 0. And sect_size_bits
can also affect this calculation, It also needs validation.
This patch add validation for sect_per_clus_bits and sect_size_bits
field of boot sector.
Fixes: 719c1e1829 ("exfat: add super block operations")
Cc: stable@vger.kernel.org # v5.9+
Reported-by: syzbot+da4fe66aaadd3c2e2d1c@syzkaller.appspotmail.com
Reviewed-by: Sungjong Seo <sj1557.seo@samsung.com>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Pull selinux updates from Paul Moore:
"We've got a good handful of patches for SELinux this time around; with
everything passing the selinux-testsuite and applying cleanly to your
tree as of a few minutes ago. The highlights are:
- Add support for labeling anonymous inodes, and extend this new
support to userfaultfd.
- Fallback to SELinux genfs file labeling if the filesystem does not
have xattr support. This is useful for virtiofs which can vary in
its xattr support depending on the backing filesystem.
- Classify and handle MPTCP the same as TCP in SELinux.
- Ensure consistent behavior between inode_getxattr and
inode_listsecurity when the SELinux policy is not loaded. This
fixes a known problem with overlayfs.
- A couple of patches to prune some unused variables from the SELinux
code, mark private variables as static, and mark other variables as
__ro_after_init or __read_mostly"
* tag 'selinux-pr-20210215' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:
fs: anon_inodes: rephrase to appropriate kernel-doc
userfaultfd: use secure anon inodes for userfaultfd
selinux: teach SELinux about anonymous inodes
fs: add LSM-supporting anon-inode interface
security: add inode_init_security_anon() LSM hook
selinux: fall back to SECURITY_FS_USE_GENFS if no xattr support
selinux: mark selinux_xfrm_refcount as __read_mostly
selinux: mark some global variables __ro_after_init
selinux: make selinuxfs_mount static
selinux: drop the unnecessary aurule_callback variable
selinux: remove unused global variables
selinux: fix inconsistency between inode_getxattr and inode_listsecurity
selinux: handle MPTCP consistently with TCP
Pull tomoyo updates from Tetsuo Handa:
"Detect kernel thread correctly, and ignore harmless data race"
* tag 'tomoyo-pr-20210215' of git://git.osdn.net/gitroot/tomoyo/tomoyo-test1:
tomoyo: recognize kernel threads correctly
tomoyo: ignore data race while checking quota
We want to use this in io_uring proper as well, for the SQPOLL thread.
Rename it from fork_thread() to io_wq_fork_thread(), and make it
available through the io-wq.h header.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The async workers are siblings of the task itself, so by definition we
have all the state that we need. Remove any of the state grabbing that
we have, and requests flagging what they need.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of using regular kthread kernel threads, create kernel threads
that are like a real thread that the task would create. This ensures that
we get all the context that we need, without having to carry that state
around. This greatly reduces the code complexity, and the risk of missing
state for a given request type.
With the move away from kthread, we can also dump everything related to
assigned state to the new threads.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
PF_IO_WORKER are kernel threads too, but they aren't PF_KTHREAD in the
sense that we don't assign ->set_child_tid with our own structure. Just
ensure that every arch sets up the PF_IO_WORKER threads like kthreads
in the arch implementation of copy_thread().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't support attach anymore, so doesn't make sense to carry the
use_refs reference count. Get rid of it.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Moving towards making the io_wq per ring per task, so we can't really
share it between rings. Which is fine, since we've now dropped some
of that fat from it.
Retain compatibility with how attaching works, so that any attempt to
attach to an fd that doesn't exist, or isn't an io_uring fd, will fail
like it did before.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When the manager thread starts up, it creates a worker per node for
the given context. Just let these get created dynamically, like we do
for adding further workers.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
