Main excitement here is Peter Zijlstra's lockless rbtree optimization to
speed module address lookup. He found some abusers of the module lock
doing that too.
A little bit of parameter work here too; including Dan Streetman's breaking
up the big param mutex so writing a parameter can load another module (yeah,
really). Unfortunately that broke the usual suspects, !CONFIG_MODULES and
!CONFIG_SYSFS, so those fixes were appended too.
Cheers,
Rusty.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=tWPQ
-----END PGP SIGNATURE-----
Merge tag 'modules-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux
Pull module updates from Rusty Russell:
"Main excitement here is Peter Zijlstra's lockless rbtree optimization
to speed module address lookup. He found some abusers of the module
lock doing that too.
A little bit of parameter work here too; including Dan Streetman's
breaking up the big param mutex so writing a parameter can load
another module (yeah, really). Unfortunately that broke the usual
suspects, !CONFIG_MODULES and !CONFIG_SYSFS, so those fixes were
appended too"
* tag 'modules-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux: (26 commits)
modules: only use mod->param_lock if CONFIG_MODULES
param: fix module param locks when !CONFIG_SYSFS.
rcu: merge fix for Convert ACCESS_ONCE() to READ_ONCE() and WRITE_ONCE()
module: add per-module param_lock
module: make perm const
params: suppress unused variable error, warn once just in case code changes.
modules: clarify CONFIG_MODULE_COMPRESS help, suggest 'N'.
kernel/module.c: avoid ifdefs for sig_enforce declaration
kernel/workqueue.c: remove ifdefs over wq_power_efficient
kernel/params.c: export param_ops_bool_enable_only
kernel/params.c: generalize bool_enable_only
kernel/module.c: use generic module param operaters for sig_enforce
kernel/params: constify struct kernel_param_ops uses
sysfs: tightened sysfs permission checks
module: Rework module_addr_{min,max}
module: Use __module_address() for module_address_lookup()
module: Make the mod_tree stuff conditional on PERF_EVENTS || TRACING
module: Optimize __module_address() using a latched RB-tree
rbtree: Implement generic latch_tree
seqlock: Introduce raw_read_seqcount_latch()
...
This patch cleans up the manual device_create_file() or
class_create_file() calls by replacing with static attribute groups.
It simplifies the code and also avoids the possible races between the
device/class registration and sysfs creations.
For the simplification, also make ubi_class a static instance with
initializers, too.
Amend a bit by Hujianyang.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Tested-by: Sheng Yong <shengyong1@huawei.com>
Signed-off-by: hujianyang <hujianyang@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
During fastmap attaching, check if a volume already exists when adding
the volume to volume tree. NOTE that the issue cannot happen, only if
the on-flash fastmap data is modified.
Signed-off-by: Sheng Yong <shengyong1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
`vol' is a newly allocated value by kzalloc. Initialize it by assignment
instead of `+='.
Signed-off-by: Sheng Yong <shengyong1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
s/fmpl1/fmpl
s/fmpl2/fmpl_wl
Add "WL" to the error message when wrong WL pool magic number is detected.
Signed-off-by: Sheng Yong <shengyong1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
This patch makes ubiblock devices have minor numbers beginning from
0, allocated dynamically independently of the ubi device/volume
number. This property becomes useful because, on 32-bit architectures
with LFS turned off in a userspace program, device minor numbers
over 8 bits cause stat to return -EOVERFLOW. If the device number is
high (>1) due to multiple MTD partitions, such an overflow will occur.
While enabling LFS is clearly a nicer solution, it's often difficult
to turn on in practice globally as many widely distributed packages
don't work with LFS on.
Other storage systems have their own workarounds, with SCSI making
multiple device majors and MMC having a config option for the number
of partitions per device. A completely dynamic minor numbering is
simpler than these. It is unlikely that anyone is depending on a
static minor number since the major is dynamic anyway. In addition,
ubiblock is still relatively new, so now is the time to make such
changes.
Signed-off-by: Dan Ehrenberg <dehrenberg@chromium.org>
Acked-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Richard Weinberger <richard@nod.at>
Most code already uses consts for the struct kernel_param_ops,
sweep the kernel for the last offending stragglers. Other than
include/linux/moduleparam.h and kernel/params.c all other changes
were generated with the following Coccinelle SmPL patch. Merge
conflicts between trees can be handled with Coccinelle.
In the future git could get Coccinelle merge support to deal with
patch --> fail --> grammar --> Coccinelle --> new patch conflicts
automatically for us on patches where the grammar is available and
the patch is of high confidence. Consider this a feature request.
Test compiled on x86_64 against:
* allnoconfig
* allmodconfig
* allyesconfig
@ const_found @
identifier ops;
@@
const struct kernel_param_ops ops = {
};
@ const_not_found depends on !const_found @
identifier ops;
@@
-struct kernel_param_ops ops = {
+const struct kernel_param_ops ops = {
};
Generated-by: Coccinelle SmPL
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Junio C Hamano <gitster@pobox.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: cocci@systeme.lip6.fr
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Block drivers are responsible for calling flush_dcache_page() on each
BIO request. This operation keeps the I$ coherent with the D$ on
architectures that don't have hardware coherency support. Without this
flush, random crashes are seen when executing user programs from an ext4
filesystem backed by a ubiblock device.
This patch is based on the change implemented in commit 2d4dc890b5
("block: add helpers to run flush_dcache_page() against a bio and a
request's pages").
Fixes: 9d54c8a33e ("UBI: R/O block driver on top of UBI volumes")
Signed-off-by: Kevin Cernekee <cernekee@chromium.org>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@imgtec.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Pull fourth vfs update from Al Viro:
"d_inode() annotations from David Howells (sat in for-next since before
the beginning of merge window) + four assorted fixes"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
RCU pathwalk breakage when running into a symlink overmounting something
fix I_DIO_WAKEUP definition
direct-io: only inc/dec inode->i_dio_count for file systems
fs/9p: fix readdir()
VFS: assorted d_backing_inode() annotations
VFS: fs/inode.c helpers: d_inode() annotations
VFS: fs/cachefiles: d_backing_inode() annotations
VFS: fs library helpers: d_inode() annotations
VFS: assorted weird filesystems: d_inode() annotations
VFS: normal filesystems (and lustre): d_inode() annotations
VFS: security/: d_inode() annotations
VFS: security/: d_backing_inode() annotations
VFS: net/: d_inode() annotations
VFS: net/unix: d_backing_inode() annotations
VFS: kernel/: d_inode() annotations
VFS: audit: d_backing_inode() annotations
VFS: Fix up some ->d_inode accesses in the chelsio driver
VFS: Cachefiles should perform fs modifications on the top layer only
VFS: AF_UNIX sockets should call mknod on the top layer only
Emulate random power cuts by switching device to ro after a number of
writes to allow simple power cut testing with nand-sim.
Maximum and minimum number of successful writes before power cut and
what kind of writes (EC header, VID header or none) to interrupt
configurable via debugfs.
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
Signed-off-by: Richard Weinberger <richard@nod.at>
If we encounter an uncorrectable ECC error while scanning for the fastmap
UBI must not fail hard. Instead fall back to scanning mode.
Reported-by: Alexander Block <Alexander.Block@continental-corporation.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
This function was added to fastmap in a very early stage
to have paranoid assertions.
With the current fastmap implementation this assert will never
trigger as fastmap PEBs are not seen by the WL sub-system.
Remove it to save us some CPU cycles.
Signed-off-by: Richard Weinberger <richard@nod.at>
Fastmap need access to various WL data structures as
fastmap tightly depends on WL.
To make the access less invasive add accessor functions.
Signed-off-by: Richard Weinberger <richard@nod.at>
Fastmap is tightly connected to the WL sub-system, many fastmap-specific
functionslive in wl.c.
To get rid of most #ifdefs in wl.c move this functions into a new file
and include it into wl.c
Signed-off-by: Richard Weinberger <richard@nod.at>
If fm_debug is set fastmap debugging is enabled by default.
This is useful if one wants to debug fastmap on an UBI device
with serves the rootfs.
The the UBI attach mechanism runs long before debugfs can be mounted
and chk_fastmap set.
Signed-off-by: Richard Weinberger <richard@nod.at>
This self check allows Fastmap to detect absent PEBs while
writing a new fastmap to the MTD device.
It will help to find implementation issues in Fastmap.
Signed-off-by: Richard Weinberger <richard@nod.at>
In some error paths the WL sub-system gives up on a PEB
and frees it's ubi_wl_entry struct but does not set
the entry in ubi->lookuptbl to NULL.
Fastmap can stumble over such a stale pointer as it uses
ubi->lookuptbl to find all PEBs.
Fix this by introducing a new helper function which free()s
a WL entry and removes the reference from the lookup table.
Signed-off-by: Richard Weinberger <richard@nod.at>
Don't update the fastmap upon detach if fastmap checking is enabled.
This is poor men's power cut testing feature. :-)
Signed-off-by: Richard Weinberger <richard@nod.at>
Using this debugfs knob fastmap self checks can be controlled.
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Tanya Brokhman <tlinder@codeaurora.org>
If UBI is unable to write the fastmap to the device
we have make sure that upon next attach UBI will fall
back to scanning mode.
In case we cannot ensure that they only thing we can do
is falling back to read-only mode.
The current error handling code is not powercut proof.
It could happen that a powercut while invalidating would
lead to a state where an too old fastmap could be used upon
attach.
This patch addresses the issue by writing a fake fastmap
super block to a fresh PEB instead of reerasing the existing one.
The fake fastmap super block will UBI case to do a full scan.
Signed-off-by: Richard Weinberger <richard@nod.at>
The current code assumes that each fastmap has the same amount of PEBs.
So far this is true but will change soon.
Signed-off-by: Richard Weinberger <richard@nod.at>
a) Rename ubi->fm_sem to ubi->fm_eba_sem as this semaphore
protects EBA changes.
b) Turn ubi->fm_mutex into a rw semaphore. It will still serialize
fastmap writes but also ensures that ubi_wl_put_peb() is not
interrupted by a fastmap write. We use a rw semaphore to allow
ubi_wl_put_peb() still to be executed in parallel if no fastmap
write is happening.
Signed-off-by: Richard Weinberger <richard@nod.at>
ubi_wl_get_peb() returns a fresh PEB which can be used by
user of UBI. Due to the pool logic fastmap will correctly
map this PEB upon attach time because it will be scanned.
If a new fastmap is written (due to heavy parallel io)
while the before the fresh PEB is assigned to the EBA table
it will not be scanned as it is no longer in the pool.
So, the race window exists between ubi_wl_get_peb()
and the EBA table assignment.
We have to make sure that no new fastmap can be written
while that.
To ensure that ubi_wl_get_peb() will grab ubi->fm_sem in read mode
and the user of ubi_wl_get_peb() has to release it after the PEB
got assigned.
Signed-off-by: Richard Weinberger <richard@nod.at>
Don't use a fixed size for the WL pool.
Make it instead 50% of the user pool.
We don't make it 100% as it is not as heavily used as the user pool.
Signed-off-by: Richard Weinberger <richard@nod.at>
This logic is in vain as we treat protected PEBs also as used, so this
case must not happen.
If a PEB is found which is in the EBA table but not known as used
has to be issued as fatal error.
Signed-off-by: Richard Weinberger <richard@nod.at>
It is legal to have PEBs left in the used list.
This can happen if UBI copies a PEB and a powercut happens
between writing a new fastmap and adding this PEB into the EBA table.
In this case the old PEB will be used.
Signed-off-by: Richard Weinberger <richard@nod.at>
This function a) requests a new PEB, b) writes data to it,
c) returns the old PEB and d) registers the new PEB in the EBA table.
For the non-fastmap case this works perfectly fine and is powercut safe.
Is fastmap enabled this can lead to issues.
If a new fastmap is written between a) and c) the freshly requested PEB
is no longer in a pool and will not be scanned upon attaching.
If now a powercut happens between c) and d) the freshly requested PEB
will not be scanned and the old one got already scheduled for erase.
After attaching the EBA table will point to a erased PEB.
Fix this issue by swapping steps c) and d).
Signed-off-by: Richard Weinberger <richard@nod.at>
There is always exactly one ubi_attach_info object allocated,
therefore we don't have to care about the name.
Signed-off-by: Richard Weinberger <richard@nod.at>
There is no need to switch to ro mode if ubi_update_fastmap() fails.
Also get rid of the ifdef.
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Tanya Brokhman <tlinder@codeaurora.org>
...such that we can implement NOP variants of some functions.
This will help to reduce fastmap specific ifdefs in other c files.
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Tanya Brokhman <tlinder@codeaurora.org>
If ubi_update_fastmap() fails notify the user.
This is not a hard error as ubi_update_fastmap() makes sure that upon failure
the current on-flash fastmap will no be used upon next UBI attach.
Signed-off-by: Richard Weinberger <richard@nod.at>
Add a ubi_fastmap_close() to free all resources used by fastmap
at WL shutdown.
Signed-off-by: Richard Weinberger <richard@nod.at>
Tested-by: Guido Martínez <guido@vanguardiasur.com.ar>
Reviewed-by: Guido Martínez <guido@vanguardiasur.com.ar>
There is no need to allocate new ones every time, we can reuse
the existing ones.
This makes the code cleaner and more easy to follow.
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Tanya Brokhman <tlinder@codeaurora.org>
Reviewed-by: Guido Martínez <guido@vanguardiasur.com.ar>
Currently ubi_refill_pools() first fills the first and then
the second one.
If only very few free PEBs are available the second pool can get
zero PEBs.
Change ubi_refill_pools() to distribute free PEBs fair between
all pools.
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Guido Martínez <guido@vanguardiasur.com.ar>
Make it two functions, wl_get_wle() and wl_get_peb().
wl_get_peb() works exactly like __wl_get_peb() but wl_get_wle()
does not call produce_free_peb().
While refilling the fastmap user pool we cannot release ubi->wl_lock
as produce_free_peb() does.
Hence the fastmap logic uses now wl_get_wle().
Signed-off-by: Richard Weinberger <richard@nod.at>
