| Age | Commit message (Collapse) | Author |
|---|
|
When we're cancelling a cow range, we don't always delete each extent
that we iterate, so we have to move icur backwards in the list to avoid
an infinite loop.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
We don't hold the ilock through the entire sequence of xfs_writepage_map
-> xfs_map_cow -> xfs_reflink_find_cow_mapping. This means that we can
race with another thread that is trying to clear the inode reflink flag,
with the result that the flag is set for the xfs_map_cow check but
cleared before we get to the assert in find_cow_mapping. When this
happens, we blow the assert even though everything is fine.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
If we try to reflink into a file with post-eof preallocations at an
offset well past the preallocations, we increase i_size as one would
expect. However, those allocations do not have page cache backing them,
so they won't get cleaned out on their own. This leads to asserts in
the collapse/insert range code and xfs_destroy_inode when they encounter
delalloc extents they weren't expecting to find.
Since there are plenty of other places where we dump those post-eof
blocks, do the same to the reflink destination file before we start
remapping extents. This was found by adding clonerange support to
fsstress and running it in write-only mode.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
Move the tracepoint in xfs_iext_insert to after the point where we've
inserted the extent because otherwise we report stale extent data in
the ftrace output.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
In e1a4e37cc7b665 ("xfs: try to avoid blowing out the transaction
reservation when bunmaping a shared extent"), we try to constrain the
amount of real extents we unmap from the data fork in a given call so
that we don't blow out transaction reservations.
However, not all bunmapi operations require a transaction -- if we're
only removing a delalloc extent, no transaction is needed, so we have to
code against that.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
While the failure handling in dvb_input_attach() has been improved lately
so any tuner failure won't result in demod driver modules with a
usecount > 0 anymore (thus requiring rmmod -f), there's still an issue
with stv090x and stv0910 based tuner modules, in that LNB driver attach
failures leave an attached demod frontend driver behind which have a
usecount of > 0 in this failure case, due to them not being detached/
released. Fix this by detaching the demod frontends if the LNB driver
fails.
Richard tested and verified the changes with STV0910 hardware, thus adding
his Tested-by.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Tested-by: Richard Scobie <rascobie@slingshot.co.nz>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
of an attribute
The new attribute leaf buffer is not held locked across the transaction
roll between the shortform->leaf modification and the addition of the
new entry. As a result, the attribute buffer modification being made is
not atomic from an operational perspective. Hence the AIL push can grab
it in the transient state of "just created" after the initial
transaction is rolled, because the buffer has been released. This leads
to xfs_attr3_leaf_verify() asserting that hdr.count is zero, treating
this as in-memory corruption, and shutting down the filesystem.
Darrick ported the original patch to 4.15 and reworked it use the
xfs_defer_bjoin helper and hold/join the buffer correctly across the
second transaction roll.
Signed-off-by: Alex Lyakas <alex@zadarastorage.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
In certain cases, defer_ops callers will lock a buffer and want to hold
the lock across transaction rolls. Similar to ijoined inodes, we want
to dirty & join the buffer with each transaction roll in defer_finish so
that afterwards the caller still owns the buffer lock and we haven't
inadvertently pinned the log.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
With CONFIG_KASAN enabled, we get a relatively large stack frame in one function
drivers/media/tuners/tda8290.c: In function 'tda8290_set_params':
drivers/media/tuners/tda8290.c:310:1: warning: the frame size of 1520 bytes is larger than 1024 bytes [-Wframe-larger-than=]
With CONFIG_KASAN_EXTRA this goes up to
drivers/media/tuners/tda8290.c: In function 'tda8290_set_params':
drivers/media/tuners/tda8290.c:310:1: error: the frame size of 3200 bytes is larger than 3072 bytes [-Werror=frame-larger-than=]
We can significantly reduce this by marking local arrays as 'static const', and
this should result in better compiled code for everyone.
[mchehab@s-opensource.com: fix a trivial merge conflict]
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The rawmidi also allows to obtaining the information via ioctl of ctl
API. It means that user can issue an ioctl to the rawmidi device even
when it's being removed as long as the control device is present.
Although the code has some protection via the global register_mutex,
its range is limited to the search of the corresponding rawmidi
object, and the mutex is already unlocked at accessing the rawmidi
object. This may lead to a use-after-free.
For avoiding it, this patch widens the application of register_mutex
to the whole snd_rawmidi_info_select() function. We have another
mutex per rawmidi object, but this operation isn't very hot path, so
it shouldn't matter from the performance POV.
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This is part of the uAPI. Add it to the documentation again,
and fix cross-references.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Acked-by: Sean Young <sean@mess.org>
|
|
CEC autorepeat is different than other protocols. Autorepeat is triggered
by the first repeated user control pressed CEC message, rather than a
fixed REP_DELAY.
This change also does away with the KEY_UP event directly after the first
KEY_DOWN event, which was used to stop autorepeat from starting.
See commit a9a249a2c997 ("media: cec: fix remote control passthrough")
for the original change.
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Another device with the 0xffdc device id, this one with 0x30 in the
config byte. Its an iMON VFD + iMON IR (it does not understand rc6).
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The pnp_irq() function returns -1 if an error occurs.
pnp_irq() error checking for zero is not correct.
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This patch changes the 32-bit time type (timeval) to the 64-bit one
(ktime_t), since 32-bit time types will break in the year 2038.
I use ktime_t instead of all uses of timeval in imon.c
This patch also changes do_gettimeofday() to ktime_get() accordingly,
since ktime_get returns a ktime_t, but do_gettimeofday returns a
struct timeval, and the other reason is that ktime_get() uses
the monotonic clock.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Once rc_unregister_device() has been called, no driver function
should be called.
This prevents some nasty race conditions with an ioctl calls
driver functions when the driver specific data has been freed.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This makes it possible for lircd to read from a lirc chardev, and not
keep it busy.
Note that this changes the default for timeout reports to on. lircd
already enables timeout reports when it opens a lirc device, leaving
them on until the next reboot.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This removes the need for include/media/lirc.h, which just includes
the uapi file.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Since removing the lirc kapi, ir-lirc-codec.c only contains lirc fops
so the file name is no longer correct. By moving its content into
lirc_dev.c the ugly extern struct lirc_fops is not longer needed,
and everything lirc related is in one file.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
If you try to store u64 in a kfifo (or a struct with u64 members),
then the buf member of __STRUCT_KFIFO_PTR will cause 4 bytes
padding due to alignment (note that struct __kfifo is 20 bytes
on 32 bit).
That in turn causes the __is_kfifo_ptr() to fail, which is caught
by kfifo_alloc(), which now returns EINVAL.
So, ensure that __is_kfifo_ptr() compares to the right structure.
Signed-off-by: Sean Young <sean@mess.org>
Acked-by: Stefani Seibold <stefani@seibold.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Now that the lirc interface supports scancodes, RC scancode devices
can also have a lirc device. The only receiving feature they will have
enabled is LIRC_CAN_REC_SCANCODE.
Note that CEC devices have no lirc device, since they can be controlled
from their /dev/cecN chardev.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Lirc supports a new mode which requires documentation.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The lirc device should get lirc repeats whether there is a keymap
match or not.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This implements LIRC_MODE_SCANCODE reading from the lirc device. The
scancode can be read from the input device too, but with this interface
you get the rc protocol, keycode, toggle and repeat status in addition
to just the scancode.
int main()
{
int fd, mode, rc;
fd = open("/dev/lirc0", O_RDWR);
mode = LIRC_MODE_SCANCODE;
if (ioctl(fd, LIRC_SET_REC_MODE, &mode)) {
// kernel too old or lirc does not support transmit
}
struct lirc_scancode scancode;
while (read(fd, &scancode, sizeof(scancode)) == sizeof(scancode)) {
printf("protocol:%d scancode:0x%x toggle:%d repeat:%d\n",
scancode.rc_proto, scancode.scancode,
!!(scancode.flags & LIRC_SCANCODE_FLAG_TOGGLE),
!!(scancode.flags & LIRC_SCANCODE_FLAG_REPEAT));
}
close(fd);
}
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
rc-core has replaced the lirc kapi many years ago, and now with the last
driver ported to rc-core, we can finally remove it.
Note this has no effect on userspace.
All future IR drivers should use the rc-core api.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This is a duplicate of rcdev->driver_name.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Replace the generic kernel lirc api with ones which use rc-core, further
reducing the lirc_dev members.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
If a lirc chardev is held open after a device is unplugged, rc_close()
will be called after rc_unregister_device(). The driver is not expecting
any calls at this point, and the iguanair driver causes an oops in
this scenario.
rc_open() can be called when the device is removed too, by calling open
on the chardev whilst the device is being removed.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This is done to further remove the lirc kernel api. Ensure that every
fops checks for this.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Since the only mode lirc devices can handle is raw IR, handle this
in a plain kfifo.
Remove lirc_buffer since this is no longer needed.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Calculate lirc features when necessary, and add LIRC_{S,G}ET_REC_MODE
cases to ir_lirc_ioctl.
This makes lirc_dev_fop_ioctl() unnecessary since all cases are
already handled by ir_lirc_ioctl().
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
For some IR protocols, some scancode values not valid, i.e. they're part
of a different protocol variant.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Ensure we reject an attempt to transmit invalid scancodes.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The lirc user interface exists as a raw decoder, which does not make
much sense for transmit-only devices.
In addition, we want to have lirc char devices for devices which do not
use raw IR, i.e. scancode only devices.
Note that rc-code, lirc_dev, ir-lirc-codec are now calling functions of
each other, so they've been merged into one module rc-core to avoid
circular dependencies.
Since ir-lirc-codec no longer exists as separate codec module, there is no
need for RC_DRIVER_IR_RAW_TX type drivers to call ir_raw_event_register().
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
When sending scancodes, load the encoder if we need it.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
If the lirc device supports it, set the carrier for the protocol.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This introduces a new lirc mode: scancode. Any device which can send raw IR
can now also send scancodes.
int main()
{
int mode, fd = open("/dev/lirc0", O_RDWR);
mode = LIRC_MODE_SCANCODE;
if (ioctl(fd, LIRC_SET_SEND_MODE, &mode)) {
// kernel too old or lirc does not support transmit
}
struct lirc_scancode scancode = {
.scancode = 0x1e3d,
.rc_proto = RC_PROTO_RC5,
};
write(fd, &scancode, sizeof(scancode));
close(fd);
}
The other fields of lirc_scancode must be set to 0.
Note that toggle (rc5, rc6) and repeats (nec) are not implemented. Nor is
there a method for holding down a key for a period.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
LIRCCODE is a lirc mode where a driver produces driver-dependent
codes for receive and transmit. No driver uses this any more. The
LIRC_GET_LENGTH ioctl was used for this mode only.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Now that lirc is no longer in the staging area, remove the entry.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The ir-kbd-i2c driver behaves like the lirc_zilog driver, except it can
send raw IR and receives scancodes rather than lirccodes.
The lirc_zilog driver only polls if the lirc chardev is opened;
similarly the ir-kbd-i2c driver only polls if the corresponding input
device is opened, or the lirc device.
Polling is disabled during IR transmission through the mutex.
The polling period is 402ms in the ir-kdb-i2c driver, and 260ms in the
lirc_zilog driver.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Currently, the SVE field in ID_AA64PFR0_EL1 is visible
unconditionally to userspace via the CPU ID register emulation,
irrespective of the kernel config. This means that if a kernel
configured with CONFIG_ARM64_SVE=n is run on SVE-capable hardware,
userspace will see SVE reported as present in the ID regs even
though the kernel forbids execution of SVE instructions.
This patch makes the exposure of the SVE field in ID_AA64PFR0_EL1
conditional on CONFIG_ARM64_SVE=y.
Since future architecture features are likely to encounter a
similar requirement, this patch adds a suitable helper macros for
use when declaring config-conditional ID register fields.
Fixes: 43994d824e84 ("arm64/sve: Detect SVE and activate runtime support")
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reported-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Cc: Suzuki Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
|
This is a fix for commit 329d88da4df9 ("[media] media: i2c: Don't export
ir-kbd-i2c module alias") that stopped the module from being loaded
automagically.
The problems described only affect the HD-PVR, so it should not affect
other hardware; also if the module happens to be loaded, the i2c IR
part of the HD-PVR will be enabled anyway.
Fixes: 329d88da4df9 ("[media] media: i2c: Don't export ir-kbd-i2c module alias")
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This code implements the transmitter which is currently implemented
in the staging lirc_zilog driver.
The new code does not need a signal database, iow. the
haup-ir-blaster.bin firmware file is no longer needed, and the driver
does not know anything about the keycodes in that file.
Instead, the new driver can send raw IR, but the hardware is limited
to few different lengths of pulse and spaces, so it is best to use
generated IR rather than recorded IR.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
These two devices ids are really just one device with multiple
addresses. Probing becomes much simpler if we simply fold this into
one i2c device with two address.
Note that this breaks the lirc_zilog driver, however we will teach
ir-kbd-i2c to do what lirc_zilog does in a later commit.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
The lirc_zilog driver only polls the device if the lirc chardev
is opened; do the same with the rc-core driver.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Use the dev_dbg dynamic infrastructure instead of rolling our own custom
debug logic.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
With the parent set for the rc device, the messages clearly state
that it is attached via i2c. The additional printk is unnecessary.
These are the old messages:
rc rc1: i2c IR (Hauppauge WinTV PVR-150 as /devices/virtual/rc/rc1
ir-kbd-i2c: i2c IR (Hauppauge WinTV PVR-150 detected at i2c-10/10-0071/ir0 [ivtv i2c driver #0]
Now we simply get:
rc rc1: Hauppauge WinTV PVR-150 as /devices/pci0000:00/0000:00:1e.0/0000:02:00.0/i2c-10/10-0071/rc/rc1
Note that we no longer copy the name. I've checked all call sites
to verfiy this is not a problem.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
When iterating through all endpoints using of_graph_get_next_endpoint(),
the refcount of the returned endpoint node is incremented and the refcount
of the node which is passed as previous endpoint is decremented.
So the caller doesn't need to call of_node_put() for each iterated node
except for error exit paths. Otherwise we get "OF: ERROR: Bad
of_node_put() on ..." messages.
Cc: Hyun Kwon <hyun.kwon@xilinx.com>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
This is odd to call 'pci_disable_device()' in an error path before a
coresponding successful 'pci_enable_device()'.
Return directly instead.
Fixes: 77e0be12100a ("V4L/DVB (4176): Bug-fix: Fix memory overflow")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
|
Currently a user can pass in an unsanitized slot number which
will lead to and out of range index into ca->slot_info. Fix this
by checking that the slot number is no more than the allowed
maximum number of slots. Seems that this bug has been in the driver
forever.
Detected by CoverityScan, CID#139381 ("Untrusted pointer read")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jasmin Jessich <jasmin@anw.at>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
|
