To detect a rx fifo overrun it's checked whether the number of elements
in the chip fifo exceeds the fifo size. This check can never return true
and is wrong.
Instead we should generate an interrupt if the fifo overruns.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When the rc_map table is created the char pointer of the name of the keymap
is copied to the rc_map->name field. However, this pointer points to memory
from the keymap module itself.
Since these keymap modules are not refcounted, that means anyone can call
rmmod to unload that module. Which is not a big deal because the contents of
the map is all copied to rc_map, except for the keymap name.
So after a keymap module is unloaded the name pointer has become stale. Unloading
the rc-core module will now cause a kernel oops in rc_dev_uevent().
The solution is to kstrdup the name so there are no more references to the
keymap module remaining.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
CONFIG_DEBUG_KOBJECT_RELEASE found this bug.
Signed-off-by: Max Kellermann <max@duempel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add USB ID of SMK RXX6000 series IR receiver. Often branded as
Lenovo receiver.
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
New USB ID for Adaptec eHome receiver in some HP laptops.
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Access to dev->initialized is atomic and dev->initialized isn't
accessed in any other code protected by this mutex.
Therefore we don't need to get the mutex here.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Passing overlapping source and destination to snprintf is
fragile. Replace with a single (mostly) equivalent call. If one wants
to preserve the space preceding udev->product whether or not there was
a manufacturer, just remove udev->manufacturer from the && expression.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
It makes the printk cleaner. As a side efect, it also fixes those smatch
warnings:
drivers/media/rc/mceusb.c:590 mceusb_dev_printdata() warn: argument 6 to %02x specifier has type 'char'
drivers/media/rc/mceusb.c:590 mceusb_dev_printdata() warn: argument 7 to %02x specifier has type 'char'
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Switch attribute wakeup_data from binary to a text attribute.
This makes it easier to handle in userspace and allows to
use the output of tools like mode2 almost as is to set a
wakeup sequence.
Changing to a text format and values in microseconds also
makes the userspace interface independent of the setting of
SAMPLE_PERIOD in the driver.
In addition document the new sysfs attribute in
Documentation/ABI/testing/sysfs-class-rc-nuvoton.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
MODULE_DEVICE_TABLE() is missing, so the module isn't auto-loading on
systems supporting infrared. This commit adds the missing line so it
works out of the box when built as a module and running on a sunxi
system with an infrared receiver.
Signed-off-by: Emilio López <emilio.lopez@collabora.co.uk>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Since rc_allocate_device() uses kmalloc, it can returns NULL,
so need to check, otherwise, NULL derefenrece can happen.
Reported-by: Insu Yun <wuninsu@gmail.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Fix AverMedia RM-KS keymap using user guide to meet LinuxTV wiki rules.
The remote command didn't seem to change in itself since its creation: it's
just to make keys more standard and remove the FIXME.
Signed-off-by: Philippe Valembois <lephilousophe@users.sourceforge.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This patch adds a binary attribute /sys/class/rc/rc?/wakeup_data which
allows to read / write the wakeup sequence.
In combination with the core extension for exposing the most recent raw
packet this allows to easily define and set a wakeup sequence.
At least on my Zotac CI321 the BIOS resets the wakeup sequence at each boot
to a factory default. Therefore I use a udev rule
SUBSYSTEM=="rc", DRIVERS=="nuvoton-cir", ACTION=="add", RUN+="<script>"
with the script basically doing
cat <stored wakeup sequence> >/sys${DEVPATH}/wakeup_data
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The device initialization completes only after calling
input_register_device(). However, rc_open() can be called while
the device is being registered by the input/evdev core. So, we
can't expect that rc_dev->initialized to be true.
Change the logic to don't require initialized == true at rc_open
and change the type of initialized to be atomic.
this way, we can check for it earlier where it is really needed,
without needing to lock the mutex just for testing it.
Tested with nuvoton_cir driver on a NUC5i7RYB with CIR integrated on it.
Reported-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
We can only unlock if mutex_lock() succeeds.
Fixes the following warning:
drivers/media/rc/lirc_dev.c:535 lirc_dev_fop_close() error: double unlock 'mutex:&lirc_dev_lock'
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Instead of having the timeouts hardcoded, and getting only the
accel value from the array, put everything in the same place.
That simplifies the logic.
As a side effect, it also cleans several smatch errors:
include/linux/jiffies.h:359:41: error: strange non-value function or array
include/linux/jiffies.h:361:42: error: strange non-value function or array
(one per time_after/time_before line)
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The mutex lock at rc_register_device() was added by commit 08aeb7c9a4
("[media] rc: add locking to fix register/show race").
It is meant to avoid race issues when trying to open a sysfs file while
the RC register didn't complete.
Adding a lock there causes troubles, as detected by the Kernel lock
debug instrumentation at the Kernel:
======================================================
[ INFO: possible circular locking dependency detected ]
4.5.0-rc3+ #46 Not tainted
-------------------------------------------------------
systemd-udevd/2681 is trying to acquire lock:
(s_active#171){++++.+}, at: [<ffffffff8171a115>] kernfs_remove_by_name_ns+0x45/0xa0
but task is already holding lock:
(&dev->lock){+.+.+.}, at: [<ffffffffa0724def>] rc_register_device+0xb2f/0x1450 [rc_core]
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (&dev->lock){+.+.+.}:
[<ffffffff8124817d>] lock_acquire+0x13d/0x320
[<ffffffff822de966>] mutex_lock_nested+0xb6/0x860
[<ffffffffa0721f2b>] show_protocols+0x3b/0x3f0 [rc_core]
[<ffffffff81cdaba5>] dev_attr_show+0x45/0xc0
[<ffffffff8171f1b3>] sysfs_kf_seq_show+0x203/0x3c0
[<ffffffff8171a6a1>] kernfs_seq_show+0x121/0x1b0
[<ffffffff81617c71>] seq_read+0x2f1/0x1160
[<ffffffff8171c911>] kernfs_fop_read+0x321/0x460
[<ffffffff815abc20>] __vfs_read+0xe0/0x3d0
[<ffffffff815ae90e>] vfs_read+0xde/0x2d0
[<ffffffff815b1d01>] SyS_read+0x111/0x230
[<ffffffff822e8636>] entry_SYSCALL_64_fastpath+0x16/0x76
-> #0 (s_active#171){++++.+}:
[<ffffffff81244f24>] __lock_acquire+0x4304/0x5990
[<ffffffff8124817d>] lock_acquire+0x13d/0x320
[<ffffffff81717d3a>] __kernfs_remove+0x58a/0x810
[<ffffffff8171a115>] kernfs_remove_by_name_ns+0x45/0xa0
[<ffffffff81721592>] remove_files.isra.0+0x72/0x190
[<ffffffff8172174b>] sysfs_remove_group+0x9b/0x150
[<ffffffff81721854>] sysfs_remove_groups+0x54/0xa0
[<ffffffff81cd97d0>] device_remove_attrs+0xb0/0x140
[<ffffffff81cdb27c>] device_del+0x38c/0x6b0
[<ffffffffa0724b8b>] rc_register_device+0x8cb/0x1450 [rc_core]
[<ffffffffa1326a7b>] dvb_usb_remote_init+0x66b/0x14d0 [dvb_usb]
[<ffffffffa1321c81>] dvb_usb_device_init+0xf21/0x1860 [dvb_usb]
[<ffffffffa13517dc>] dib0700_probe+0x14c/0x410 [dvb_usb_dib0700]
[<ffffffff81dbb1dd>] usb_probe_interface+0x45d/0x940
[<ffffffff81ce7e7a>] driver_probe_device+0x21a/0xc30
[<ffffffff81ce89b1>] __driver_attach+0x121/0x160
[<ffffffff81ce21bf>] bus_for_each_dev+0x11f/0x1a0
[<ffffffff81ce6cdd>] driver_attach+0x3d/0x50
[<ffffffff81ce5df9>] bus_add_driver+0x4c9/0x770
[<ffffffff81cea39c>] driver_register+0x18c/0x3b0
[<ffffffff81db6e98>] usb_register_driver+0x1f8/0x440
[<ffffffffa074001e>] dib0700_driver_init+0x1e/0x1000 [dvb_usb_dib0700]
[<ffffffff810021b1>] do_one_initcall+0x141/0x300
[<ffffffff8144d8eb>] do_init_module+0x1d0/0x5ad
[<ffffffff812f27b6>] load_module+0x6666/0x9ba0
[<ffffffff812f5fe8>] SyS_finit_module+0x108/0x130
[<ffffffff822e8636>] entry_SYSCALL_64_fastpath+0x16/0x76
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&dev->lock);
lock(s_active#171);
lock(&dev->lock);
lock(s_active#171);
*** DEADLOCK ***
3 locks held by systemd-udevd/2681:
#0: (&dev->mutex){......}, at: [<ffffffff81ce8933>] __driver_attach+0xa3/0x160
#1: (&dev->mutex){......}, at: [<ffffffff81ce8941>] __driver_attach+0xb1/0x160
#2: (&dev->lock){+.+.+.}, at: [<ffffffffa0724def>] rc_register_device+0xb2f/0x1450 [rc_core]
In this specific case, some error happened during device init,
causing IR to be disabled.
Let's fix it by adding a var that will tell when the device is
initialized. Any calls before that will return a -EINVAL.
That should prevent the race issues.
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
nvt_disable_cir calls nvt_disable_logical_dev (that may sleep) and is
called from contexts holding a spinlock.
Fix this and remove the unneeded clearing of CIR_IREN as this is done
in nvt_cir_disable already.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
nvt_enable_wake calls nvt_select_logical_dev (that may sleep) and is called
from contexts holding a spinlock. Fix this.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
nvt_enable_cir calls nvt_enable_logical_dev (that may sleep)
while holding a spinlock.
This patch fixes this and moves the content of nvt_enable_cir
to nvt_open as this is the only caller.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Extend the locking to protect more critical actions like register accesses
in the interrupt handlers.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Only handle an interrupt if at least one combination of event bit
and related interrupt bit is set.
Previously it was just checked that at least one event bit and
at least one interrupt bit are set.
This fixes issues like the following which was caused by
interrupt sharing:
An interrupt intended for nvt_cir_isr was handled by nvt_cir_wake_isr
first and because status bit CIR_WAKE_IRSTS_IR_PENDING was set
the wake fifo was accidently cleared.
This patch also fixes the bug that nvt_cir_wake_isr returned IRQ_HANDLED
even if it detected that the (shared) interrupt was meant for another
handler.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Check for the case that no Nuvoton chip is found on either EFM port.
Also move the position of nvt_efm_disable to reduce the time the
EFM ports are locked.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
At least on Zotac CI321 ACPI provides an ioport range for the wake up part
but accessing these ioports has no effect.
Instead the ioport base address is set to another value already
(0xa20 in my case) and accessing this ioport range works.
Therefore set a new ioport base address only if the current ioport base
address is 0 (register reset default).
The need to use the existing base address instead of trying to set
an own one doesn't seem to be limited to this specific device as other
drivers like hwmon/nct6775 do it the same way.
This change was successfully tested on the mentioned device.
And the change should be generic enough to not break the driver for
other chips (however due to lack of appropriate hardware I wasn't
able to test this).
[mchehab@osg.samsung.com: Tested on Intel NUC NUC5i7RYB with BIOS
version RYBDWi35.86A.0350.2015.0812.1722]
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Tested-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Calling nvt_set_cir_iren separately is not needed as this is done
by nvt_cir_regs_init.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Selecting the logical device in the interrupt handler is not needed
as no configuration register is accessed.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Only enable the logical devices after the registers have been initialized.
The call to nvt_enable_logical_dev in nvt_resume is not needed as this is
done implicitely by nvt_cir_regs_init now.
[mchehab@osg.samsung.com: fixed multiline comment to kernel CodingStyle]
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Switch to using the recently introduced IR default timeout value
(IR_DEFAULT_TIMEOUT) and consider the value of SAMPLE_PERIOD
when calculating the limit count register value.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
At least on NVT6779D clearing the wake fifo works in learning mode only
(although this condition is not mentioned in the chip spec).
Setting the clear fifo bit has no effect in wake up mode.
Even if clearing the wake fifo should work in wake up mode on other
chips this workaround doesn't hurt.
If needed the caller of nvt_clear_cir_wake_fifo has to take care
of locking.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The two EFM ioports are accessed by drivers for other parts of the Nuvoton
Super-IO chips too. Therefore access to these ioports needs to be
protected by using request_muxed_region (like it's implemented e.g. in
hwmon/nct6775 already).
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The FIFO is used for ir_raw_event records, however for some historic
reason the FIFO is used on a per byte basis. IMHO this adds unneeded
complexity. Therefore set up the FIFO for ir_raw_event records.
This also allows to define the FIFO statically as part of
ir_raw_event_ctrl instead of having to allocate the FIFO dynamically.
In addition:
- When writing into the FIFO and it's full return ENOSPC instead of
ENOMEM thus making it easier to tell between "FIFO full" and
"Dynamic memory allocation failed" when the error is propagated to
a higher level.
Also add an error message.
- When reading from the FIFO check whether it's empty.
This is not strictly needed here but kfifo_out is annotated
"must check" anyway.
Successfully tested it with the nuvoton-cir driver.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
When requesting resources use different names for the normal and
the wakeup part. This makes it easier to interpret the output
of e.g. /proc/interrupts and /proc/ioports.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The driver allocates the spinlock but fails to initialize it correctly.
The kernel reports a BUG indicating bad spinlock magic when spinlock
debugging is enabled.
Call spin_lock_init() on it to initialize it correctly.
Fixes: b4e3e59fb5 ("[media] rc: add sunxi-ir driver")
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Now that that the decoder modules are loaded on-demand we can move
loading the lirc module to rc_register_device directly and remove
unneeded functions and comments.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Checking whether a protocol is enabled and calling the related decoder
functions should be done by the rc core, not the protocol handlers.
Properly handle lirc considering that no protocol bit is set for lirc.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Declare struct proto_names and its member name as const.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Prepare on-demand decoder module loading by adding a module_name member
to struct proto_names and introducing the related load function.
After this patch of the series the decoder modules are still loaded
unconditionally.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Currently, if a decoder module is unloadad, the respective protocol
is still shown as enabled (if it was enabled before).
Fix this by resetting the respective protocol bits if a decoder
module is unloaded.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Simplify the logging.
I had some doubts about using the elvis operator as it's GNU extension.
However GNU extensions are explicitely allowed and this operator is
used at several places in the kernel code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Switch the info about the detected chip type from debug to info level
as it might be useful not only for debugging purposes.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Instead of explicitely checking debug use nvt_dbg like in other parts
of the driver thus simplifying the code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
