linux-stable.git/drivers/counter, branch v6.6.86

counter: microchip-tcb-capture: Fix undefined counter channel state on probe

2025-04-07T08:06:37+00:00

commit c0c9c73434666dc99ee156b25e7e722150bee001 upstream.

Hardware initialize of the timer counter channel does not occur on probe
thus leaving the Count in an undefined state until the first
function_write() callback is executed. Fix this by performing the proper
hardware initialization during probe.

Fixes: 106b104137fd ("counter: Add microchip TCB capture counter")
Reported-by: Csókás Bence 
Closes: https://lore.kernel.org/all/bfa70e78-3cc3-4295-820b-3925c26135cb@prolan.hu/
Link: https://lore.kernel.org/r/20250305-preset-capture-mode-microchip-tcb-capture-v1-1-632c95c6421e@kernel.org
Signed-off-by: William Breathitt Gray 
Signed-off-by: Greg Kroah-Hartman

counter: stm32-lptimer-cnt: fix error handling when enabling

2025-04-07T08:06:37+00:00

commit 8744dcd4fc7800de2eb9369410470bb2930d4c14 upstream.

In case the stm32_lptim_set_enable_state() fails to update CMP and ARR,
a timeout error is raised, by regmap_read_poll_timeout. It may happen,
when the lptimer runs on a slow clock, and the clock is gated only
few times during the polling.

Badly, when this happen, STM32_LPTIM_ENABLE in CR register has been set.
So the 'enable' state in sysfs wrongly lies on the counter being
correctly enabled, due to CR is read as one in stm32_lptim_is_enabled().

To fix both issues:
- enable the clock before writing CMP, ARR and polling ISR bits. It will
avoid the possible timeout error.
- clear the ENABLE bit in CR and disable the clock in the error path.

Fixes: d8958824cf07 ("iio: counter: Add support for STM32 LPTimer")
Signed-off-by: Fabrice Gasnier 
Link: https://lore.kernel.org/r/20250224170657.3368236-1-fabrice.gasnier@foss.st.com
Signed-off-by: William Breathitt Gray 
Signed-off-by: Greg Kroah-Hartman

counter: ti-ecap-capture: Add check for clk_enable()

2024-12-09T09:32:37+00:00

[ Upstream commit 1437d9f1c56fce9c24e566508bce1d218dd5497a ]

Add check for the return value of clk_enable() in order to catch the
potential exception.

Fixes: 4e2f42aa00b6 ("counter: ti-ecap-capture: capture driver support for ECAP")
Reviewed-by: Julien Panis 
Signed-off-by: Jiasheng Jiang 
Link: https://lore.kernel.org/r/20241104194059.47924-1-jiashengjiangcool@gmail.com
Signed-off-by: William Breathitt Gray 
Signed-off-by: Sasha Levin

counter: stm32-timer-cnt: Add check for clk_enable()

2024-12-09T09:32:37+00:00

[ Upstream commit 842c3755a6bfbfcafa4a1438078d2485a9eb1d87 ]

Add check for the return value of clk_enable() in order to catch the
potential exception.

Fixes: c5b8425514da ("counter: stm32-timer-cnt: add power management support")
Fixes: ad29937e206f ("counter: Add STM32 Timer quadrature encoder")
Signed-off-by: Jiasheng Jiang 
Link: https://lore.kernel.org/r/20241104191825.40155-1-jiashengjiangcool@gmail.com
Signed-off-by: William Breathitt Gray 
Signed-off-by: Sasha Levin

counter: ti-eqep: enable clock at probe

2024-07-05T07:33:56+00:00

[ Upstream commit 0cf81c73e4c6a4861128a8f27861176ec312af4e ]

The TI eQEP clock is both a functional and interface clock. Since it is
required for the device to function, we should be enabling it at probe.

Up to now, we've just been lucky that the clock was enabled by something
else on the system already.

Fixes: f213729f6796 ("counter: new TI eQEP driver")
Reviewed-by: Judith Mendez 
Signed-off-by: David Lechner 
Link: https://lore.kernel.org/r/20240621-ti-eqep-enable-clock-v2-1-edd3421b54d4@baylibre.com
Signed-off-by: William Breathitt Gray 
Signed-off-by: Sasha Levin

Merge tag 'counter-fixes-for-6.6a' of git://git.kernel.org/pub/scm/linux/kernel/git/wbg/counter into char-misc-linus

2023-10-02T11:13:15+00:00

William writes:

First set of Counter fixes for 6.6

The counter_get_ext() function would incorrectly refer to the first
element of the extensions array to handle component array extensions
when they are located at a different index; a fix is provided to index
to the correct element in the array for this case. A fix for the
microchip-tcb-capture is provided as well to correct an inverted
internal GCLK logic for clock selection.

* tag 'counter-fixes-for-6.6a' of git://git.kernel.org/pub/scm/linux/kernel/git/wbg/counter:
  counter: microchip-tcb-capture: Fix the use of internal GCLK logic
  counter: chrdev: fix getting array extensions

counter: microchip-tcb-capture: Fix the use of internal GCLK logic

2023-09-05T16:25:14+00:00

As per the datasheet, the clock selection Bits 2:0 – TCCLKS[2:0] should
be set to 0 while using the internal GCLK (TIMER_CLOCK1).

Fixes: 106b104137fd ("counter: Add microchip TCB capture counter")
Signed-off-by: Dharma Balasubiramani 
Link: https://lore.kernel.org/r/20230905100835.315024-1-dharma.b@microchip.com
Signed-off-by: William Breathitt Gray

Merge tag 'mfd-next-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd

2023-09-04T20:47:59+00:00

Pull NFD updates from Lee Jones:
 "New Drivers:
   - Add support for the Cirrus Logic CS42L43 Audio CODEC

  Fix-ups:
   - Make use of specific printk() format tags for various optimisations
   - Kconfig / module modifications / tweaking
   - Simplify obtaining resources (memory, device data) using unified
     API helpers
   - Bunch of Device Tree additions, conversions and adaptions
   - Convert a bunch of Regmap configurations to use the Maple Tree
     cache
   - Ensure correct includes are present and remove some that are not
     required
   - Remove superfluous code
   - Reduce amount of cycles spent in critical sections
   - Omit the use of redundant casts and if relevant replace with better
     ones
   - Swap out raw_spin_{un}lock_irq{save,restore}() for
     spin_{un}lock_irq{save,restore}()

  Bug Fixes:
   - Repair theoretical deadlock situation
   - Fix some link-time dependencies
   - Use more appropriate datatype when casting"

* tag 'mfd-next-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (70 commits)
  mfd: mc13xxx: Simplify device data fetching in probe()
  mfd: rz-mtu3: Replace raw_spin_lock->spin_lock()
  mfd: rz-mtu3: Reduce critical sections
  mfd: mxs-lradc: Fix Wvoid-pointer-to-enum-cast warning
  mfd: wm31x: Fix Wvoid-pointer-to-enum-cast warning
  mfd: wm8994: Fix Wvoid-pointer-to-enum-cast warning
  mfd: tc3589: Fix Wvoid-pointer-to-enum-cast warning
  mfd: lp87565: Fix Wvoid-pointer-to-enum-cast warning
  mfd: hi6421-pmic: Fix Wvoid-pointer-to-enum-cast warning
  mfd: max77541: Fix Wvoid-pointer-to-enum-cast warning
  mfd: max14577: Fix Wvoid-pointer-to-enum-cast warning
  mfd: stmpe: Fix Wvoid-pointer-to-enum-cast warning
  mfd: rn5t618: Remove redundant of_match_ptr()
  mfd: lochnagar-i2c: Remove redundant of_match_ptr()
  mfd: stpmic1: Remove redundant of_match_ptr()
  mfd: act8945a: Remove redundant of_match_ptr()
  mfd: rsmu_spi: Remove redundant of_match_ptr()
  mfd: altera-a10sr: Remove redundant of_match_ptr()
  mfd: rsmu_i2c: Remove redundant of_match_ptr()
  mfd: tc3589x: Remove redundant of_match_ptr()
  ...

counter: chrdev: fix getting array extensions

2023-09-04T18:56:27+00:00

When trying to watch a component array extension, and the array isn't the
first extended element, it fails as the type comparison is always done on
the 1st element. Fix it by indexing the 'ext' array.

Example on a dummy struct counter_comp:
static struct counter_comp dummy[] = {
	COUNTER_COMP_DIRECTION(..),
	...,
	COUNTER_COMP_ARRAY_CAPTURE(...),
};
static struct counter_count dummy_cnt = {
	...
	.ext = dummy,
	.num_ext = ARRAY_SIZE(dummy),
}

Currently, counter_get_ext() returns -EINVAL when trying to add a watch
event on one of the capture array element in such example.

Fixes: d2011be1e22f ("counter: Introduce the COUNTER_COMP_ARRAY component type")
Signed-off-by: Fabrice Gasnier 
Link: https://lore.kernel.org/r/20230829134029.2402868-2-fabrice.gasnier@foss.st.com
Signed-off-by: William Breathitt Gray

mfd: rz-mtu3: Link time dependencies

2023-08-18T20:48:04+00:00

The new set of drivers for RZ/G2L MTU3a tries to enable compile-testing the
individual client drivers even when the MFD portion is disabled but gets it
wrong, causing a link failure when the core is in a loadable module but the
other drivers are built-in:

x86_64-linux-ld: drivers/pwm/pwm-rz-mtu3.o: in function `rz_mtu3_pwm_apply':
pwm-rz-mtu3.c:(.text+0x4bf): undefined reference to `rz_mtu3_8bit_ch_write'
x86_64-linux-ld: pwm-rz-mtu3.c:(.text+0x509): undefined reference to `rz_mtu3_disable'

arm-linux-gnueabi-ld: drivers/counter/rz-mtu3-cnt.o: in function `rz_mtu3_cascade_counts_enable_get':
rz-mtu3-cnt.c:(.text+0xbec): undefined reference to `rz_mtu3_shared_reg_read'

It seems better not to add the extra complexity here but instead just use
a normal hard dependency, so remove the #else portion in the header along
with the "|| COMPILE_TEST". This could also be fixed by having slightly more
elaborate Kconfig dependencies or using the cursed 'IS_REACHABLE()' helper,
but in practice it's already possible to compile-test all these drivers
by enabling the mtd portion.

Fixes: 254d3a727421c ("pwm: Add Renesas RZ/G2L MTU3a PWM driver")
Fixes: 0be8907359df4 ("counter: Add Renesas RZ/G2L MTU3a counter driver")
Fixes: 654c293e1687b ("mfd: Add Renesas RZ/G2L MTU3a core driver")
Signed-off-by: Arnd Bergmann 
Acked-by: Thierry Reding 
Reviewed-by: Biju Das 
Link: https://lore.kernel.org/r/20230719090430.1925182-1-arnd@kernel.org
Signed-off-by: Lee Jones