When adding the dedicated gs101_wkup_irq_chip struct to support the eint
wakeup mask the .eint_con, eint_mask and .eint_pend fields were missed. The
result is that irqs on gs101 for the buttons etc are broken.
Reported-by: André Draszik <andre.draszik@linaro.org>
Fixes: 2642f55d44 ("pinctrl: samsung: add support for gs101 wakeup mask programming")
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Tested-by: André Draszik <andre.draszik@linaro.org>
Link: https://lore.kernel.org/r/20250702-fix-gs101-irqchip-v1-1-ccc84b44ad72@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
gs101 differs to other currently supported SoCs in that it has 3 wakeup
mask registers for the 67 external wakeup interrupt pins in alive and
far_alive.
EINT_WAKEUP_MASK 0x3A80 EINT[31:0]
EINT_WAKEUP_MASK2 0x3A84 EINT[63:32]
EINT_WAKEUP_MASK3 0x3A88 EINT[66:64]
Add gs101 specific callbacks and a dedicated gs101_wkup_irq_chip struct to
handle these differences.
The current wakeup mask with upstream is programmed as
WAKEUP_MASK0[0x3A80] value[0xFFFFFFFF]
WAKEUP_MASK1[0x3A84] value[0xF2FFEFFF]
WAKEUP_MASK2[0x3A88] value[0xFFFFFFFF]
Which corresponds to the following wakeup sources:
gpa7-3 vol down
gpa8-1 vol up
gpa10-1 power
gpa8-2 typec-int
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Link: https://lore.kernel.org/r/20250619-gs101-eint-mask-v1-2-89438cfd7499@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Newer Exynos based SoCs have a filter selection bitfield in the filter
configuration registers on alive bank pins. This allows the selection of
a digital or analog delay filter for each pin. Add support for selecting
and enabling the filter.
On suspend we set the analog filter to all pins in the bank (as the
digital filter relies on a clock). On resume the digital filter is
reapplied to all pins in the bank. The digital filter is working via
a clock and has an adjustable filter delay flt_width bitfield, whereas
the analog filter uses a fixed delay.
The filter determines to what extent signal fluctuations received through
the pad are considered glitches.
The code path can be exercised using
echo mem > /sys/power/state
And then wake the device using a eint gpio
Reviewed-by: André Draszik <andre.draszik@linaro.org>
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Link: https://lore.kernel.org/r/20250402-pinctrl-fltcon-suspend-v6-4-78ce0d4eb30c@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
gs101 differs to other SoCs in that fltcon1 register doesn't
always exist. Additionally the offset of fltcon0 is not fixed
and needs to use the newly added eint_fltcon_offset variable.
Fixes: 4a8be01a1a ("pinctrl: samsung: Add gs101 SoC pinctrl configuration")
Cc: stable@vger.kernel.org # depends on the previous three patches
Reviewed-by: André Draszik <andre.draszik@linaro.org>
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Link: https://lore.kernel.org/r/20250402-pinctrl-fltcon-suspend-v6-3-78ce0d4eb30c@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Refactor the existing platform specific suspend/resume callback
so that each SoC variant has it's own callback containing the
SoC specific logic.
This allows exynosautov920 to have a dedicated function for using
eint_con_offset and eint_mask_offset. Also it is easily extendable
for gs101 which will need dedicated logic for handling the varying
register offset of fltcon0 via eint_fltcon_offset.
Reviewed-by: André Draszik <andre.draszik@linaro.org>
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Link: https://lore.kernel.org/r/20250402-pinctrl-fltcon-suspend-v6-2-78ce0d4eb30c@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
This enables the clk_enable() and clk_disable() logic to be removed
from each callback, but otherwise should have no functional impact.
It is a prepatory patch so that the callbacks can become SoC
specific.
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Link: https://lore.kernel.org/r/20250402-pinctrl-fltcon-suspend-v6-1-78ce0d4eb30c@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Replace ternary (condition ? "enable" : "disable") syntax with helpers
from string_choices.h because:
1. Simple function call with one argument is easier to read. Ternary
operator has three arguments and with wrapping might lead to quite
long code.
2. Is slightly shorter thus also easier to read.
3. It brings uniformity in the text - same string.
4. Allows deduping by the linker, which results in a smaller binary
file.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Antonio Borneo <antonio.borneo@foss.st.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/20250114203602.1013275-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
chained_irq_enter(() should be paired with a corresponding
chained_irq_exit().
Here, if clk_enable() fails, a early return occurs and chained_irq_exit()
is not called.
Add a new label and a goto for fix it.
Fixes: f9c7447479 ("pinctrl: samsung: support a bus clock")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: André Draszik <andre.draszik@linaro.org>
Link: https://lore.kernel.org/r/f148d823acfb3326a115bd49a0eed60f2345f909.1731844995.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
On some Samsung-based SoCs there are separate bus clocks / gates each
for each pinctrl instance. To be able to access each pinctrl instance's
registers, this bus clock needs to be running, otherwise register
access will hang. Google Tensor gs101 is one example for such an
implementation.
Update the driver to handle this optional bus clock:
* handle an optional bus clock from DT
* prepare it during driver probe
* enclose all relevant register accesses with a clock enable & disable
Signed-off-by: André Draszik <andre.draszik@linaro.org>
Link: https://lore.kernel.org/r/20240426-samsung-pinctrl-busclock-v3-2-adb8664b8a7e@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Constify few pointers to iomem, where the destination memory is not
modified, for code safety and readability.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20231223191902.22857-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Add pinctrl data for ExynosAutov920 SoC.
It has a newly applied pinctrl register layer for ExynosAuto series.
Pinctrl data for ExynosAutoV920 SoC.
- GPA0,GPA1 (10): External wake up interrupt
- GPQ0 (2): SPMI (PMIC I/F)
- GPB0,GPB1,GPB2,GPB3,GPB4,GPB5,GPB6 (47): I2S Audio
- GPH0,GPH1,GPH2,GPH3,GPH4,GPH5,GPH6,GPH8 (49): PCIE, UFS, Ethernet
- GPG0,GPG1,GPG2,GPG3,GPG4,GPG5 (29): General purpose
- GPP0,GPP1,GPP2,GPP3,GPP4,GPP5,GPP6,GPP7,GPP8,GPP9,GPP10 (77): USI
Signed-off-by: Jaewon Kim <jaewon02.kim@samsung.com>
Link: https://lore.kernel.org/r/20231211114145.106255-3-jaewon02.kim@samsung.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
New ExynosAuto series GPIO have a different register structure.
In the existing Exynos series, EINT control register is enumerated after
a specific offset (e.g EXYNOS_GPIO_ECON_OFFSET, EXYNOS_GPIO_EMASK_OFFSET).
However, from ExynosAutov920 SoC, the register that controls EINT belongs
to each GPIO bank, and each GPIO bank has 0x1000 align.
This is a structure to protect the GPIO bank using S2MPU in VM environment,
and will only be applied in ExynosAuto series SoCs.
--------------------------------------------------------------
| Original Exynos | ExynosAuto |
|------------------------------------------------------------|
| 0x0 GPIO_CON | 0x0 GPIO_CON |
| 0x4 GPIO_DAT | 0x4 GPIO_DAT |
| 0x8 GPIO_PUD | 0x8 GPIO_PUD |
| 0xc GPIO_DRV | 0xc GPIO_DRV |
| 0x10 GPIO_CONPDN | 0x10 GPIO_CONPDN |
| 0x14 GPIO_PUDPDN | 0x14 GPIO_PUDPDN |
|----------------------------| 0x18 EINT_CON (per_bank) |
| ... | 0x1c EINT_FLTCON0 (per_bank) |
| ... | 0x20 EINT_FLTCON1 (per_bank) |
| ... | 0x24 EINT_MASK (per_bank) |
| ... | 0x28 EINT_PEND (per_bank) |
|----------------------------|-------------------------------|
| 0x700 EINT_CON (global) | ... |
| 0x800 EINT_FLTCON (global) | ... |
| 0x900 EINT_MASK (global) | ... |
| 0xa00 EINT_FEND (global) | ... |
--------------------------------------------------------------
Signed-off-by: Jaewon Kim <jaewon02.kim@samsung.com>
Link: https://lore.kernel.org/r/20231211114145.106255-2-jaewon02.kim@samsung.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
To support affinity setting for non wake up external gpio interrupt, add
irq_set_affinity callback using irq number from pinctrl driver data.
Before this patch, changing the irq affinity of gpio interrupt is not
possible:
# cat /proc/irq/418/smp_affinity
3ff
# echo 00f > /proc/irq/418/smp_affinity
# cat /proc/irq/418/smp_affinity
3ff
# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3 ...
418: 3631 0 0 0 ...
With this patch applied, it's possible to change irq affinity of gpio
interrupt:
# cat /proc/irq/418/smp_affinity
3ff
# echo 00f > /proc/irq/418/smp_affinity
# cat /proc/irq/418/smp_affinity
00f
# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3 ...
418: 3893 201 181 188 ...
Signed-off-by: Youngmin Nam <youngmin.nam@samsung.com>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Tested-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20231126094618.2545116-1-youngmin.nam@samsung.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct
exynos_muxed_weint_data. Additionally, since the element count member
must be set before accessing the annotated flexible array member, move
its initialization earlier.
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Alim Akhtar <alim.akhtar@samsung.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-gpio@vger.kernel.org
Cc: linux-hardening@vger.kernel.org
Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci [1]
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20231006201707.work.405-kees@kernel.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
The Samsung SoC pin controller driver uses only three defines from the
bindings header with pin configuration register values, which proves
the point that this header is not a proper bindings-type abstraction
layer with IDs.
Define the needed register values directly in the driver and stop using
the bindings header.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Chanho Park <chanho61.park@samsung.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20220605160508.134075-8-krzysztof.kozlowski@linaro.org
Link: https://lore.kernel.org/r/20220624081022.32384-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Switch the code to use for_each_gpiochip_node() helper.
While at it, in order to avoid additional churn in the future,
switch to fwnode APIs where it makes sense.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Improve the wake irq message by also printing the bank name and hwirq
number that matches this irq number.
Signed-off-by: Martin Jücker <martin.juecker@gmail.com>
Link: https://lore.kernel.org/r/20220130232122.GA119248@adroid
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
It seems that newer ARMv8 Exynos SoC like Exynos850 and
ExynosAutov9 have differences of their pin controller node capable of
external wake-up interrupts:
1. No multiplexed external wake-up interrupt, only direct,
2. More than one pin controller capable of external wake-up interrupts.
Add support for dedicated Exynos850 and ExynosAutov9 compatibles.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Chanho Park <chanho61.park@samsung.com>
Tested-by: Chanho Park <chanho61.park@samsung.com>
Link: https://lore.kernel.org/r/20220111201722.327219-20-krzysztof.kozlowski@canonical.com
Wherever possible, replace constructs that match either
generic_handle_irq(irq_find_mapping()) or
generic_handle_irq(irq_linear_revmap()) to a single call to
generic_handle_domain_irq().
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
The Special Function Registers on all Exynos SoC, including ARM64, are
32-bit wide, so entire driver uses matching functions like readl() or
writel(). On 64-bit ARM using unsigned long for register masks:
1. makes little sense as immediately after bitwise operation it will be
cast to 32-bit value when calling writel(),
2. is actually error-prone because it might promote other operands to
64-bit.
Addresses-Coverity: Unintentional integer overflow
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Link: https://lore.kernel.org/r/20210408195029.69974-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This patch converts spin_[lock|unlock] functions of pin bank to
raw_spinlock to support preempt-rt. This can avoid BUG() assertion when
irqchip callbacks are triggerred. Spinlocks can be converted rt_mutex
which is preemptible when we apply preempt-rt patches.
According to "Documentation/driver-api/gpio/driver.rst",
"Realtime considerations: a realtime compliant GPIO driver should not
use spinlock_t or any sleepable APIs (like PM runtime) as part of its
irqchip implementation.
- spinlock_t should be replaced with raw_spinlock_t.[1]
"
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Chanho Park <chanho61.park@samsung.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Link: https://lore.kernel.org/r/20210121030009.25673-1-chanho61.park@samsung.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Use the bank name as the irqchip name. This name is later visible in
/proc/interrupts, what makes it possible to easily identify each
GPIO interrupt.
/proc/interrupts before this patch:
143: 0 exynos4210_wkup_irq_chip 7 Edge hdmi
144: 0 exynos4210_wkup_irq_chip 6 Level wm8994
145: 1 exynos4210_wkup_irq_chip 7 Edge max77686-pmic, max77686-rtc
146: 1 exynos_gpio_irq_chip 3 Edge 3-0048
/proc/interrupts after this patch:
143: 0 gpx3 7 Edge hdmi
144: 0 gpx3 6 Level wm8994
145: 1 gpx0 7 Edge max77686-pmic, max77686-rtc
146: 1 gpm2 3 Edge 3-0048
Handling of the eint_wake_mask_value has been reworked, because each bank
has now its own exynos_irq_chip structure allocated.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Link: https://lore.kernel.org/r/20200720145412.24221-1-krzk@kernel.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Currently, for EINT_TYPE GPIOs, the CON and FLTCON registers
are saved and restored over a suspend/resume cycle. However, the
EINT_MASK registers are not.
On S5PV210 at the very least, these registers are not retained over
suspend, leading to the interrupts remaining masked upon resume and
therefore no interrupts being triggered for the device. There should
be no effect on any SoCs that do retain these registers as theoretically
we would just be re-writing what was already there.
Fixes: 7ccbc60cd9 ("pinctrl: exynos: Handle suspend/resume of GPIO EINT registers")
Cc: <stable@vger.kernel.org>
Signed-off-by: Jonathan Bakker <xc-racer2@live.ca>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Commit a8be2af021 ("pinctrl: samsung: Write external wakeup interrupt
mask") started writing the eint wakeup mask from the pinctrl driver.
Unfortunately, it made the assumption that the private retention data
was always a regmap while in the case of s5pv210 it is a raw pointer
to the clock base (as the eint wakeup mask not in the PMU as with newer
Exynos platforms).
Fixes: a8be2af021 ("pinctrl: samsung: Write external wakeup interrupt mask")
Cc: <stable@vger.kernel.org>
Signed-off-by: Jonathan Bakker <xc-racer2@live.ca>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
In exynos_eint_wkup_init() the for_each_child_of_node() loop is used
with a break to find a matching child node. Although each iteration of
for_each_child_of_node puts the previous node, but early exit from loop
misses it. This leads to leak of device node.
Cc: <stable@vger.kernel.org>
Fixes: 43b169db18 ("pinctrl: add exynos4210 specific extensions for samsung pinctrl driver")
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Each iteration of for_each_child_of_node puts the previous node, but in
the case of a return from the middle of the loop, there is no put, thus
causing a memory leak. Hence add an of_node_put before the return of
exynos_eint_wkup_init() error path.
Issue found with Coccinelle.
Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
Cc: <stable@vger.kernel.org>
Fixes: 14c255d35b ("pinctrl: exynos: Add irq_chip instance for Exynos7 wakeup interrupts")
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Remove the legacy, ugly API of exposing the static value of external
wakeup interrupts mask, because all arch-machine users where converted
to use generic implementation from pinctrl driver.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Sylwester Nawrocki <snawrocki@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
The pinctrl driver defines an IRQ chip which handles external wakeup
interrupts, therefore from logical point of view, it is the owner of
external interrupt mask. The register controlling the mask belongs to
Power Management Unit address space so it has to be accessed with PMU
syscon regmap handle.
This mask should be written to hardware during system suspend. Till now
ARMv7 machine code was responsible for this which created a dependency
between pin controller driver and arch/arm/mach code.
Try to rework this dependency so the pinctrl driver will write external
wakeup interrupt mask during late suspend.
Impact on ARMv7 designs (S5Pv210 and Exynos)
============================================
This duplicates setting mask with existing machine code
arch/arm/mach-exynos/suspend.c and arch/arm/mach-s5pv210/pm.c but it is
not a problem - the wakeup mask register will be written twice. The
machine code will be cleaned up later.
The difference between implementation here and ARMv7 machine code
(arch/arm/mach-*) is the time of writing the mask:
1. The machine code is writing the mask quite late during system suspend
path, after offlining secondary CPUs and just before doing actual
suspend.
2. The implementation in pinctrl driver uses late suspend ops, therefore it
will write the mask much earlier. Hopefully late enough, after all
drivers will enable or disable their interrupt wakeups
(enable_irq_wake() etc).
Impact on ARMv8 designs (Exynos5433 and Exynos7)
================================================
The Suspend to RAM was not supported and external wakeup interrupt mask
was not written to HW. This change brings us one step closer to
supporting Suspend to RAM.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Sylwester Nawrocki <snawrocki@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
The S5Pv210 external wakeup interrupts differ from Exynos therefore
separate compatible is needed. Duplicate existing flavor specific data
from exynos4210_wkup_irq_chip and add new compatible for S5Pv210.
At this point this new compatible does not bring anything new and works
exactly as existing "samsung,exynos4210-wakeup-eint".
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Sylwester Nawrocki <snawrocki@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Two local variables (shift and reg_con) were initialized to unused
values - they were overwritten just few lines after. Getting rid of
this unused initialization allows dropping other variables and
compacting slightly the code.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
We already have macros for values used by driver and Device Tree
sources for pin mux configuration. Use them instead of duplicating
defines.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
When setting the pin function for external interrupts, the driver used
wrong IO memory address base. The pin function register is always under
pctl_base, not the eint_base.
By updating wrong register, the external interrupts for chosen GPIO
would not work at all and some other GPIO might be configured to wrong
value. For example on Exynos5433-based boards, the external interrupts
for gpf{1-5}-X GPIOs should not work at all (driver toggled reserved
registers from ALIVE bank instead).
Platforms other than Exynos5433 should not be affected as eint_base
equals pctl_base in such case.
Fixes: 8b1bd11c1f ("pinctrl: samsung: Add the support the multiple IORESOURCE_MEM for one pin-bank")
Cc: <stable@vger.kernel.org>
Reported-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Tested-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
The irq chip callbacks irq_request/release_resources() have absolutely no
business with masking and unmasking the irq.
The core code unmasks the interrupt after complete setup and masks it
before invoking irq_release_resources().
The unmask is actually harmful as it happens before the interrupt is
completely initialized in __setup_irq().
Remove it.
Fixes: f6a8249f9e ("pinctrl: exynos: Lock GPIOs as interrupts when used as EINTs")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Kukjin Kim <kgene@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-gpio@vger.kernel.org
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Check if kmemdup failed during wakeup banks initialization. Otherwise
NULL pointer would be stored under "irq_chip" member of bank and later
dereferenced in interrupt handler.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Static exynos_irq_chip structures, containing driver specific data, are
referenced only through opaque data pointer in const of_device_id table.
The contents of pointed memory (exynos_irq_chip structure itself) is
then copied with kmemdup() during wakeup initialization so
exynos_irq_chip can be made const for code safenes.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Exynos pinctrl drivers contain pretty big per-SoC data structures. The
pinctrl-exynos object file contained code and data for both ARMv7 and
ARMv8 SoCs thus it grew big. There will not be a shared image between
ARMv7 and ARMv8 so there is no need to combine all of this into one
driver.
Splitting the data allows to make it more granular (e.g. code related to
ARMv8 Exynos is self-contained), slightly speed up the compilation and
reduce the effective size of compiled kernel.
The common data structures and functions reside still in existing
pinctrl-exynos.c. Only the SoC-specific parts were moved out to new
files. Except marking few functions non-static and adding them to
header, there were no functional changes in the code.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Fixups here tend to be more all over the map vs. some of the other
repeated/systematic ones we've seen elsewhere.
We remove module.h from code that isn't doing anything modular at
all; if they have __init sections, then replace it with init.h
A couple drivers have module_exit() code that is essentially orphaned,
and so we remove that.
There are no module_init replacements, so we have no concerns wrt.
initcall ordering changes as per some of the other cleanups.
Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.
We also delete the MODULE_LICENSE tag etc. since all that information
was (or is now) contained at the top of the file in the comments.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Core changes:
- Add bi-directional and output-enable pin configurations to
the generic bindings and generic pin controlling core.
New drivers or subdrivers:
- Armada 37xx SoC pin controller and GPIO support.
- Axis ARTPEC-6 SoC pin controller support.
- AllWinner A64 R_PIO controller support, and opening up the
AllWinner sunxi driver for ARM64 use.
- Rockchip RK3328 support.
- Renesas R-Car H3 ES2.0 support.
- STM32F469 support in the STM32 driver.
- Aspeed G4 and G5 pin controller support.
Improvements:
- A whole slew of realtime improvements to drivers implementing
irqchips: BCM, AMD, SiRF, sunxi, rockchip.
- Switch meson driver to get the GPIO ranges from the device
tree.
- Input schmitt trigger support on the Rockchip driver.
- Enable the sunxi (AllWinner) driver to also be used on ARM64
silicon.
- Name the Qualcomm QDF2xxx GPIO lines.
- Support GMMR GPIO regions on the Intel Cherryview. This
fixes a serialization problem on these platforms.
- Pad retention support for the Samsung Exynos 5433.
- Handle suspend-to-ram in the AT91-pio4 driver.
- Pin configuration support in the Aspeed driver.
Cleanups:
- The final name of Rockchip RK1108 was RV1108 so rename the
driver and variables to stay consistent.
-----BEGIN PGP SIGNATURE-----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=BCaA
-----END PGP SIGNATURE-----
Merge tag 'pinctrl-v4.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl
Pull pin control updates from Linus Walleij:
"This is the bulk of pin control changes for the v4.12 cycle.
The extra week before the merge window actually resulted in some of
the type of fixes that usually arrive after the merge window already
starting to trickle in from eager developers using -next, I'm
impressed.
I have recruited a Samsung subsubsystem maintainer (Krzysztof) to deal
with the onset of Samsung patches. It works great.
Apart from that it is a boring round, just incremental updates and
fixes all over the place, no serious core changes or anything exciting
like that. The most pleasing to see is Julia Cartwrights work to audit
the irqchip-providing drivers for realtime locking compliance. It's
one of those "I should really get around to looking into that" things
that have been on my TODO list since forever.
Summary:
Core changes:
- add bi-directional and output-enable pin configurations to the
generic bindings and generic pin controlling core.
New drivers or subdrivers:
- Armada 37xx SoC pin controller and GPIO support.
- Axis ARTPEC-6 SoC pin controller support.
- AllWinner A64 R_PIO controller support, and opening up the
AllWinner sunxi driver for ARM64 use.
- Rockchip RK3328 support.
- Renesas R-Car H3 ES2.0 support.
- STM32F469 support in the STM32 driver.
- Aspeed G4 and G5 pin controller support.
Improvements:
- a whole slew of realtime improvements to drivers implementing
irqchips: BCM, AMD, SiRF, sunxi, rockchip.
- switch meson driver to get the GPIO ranges from the device tree.
- input schmitt trigger support on the Rockchip driver.
- enable the sunxi (AllWinner) driver to also be used on ARM64
silicon.
- name the Qualcomm QDF2xxx GPIO lines.
- support GMMR GPIO regions on the Intel Cherryview. This fixes a
serialization problem on these platforms.
- pad retention support for the Samsung Exynos 5433.
- handle suspend-to-ram in the AT91-pio4 driver.
- pin configuration support in the Aspeed driver.
Cleanups:
- the final name of Rockchip RK1108 was RV1108 so rename the driver
and variables to stay consistent"
* tag 'pinctrl-v4.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (80 commits)
pinctrl: mediatek: Add missing pinctrl bindings for mt7623
pinctrl: artpec6: Fix return value check in artpec6_pmx_probe()
pinctrl: artpec6: Remove .owner field for driver
pinctrl: tegra: xusb: Silence sparse warnings
ARM: at91/at91-pinctrl documentation: fix spelling mistake: "contoller" -> "controller"
pinctrl: make artpec6 explicitly non-modular
pinctrl: aspeed: g5: Add pinconf support
pinctrl: aspeed: g4: Add pinconf support
pinctrl: aspeed: Add core pinconf support
pinctrl: aspeed: Document pinconf in devicetree bindings
pinctrl: Add st,stm32f469-pinctrl compatible to stm32-pinctrl
pinctrl: stm32: Add STM32F469 MCU support
Documentation: dt: Remove ngpios from stm32-pinctrl binding
pinctrl: stm32: replace device_initcall() with arch_initcall()
pinctrl: stm32: add possibility to use gpio-ranges to declare bank range
pinctrl: armada-37xx: Add gpio support
pinctrl: armada-37xx: Add pin controller support for Armada 37xx
pinctrl: dt-bindings: Add documentation for Armada 37xx pin controllers
pinctrl: core: Make pinctrl_init_controller() static
pinctrl: generic: Add bi-directional and output-enable
...
The commit 1259feddd0f8("pinctrl: samsung: Fix the width of
PINCFG_TYPE_DRV bitfields for Exynos5433") already fixed
the different width of PINCFG_TYPE_DRV from previous Exynos SoC.
However wrong merge conflict resolution was chosen in commit
7f36f5d11c ("Merge tag 'v4.10-rc6' into devel") effectively dropping
the changes for PINCFG_TYPE_DRV. Re-do them here.
The macro EXYNOS_PIN_BANK_EINTW is no longer used so remove it.
Fixes: 7f36f5d11c ("Merge tag 'v4.10-rc6' into devel")
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This patch adds support for retention control for Exynos5433 SoCs. Three
groups of pins has been defined for retention control: common shared group
for ALIVE, CPIF, eSE, FINGER, IMEM, NFC, PERIC, TOUCH pin banks and
separate control for FSYS and AUD pin banks, for which PMU retention
registers match whole banks.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
When pin controller device is a part of power domain, there is no guarantee
that the power domain was not turned off and then on during boot process
before probing of the pin control driver. If it happened, then pin control
driver should ensure that pad retention is turned off during its probe call.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
This patch moves pad retention control from S5PV210 machine code to
Exynos pin controller driver. This helps to avoid possible ordering
and logical dependencies between machine and pin control code. Till
now it worked fine only because sys_ops for machine code and pin
controller were called in registration order.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
For mach-s5pv210:
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This patch moves pad retention control from PMU driver to Exynos pin
controller driver. This helps to avoid possible ordering and logical
dependencies between machine, PMU and pin control code. Till now it
worked fine only because sys_ops for PMU and pin controller were called
in registration order.
This is also a preparation for adding new features to Exynos pin
controller driver, like runtime power management and suspending
individual pin controllers, which might be a part of some power domain.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Exynos5433 support has been added in parallel to adding initconst
annotation to most of the init data structures, so add those
annotations also to Exynos5433 structures.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Memory subsystem already prints message about failed memory
allocation, there is no need to do it in the drivers.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
