kernel/linux.git/drivers/base/regmap, branch linux-6.0.y

regmap-irq: Use the new num_config_regs property in regmap_add_irq_chip_fwnode

2022-12-31T12:25:56+00:00

[ Upstream commit 84498d1fb35de6ab71bdfdb6270a464fb4a0951b ] Commit faa87ce9196d ("regmap-irq: Introduce config registers for irq types") added the num_config_regs, then commit 9edd4f5aee84 ("regmap-irq: Deprecate type registers and virtual registers") suggested to replace num_type_reg with it. However, regmap_add_irq_chip_fwnode wasn't modified to use the new property. Later on, commit 255a03bb1bb3 ("ASoC: wcd9335: Convert irq chip to config regs") removed the old num_type_reg property from the WCD9335 driver's struct regmap_irq_chip, causing a null pointer dereference in regmap_irq_set_type when it tried to index d->type_buf as it was never allocated in regmap_add_irq_chip_fwnode: [ 39.199374] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000 [ 39.200006] Call trace: [ 39.200014] regmap_irq_set_type+0x84/0x1c0 [ 39.200026] __irq_set_trigger+0x60/0x1c0 [ 39.200040] __setup_irq+0x2f4/0x78c [ 39.200051] request_threaded_irq+0xe8/0x1a0 Use num_config_regs in regmap_add_irq_chip_fwnode instead of num_type_reg, and fall back to it if num_config_regs isn't defined to maintain backward compatibility. Fixes: faa87ce9196d ("regmap-irq: Introduce config registers for irq types") Signed-off-by: Yassine Oudjana Link: https://lore.kernel.org/r/20221107202114.823975-1-y.oudjana@protonmail.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

regmap: spi: Reserve space for register address/padding

2022-08-18T14:02:05+00:00

Currently the max_raw_read and max_raw_write limits in regmap_spi struct do not take into account the additional size of the transmitted register address and padding. This may result in exceeding the maximum permitted SPI message size, which could cause undefined behaviour, e.g. data corruption. Fix regmap_get_spi_bus() to properly adjust the above mentioned limits by reserving space for the register address/padding as set in the regmap configuration. Fixes: f231ff38b7b2 ("regmap: spi: Set regmap max raw r/w from max_transfer_size") Signed-off-by: Cristian Ciocaltea Reviewed-by: Lucas Tanure Link: https://lore.kernel.org/r/20220818104851.429479-1-cristian.ciocaltea@collabora.com Signed-off-by: Mark Brown

regmap: permit to set reg_update_bits with bulk implementation

2022-07-18T12:45:28+00:00

A regmap may still require to set a custom reg_update_bits instead of relying to the regmap_bus_read/write general function. Permit to set it in the map if provided by the regmap config. Signed-off-by: Christian Marangi Link: https://lore.kernel.org/r/20220715201032.19507-1-ansuelsmth@gmail.com Signed-off-by: Mark Brown

regmap: add WARN_ONCE when invalid mask is provided to regmap_field_init()

2022-07-08T10:51:56+00:00

In regmap_field_init() when a invalid mask is provided it still initializes with any warnings. An example of this is when the LSB is greater than MSB a mask of zero is produced. WARN_ONCE() is not ideal for this but requires less changes to core regmap code. Cc: Mark Brown Cc: Nishanth Menon Signed-off-by: Matt Ranostay Link: https://lore.kernel.org/r/20220708013125.313892-1-mranostay@ti.com Signed-off-by: Mark Brown

regmap-irq: Fix bug in regmap_irq_get_irq_reg_linear()

2022-07-05T12:00:55+00:00

irq_reg_stride in struct regmap_irq_chip is often 0, but that actually means to use the default stride of 1. The effective stride is stored in struct regmap_irq_chip_data->irq_reg_stride and will get the corrected default value. The default ->get_irq_reg() callback was using the stride from the chip definition, which is wrong; fix it to use the effective stride from the chip data instead. Link: https://lore.kernel.org/lkml/acaaf77f-3282-8544-dd3c-7915fc1a6a4f@samsung.com/ Signed-off-by: Aidan MacDonald Link: https://lore.kernel.org/r/20220704112847.23844-1-aidanmacdonald.0x0@gmail.com Signed-off-by: Mark Brown

regmap-irq cleanups and refactoring

2022-06-30T17:26:37+00:00

Merge series from Aidan MacDonald : This series is an attempt at cleaning up the regmap-irq API in order to simplify things and consolidate existing features, while at the same time generalizing it to support a wider range of hardware. There is a new system for IRQ type configuration, some tweaks to unmask registers so they're more intuitive and useful, and a new callback for calculating register addresses. There's also a few minor code cleanups in here. In v2 I've taken the approach of adding new features and deprecating existing ones rather than removing them aggressively. Warnings will be issued for any drivers that use deprecated features, but they'll otherwise continue to function normally. One important caveat: not all of these changes are tested beyond compile testing, since I don't have hardware to exercise all of the features.

regmap: cache: Add extra parameter check in regcache_init

2022-06-30T12:33:21+00:00

When num_reg_defaults > 0 but reg_defaults is NULL, there will be a NULL pointer exception. Current code has no such usage, but as additional hardening, also check this to prevent any chance of crashing. Signed-off-by: Schspa Shi Link: https://lore.kernel.org/r/20220629130951.63040-1-schspa@gmail.com Signed-off-by: Mark Brown

regmap-irq: Deprecate the not_fixed_stride flag

2022-06-29T17:13:17+00:00

This flag is a bit of a hack and the same thing can be accomplished using a custom ->get_irq_reg() callback. Add a warning to catch any use of the flag. Signed-off-by: Aidan MacDonald Link: https://lore.kernel.org/r/20220623211420.918875-13-aidanmacdonald.0x0@gmail.com Signed-off-by: Mark Brown

regmap-irq: Add get_irq_reg() callback

2022-06-29T17:13:16+00:00

Replace the internal sub_irq_reg() function with a public callback that drivers can use when they have more complex register layouts. The default implementation is regmap_irq_get_irq_reg_linear(), used if the chip doesn't provide its own callback. Signed-off-by: Aidan MacDonald Link: https://lore.kernel.org/r/20220623211420.918875-12-aidanmacdonald.0x0@gmail.com Signed-off-by: Mark Brown

regmap-irq: Fix inverted handling of unmask registers

2022-06-29T17:13:15+00:00

To me "unmask" suggests that we write 1s to the register when an interrupt is enabled. This also makes sense because it's the opposite of what the "mask" register does (write 1s to disable an interrupt). But regmap-irq does the opposite: for a disabled interrupt, it writes 1s to "unmask" and 0s to "mask". This is surprising and deviates from the usual way mask registers are handled. Additionally, mask_invert didn't interact with unmask registers properly -- it caused them to be ignored entirely. Fix this by making mask and unmask registers orthogonal, using the following behavior: * Mask registers are written with 1s for disabled interrupts. * Unmask registers are written with 1s for enabled interrupts. This behavior supports both normal or inverted mask registers and separate set/clear registers via different combinations of mask_base/unmask_base. The old unmask register behavior is deprecated. Drivers need to opt-in to the new behavior by setting mask_unmask_non_inverted. Warnings are issued if the driver relies on deprecated behavior. Chips that only set one of mask_base/unmask_base don't have to use the mask_unmask_non_inverted flag because that use case was previously not supported. The mask_invert flag is also deprecated in favor of describing inverted mask registers as unmask registers. Signed-off-by: Aidan MacDonald Link: https://lore.kernel.org/r/20220623211420.918875-11-aidanmacdonald.0x0@gmail.com Signed-off-by: Mark Brown