user/sven/linux.git/drivers/base/regmap, branch v6.15.1

Merge tag 'regmap-v6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap

2025-03-29T21:31:39Z

Pull regmap updates from Mark Brown: "Only a couple of small patches this release, one refactoring struct regmap to pack it more efficiently and another which makes our way of setting all bits consistent in the regmap-irq code" * tag 'regmap-v6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap: regmap: irq: Use one way of setting all bits in the register regmap: Reorder 'struct regmap'

regcache: Add support for sorting defaults arrays

2025-02-27T13:09:11Z

The defaults array in regcache must be sorted into ascending register address order, because binary search is used to locate values in the array. Add a helper to sort the register defaults array which can be useful for systems that dynamically create a defaults array based on external information. Signed-off-by: Charles Keepax Reviewed-by: Pierre-Louis Bossart Link: https://patch.msgid.link/20250217140159.2288784-2-ckeepax@opensource.cirrus.com Signed-off-by: Mark Brown

regmap: irq: Use one way of setting all bits in the register

2025-02-06T20:46:09Z

Currently there are two ways of how we represent all bits set, i.e. UINT_MAX and GENMASK(31, 0). Use the former as the single way of doing that, which is crystal clear on how we fill the unsigned int value. Signed-off-by: Andy Shevchenko Link: https://patch.msgid.link/20250206191644.1132869-1-andriy.shevchenko@linux.intel.com Signed-off-by: Mark Brown

regmap-irq: Add missing kfree()

2025-02-05T11:55:42Z

Add kfree() for "d->main_status_buf" to the error-handling path to prevent a memory leak. Fixes: a2d21848d921 ("regmap: regmap-irq: Add main status register support") Cc: stable@vger.kernel.org # v5.1+ Signed-off-by: Jiasheng Jiang Link: https://patch.msgid.link/20250205004343.14413-1-jiashengjiangcool@gmail.com Signed-off-by: Mark Brown

regmap: Reorder 'struct regmap'

2025-02-03T00:38:32Z

On a x86_64, with configured with allmodconfig, pahole states that the regmap structure is: /* size: 1048, cachelines: 17, members: 78 */ /* sum members: 1006, holes: 9, sum holes: 35 */ /* padding: 7 */ /* member types with holes: 2, total: 2 */ /* last cacheline: 24 bytes */ So, when such a struct is allocated, 2048 bytes are allocated, with most of this space being wasted. Move a few bools so that the size is reduced to 1024. After this change, pahole gives: /* size: 1024, cachelines: 16, members: 78 */ /* sum members: 1006, holes: 6, sum holes: 18 */ /* member types with holes: 2, total: 2 */ Signed-off-by: Christophe JAILLET Link: https://patch.msgid.link/f01f900d15633d5cda5f27763723acb307c0d22f.1737725820.git.christophe.jaillet@wanadoo.fr Signed-off-by: Mark Brown

Expand SoundWire MBQ register map support

2025-01-07T23:28:07Z

Merge series from Charles Keepax : The current SDCA MBQ (Multi-Byte Quantities) register map only supports 16-bit types, add support for more sizes and then update the rt722 driver to use the new support. We also add support for the deferring feature of MBQs to allow hardware to indicate it is not currently ready to service a read/write. Afraid I don't have hardware to test the rt722 change so it is only build tested, but I thought it good to include a change to demonstrate the new features in use.

regmap: sdw-mbq: Add support for SDCA deferred controls

2025-01-07T20:21:00Z

The SDCA specification allows for controls to be deferred. In the case of a deferred control the device will return COMMAND_IGNORED to the 8-bit operation that would cause the value to commit. Which is the final 8-bits on a write, or the first 8-bits on a read. In the case of receiving a defer, the regmap will poll the SDCA function busy bit, after which the transaction will be retried, returning an error if the function busy does not clear within a chip specific timeout. Since this is common SDCA functionality which is the 99% use-case for MBQs it makes sense to incorporate this functionality into the register map. If no MBQ configuration is specified, the behaviour will default to the existing behaviour. Signed-off-by: Charles Keepax Link: https://patch.msgid.link/20250107154408.814455-5-ckeepax@opensource.cirrus.com Reviewed-by: Pierre-Louis Bossart Signed-off-by: Mark Brown

regmap: sdw-mbq: Add support for further MBQ register sizes

2025-01-07T20:20:59Z

SoundWire MBQ register maps typically contain a variety of register sizes, which doesn't map ideally to the regmap abstraction which expects register maps to have a consistent size. Currently the MBQ register map only allows 16-bit registers to be defined, however this leads to complex CODEC driver implementations with an 8-bit register map and a 16-bit MBQ, every control will then have a custom get and put handler that allows them to access different register maps. Further more 32-bit MBQ quantities are not currently supported. Add support for additional MBQ sizes and to avoid the complexity of multiple register maps treat the val_size as a maximum size for the register map. Within the regmap use an ancillary callback to determine how many bytes to actually read/write to the hardware for a specific register. In the case that no callback is defined the behaviour defaults back to the existing behaviour of a fixed size register map. Signed-off-by: Charles Keepax Link: https://patch.msgid.link/20250107154408.814455-4-ckeepax@opensource.cirrus.com Reviewed-by: Pierre-Louis Bossart Signed-off-by: Mark Brown

regmap: regmap_multi_reg_read(): make register list const

2024-12-11T14:24:09Z

Mark the list of registers passed into regmap_multi_reg_read() as a pointer to const. This allows the caller to define the register list as const data. This requires making the same change to _regmap_bulk_read(), which is called by regmap_multi_reg_read(). Signed-off-by: Richard Fitzgerald Link: https://patch.msgid.link/20241211133558.884669-1-rf@opensource.cirrus.com Signed-off-by: Mark Brown

regmap: Merge up v6.12-rc2

2024-12-09T12:50:45Z

This has fixes for several boards which help my testing a lot.