| Message ID | 20210122070634.12825-1-yung-chuan.liao@linux.intel.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | soundwire/regmap: use _no_pm routines | expand |
On 22-01-21, 15:06, Bard Liao wrote: > When a Slave device is resumed, it may resume the bus and restart the > enumeration. And Slave drivers will wait for initialization_complete > complete in their resume function, however initialization_complete will > complete after sdw_update_slave_status function is finished and codec > driver usually call some IO functions in the update_status callback > function. > It will become a deadlock if we use regular read/write routines during > the resuming process. Applied all sdw patches, thanks
On Sat, Feb 06, 2021 at 03:56:44PM +0530, Vinod Koul wrote:
> Applied all sdw patches, thanks
Is there a tag I can pull the new APIs from?
Hello Mark, On 08-02-21, 12:50, Mark Brown wrote: > On Sat, Feb 06, 2021 at 03:56:44PM +0530, Vinod Koul wrote: > > > Applied all sdw patches, thanks > > Is there a tag I can pull the new APIs from? Yes, please pull from below: git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire.git tags/soundwire-5.12-rc1 Thanks
On Mon, Feb 08, 2021 at 08:34:49PM +0530, Vinod Koul wrote: > On 08-02-21, 12:50, Mark Brown wrote: > > Is there a tag I can pull the new APIs from? > Yes, please pull from below: > git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire.git tags/soundwire-5.12-rc1 Ah, that's the entire Soundwire pull request including driver updates - that's a bit more than is ideal. Instead if you want to pull in the regmap bits that might be easier: Acked-by: Mark Brown <broonie@kernel.org>
On 10-02-21, 19:58, Mark Brown wrote: > On Mon, Feb 08, 2021 at 08:34:49PM +0530, Vinod Koul wrote: > > On 08-02-21, 12:50, Mark Brown wrote: > > > > Is there a tag I can pull the new APIs from? > > > Yes, please pull from below: > > > git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire.git tags/soundwire-5.12-rc1 > > Ah, that's the entire Soundwire pull request including driver updates - > that's a bit more than is ideal. Instead if you want to pull in the > regmap bits that might be easier: > > Acked-by: Mark Brown <broonie@kernel.org> Since I had a fix to send to Greg, I am adding these as well.. Thanks
When a Slave device is resumed, it may resume the bus and restart the enumeration. And Slave drivers will wait for initialization_complete complete in their resume function, however initialization_complete will complete after sdw_update_slave_status function is finished and codec driver usually call some IO functions in the update_status callback function. It will become a deadlock if we use regular read/write routines during the resuming process. This series touches both soundwire and regmap trees. commit fb5103f9d6ce ("regmap/SoundWire: sdw: add support for SoundWire 1.2 MBQ") is needed for soundwire tree to complie. On the other hands, commit 6e06a85556f9 ("soundwire: bus: add comments to explain interrupt loop filter") to commit 47b8520997a8 ("soundwire: bus: only clear valid DPN interrupts") are needed for regmap tree. v2: - Separate commits according to maintainer's comments. Bard Liao (4): soundwire: export sdw_write/read_no_pm functions regmap: sdw: use _no_pm functions in regmap_read/write regmap: sdw: use no_pm routines for SoundWire 1.2 MBQ regmap: sdw-mbq: use MODULE_LICENSE("GPL") Pierre-Louis Bossart (5): soundwire: bus: use sdw_update_no_pm when initializing a device soundwire: bus: use sdw_write_no_pm when setting the bus scale registers soundwire: bus: use no_pm IO routines for all interrupt handling soundwire: bus: fix confusion on device used by pm_runtime soundwire: bus: clarify dev_err/dbg device references drivers/base/regmap/regmap-sdw-mbq.c | 10 +- drivers/base/regmap/regmap-sdw.c | 4 +- drivers/soundwire/bus.c | 136 +++++++++++++++------------ include/linux/soundwire/sdw.h | 2 + 4 files changed, 85 insertions(+), 67 deletions(-)