[00/10] i2c: move handling of suspended adapters to the core

Message

Wolfram Sang Dec. 19, 2018, 4:48 p.m. UTC

Finally, here is the implementation Hans and I agreed on. Plus, all potential
users I could spot already converted. Renesas R-Car driver was added on top.
This series was tested on a Renesas Lager board (R-Car H2). I had to hack some
error cases into the code to verify the workings. Thanks for all the tests so
far. Of course, more testing never hurts ;)

Please comment, review, test... a branch can be found here:

git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git renesas/i2c/core-pm-helpers

Thanks,

   Wolfram

Changes since RFC:

* don't use bitfields anymore but an unsigned long flags variable which is only
  meant for flags to be changed when the adapter lock is held

* split the genereic accessor into _mark_suspended() and _mark_resumed()

* added kerneldoc and noted that using these helpers is optional

* documented -ESHUTDOWN as the error code when trying to transfer using an
  already suspended adapter

* added acks from last series. Changes were minor enough to keep them. Please let
  me know if you disagree.


Wolfram Sang (10):
  i2c: add suspended flag and accessors for i2c adapters
  i2c: reject new transfers when adapters are suspended
  i2c: synquacer: remove unused is_suspended flag
  i2c: brcmstb: use core helper to mark adapter suspended
  i2c: zx2967: use core helper to mark adapter suspended
  i2c: sprd: don't use pdev as variable name for struct device *
  i2c: sprd: use core helper to mark adapter suspended
  i2c: exynos5: use core helper to mark adapter suspended
  i2c: s3c2410: use core helper to mark adapter suspended
  i2c: rcar: add suspend/resume support

 Documentation/i2c/fault-codes      |  4 ++++
 drivers/i2c/busses/i2c-brcmstb.c   | 13 ++-----------
 drivers/i2c/busses/i2c-exynos5.c   | 11 ++---------
 drivers/i2c/busses/i2c-rcar.c      | 25 +++++++++++++++++++++++++
 drivers/i2c/busses/i2c-s3c2410.c   |  8 ++------
 drivers/i2c/busses/i2c-sprd.c      | 34 ++++++++++++----------------------
 drivers/i2c/busses/i2c-synquacer.c |  5 -----
 drivers/i2c/busses/i2c-zx2967.c    |  8 ++------
 drivers/i2c/i2c-core-base.c        |  3 +++
 include/linux/i2c.h                | 34 ++++++++++++++++++++++++++++++++++
 10 files changed, 86 insertions(+), 59 deletions(-)

Comments

Wolfram Sang Jan. 8, 2019, 8:09 p.m. UTC | #1

On Wed, Dec 19, 2018 at 05:48:16PM +0100, Wolfram Sang wrote:
> Finally, here is the implementation Hans and I agreed on. Plus, all potential
> users I could spot already converted. Renesas R-Car driver was added on top.
> This series was tested on a Renesas Lager board (R-Car H2). I had to hack some
> error cases into the code to verify the workings. Thanks for all the tests so
> far. Of course, more testing never hurts ;)
> 
> Please comment, review, test... a branch can be found here:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git renesas/i2c/core-pm-helpers
> 
> Thanks,
> 
>    Wolfram
> 
> Changes since RFC:
> 
> * don't use bitfields anymore but an unsigned long flags variable which is only
>   meant for flags to be changed when the adapter lock is held
> 
> * split the genereic accessor into _mark_suspended() and _mark_resumed()
> 
> * added kerneldoc and noted that using these helpers is optional
> 
> * documented -ESHUTDOWN as the error code when trying to transfer using an
>   already suspended adapter
> 
> * added acks from last series. Changes were minor enough to keep them. Please let
>   me know if you disagree.
> 
> 
> Wolfram Sang (10):
>   i2c: add suspended flag and accessors for i2c adapters
>   i2c: reject new transfers when adapters are suspended
>   i2c: synquacer: remove unused is_suspended flag
>   i2c: brcmstb: use core helper to mark adapter suspended
>   i2c: zx2967: use core helper to mark adapter suspended
>   i2c: sprd: don't use pdev as variable name for struct device *
>   i2c: sprd: use core helper to mark adapter suspended
>   i2c: exynos5: use core helper to mark adapter suspended
>   i2c: s3c2410: use core helper to mark adapter suspended
>   i2c: rcar: add suspend/resume support

After a long road, I applied this series now to for-next. I only
squashed patches 1+2 to have the functionality in one place. And fixed a
typo in patch 1. Thanks to the testers and reviewers, and especially
Hans, for the input and discussions.

This will now get a full test cycle in -next. Let's see what happens...

And let's move on to how to add 'very late transfers without irqs' on
top of this series.

Happy hacking,

   Wolfram