mbox series

[v2,0/4] i2c: Replace lists of special clients with flagging of such clients

Message ID 3982b2a6-975e-40d2-bf02-2155e5c36c14@gmail.com
Headers show
Series i2c: Replace lists of special clients with flagging of such clients | expand

Message

Heiner Kallweit Aug. 22, 2024, 7:45 p.m. UTC
So far lists are used to track special clients, i.e. auto-detected and
userspace-created clients. The same functionality can be achieved much
simpler by flagging such clients.

v2:
- The i2c_driver.clients list is core-internal, however there's an ALSA
  driver using it. So add patch 1 to address this first.

Heiner Kallweit (4):
  ALSA: ppc: Remove i2c client removal hack
  i2c: Replace list-based mechanism for handling auto-detected clients
  i2c: Replace list-based mechanism for handling userspace-created
    clients
  i2c: core: Remove obsolete members of i2c_adapter and i2c_client

 drivers/i2c/i2c-core-base.c | 108 +++++++++++-------------------------
 include/linux/i2c.h         |  10 +---
 sound/ppc/keywest.c         |   7 +--
 3 files changed, 36 insertions(+), 89 deletions(-)

Comments

Wolfram Sang Sept. 1, 2024, 9:30 a.m. UTC | #1
On Thu, Aug 22, 2024 at 09:45:37PM +0200, Heiner Kallweit wrote:
> So far lists are used to track special clients, i.e. auto-detected and
> userspace-created clients. The same functionality can be achieved much
> simpler by flagging such clients.

This looks promising and I like the idea from a high-level perspective.
Need to dive into the details. However, I think this is 6.13 material. I
want to let it cook in linux-next for a full cycle.
Heiner Kallweit Sept. 1, 2024, 8:08 p.m. UTC | #2
On 01.09.2024 11:30, Wolfram Sang wrote:
> On Thu, Aug 22, 2024 at 09:45:37PM +0200, Heiner Kallweit wrote:
>> So far lists are used to track special clients, i.e. auto-detected and
>> userspace-created clients. The same functionality can be achieved much
>> simpler by flagging such clients.
> 
> This looks promising and I like the idea from a high-level perspective.
> Need to dive into the details. However, I think this is 6.13 material. I
> want to let it cook in linux-next for a full cycle.
> 
Fine with me.
Heiner Kallweit Sept. 30, 2024, 10:10 a.m. UTC | #3
On 01.09.2024 11:30, Wolfram Sang wrote:
> On Thu, Aug 22, 2024 at 09:45:37PM +0200, Heiner Kallweit wrote:
>> So far lists are used to track special clients, i.e. auto-detected and
>> userspace-created clients. The same functionality can be achieved much
>> simpler by flagging such clients.
> 
> This looks promising and I like the idea from a high-level perspective.
> Need to dive into the details. However, I think this is 6.13 material. I
> want to let it cook in linux-next for a full cycle.
> 
Now that 6.12-rc1 is out: Are you going to push this to linux-next?
Heiner Kallweit Oct. 8, 2024, 11:31 a.m. UTC | #4
On 08.10.2024 10:43, Wolfram Sang wrote:
> 
>> Yes, hopefully this week.
> 
> Reviewing and testing now. What tests did you run?
> 
Of course compile-tested the series. Functional test cases:
- Create device from userspace
- Delete device from userspace
- Delete adapter (by removing adapter driver module) and check that
  userspace-created client device has been auto-removed

Driver auto-detection I couldn't check due to missing hw.
Heiner Kallweit Oct. 29, 2024, 6:44 a.m. UTC | #5
On 30.09.2024 12:59, Wolfram Sang wrote:
> 
>> Now that 6.12-rc1 is out: Are you going to push this to linux-next?
> 
> Yes, hopefully this week.
> 
Now we're at rc5. Is it still something for 6.13 or better postpone
to have a full cycle in linux-next?