[PATCH v2 0/2] add hooks for usb suspend and resume

Puma Hsu posted 2 patches 2 years, 9 months ago
drivers/usb/core/driver.c              | 36 +++++++++++++
drivers/usb/core/usb-hooks-impl-goog.c | 72 ++++++++++++++++++++++++++
drivers/usb/core/usb.h                 |  5 ++
3 files changed, 113 insertions(+)
create mode 100644 drivers/usb/core/usb-hooks-impl-goog.c
[PATCH v2 0/2] add hooks for usb suspend and resume
Posted by Puma Hsu 2 years, 9 months ago
In mobile, a co-processor can be used for USB audio. When the co-processor
is working for USB audio, the co-processor is the user/owner of the USB
driver, and the ACPU is able to sleep in such condition to improve power
consumption. In order to support this, we need to create hooks in suspend
and resume functions. We also upload our implementations for reviewing.

Puma Hsu (2):
  usb: core: add hooks for usb suspend and resume
  usb: core: add implementations for usb suspend/resume hooks

 drivers/usb/core/driver.c              | 36 +++++++++++++
 drivers/usb/core/usb-hooks-impl-goog.c | 72 ++++++++++++++++++++++++++
 drivers/usb/core/usb.h                 |  5 ++
 3 files changed, 113 insertions(+)
 create mode 100644 drivers/usb/core/usb-hooks-impl-goog.c

---
Changes in v2:
- Remove the wrong input in the Makefile
- Change description in commit message

-- 
2.39.0.rc1.256.g54fd8350bd-goog
Re: [PATCH v2 0/2] add hooks for usb suspend and resume
Posted by Oliver Neukum 2 years, 9 months ago

On 14.12.22 09:14, Puma Hsu wrote:
> In mobile, a co-processor can be used for USB audio. When the co-processor
> is working for USB audio, the co-processor is the user/owner of the USB
> driver, and the ACPU is able to sleep in such condition to improve power
> consumption. In order to support this, we need to create hooks in suspend
> and resume functions. We also upload our implementations for reviewing.

Ok, before this gets hopelessly unproductive, please describe what
you are aiming at and operating on in greater detail.

It looks to me like you have an audio device that is connected
to the host by USB _and_ another bus. Is that correct?
Will you submit the "subdriver" that drives the device over
that secondary bus?
The operation over the secondary bus requires a hook in USB
power management. Why?

Secondly, the naming is atrocious.

Now could we please first define the technical nature of the issue,
so that we fully understand it before we debate the desirability?

	Regards
		Oliver
Re: [PATCH v2 0/2] add hooks for usb suspend and resume
Posted by Greg KH 2 years, 9 months ago
On Wed, Dec 14, 2022 at 04:14:54PM +0800, Puma Hsu wrote:
> In mobile, a co-processor can be used for USB audio. When the co-processor
> is working for USB audio, the co-processor is the user/owner of the USB
> driver, and the ACPU is able to sleep in such condition to improve power
> consumption. In order to support this, we need to create hooks in suspend
> and resume functions. We also upload our implementations for reviewing.
> 
> Puma Hsu (2):
>   usb: core: add hooks for usb suspend and resume
>   usb: core: add implementations for usb suspend/resume hooks
> 
>  drivers/usb/core/driver.c              | 36 +++++++++++++
>  drivers/usb/core/usb-hooks-impl-goog.c | 72 ++++++++++++++++++++++++++
>  drivers/usb/core/usb.h                 |  5 ++
>  3 files changed, 113 insertions(+)
>  create mode 100644 drivers/usb/core/usb-hooks-impl-goog.c
> 
> ---
> Changes in v2:
> - Remove the wrong input in the Makefile
> - Change description in commit message
> 
> -- 
> 2.39.0.rc1.256.g54fd8350bd-goog
> 

My review comments on v1 still pertain here.