mbox series

[v2,0/7] ALSA: hda: Improvements to hda_component

Message ID 20240603103524.32442-1-simont@opensource.cirrus.com
Headers show
Series ALSA: hda: Improvements to hda_component | expand

Message

Simon Trimmer June 3, 2024, 10:35 a.m. UTC
This series of patches makes sure that the existing consumers of
the infrastructure unbind their interface before they begin
destroying the driver, moves duplicated members from the
instanced component structures into a new parent structure and
introduces locking so that consumers of the interface do not use
stale data.

v2:
 - Correct application of sizeof to pointer flagged by kernel test robot

Simon Trimmer (7):
  ALSA: hda: cs35l56: Component should be unbound before deconstruction
  ALSA: hda: cs35l41: Component should be unbound before deconstruction
  ALSA: hda/tas2781: Component should be unbound before deconstruction
  ALSA: hda: hda_component: Introduce component parent structure
  ALSA: hda: hda_component: Change codecs to use component parent
    structure
  ALSA: hda: hda_component: Move codec field into the parent
  ALSA: hda: hda_component: Protect shared data with a mutex

 sound/pci/hda/cs35l41_hda.c     | 47 ++++++++++++---------
 sound/pci/hda/cs35l56_hda.c     | 29 +++++++------
 sound/pci/hda/hda_component.c   | 75 ++++++++++++++++++++-------------
 sound/pci/hda/hda_component.h   | 48 ++++++++++++++-------
 sound/pci/hda/patch_realtek.c   | 17 ++++----
 sound/pci/hda/tas2781_hda_i2c.c | 37 ++++++++--------
 6 files changed, 147 insertions(+), 106 deletions(-)

Comments

Takashi Iwai June 13, 2024, 12:06 p.m. UTC | #1
On Mon, 03 Jun 2024 12:35:17 +0200,
Simon Trimmer wrote:
> 
> This series of patches makes sure that the existing consumers of
> the infrastructure unbind their interface before they begin
> destroying the driver, moves duplicated members from the
> instanced component structures into a new parent structure and
> introduces locking so that consumers of the interface do not use
> stale data.
> 
> v2:
>  - Correct application of sizeof to pointer flagged by kernel test robot
> 
> Simon Trimmer (7):
>   ALSA: hda: cs35l56: Component should be unbound before deconstruction
>   ALSA: hda: cs35l41: Component should be unbound before deconstruction
>   ALSA: hda/tas2781: Component should be unbound before deconstruction
>   ALSA: hda: hda_component: Introduce component parent structure
>   ALSA: hda: hda_component: Change codecs to use component parent
>     structure
>   ALSA: hda: hda_component: Move codec field into the parent
>   ALSA: hda: hda_component: Protect shared data with a mutex

The first three patches look rather like independent fixes.
Could you split those out and add proper Fixes tags, so that stable
trees can pick up?

The rest are a code refactoring and additional protection of the
mutex.


thanks,

Takashi

>  sound/pci/hda/cs35l41_hda.c     | 47 ++++++++++++---------
>  sound/pci/hda/cs35l56_hda.c     | 29 +++++++------
>  sound/pci/hda/hda_component.c   | 75 ++++++++++++++++++++-------------
>  sound/pci/hda/hda_component.h   | 48 ++++++++++++++-------
>  sound/pci/hda/patch_realtek.c   | 17 ++++----
>  sound/pci/hda/tas2781_hda_i2c.c | 37 ++++++++--------
>  6 files changed, 147 insertions(+), 106 deletions(-)
> 
> -- 
> 2.34.1
>