[PATCH RESEND v3 0/4] ASoC: fsl: fsl_qmc_audio: Reduce amount of interrupts

[Christophe Leroy]

Le 18/09/2025 à 17:34, Christophe Leroy a écrit :
> This is a RESEND of v3 sent one month ago, see:
> https://lore.kernel.org/all/cover.1754993232.git.christophe.leroy@csgroup.eu/

Sorry for the noise, v3 was 
https://lore.kernel.org/all/cover.1755504428.git.christophe.leroy@csgroup.eu/
The resent is correct though, only this reference in the cover is wrong.

Christophe

> 
> This series reduces significantly the amount of interrupts on
> fsl_qmc_audio device.
> 
> Patches 1 and 2 are preparatory patches.
> Patch 3 is the main change
> Patch 4 is a cleanup which is enabled by previous patch
> 
> Changes in v3:
> - Properly check the buffer descriptor is unused (Patch 1, comment from Herve Codina)
> - Fixed copy/paste error (patch 2, comment from Herve Codina)
> - Fixed build failure (patch 2, comment from Herve Codina and Test robot)
> 
> Changes in v2:
> - Don't remove UB bit (Patch 1, comment from Herve Codina)
> - Make sure audio channels are ordered on TDM bus (Patch 2, new patch, comment from Herve Codina)
> - Drop struct qmc_dai_chan  (patch 4, new patch)
> 
> Backgroup (copied from patch 3):
> 
> In non-interleaved mode, several QMC channels are used in sync.
> More details can be found in commit 188d9cae5438 ("ASoC: fsl:
> fsl_qmc_audio: Add support for non-interleaved mode.")
> At the time being, an interrupt is requested on each channel to
> perform capture/playback completion, allthough the completion is
> really performed only once all channels have completed their work.
> 
> This leads to a lot more interrupts than really needed. Looking at
> /proc/interrupts shows ~3800 interrupts per second when using
> 4 capture and 4 playback devices with 5ms periods while
> only 1600 (200 x 4 + 200 x 4) periods are processed during one second.
> 
> The QMC channels work in sync, the one started first is the one
> finishing first and the one started last is the one finishing last,
> so when the last one finishes it is guaranteed that the other ones are
> finished as well. Therefore only request completion processing on the
> last QMC channel.
> 
> On my board with the above exemple, on a kernel started with
> 'threadirqs' option, the QMC irq thread uses 16% CPU time with this
> patch while it uses 26% CPU time without this patch.
> 
> Christophe Leroy (4):
>    soc: fsl: qmc: Only set completion interrupt when needed
>    ASoc: fsl: fsl_qmc_audio: Ensure audio channels are ordered in TDM bus
>    ASoC: fsl: fsl_qmc_audio: Only request completion on last channel
>    ASoc: fsl: fsl_qmc_audio: Drop struct qmc_dai_chan
> 
>   drivers/soc/fsl/qe/qmc.c      |  44 +++++++++---
>   sound/soc/fsl/fsl_qmc_audio.c | 125 +++++++++++++++-------------------
>   2 files changed, 87 insertions(+), 82 deletions(-)
>