mbox series

[v3,0/2] mmc: meson-gx: add SDIO interrupt support

Message ID 72459a46-fd20-60d8-a7e7-076d8f321816@gmail.com (mailing list archive)
Headers show
Series mmc: meson-gx: add SDIO interrupt support | expand

Message

Heiner Kallweit Aug. 25, 2022, 7:52 p.m. UTC
Add SDIO interrupt support. Successfully tested on a S905X4-based
system (V3 register layout) with a BRCM4334 SDIO wifi module
(brcmfmac driver).

v2:
- use new SDIO interrupt API
v3:
- don't duplicate checking mmc->sdio_irq_pending

Heiner Kallweit (2):
  mmc: meson: adjust and re-use constant IRQ_EN_MASK
  mmc: meson-gx: add SDIO interrupt support

 drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
 1 file changed, 65 insertions(+), 19 deletions(-)

Comments

Ulf Hansson Aug. 26, 2022, 9:16 a.m. UTC | #1
On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
>
> Add SDIO interrupt support. Successfully tested on a S905X4-based
> system (V3 register layout) with a BRCM4334 SDIO wifi module
> (brcmfmac driver).
>
> v2:
> - use new SDIO interrupt API
> v3:
> - don't duplicate checking mmc->sdio_irq_pending
>
> Heiner Kallweit (2):
>   mmc: meson: adjust and re-use constant IRQ_EN_MASK
>   mmc: meson-gx: add SDIO interrupt support
>
>  drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
>  1 file changed, 65 insertions(+), 19 deletions(-)
>

Applied for next, thanks!

Kind regards
Uffe
Neil Armstrong Aug. 29, 2022, 7:42 a.m. UTC | #2
Hi Uffe, Heiner,

On 26/08/2022 11:16, Ulf Hansson wrote:
> On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> Add SDIO interrupt support. Successfully tested on a S905X4-based
>> system (V3 register layout) with a BRCM4334 SDIO wifi module
>> (brcmfmac driver).
>>
>> v2:
>> - use new SDIO interrupt API
>> v3:
>> - don't duplicate checking mmc->sdio_irq_pending
>>
>> Heiner Kallweit (2):
>>    mmc: meson: adjust and re-use constant IRQ_EN_MASK
>>    mmc: meson-gx: add SDIO interrupt support
>>
>>   drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
>>   1 file changed, 65 insertions(+), 19 deletions(-)
>>
> 
> Applied for next, thanks!

Thanks for applying, but I was off (like many of reviewers) and I would have like
this patchset to be tested on earlier SoCs aswell, or only enabled on S4 for now.

Neil

> 
> Kind regards
> Uffe
Ulf Hansson Aug. 29, 2022, 12:07 p.m. UTC | #3
On Mon, 29 Aug 2022 at 09:42, Neil Armstrong <narmstrong@baylibre.com> wrote:
>
> Hi Uffe, Heiner,
>
> On 26/08/2022 11:16, Ulf Hansson wrote:
> > On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
> >>
> >> Add SDIO interrupt support. Successfully tested on a S905X4-based
> >> system (V3 register layout) with a BRCM4334 SDIO wifi module
> >> (brcmfmac driver).
> >>
> >> v2:
> >> - use new SDIO interrupt API
> >> v3:
> >> - don't duplicate checking mmc->sdio_irq_pending
> >>
> >> Heiner Kallweit (2):
> >>    mmc: meson: adjust and re-use constant IRQ_EN_MASK
> >>    mmc: meson-gx: add SDIO interrupt support
> >>
> >>   drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
> >>   1 file changed, 65 insertions(+), 19 deletions(-)
> >>
> >
> > Applied for next, thanks!
>
> Thanks for applying, but I was off (like many of reviewers) and I would have like
> this patchset to be tested on earlier SoCs aswell, or only enabled on S4 for now.

To actually make use of the new code, an update to the corresponding
DTS files are needed to set the cap-sdio-irq property.

If that isn't sufficient, I prefer if we make a patch on top of the
$subject series.

Kind regards
Uffe
Neil Armstrong Aug. 29, 2022, 1:29 p.m. UTC | #4
On 29/08/2022 14:07, Ulf Hansson wrote:
> On Mon, 29 Aug 2022 at 09:42, Neil Armstrong <narmstrong@baylibre.com> wrote:
>>
>> Hi Uffe, Heiner,
>>
>> On 26/08/2022 11:16, Ulf Hansson wrote:
>>> On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>>>
>>>> Add SDIO interrupt support. Successfully tested on a S905X4-based
>>>> system (V3 register layout) with a BRCM4334 SDIO wifi module
>>>> (brcmfmac driver).
>>>>
>>>> v2:
>>>> - use new SDIO interrupt API
>>>> v3:
>>>> - don't duplicate checking mmc->sdio_irq_pending
>>>>
>>>> Heiner Kallweit (2):
>>>>     mmc: meson: adjust and re-use constant IRQ_EN_MASK
>>>>     mmc: meson-gx: add SDIO interrupt support
>>>>
>>>>    drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
>>>>    1 file changed, 65 insertions(+), 19 deletions(-)
>>>>
>>>
>>> Applied for next, thanks!
>>
>> Thanks for applying, but I was off (like many of reviewers) and I would have like
>> this patchset to be tested on earlier SoCs aswell, or only enabled on S4 for now.
> 
> To actually make use of the new code, an update to the corresponding
> DTS files are needed to set the cap-sdio-irq property.

I know but it changed the IRQ handling, let's hope we don't see any regressions.

> 
> If that isn't sufficient, I prefer if we make a patch on top of the
> $subject series.

OK, let's wait and see

> 
> Kind regards
> Uffe

Thanks,
Neil
Ulf Hansson Aug. 29, 2022, 2:21 p.m. UTC | #5
On Mon, 29 Aug 2022 at 15:29, Neil Armstrong <narmstrong@baylibre.com> wrote:
>
> On 29/08/2022 14:07, Ulf Hansson wrote:
> > On Mon, 29 Aug 2022 at 09:42, Neil Armstrong <narmstrong@baylibre.com> wrote:
> >>
> >> Hi Uffe, Heiner,
> >>
> >> On 26/08/2022 11:16, Ulf Hansson wrote:
> >>> On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
> >>>>
> >>>> Add SDIO interrupt support. Successfully tested on a S905X4-based
> >>>> system (V3 register layout) with a BRCM4334 SDIO wifi module
> >>>> (brcmfmac driver).
> >>>>
> >>>> v2:
> >>>> - use new SDIO interrupt API
> >>>> v3:
> >>>> - don't duplicate checking mmc->sdio_irq_pending
> >>>>
> >>>> Heiner Kallweit (2):
> >>>>     mmc: meson: adjust and re-use constant IRQ_EN_MASK
> >>>>     mmc: meson-gx: add SDIO interrupt support
> >>>>
> >>>>    drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
> >>>>    1 file changed, 65 insertions(+), 19 deletions(-)
> >>>>
> >>>
> >>> Applied for next, thanks!
> >>
> >> Thanks for applying, but I was off (like many of reviewers) and I would have like
> >> this patchset to be tested on earlier SoCs aswell, or only enabled on S4 for now.
> >
> > To actually make use of the new code, an update to the corresponding
> > DTS files are needed to set the cap-sdio-irq property.
>
> I know but it changed the IRQ handling, let's hope we don't see any regressions.

Right. Let's encourage people to test this!

I noticed that $subject patch sets the CLK_IRQ_SDIO_SLEEP bit, no
matter whether the SDIO IRQs has been enabled. Don't know if that is
intentional and/or could be a problem? Heiner?

[...]

Kind regards
Uffe
Heiner Kallweit Aug. 29, 2022, 6:23 p.m. UTC | #6
On 29.08.2022 16:21, Ulf Hansson wrote:
> On Mon, 29 Aug 2022 at 15:29, Neil Armstrong <narmstrong@baylibre.com> wrote:
>>
>> On 29/08/2022 14:07, Ulf Hansson wrote:
>>> On Mon, 29 Aug 2022 at 09:42, Neil Armstrong <narmstrong@baylibre.com> wrote:
>>>>
>>>> Hi Uffe, Heiner,
>>>>
>>>> On 26/08/2022 11:16, Ulf Hansson wrote:
>>>>> On Thu, 25 Aug 2022 at 21:52, Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>>>>>
>>>>>> Add SDIO interrupt support. Successfully tested on a S905X4-based
>>>>>> system (V3 register layout) with a BRCM4334 SDIO wifi module
>>>>>> (brcmfmac driver).
>>>>>>
>>>>>> v2:
>>>>>> - use new SDIO interrupt API
>>>>>> v3:
>>>>>> - don't duplicate checking mmc->sdio_irq_pending
>>>>>>
>>>>>> Heiner Kallweit (2):
>>>>>>     mmc: meson: adjust and re-use constant IRQ_EN_MASK
>>>>>>     mmc: meson-gx: add SDIO interrupt support
>>>>>>
>>>>>>    drivers/mmc/host/meson-gx-mmc.c | 84 +++++++++++++++++++++++++--------
>>>>>>    1 file changed, 65 insertions(+), 19 deletions(-)
>>>>>>
>>>>>
>>>>> Applied for next, thanks!
>>>>
>>>> Thanks for applying, but I was off (like many of reviewers) and I would have like
>>>> this patchset to be tested on earlier SoCs aswell, or only enabled on S4 for now.
>>>
>>> To actually make use of the new code, an update to the corresponding
>>> DTS files are needed to set the cap-sdio-irq property.
>>
>> I know but it changed the IRQ handling, let's hope we don't see any regressions.
> 
> Right. Let's encourage people to test this!
> 
> I noticed that $subject patch sets the CLK_IRQ_SDIO_SLEEP bit, no
> matter whether the SDIO IRQs has been enabled. Don't know if that is
> intentional and/or could be a problem? Heiner?
> 
This bit enables SDIO interrupts in sleep mode too. I experienced no difference
in behavior if SDIO interrupts aren't enabled.

> [...]
> 
> Kind regards
> Uffe