Re: [PATCHv1 0/6] leds: pca9653x: support inverted outputs and cleanups

[Olliver Schinagl]

Hey Ricardo,

On 19-04-16 13:18, Ricardo Ribalda Delgado wrote:
> Hi Ollivier
>
> Sorry to not reply to the patches, but I am not subscribed to linux-leds
>
> Regarding:
> [PATCH 2/6] leds: pca963x: Lock i2c r/w access
>
> I am not sure why this patch is needed. the only thing that should be
> protected is the write to ledout.
>
> It seems that mode2 needs to be set to PCA963X_MODE2_DMBLNK, or at
> least, the driver never clears that bit. Couldnt we just set it at
> probe time and remove the read/write of it? I do not have the hardware
> at the moment, so it should be something that you need to test.
Without actually looking at the code right now, but the driver does a 
read/modify/write on the register, and a register is shared among 
several leds. So in that regard, it makes sense and I don't think it's 
very expensive to move the lock, since we have to lock for the write a 
few lines down anyway.
> [PATCH 4/6] leds: pca963x: Reduce magic values
> Maybe you want to create the inverse macro of PCA963X_LEDOUT_LDR, so
> you can do something linke
>
> PCA963X_LEDOUT_LDR_INV(ledout, pca963x->led_num) != PCA963X_LEDOUT_LED_GRP_PWM
Good point, I'll add it, I like it.
> [PATCH 3/6] leds: pca963x: Add defines and remove some magic values
>
> I am not big fan of defining things that are not used. and the magic
> assigment to n_leds is perfectly fine IMHO
Well i needed some of the defines for the invert part and then I figured 
just add everything that the datasheet defines to make everything 
exlusive/easy to use.

But I can remove unused defines if desired.
> For PCA963X_LEDOUT_LDR.  Do not forget the parenthesis around led_num.
> Also replace %4 with &3 to be consisten.t
Yeah, i'll check and fix that.
> Regards!
>
> On Tue, Apr 19, 2016 at 11:39 AM, Olliver Schinagl <oliver@xxxxxxxxxxx> wrote:
> > On 19-04-16 11:23, Jacek Anaszewski wrote:
> > > Hi Olliver,
> > >
> > > Thanks for the patches.
> > > Adding driver authors on cc.
> > Ah sorry about that, thanks. I guess get_maintainers doesn't do that.
> >
> > As for the compile bug, I'll fix that with v2, it only applies on the
> > intermediate patches, not on the whole set.
> >
> > Olliver
> >
> > > On 04/19/2016 09:40 AM, Olliver Schinagl wrote:
> > > > Using the pca963x for a while, I noticed something that may look like
> > > > some
> > > > i2c accessing issues where sometimes data was incorrectly written to the
> > > > bus,
> > > > possibly because we where not properly locking the i2c reads. Though I'm
> > > > not
> > > > familiar enough with the i2c framework to be certain reads need to be
> > > > locked
> > > > at all. A patch was added to properly lock i2c access more tightly.
> > > >
> > > > Furthermore there was no method to support inverted outputs. This series
> > > > adds a property to the device tree to inform the driver that the output
> > > > is inverted (active-high vs active-low).
> > > >
> > > > Additionally, this patch set does some cleanups to please checkpatch, and
> > > > removes a few magic values.
> > > >
> > > > Olliver Schinagl (6):
> > > >     leds: pca963x: Alphabetize headers
> > > >     leds: pca963x: Lock i2c r/w access
> > > >     leds: pca963x: Add defines and remove some magic values
> > > >     leds: pca963x: Reduce magic values
> > > >     leds: pca963x: Inform the output that it is inverted
> > > >     leds: pca963x: Remove whitespace and checkpatch problems
> > > >
> > > >    Documentation/devicetree/bindings/leds/pca963x.txt |   1 +
> > > >    drivers/leds/leds-pca963x.c                        | 243
> > > > ++++++++++++++-------
> > > >    include/linux/platform_data/leds-pca963x.h         |   1 +
> > > >    3 files changed, 171 insertions(+), 74 deletions(-)