From: Miquel Raynal <miquel.raynal@bootlin.com>
To: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Cc: Jonathan Cameron <jic23@kernel.org>,
Alexandru Ardelean <ardeleanalex@gmail.com>,
linux-iio <linux-iio@vger.kernel.org>,
Lars-Peter Clausen <lars@metafoo.de>,
Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
Olivier MOYSAN <olivier.moysan@foss.st.com>
Subject: Re: [PATCH 04/10] iio: adc: stm32-dfsdm: Avoid dereferencing ->currentmode
Date: Wed, 2 Feb 2022 10:33:59 +0100 [thread overview]
Message-ID: <20220202103359.3d4c5fb7@xps13> (raw)
In-Reply-To: <ff1ddd2b-acbb-3154-5712-87c1d9a7f8b7@foss.st.com>
Hi Fabrice,
fabrice.gasnier@foss.st.com wrote on Tue, 1 Feb 2022 09:41:03 +0100:
> On 1/28/22 4:04 PM, Miquel Raynal wrote:
> > Hi Jonathan,
> >
> > jic23@kernel.org wrote on Sat, 15 Jan 2022 16:06:19 +0000:
> >
> >> On Thu, 16 Dec 2021 09:22:35 +0100
> >> Miquel Raynal <miquel.raynal@bootlin.com> wrote:
> >>
> >>> Hi Alexandru,
> >>>
> >>> ardeleanalex@gmail.com wrote on Thu, 16 Dec 2021 08:47:02 +0200:
> >>>
> >>>> On Wed, Dec 15, 2021 at 10:03 PM Miquel Raynal
> >>>> <miquel.raynal@bootlin.com> wrote:
> >>>>>
> >>>>> This is an internal variable of the core, let's use the
> >>>>> iio_buffer_enabled() helper which is exported for the following purpose:
> >>>>> telling if the current mode is a buffered mode, which is precisely what
> >>>>> this driver looks for.
> >>>>>
> >>>>> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
> >>>>> ---
> >>>>> drivers/iio/adc/stm32-dfsdm-adc.c | 5 ++---
> >>>>> 1 file changed, 2 insertions(+), 3 deletions(-)
> >>>>>
> >>>>> diff --git a/drivers/iio/adc/stm32-dfsdm-adc.c b/drivers/iio/adc/stm32-dfsdm-adc.c
> >>>>> index 1cfefb3b5e56..a3b8827d3bbf 100644
> >>>>> --- a/drivers/iio/adc/stm32-dfsdm-adc.c
> >>>>> +++ b/drivers/iio/adc/stm32-dfsdm-adc.c
> >>>>> @@ -466,8 +466,7 @@ static int stm32_dfsdm_channels_configure(struct iio_dev *indio_dev,
> >>>>> * In continuous mode, use fast mode configuration,
> >>>>> * if it provides a better resolution.
> >>>>> */
> >>>>> - if (adc->nconv == 1 && !trig &&
> >>>>> - (indio_dev->currentmode & INDIO_BUFFER_SOFTWARE)) {
> >>>>> + if (adc->nconv == 1 && !trig && iio_buffer_enabled(indio_dev)) {
> >>>>
> >>>> This may become tricky if other modes get added later.
> >>>> STM does a relatively good job in updating and re-using their drivers
> >>>> (even if some of them do look quirky sometimes).
> >>
> >> Their hardware is crazy/complicated so tends to push the limits!
> >>
> >>>>
> >>>> So, the question here would be: is "iio_buffer_enabled(indio_dev)"
> >>>> going to be valid [in this place] once INDIO_BUFFER_TRIGGERED or
> >>>> INDIO_BUFFER_HARDWARE get added?
> >>>
> >>> I would argue, is this a real problem? Today iio_buffer_enabled() seem
> >>> to handle well what this driver is expecting. If tomorrow someone adds
> >>> another mode, that is his/her responsibility to state "okay, this
> >>> section is not common to all buffer styles *anymore*, so we need to do
> >>> a more fine grained check against ->currentmodes than
> >>> iio_buffer_enabled() does". In that case using the ->currentmodes
> >>> getter would be the right way to go, but only at that particular
> >>> moment, not today.
> >>
> >> It should be isolated to this driver, so I think it is fine to use
> >> the broader check today, but I'll leave this to the st folks as
> >> it's their driver and I don't feel that strongly about it.
>
> Hi Miquel, Alexandru, Jonathan, all,
>
> First, sorry for the delay.
>
> Indeed, I don't expect any functional changes here by using
> iio_buffer_enabled(indio_dev).
> So it should be fine to use it. You're right, the driver looks for
> buffer mode in both places where this gets used.
>
> Just an additional statement is: the driver also checks for no trigger,
> and single channel in both places (to select desired mode in the dfsdm).
> As I see, only INDIO_BUFFER_SOFTWARE is expected then.
Ok, thanks for the validation, do not hesitate to drop a Reviewed-by to
the next version of this series if you agree with the changes made here.
> For my own understanding (I'm just asking), why not using the
> currentmodes getter routine ?
>
> I've had a look at the whole series [1], It seems used elsewhere. I may
> miss something... It would be 100% equivalent to current code to use:
> iio_get_internal_mode(indio_dev) & INDIO_BUFFER_SOFTWARE ?
>
> This would be safe in case new modes gets introduced later ?
> (another note: unless these new modes gets set by default in the 'modes'
> field, this should have no impact here as well anyway ?)
I would argue that this is more a conceptual change. IMHO:
- currentmode is a variable that should have been kept internal
- checking against its value directly is kind of a hack and should be
avoided when possible because we want the core to have full freedom
over the way it manages these flags
- if you want to verify if buffers are enabled, then the core offers
you a dedicated helper that does exactly this, and will do it better
than if hardcoded by individual writers, generally
And it's not "used elsewhere" anymore thanks to this series :) only two
drivers _really_ need to check the actual current mode to do specific
actions, but that's all.
I hope it clarifies a bit.
Thanks,
Miquèl
next prev parent reply other threads:[~2022-02-02 9:34 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-15 15:13 [PATCH 00/10] Light core IIO enhancements Miquel Raynal
2021-12-15 15:13 ` [PATCH 01/10] iio: core: Fix the kernel doc regarding the currentmode iio_dev entry Miquel Raynal
2021-12-16 6:23 ` Alexandru Ardelean
2022-01-15 15:47 ` Jonathan Cameron
2021-12-15 15:13 ` [PATCH 02/10] iio: core: Enhance the kernel doc of modes and currentmodes iio_dev entries Miquel Raynal
2021-12-16 6:24 ` Alexandru Ardelean
2021-12-16 8:15 ` Miquel Raynal
2022-01-15 15:51 ` Jonathan Cameron
2022-01-28 14:56 ` Miquel Raynal
2021-12-15 15:13 ` [PATCH 03/10] iio: magnetometer: rm3100: Stop abusing the ->currentmode Miquel Raynal
2021-12-16 6:40 ` Alexandru Ardelean
2021-12-16 8:18 ` Miquel Raynal
2022-01-15 15:58 ` Jonathan Cameron
2022-01-28 15:00 ` Miquel Raynal
2021-12-15 15:13 ` [PATCH 04/10] iio: adc: stm32-dfsdm: Avoid dereferencing ->currentmode Miquel Raynal
2021-12-16 6:47 ` Alexandru Ardelean
2021-12-16 8:22 ` Miquel Raynal
2022-01-15 16:06 ` Jonathan Cameron
2022-01-28 15:04 ` Miquel Raynal
2022-02-01 8:41 ` Fabrice Gasnier
2022-02-02 9:33 ` Miquel Raynal [this message]
2021-12-15 15:13 ` [PATCH 05/10] iio: st_sensors: Use iio_device_claim/release_direct_mode() when relevant Miquel Raynal
2021-12-16 7:16 ` Alexandru Ardelean
2021-12-16 8:32 ` Miquel Raynal
2022-01-15 16:38 ` Jonathan Cameron
2021-12-15 15:13 ` [PATCH 06/10] iio: core: Hide read accesses to iio_dev->currentmode Miquel Raynal
2021-12-16 7:38 ` Alexandru Ardelean
2021-12-16 7:41 ` Alexandru Ardelean
2021-12-16 8:37 ` Miquel Raynal
2022-01-15 16:51 ` Jonathan Cameron
2021-12-15 15:13 ` [PATCH 07/10] iio: core: Hide write " Miquel Raynal
2022-01-15 16:56 ` Jonathan Cameron
2022-02-02 8:16 ` Miquel Raynal
2021-12-15 15:13 ` [PATCH 08/10] iio: core: Move iio_dev->currentmode to the opaque structure Miquel Raynal
2021-12-16 7:48 ` Alexandru Ardelean
2021-12-16 8:38 ` Miquel Raynal
2022-01-15 17:00 ` Jonathan Cameron
2021-12-15 15:13 ` [PATCH 09/10] iio: core: Simplify the registration of kfifo buffers Miquel Raynal
2021-12-16 7:52 ` Alexandru Ardelean
2022-01-15 17:12 ` Jonathan Cameron
2022-02-02 8:09 ` Miquel Raynal
2022-02-05 18:50 ` Jonathan Cameron
2021-12-15 15:13 ` [PATCH 10/10] iio: core: Clarify the modes Miquel Raynal
2022-01-15 17:30 ` Jonathan Cameron
2022-02-02 13:46 ` Miquel Raynal
2022-02-05 18:56 ` Jonathan Cameron
2022-02-06 13:21 ` Miquel Raynal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220202103359.3d4c5fb7@xps13 \
--to=miquel.raynal@bootlin.com \
--cc=ardeleanalex@gmail.com \
--cc=fabrice.gasnier@foss.st.com \
--cc=jic23@kernel.org \
--cc=lars@metafoo.de \
--cc=linux-iio@vger.kernel.org \
--cc=olivier.moysan@foss.st.com \
--cc=thomas.petazzoni@bootlin.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).