Re: [PATCH 1/3] hwmon: (nct6775) Use nct6775_*() lock function pointers in nct6775_data.
From: Denis Pauk
Date: Thu Nov 25 2021 - 16:11:29 EST
On Wed, 24 Nov 2021 18:03:25 +0200
Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote:
> On Mon, Nov 22, 2021 at 11:29 PM Denis Pauk <pauk.denis@xxxxxxxxx>
> wrote:
>
> Better subject line (after prefix): Use lock function pointers in
> nct6775_data (note no period and drop of redundancy)
>
> > Prepare for platform specific callbacks usage:
> > * Use nct6775 lock function pointers in struct nct6775_data instead
> > direct calls.
>
> ...
>
> > +static int nct6775_lock(struct nct6775_data *data)
> > +{
> > + mutex_lock(&data->update_lock);
> > +
> > + return 0;
> > +}
> > +
> > +static void nct6775_unlock(struct nct6775_data *data, struct
> > device *dev) +{
> > + mutex_unlock(&data->update_lock);
> > +}
>
> Have you run `sparse` against this?
> Install `sparse` in your distribution and make kernel with
> `make W=1 C=1 CF=-D__CHECK_ENDIAN__ ...`
>
> It might require using special annotations to these functions to make
> static analysers happy.
>
Thank you, I will validate my patches before sending with sparse also.
I have tried with sparse==0.6.4:
---
$ make CC="ccache gcc" W=1 C=2 CF=-D__CHECK_ENDIAN__ 2>&1 | grep
nct6775 -n5
....
27219- CHECK drivers/hwmon/nct6683.c
27220: CHECK drivers/hwmon/nct6775.c
27221- CHECK drivers/hwmon/nct7802.c
....
---
It has not showed any warnings. Have I missed some flag?
Best regards,
Denis.