[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YUGG8iPWMLx5vJ1f@kroah.com>
Date: Wed, 15 Sep 2021 07:38:58 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: Min Li <min.li.xe@...esas.com>
Cc: "derek.kiernan@...inx.com" <derek.kiernan@...inx.com>,
"dragan.cvetic@...inx.com" <dragan.cvetic@...inx.com>,
"arnd@...db.de" <arnd@...db.de>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH misc] misc: Add Renesas Synchronization Management Unit
(SMU) support
On Tue, Sep 14, 2021 at 08:43:44PM +0000, Min Li wrote:
> Hi Greg
>
> Thanks for the review
>
> > > drivers/misc/Kconfig | 9 ++
> > > drivers/misc/Makefile | 2 +
> > > drivers/misc/rsmu_cdev.c | 239
> > > ++++++++++++++++++++++++++++++++++++++++++++++
> > > drivers/misc/rsmu_cdev.h | 77 +++++++++++++++
> > > drivers/misc/rsmu_cm.c | 164 +++++++++++++++++++++++++++++++
> > > drivers/misc/rsmu_sabre.c | 133 ++++++++++++++++++++++++++
> >
> > If you make this all one .c file, the .h file can go away and it will be much
> > simpler in the end. And will get rid of the global symbols.
> >
>
> That is doable. But <linux/mfd/idt8a340_reg.h> and <linux/mfd/idt82p33_reg.h> have naming confliction.
> To make this change one file, I have to include both of them and therefore I have to change them to resolve
> Conflicts. Can I include this
Yes, that should be fine, do it as patch 1/X in a series.
> > >
> > > +config RSMU
> > > + tristate "Renesas Synchronization Management Unit (SMU)"
> > > + help
> > > + This option enables support for the IDT ClockMatrix(TM) and
> > 82P33xxx
> > > + families of timing and synchronization devices. It will be used by
> > > + Renesas PTP Clock Manager for Linux (pcm4l) software to provide
> > support
> > > + for GNSS assisted partial timing support (APTS) and other
> > networking
> > > + timing functions.
> >
> > No driver name listed?
>
> Sorry, what do you mean by driver name in this context?
Most Kconfig help texts say the driver name if it is built as a module.
> > > +
> > > +/**
> > > + * struct rsmu_cdev - Driver data for RSMU character device
> > > + * @name: rsmu device name as rsmu[index]
> > > + * @dev: pointer to device
> > > + * @mfd: pointer to MFD device
> > > + * @miscdev: character device handle
> > > + * @regmap: I2C/SPI regmap handle
> > > + * @lock: mutex to protect operations from being interrupted
> > > + * @type: rsmu device type, passed through platform data
> > > + * @ops: rsmu device methods
> > > + * @index: rsmu device index
> > > + */
> > > +struct rsmu_cdev {
> > > + char name[16];
> > > + struct device *dev;
> >
> > What device is this pointing to?
>
> It is the platform device from rsmu_probe(struct platform_device *pdev)
Then why not just point to the platform device?
> >
> > > + struct device *mfd;
> >
> > What is this for?
>
> It is the multi-functional device from driver/mfd/rsmu_core.c
> The mfd driver is responsible for spawn this platform device and spi/i2c
> bus access
Why do you need to mess with that if you have a pointer to the platform
device instead?
> > > +/*
> > > + * RSMU IOCTL List
> > > + */
> > > +#define RSMU_MAGIC '?'
> >
> > Where did you get this value from?
> >
> > Where did you reserve it?
>
> No I didn't reserve it. I checked other code and they all seem to use a random character
That's not the best way to do this.
Why do you need ioctls at all anyway? What userspace tools will be
accessing this driver? Do you have a link to where they are located at?
> > > +
> > > +/**
> > > + * @Description
> >
> > What is this format? It is not kernel-doc :(
> >
> > > + * ioctl to set SMU combo mode.Combo mode provides physical layer
> > > + frequency
> > > + * support from the Ethernet Equipment Clock to the PTP clock
> > > + *
> > > + * @Parameters
> >
> > Same here and elsewhere in this file.
>
> I was copying the format from xilinx_sdfec.h
>
> Is there a place that tells me how to properly document ioctl or can you give me an code example?
The kerneldoc format should be described in Documentation/ somewhere...
thanks,
greg k-h
Powered by blists - more mailing lists