Re: [PATCH v4 1/3] Staging: IndustryPack bus for the Linux Kernel

["Samuel Iglesias Gonsálvez" <siglesias@igalia.com>]

On Wed, 2012-05-09 at 14:13 -0700, Greg Kroah-Hartman wrote:
> On Wed, May 09, 2012 at 03:27:19PM +0200, Samuel Iglesias Gonsalvez
wrote:
> > Add IndustryPack bus support for the Linux Kernel.
> > 
> > This is a virtual bus that allows to perform all the operations
between
> > carrier and mezzanine boards.
> 
> Note, I've accepted this patch, just a few comments that you might
want
> to fix up in future patches you send to me:
> 

I am learning a lot thanks to your comments. 

I will fix them and send you the corresponding patches. I can also fix
these patches directly and resend them, if you prefer that.

> > +++ b/drivers/staging/Kconfig
> > @@ -24,6 +24,8 @@ menuconfig STAGING
> >  
> >  if STAGING
> >  
> > +source "drivers/staging/ipack/Kconfig"
> > +
> >  source "drivers/staging/et131x/Kconfig"
> >  
> >  source "drivers/staging/slicoss/Kconfig"
> 
> Why put yourself at the front of the list, and not at the end?


My fault. I don't know if it is better to fix the patch or send another
one adding myself at the end.

What do you think?

> 
> > diff --git a/drivers/staging/Makefile b/drivers/staging/Makefile
> > index ffe7d44..23eb56b 100644
> > --- a/drivers/staging/Makefile
> > +++ b/drivers/staging/Makefile
> > @@ -30,6 +30,7 @@ obj-$(CONFIG_OCTEON_ETHERNET)     += octeon/
> >  obj-$(CONFIG_VT6655)               += vt6655/
> >  obj-$(CONFIG_VT6656)               += vt6656/
> >  obj-$(CONFIG_VME_BUS)              += vme/
> > +obj-$(CONFIG_IPACK_BUS)            += ipack/
> >  obj-$(CONFIG_DX_SEP)            += sep/
> >  obj-$(CONFIG_IIO)          += iio/
> >  obj-$(CONFIG_ZRAM)         += zram/
> 
> Same here, why not at the end?
> 
> > diff --git a/drivers/staging/ipack/Kconfig
b/drivers/staging/ipack/Kconfig
> > new file mode 100644
> > index 0000000..e20187f
> > --- /dev/null
> > +++ b/drivers/staging/ipack/Kconfig
> > @@ -0,0 +1,9 @@
> > +#
> > +# IPACK configuration.
> > +#
> > +
> > +menuconfig IPACK_BUS
> > +   tristate "IndustryPack bus support"
> > +   ---help---
> > +     If you say Y here you get support for the IndustryPack
Framework.
> > +
> 
> You are going to have to flush this help out more, you do know that,
> right?
> 

Yes, I have done one patch adding a proper description in the help area.

> 
> > diff --git a/drivers/staging/ipack/Makefile
b/drivers/staging/ipack/Makefile
> > new file mode 100644
> > index 0000000..56e2340
> > --- /dev/null
> > +++ b/drivers/staging/ipack/Makefile
> > @@ -0,0 +1,4 @@
> > +#
> > +# Makefile for the IPACK bridge device drivers.
> > +#
> > +obj-$(CONFIG_IPACK_BUS)            += ipack.o
> > diff --git a/drivers/staging/ipack/TODO b/drivers/staging/ipack/TODO
> > new file mode 100644
> > index 0000000..167ae4d
> > --- /dev/null
> > +++ b/drivers/staging/ipack/TODO
> > @@ -0,0 +1,21 @@
> > +                           TODO
> > +                           ====
> > +Introduction
> > +============
> > +
> > +These drivers add support for IndustryPack devices: carrier and
mezzanine
> > +boards.
> > +
> > +The ipack driver is just an abstraction of the bus providing the
common
> > +operations between the two kind of boards.
> > +
> > +TODO
> > +====
> > +
> > +Ipack
> > +-----
> > +
> > +* The structures and API exported can be improved a lot. For
example, the
> > +  way to unregistering mezzanine devices, doing the mezzanine
driver a call to
> > +  remove_device() to notify the carrier driver, or the opposite
with the call to
> > +  the ipack_driver_ops' remove() function could be improved.
> 
> I need an email address in this file for who to cc: patches to, look
at
> the other TODO files in drivers/staging/ for examples of what needs to
> be here.
> 

OK.

> 
> > diff --git a/drivers/staging/ipack/ipack.c
b/drivers/staging/ipack/ipack.c
> > new file mode 100644
> > index 0000000..a54bfd7
> > --- /dev/null
> > +++ b/drivers/staging/ipack/ipack.c
> > @@ -0,0 +1,175 @@
> > +/*
> > + * Industry-pack bus support functions.
> > + *
> > + * (C) 2011 Samuel Iglesias Gonsalvez <siglesia@cern.ch>, CERN
> > + * (C) 2012 Samuel Iglesias Gonsalvez <siglesias@igalia.com>,
Igalia
> > + *
> > + * This program is free software; you can redistribute it and/or
modify it
> > + * under the terms of the GNU General Public License as published
by the Free
> > + * Software Foundation; either version 2 of the License, or (at
your option)
> > + * any later version.
> 
> Do you really mean "any later version"?  If so, fine, just be aware of
> this please.
> 

OK. I will change it to explicitly and only GPL v2.

> > + */
> > +
> > +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> > +
> > +#include <linux/module.h>
> > +#include <linux/mutex.h>
> > +#include "ipack.h"
> > +
> > +#define to_ipack_dev(device) container_of(device, struct
ipack_device, dev)
> > +#define to_ipack_driver(drv) container_of(drv, struct ipack_driver,
driver)
> > +
> > +/* used when allocating bus numbers */
> > +#define IPACK_MAXBUS              64
> > +
> > +static DEFINE_MUTEX(ipack_mutex);
> > +
> > +struct ipack_busmap {
> > +   unsigned long busmap[IPACK_MAXBUS / (8*sizeof(unsigned long))];
> > +};
> > +static struct ipack_busmap busmap;
> > +
> > +static int ipack_bus_match(struct device *device, struct
device_driver *driver)
> > +{
> > +   int ret;
> > +   struct ipack_device *dev = to_ipack_dev(device);
> > +   struct ipack_driver *drv = to_ipack_driver(driver);
> > +
> > +   if (!drv->ops->match)
> > +           return -EINVAL;
> > +
> > +   ret = drv->ops->match(dev);
> > +   if (ret)
> > +           dev->driver = drv;
> > +
> > +   return 0;
> > +}
> > +
> > +static int ipack_bus_probe(struct device *device)
> > +{
> > +   struct ipack_device *dev = to_ipack_dev(device);
> > +
> > +   if (!dev->driver->ops->probe)
> > +           return -EINVAL;
> > +
> > +   return dev->driver->ops->probe(dev);
> > +}
> > +
> > +static int ipack_bus_remove(struct device *device)
> > +{
> > +   struct ipack_device *dev = to_ipack_dev(device);
> > +
> > +   if (!dev->driver->ops->remove)
> > +           return -EINVAL;
> > +
> > +   dev->driver->ops->remove(dev);
> > +   return 0;
> > +}
> > +
> > +static struct bus_type ipack_bus_type = {
> > +   .name  = "ipack",
> > +   .probe = ipack_bus_probe,
> > +   .match = ipack_bus_match,
> > +   .remove = ipack_bus_remove,
> > +};
> > +
> > +static int ipack_assign_bus_number(void)
> > +{
> > +   int busnum;
> > +
> > +   mutex_lock(&ipack_mutex);
> > +   busnum = find_next_zero_bit(busmap.busmap, IPACK_MAXBUS, 1);
> 
> Nice, but you also can use the ics interface as well, that keeps you
> from having to have MAXBUS busses, if you want to.
> 

I didn't know about the ics interface. The find_next_zero_bit and busmap
code was taken from drivers/usb/core/hcd.c because it does the same I
want to.

> 
> > +
> > +   if (busnum >= IPACK_MAXBUS) {
> > +           pr_err("too many buses\n");
> > +           busnum = -1;
> > +           goto error_find_busnum;
> > +   }
> > +
> > +   set_bit(busnum, busmap.busmap);
> > +
> > +error_find_busnum:
> > +   mutex_unlock(&ipack_mutex);
> > +   return busnum;
> > +}
> > +
> > +int ipack_bus_register(struct ipack_bus_device *bus)
> > +{
> > +   int bus_nr;
> > +
> > +   bus_nr = ipack_assign_bus_number();
> > +   if (bus_nr < 0)
> > +           return -1;
> > +
> > +   bus->bus_nr = bus_nr;
> > +   return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(ipack_bus_register);
> 
> Don't you need to register the bus with the driver core here?
> 

Oops, you are right. I am preparing a patch fixing that.

> > +
> > +int ipack_bus_unregister(struct ipack_bus_device *bus)
> > +{
> > +   mutex_lock(&ipack_mutex);
> > +   clear_bit(bus->bus_nr, busmap.busmap);
> > +   mutex_unlock(&ipack_mutex);
> > +   return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(ipack_bus_unregister);
> 
> And unregister it here?
> 
> > +
> > +int ipack_driver_register(struct ipack_driver *edrv)
> > +{
> > +   edrv->driver.bus = &ipack_bus_type;
> > +   return driver_register(&edrv->driver);
> > +}
> > +EXPORT_SYMBOL_GPL(ipack_driver_register);
> > +
> > +void ipack_driver_unregister(struct ipack_driver *edrv)
> > +{
> > +   driver_unregister(&edrv->driver);
> > +}
> > +EXPORT_SYMBOL_GPL(ipack_driver_unregister);
> > +
> > +static void ipack_device_release(struct device *dev)
> > +{
> > +}
> 
> Weee.  As per the in-kernel documentation, I get to publically mock
you
> for doing something as foolish as thinking you are smarter than the
> kernel by just creating an empty function for the release callback.
> 
> Did you think this really is the solution for when the kernel is
> complaining to you about the fact that you need a callback function
> here?  Surely I didn't just put that logic in the core for no good
> reason now, right?
> 
> Please fix this up NOW.

OK, I will fix it. However reading my code, I don't see the need to
kfree anything here, like in other drivers, for example.

Is it OK to have a pr_info notifying the release of the device or should
I think again about it? 

> 
> > +++ b/drivers/staging/ipack/ipack.h
> > @@ -0,0 +1,183 @@
> > +/*
> > + * Industry-pack bus.
> > + *
> > + * (C) 2011 Samuel Iglesias Gonsalvez <siglesia@cern.ch>, CERN
> > + * (C) 2012 Samuel Iglesias Gonsalvez <siglesias@igalia.com>,
Igalia
> > + *
> > + * This program is free software; you can redistribute it and/or
modify it
> > + * under the terms of the GNU General Public License as published
by the Free
> > + * Software Foundation; either version 2 of the License, or (at
your option)
> > + * any later version.
> 
> Again, "any later version", are you sure?  Be very sure about this
> please.
> 
> > +struct ipack_device {
> > +   char board_name[IPACK_BOARD_NAME_SIZE];
> 
> Why not use dev->name?

May I be wrong, do you refer rename it to "name"?

> 
> > +   char bus_name[IPACK_BOARD_NAME_SIZE];
> > +   unsigned int bus_nr;
> > +   unsigned int slot;
> > +   unsigned int irq;
> > +   struct ipack_driver *driver;
> > +   struct ipack_bus_ops *ops;
> > +   struct ipack_addr_space id_space;
> > +   struct ipack_addr_space io_space;
> > +   struct ipack_addr_space mem_space;
> > +   struct device dev;
> > +};
> 
> thanks,

Thanks to you. 

As I told at the beginning of the email, it could worthy to resend the
patches with all the things fixed. Or I can send patches fixing all the
things, I don't know what do you prefer.

Best regards,

Sam