[PATCH v2 3/5] PCI: st: Provide support for the sti PCIe controller

Bjorn Helgaas bhelgaas at google.com
Thu Apr 9 05:43:30 PDT 2015


On Tue, Mar 31, 2015 at 11:11:44AM +0200, Gabriel Fernandez wrote:
> Hi Bjorn,
> 
> pci-st.c driver could be modular with modification of pcie-designware
> core driver. But as Fabrice said  it should be another patchset.
> 
> What do you prefer ?
> 
> drop all the module related macros as mentioned by Paul ?
> or
> keep macros like other vendors do ?

I don't think this is a big deal either way.  You made it consistent with
the style of the existing PCI host drivers, which is exactly what you
should do.

If somebody wants to do the work of making them all modular, or leaving
them non-modular and removing all the MODULE_*() annotations, that can be
done later.

> On 18 March 2015 at 11:35, Paul Bolle <pebolle at tiscali.nl> wrote:
> > Hi Fabrice,
> >
> > Fabrice Gasnier schreef op wo 18-03-2015 om 09:49 [+0100]:
> >> On 03/16/2015 04:11 PM, Paul Bolle wrote:
> >> >> +config PCI_ST
> >> >> +  bool "ST PCIe controller"
> >> > You add a bool Kconfig symbol. A week or two ago I saw some patches fly
> >> > by that - I think - allowed PCIe controllers to be built modular.
> >>
> >> Thanks for your review.
> >>
> >> Are you talking about "PCI: Export symbols of PCI functions" patch, that
> >> is part of a series
> >> named "pci: iproc: Add Broadcom iProc PCIe support" ?
> >
> > Yes, that is the series I was thinking about. (I made you search lkml,
> > and that was a bit rude. But you found the patch anyhow.)
> >
> >> This controller doesn't look like to be based on pcie-designware core
> >> driver.
> >> Other vendors that are using "pcie-designware" core driver are also make
> >> it bool.
> >> The current core driver doesn't support module loading/unloading as I
> >> see it.
> >> If this is required, I also think this should be part of another patchset.
> >>
> >> What do you think ?
> >
> > I wouldn't know whether your driver might work as a loadable module, but
> > other people reading this surely will. But if it can't work as a module
> > you should drop all the module related macros etc. I spotted. Because
> > then they serve no purpose.
> >
> >
> > Paul Bolle
> >



More information about the linux-arm-kernel mailing list