[PATCH v2] PCI: layerscape: Add PCIe support for LS1043a and LS2080a

[Lian M.H.]

Hi Bjorn,

Thanks for your comments
I will split the patch.

Thanks,
Minghuan

> -----Original Message-----
> From: Bjorn Helgaas [mailto:helgaas at kernel.org]
> Sent: Monday, October 12, 2015 11:27 PM
> To: Arnd Bergmann <arnd at arndb.de>
> Cc: Lian Minghuan-B31939 <Minghuan.Lian at freescale.com>;
> linux-pci at vger.kernel.org; linux-arm-kernel at lists.infradead.org; Zang
> Roy-R61911 <tie-fei.zang at freescale.com>; Hu Mingkai-B21284
> <Mingkai.Hu at freescale.com>; Yoder Stuart-B08248
> <stuart.yoder at freescale.com>; Li Yang-Leo-R58472 <LeoLi at freescale.com>;
> Bjorn Helgaas <bhelgaas at google.com>; Jingoo Han <jg1.han at samsung.com>;
> Zhou Wang <wangzhou1 at hisilicon.com>
> Subject: Re: [PATCH v2] PCI: layerscape: Add PCIe support for LS1043a and
> LS2080a
> 
> On Mon, Oct 12, 2015 at 02:36:29PM +0200, Arnd Bergmann wrote:
> > On Wednesday 07 October 2015 12:57:25 Bjorn Helgaas wrote:
> > > > diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> > > > index ae873be..38fe8a8 100644
> > > > --- a/drivers/pci/host/Kconfig
> > > > +++ b/drivers/pci/host/Kconfig
> > > > @@ -105,7 +105,7 @@ config PCI_XGENE_MSI
> > > >
> > > >  config PCI_LAYERSCAPE
> > > >       bool "Freescale Layerscape PCIe controller"
> > > > -     depends on OF && ARM
> > > > +     depends on OF && (ARM || ARM64)
> > >
> > > It seems like there are a couple things going on here, and I wonder
> > > if you can split them out into separate patches.
> > >
> > > 1) Making this work on ARM64 as well as on ARM.  This may be of
> > > interest for other DesignWare-based drivers, so if you split this
> > > out, maybe it would be a useful template for converting the other
> > > drivers, too.
> >
> > The Kconfig change apparently made it into linux-next now, but it
> > doesn't actually build on arm64, because the dependency that was
> > mentioned in the cover letter [0] is not there.
> > ...
> > [0] This patch is based on v4.3-rc1 and [PATCH v9 0/6]
> > PCI: hisi: Add PCIe host support for HiSilicon SoC Hip05 patchset from
> > Zhou Wang.
> 
> Oops, my fault.  I read that as "this patch is derived from, e.g., it a newer
> version of, the HiSilicon patchset."  But that's obviously a silly way to read it.
> I re-did the merge without the Layerscape patch and repushed it to my "next"
> branch.
> 
> I'll work on Zhou's series, then revisit Minghuan's patch(es).  I'm hoping the
> Layerscape patch can be split up a bit, both to split out the ARM64 config
> change and to make a cleaner "LS1043a and LS2080a"
> support patch.  The current patch changes several things that are not
> obviously directly related to LS1043a and LS2080a.  It would be ideal to have a
> series of things like:
> 
>   - add ARM64 build support (maybe just the Kconfig change)
>   - make the ls_pcie_probe() changes that are apparently generic
>     across all the LS devices
>   - add driver data to lc_pcie_of_match[] (this wouldn't change any
>     behavior at all; it would just add the ls1021_drvdata that works
>     for the already-supported devices)
>   - add ls1032_drvdata and ls2080_drvdata for the new devices
>   - add the "msi-parent" diagnostic (this seems to be basically new
>     functionality that applies to all devices, not just the new ones)
> 
> That way, each individual change would be easier to review.
> 
> Bjorn