RE: [PATCHv3 17/27] PCI: mobiveil: fix the checking of valid device

From: Z.q. Hou
Date: Mon Feb 18 2019 - 02:15:30 EST


Hi Bjorn,

Thanks a lot for your comments!

> -----Original Message-----
> From: Bjorn Helgaas <helgaas@xxxxxxxxxx>
> Sent: 2019年2月8日 22:13
> To: Subrahmanya Lingappa <l.subrahmanya@xxxxxxxxxxxxxx>
> Cc: Z.q. Hou <zhiqiang.hou@xxxxxxx>; mark.rutland@xxxxxxx;
> devicetree@xxxxxxxxxxxxxxx; lorenzo.pieralisi@xxxxxxx; Xiaowei Bao
> <xiaowei.bao@xxxxxxx>; linux-pci@xxxxxxxxxxxxxxx; will.deacon@xxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; Leo Li <leoyang.li@xxxxxxx>; M.h. Lian
> <minghuan.lian@xxxxxxx>; robh+dt@xxxxxxxxxx; Mingkai Hu
> <mingkai.hu@xxxxxxx>; catalin.marinas@xxxxxxx; shawnguo@xxxxxxxxxx;
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; Thomas Petazzoni
> <thomas.petazzoni@xxxxxxxxxxx>
> Subject: Re: [PATCHv3 17/27] PCI: mobiveil: fix the checking of valid device
>
> [+cc Thomas]
>
> On Fri, Feb 08, 2019 at 06:11:15PM +0530, Subrahmanya Lingappa wrote:
> > On Tue, Jan 29, 2019 at 1:40 PM Z.q. Hou <zhiqiang.hou@xxxxxxx> wrote:
> > >
> > > From: Hou Zhiqiang <Zhiqiang.Hou@xxxxxxx>
> > >
> > > Avoid to issue CFG transactions to link partner when the PCIe link
> > > is not up. And allow CFG transactions to all functions of Endpoint
> > > implemented multiple functions.
> > >
> > > Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge
> > > IP driver")
> > > Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@xxxxxxx>
> > > Reviewed-by: Minghuan Lian <Minghuan.Lian@xxxxxxx>
> > > ---
> > > V3:
> > > - No change
> > >
> > > drivers/pci/controller/mobiveil/pcie-mobiveil-host.c | 6 +++++-
> > > 1 file changed, 5 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/pci/controller/mobiveil/pcie-mobiveil-host.c
> > > b/drivers/pci/controller/mobiveil/pcie-mobiveil-host.c
> > > index dc5324d94466..1ae82e790562 100644
> > > --- a/drivers/pci/controller/mobiveil/pcie-mobiveil-host.c
> > > +++ b/drivers/pci/controller/mobiveil/pcie-mobiveil-host.c
> > > @@ -29,6 +29,10 @@ static bool mobiveil_pcie_valid_device(struct
> > > pci_bus *bus, unsigned int devfn) {
> > > struct mobiveil_pcie *pcie = bus->sysdata;
> > >
> > > + /* If there is no link, then there is no device */
> > > + if (bus->number > pcie->rp.root_bus_nr
> && !mobiveil_pcie_link_up(pcie))
> > > + return false;
> > > +
> > > /* Only one device down on each root port */
> > > if ((bus->number == pcie->rp.root_bus_nr) && (devfn > 0))
> > > return false;
> > > @@ -37,7 +41,7 @@ static bool mobiveil_pcie_valid_device(struct
> pci_bus *bus, unsigned int devfn)
> > > * Do not read more than one device on the bus directly
> > > * attached to RC
> > > */
> > > - if ((bus->primary == pcie->rp.root_bus_nr) && (devfn > 0))
> > > + if ((bus->primary == pcie->rp.root_bus_nr) &&
> > > + (PCI_SLOT(devfn) > 0))
>
> > here change "primary" to "number", as it's a bug in the original driver too.
>
> This looks like it should be split into two patches: (1) checking for link up, and
> (2) checking root_bus_nr.

Yes, will split this patch in next version.

>
> And if you mean "bus->primary == pcie->rp.root_bus_nr" is a bug in
> pci-aardvark.c, too, it is imperative to fix that bug also (with a separate patch).

It is not a bug, but the bug is to limit multiple function EP (devfn > 0).

Thanks,
Zhiqiang