[PATCH v2] PCI: designware: move remaining rc setup code to dw_pcie_setup_rc()
Bjorn Helgaas
helgaas at kernel.org
Fri Apr 8 09:01:45 PDT 2016
On Fri, Apr 08, 2016 at 01:33:28PM +0000, Gabriele Paoloni wrote:
> Hi Bjorn
>
> Many thanks for your reply
>
> > -----Original Message-----
> > From: Bjorn Helgaas [mailto:helgaas at kernel.org]
> > Sent: 07 April 2016 15:06
> > To: Gabriele Paoloni
> > Cc: Jisheng Zhang; jingoohan1 at gmail.com; pratyush.anand at gmail.com;
> > bhelgaas at google.com; linux-pci at vger.kernel.org; linux-
> > kernel at vger.kernel.org; linux-arm-kernel at lists.infradead.org
> > Subject: Re: [PATCH v2] PCI: designware: move remaining rc setup code
> > to dw_pcie_setup_rc()
> >
> > On Thu, Apr 07, 2016 at 10:06:45AM +0000, Gabriele Paoloni wrote:
> > > Hi Jisheng
> > >
> > > > -----Original Message-----
> > > > From: Jisheng Zhang [mailto:jszhang at marvell.com]
> > > > Sent: 07 April 2016 09:35
> > > > To: Gabriele Paoloni
> > > > Cc: jingoohan1 at gmail.com; pratyush.anand at gmail.com;
> > > > bhelgaas at google.com; linux-pci at vger.kernel.org; linux-
> > > > kernel at vger.kernel.org; linux-arm-kernel at lists.infradead.org
> > > > Subject: Re: [PATCH v2] PCI: designware: move remaining rc setup
> > code
> > > > to dw_pcie_setup_rc()
> > > >
> > > > Hi Gabriele,
> > > >
> > > > On Thu, 7 Apr 2016 08:20:28 +0000 Gabriele Paoloni wrote:
> > > >
> > > > > Hi Jisheng
> > > > >
> > > > > Thanks for your reply
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Jisheng Zhang [mailto:jszhang at marvell.com]
> > > > > > Sent: 07 April 2016 03:38
> > > > > > To: Gabriele Paoloni; jingoohan1 at gmail.com;
> > > > pratyush.anand at gmail.com;
> > > > > > bhelgaas at google.com
> > > > > > Cc: linux-pci at vger.kernel.org; linux-kernel at vger.kernel.org;
> > linux-
> > > > arm-
> > > > > > kernel at lists.infradead.org
> > > > > > Subject: Re: [PATCH v2] PCI: designware: move remaining rc
> > setup
> > > > code
> > > > > > to dw_pcie_setup_rc()
> > > > > >
> > > > > > Hi Gabriele,
> > > > > >
> > > > > > On Wed, 6 Apr 2016 14:50:29 +0000 Gabriele Paoloni wrote:
> > > > > >
> > > > > > > Hi, sorry to be late on this
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: linux-kernel-owner at vger.kernel.org [mailto:linux-
> > kernel-
> > > > > > > > owner at vger.kernel.org] On Behalf Of Jisheng Zhang
> > > > > > > > Sent: 16 March 2016 11:41
> > > > > > > > To: jingoohan1 at gmail.com; pratyush.anand at gmail.com;
> > > > > > bhelgaas at google.com
> > > > > > > > Cc: linux-pci at vger.kernel.org; linux-
> > kernel at vger.kernel.org;
> > > > linux-
> > > > > > arm-
> > > > > > > > kernel at lists.infradead.org; Jisheng Zhang
> > > > > > > > Subject: [PATCH v2] PCI: designware: move remaining rc
> > setup
> > > > code
> > > > > > to
> > > > > > > > dw_pcie_setup_rc()
> > > > > > > >
> > > > > > > > dw_pcie_setup_rc(), as its name indicates, setups the RC.
> > But
> > > > > > current
> > > > > > > > dw_pcie_host_init() also contains some necessary rc setup
> > code.
> > > > > > > >
> > > > > > > > Another reason: the host may lost power during suspend to
> > ram,
> > > > the
> > > > > > RC
> > > > > > > > need to be re-setup after resume. The rc can't be correctly
> > > > resumed
> > > > > > > > without the rc setup code in dw_pcie_host_init().
> > > > > > > >
> > > > > > > > So this patch moves the code to dw_pcie_setup_rc() to
> > address
> > > > the
> > > > > > above
> > > > > > > > two issues. After this patch, each pcie designware driver
> > users
> > > > > > could
> > > > > > > > call dw_pcie_setup_rc() to re-setup rc when resume back.
> > > > > > >
> > > > > > > I think this patch breaks the Hisilicon driver...
> > > > > > >
> > > > > > > Our driver performs linkup setup in UEFI therefore we do not
> > call
> > > > > > > dw_pcie_setup_rc(), we only call dw_pcie_host_init().
> > > > > >
> > > > > > Thanks for the information. So pcie-hisi rely on UEFI to do
> > > > something
> > > > > > similar
> > > > > > in dw_pcie_setup_rc(), this comes to a common driver implement
> > > > > > question: should
> > > > > > linux device driver rely on bootloader to configure HW device?
> > > > >
> > > > > I don't see any issue with this...
> > > > >
> > > > > >
> > > > > > Is it acceptable that pcie-hisi adds a call to
> > dw_pcie_setup_rc()
> > > > in
> > > > > > hisi_add_pcie_port()?
> > > > >
> > > > > I don't think so...that would try to overwrite what is already
> > set by
> > > > > the bootloader; so it is wrong in principle and maybe it can lead
> > to
> > > > > undefined behaviours...
> > > >
> > > > make sense! This commit is intend to re-setup the rc when waken
> > from
> > > > s2ram (in
> > > > s2ram state, the host lost power)
> > > >
> > > > I have no good solution but to introduce one function e.g
> > > > dw_pcie_setup_rc_after_linkup(), then move related code from
> > > > dw_pcie_host_init
> > > > to it, then let my host driver resume hook to call.
> > > >
> > > > Hi Pratyush, Jingoo and Bjorn etc.
> > > >
> > > > any suggestions are appreciated!
> > >
> > > What about:
> > >
> > > diff --git a/drivers/pci/host/pcie-designware.c
> > b/drivers/pci/host/pcie-designware.c
> > > index a4cccd3..e461f5d 100644
> > > --- a/drivers/pci/host/pcie-designware.c
> > > +++ b/drivers/pci/host/pcie-designware.c
> > > @@ -434,7 +434,6 @@ int dw_pcie_host_init(struct pcie_port *pp)
> > > struct platform_device *pdev = to_platform_device(pp->dev);
> > > struct pci_bus *bus, *child;
> > > struct resource *cfg_res;
> > > - u32 val;
> > > int i, ret;
> > > LIST_HEAD(res);
> > > struct resource_entry *win;
> > > @@ -544,25 +543,6 @@ int dw_pcie_host_init(struct pcie_port *pp)
> > > if (pp->ops->host_init)
> > > pp->ops->host_init(pp);
> > >
> > > - /*
> > > - * If the platform provides ->rd_other_conf, it means the
> > platform
> > > - * uses its own address translation component rather than ATU, so
> > > - * we should not program the ATU here.
> > > - */
> > > - if (!pp->ops->rd_other_conf)
> > > - dw_pcie_prog_outbound_atu(pp, PCIE_ATU_REGION_INDEX1,
> > > - PCIE_ATU_TYPE_MEM, pp->mem_base,
> > > - pp->mem_bus_addr, pp->mem_size);
> > > -
> > > - dw_pcie_wr_own_conf(pp, PCI_BASE_ADDRESS_0, 4, 0);
> > > -
> > > - /* program correct class for RC */
> > > - dw_pcie_wr_own_conf(pp, PCI_CLASS_DEVICE, 2,
> > PCI_CLASS_BRIDGE_PCI);
> > > -
> > > - dw_pcie_rd_own_conf(pp, PCIE_LINK_WIDTH_SPEED_CONTROL, 4, &val);
> > > - val |= PORT_LOGIC_SPEED_CHANGE;
> > > - dw_pcie_wr_own_conf(pp, PCIE_LINK_WIDTH_SPEED_CONTROL, 4, val);
> > > -
> > > pp->root_bus_nr = pp->busn->start;
> > > if (IS_ENABLED(CONFIG_PCI_MSI)) {
> > > bus = pci_scan_root_bus_msi(pp->dev, pp->root_bus_nr,
> > > @@ -725,6 +705,29 @@ static struct pci_ops dw_pcie_ops = {
> > > .write = dw_pcie_wr_conf,
> > > };
> > >
> > > +void dw_pcie_setup_own_cfg(struct pcie_port *pp)
> > > +{
> > > + u32 val;
> > > + /*
> > > + * If the platform provides ->rd_other_conf, it means the
> > platform
> > > + * uses its own address translation component rather than ATU, so
> > > + * we should not program the ATU here.
> > > + */
> > > + if (!pp->ops->rd_other_conf)
> > > + dw_pcie_prog_outbound_atu(pp, PCIE_ATU_REGION_INDEX1,
> > > + PCIE_ATU_TYPE_MEM, pp->mem_base,
> > > + pp->mem_bus_addr, pp->mem_size);
> > > +
> > > + dw_pcie_wr_own_conf(pp, PCI_BASE_ADDRESS_0, 4, 0);
> > > +
> > > + /* program correct class for RC */
> > > + dw_pcie_wr_own_conf(pp, PCI_CLASS_DEVICE, 2,
> > PCI_CLASS_BRIDGE_PCI);
> > > +
> > > + dw_pcie_rd_own_conf(pp, PCIE_LINK_WIDTH_SPEED_CONTROL, 4, &val);
> > > + val |= PORT_LOGIC_SPEED_CHANGE;
> > > + dw_pcie_wr_own_conf(pp, PCIE_LINK_WIDTH_SPEED_CONTROL, 4, val);
> > > +}
> > > +
> > > void dw_pcie_setup_rc(struct pcie_port *pp)
> > > {
> > > u32 val;
> > > @@ -800,6 +803,8 @@ void dw_pcie_setup_rc(struct pcie_port *pp)
> > > val |= PCI_COMMAND_IO | PCI_COMMAND_MEMORY |
> > > PCI_COMMAND_MASTER | PCI_COMMAND_SERR;
> > > dw_pcie_writel_rc(pp, val, PCI_COMMAND);
> > > +
> > > + dw_pcie_setup_own_cfg(pp);
> > > }
> > >
> > > MODULE_AUTHOR("Jingoo Han <jg1.han at samsung.com>");
> > > diff --git a/drivers/pci/host/pcie-designware.h
> > b/drivers/pci/host/pcie-designware.h
> > > index f437f9b..caf0f5d 100644
> > > --- a/drivers/pci/host/pcie-designware.h
> > > +++ b/drivers/pci/host/pcie-designware.h
> > > @@ -85,5 +85,6 @@ int dw_pcie_wait_for_link(struct pcie_port *pp);
> > > int dw_pcie_link_up(struct pcie_port *pp);
> > > void dw_pcie_setup_rc(struct pcie_port *pp);
> > > int dw_pcie_host_init(struct pcie_port *pp);
> > > +void dw_pcie_setup_own_cfg(struct pcie_port *pp);
> > >
> > > #endif /* _PCIE_DESIGNWARE_H */
> > > diff --git a/drivers/pci/host/pcie-hisi.c b/drivers/pci/host/pcie-
> > hisi.c
> > > index 3e98d4e..8da29b2 100644
> > > --- a/drivers/pci/host/pcie-hisi.c
> > > +++ b/drivers/pci/host/pcie-hisi.c
> > > @@ -164,6 +164,7 @@ static int hisi_add_pcie_port(struct pcie_port
> > *pp,
> > > dev_err(&pdev->dev, "failed to initialize host\n");
> > > return ret;
> > > }
> > > + dw_pcie_setup_own_cfg(pp);
> > >
> > > return 0;
> > > }
> >
> > What's the hisi plan for resuming after suspend-to-RAM? How does the
> > RC get reprogrammed after it loses all its state?
>
> PM is not part of the driver yet. This is planned for near
> future release so haven't made such considerations yet
> >
> > What would break if hisi did call dw_pcie_setup_rc()? I know you said
> > it would overwrite what the bootloader already did, which is true.
>
> I am try to figure this out now with our HW team.
>
> >
> > But hisi does call dw_pcie_host_init(), so it reads pp->mem (which
> > determines pp->mem_base) and pp->lanes from the DT. Other drivers
> > then call dw_pcie_setup_rc() which programs the RC based on
> > pp->mem_base and pp->lanes. So hisi assumes UEFI programmed the RC to
> > match the DT, while the other drivers read the DT and program the RC
> > to match. The latter seems more robust because it enforces the
> > consistency rather than relying on it.
>
> Yes I agree with you, however we have preferred to move RC config to
> BIOS to have a single driver to support multiple versions of the
> same SoC.
I think there are two reasonable approaches:
1) A single generic driver that doesn't have any knowledge about the
chipset registers; it uses run-time firmware interfaces to manage
the bridge. The ACPI pci_root.c driver is the best example so far
and works very well. It supports basically all x86 and ia64
chipsets and requires no kernel work for new ones.
2) Native drivers specific to each chipset. These may get
configuration information from DT, but they do their own
register-level programming of the device without run-time help from
firmware.
I think hisi is a native driver because it uses hip05/hip06 registers
to check link state and perform config operations. And apparently you
rely on the ATU, BAR, class, and link width programming currently done
in dw_pcie_host_init(). But you want to rely on pre-boot firmware to
set up the link. That doesn't make sense to me -- if the driver wants
to twiddle the registers, it should know how to do it all. I don't
see how you can reasonably manage this half-way approach.
> The patch I proposed above does the same job as the original patch
> proposed by Jisheng and also allows hisi driver to call the moved
> code.
>
> Do you see anything wrong with it?
Only that it makes the structure more complicated and we haven't
identified a corresponding benefit yet.
Bjorn
More information about the linux-arm-kernel
mailing list