Re: [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources.

From: Lorenzo Pieralisi
Date: Wed Apr 27 2016 - 12:09:34 EST

Next message: Jassi Brar: "Re: [PATCH] dmaengine: pl330: Fix race in pl330_get_desc()"
Previous message: Frederic Weisbecker: "Re: [PATCH/RFC v3] perf core: Allow setting up max frame stack depth via sysctl"
In reply to: Liviu . Dudau: "Re: [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources."
Next in thread: Bjorn Helgaas: "Re: [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Apr 27, 2016 at 04:10:36PM +0100, Liviu.Dudau@xxxxxxx wrote:
> On Wed, Apr 27, 2016 at 03:26:59PM +0100, Lorenzo Pieralisi wrote:
> > On Tue, Apr 26, 2016 at 09:39:16PM -0500, Bjorn Helgaas wrote:
> > > On Fri, Apr 15, 2016 at 07:06:40PM +0200, Tomasz Nowicki wrote:
> > > > Platforms that have memory mapped IO port (such as ARM64) need special
> > > > handling for PCI I/O resources. For host bridge's resource probing case
> > > > these resources need to be fixed up with pci_register_io_range/pci_remap_iospace etc.
> > >
> > > ia64 also has memory-mapped I/O port space. It would be ideal to find
> > > some way to handle ia64 and ARM64 similarly. At the very least, we
> > > have to make sure that this doesn't break ia64. The ia64 dense/sparse
> > > I/O spaces complicate things; I don't know if ARM64 has something
> > > similar or not.
> >
> > No it does not, and that's exactly the same problem we faced with
> > the DT generic version of of_pci_range_to_resource() which basically
> > relies on PCI_IOBASE to be defined to add code that creates IO port
> > resources out of the MMIO resource describing how IO port space is
> > mapped to MMIO (physical) address space.
> >
> > IIRC everything hinges on PCI_IOBASE definition to make sure that
> > of_pci_range_to_resource() *works*, which means that if PCI_IOBASE is
> > not defined (ie IA64) that code - acpi_pci_root_remap_iospace() in this
> > case - does nothing.
> >
> > So acpi_pci_root_remap_iospace() is of_pci_range_to_resource() ACPI
> > equivalent + the pci_remap_iospace() call (I have to dig into the
> > logs to check why Liviu did not add a call to pci_remap_iospace()
> > in of_pci_get_host_bridge_resources() - I want to do that actually).
>
> Because of_pci_get_host_bridge_resources() only gives you a list of
> resources, it doesn't allocate them. An arch or platform could add
> further filtering to that list before it gets requested (in our case
> it is done in pci-host-common.c)

Well, it does register the IO cpu physical address in pci_register_io_range()
though, if pci_remap_iospace() fails in arch/platform code we can delete the
resource but we must also unregister the corresponding cpu address from
the IO ranges otherwise we end up with stale entries in the io_range_list.

Anyway, it is not related to this thread, I will see what I can do
to improve that API from this standpoint.

Thanks !
Lorenzo

>
> Best regards,
> Liviu
>
> >
> > The point here is: IO space (in DT and ACPI) handling is arch specific.
> >
> > For DT, by relying on PCI_IOBASE, we left that code in drivers/of and
> > it works (well, with some niggles - see the thread with Murali on IO
> > space on TI keystone) for ARM/ARM64.
> >
> > http://www.spinics.net/lists/linux-pci/msg49725.html
> >
> > What are we going to do with the ACPI version ?
> >
> > Do we want to add an arch specific call that takes the raw resource
> > describing IO space and creates an IO port resource (and the MMIO
> > equivalent - that's what add_io_space() does in IA64) and use that
> > in generic ACPI parsing code ?
> >
> > Or we just do what Tomasz does, which is basically the approach we took
> > for DT ?
> >
> > > > Furthermore, the same I/O resources need to be released after hotplug
> > > > removal so that it can be re-added back by the pci_remap_iospace
> > > > function during insertion. Therefore we implement new pci_unmap_iospace call
> > > > which unmaps I/O space as the symmetry to pci_remap_iospace.
> > >
> > > "Furthermore" is a hint that we should check to see if this can be
> > > split into two patches.
> > >
> > > We already have a pci_remap_iospace(), and you're adding
> > > pci_unmap_iospace(), which will be used for hotplug removal. So let's
> > > add pci_unmap_iospace() first in a patch by itself because that's
> > > potentially useful for other callers of pci_remap_iospace(), even if
> > > they don't need the acpi_pci_root_remap_iospace() stuff.
> >
> > I agree.
> >
> > Thanks,
> > Lorenzo
> >
> > > > Signed-off-by: Jayachandran C <jchandra@xxxxxxxxxxxx>
> > > > Signed-off-by: Sinan Kaya <okaya@xxxxxxxxxxxxxx>
> > > > Signed-off-by: Tomasz Nowicki <tn@xxxxxxxxxxxx>
> > > > ---
> > > > drivers/acpi/pci_root.c | 33 +++++++++++++++++++++++++++++++++
> > > > drivers/pci/pci.c | 24 ++++++++++++++++++++++++
> > > > include/linux/pci.h | 1 +
> > > > 3 files changed, 58 insertions(+)
> > > >
> > > > diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> > > > index d9a70c4..815b6ca 100644
> > > > --- a/drivers/acpi/pci_root.c
> > > > +++ b/drivers/acpi/pci_root.c
> > > > @@ -742,6 +742,34 @@ next:
> > > > resource_list_add_tail(entry, resources);
> > > > }
> > > > }
> > > > +static void acpi_pci_root_remap_iospace(struct resource_entry *entry)
> > > > +{
> > > > +#ifdef PCI_IOBASE
> > > > + struct resource *res = entry->res;
> > > > + resource_size_t cpu_addr = res->start;
> > > > + resource_size_t pci_addr = cpu_addr - entry->offset;
> > > > + resource_size_t length = resource_size(res);
> > > > + unsigned long port;
> > > > +
> > > > + if (pci_register_io_range(cpu_addr, length))
> > > > + goto err;
> > > > +
> > > > + port = pci_address_to_pio(cpu_addr);
> > > > + if (port == (unsigned long)-1)
> > > > + goto err;
> > > > +
> > > > + res->start = port;
> > > > + res->end = port + length - 1;
> > > > + entry->offset = port - pci_addr;
> > > > +
> > > > + if (pci_remap_iospace(res, cpu_addr) < 0)
> > > > + goto err;
> > > > + pr_info("Remapped I/O %pa to %pR\n", &cpu_addr, res);
> > > > + return;
> > > > +err:
> > > > + res->flags |= IORESOURCE_DISABLED;
> > > > +#endif
> > > > +}
> > > >
> > > > int acpi_pci_probe_root_resources(struct acpi_pci_root_info *info)
> > > > {
> > > > @@ -763,6 +791,9 @@ int acpi_pci_probe_root_resources(struct acpi_pci_root_info *info)
> > > > "no IO and memory resources present in _CRS\n");
> > > > else {
> > > > resource_list_for_each_entry_safe(entry, tmp, list) {
> > > > + if (entry->res->flags & IORESOURCE_IO)
> > > > + acpi_pci_root_remap_iospace(entry);
> > > > +
> > > > if (entry->res->flags & IORESOURCE_DISABLED)
> > > > resource_list_destroy_entry(entry);
> > > > else
> > > > @@ -834,6 +865,8 @@ static void acpi_pci_root_release_info(struct pci_host_bridge *bridge)
> > > >
> > > > resource_list_for_each_entry(entry, &bridge->windows) {
> > > > res = entry->res;
> > > > + if (res->flags & IORESOURCE_IO)
> > > > + pci_unmap_iospace(res);
> > > > if (res->parent &&
> > > > (res->flags & (IORESOURCE_MEM | IORESOURCE_IO)))
> > > > release_resource(res);
> > > > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > > > index 89e9996..c0f8a4e 100644
> > > > --- a/drivers/pci/pci.c
> > > > +++ b/drivers/pci/pci.c
> > > > @@ -26,6 +26,7 @@
> > > > #include <linux/device.h>
> > > > #include <linux/pm_runtime.h>
> > > > #include <linux/pci_hotplug.h>
> > > > +#include <linux/vmalloc.h>
> > > > #include <asm/setup.h>
> > > > #include <linux/aer.h>
> > > > #include "pci.h"
> > > > @@ -3168,6 +3169,29 @@ int __weak pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr)
> > > > #endif
> > > > }
> > > >
> > > > +/**
> > > > + * pci_unmap_iospace - Unmap the memory mapped I/O space
> > > > + * @res: resource to be unmapped
> > > > + *
> > > > + * Unmap the CPU virtual address @res from virtual address space.
> > > > + * Only architectures that have memory mapped IO functions defined
> > > > + * (and the PCI_IOBASE value defined) should call this function.
> > > > + */
> > > > +void pci_unmap_iospace(struct resource *res)
> > > > +{
> > > > +#if defined(PCI_IOBASE) && defined(CONFIG_MMU)
> > > > + unsigned long vaddr = (unsigned long)PCI_IOBASE + res->start;
> > > > +
> > > > + unmap_kernel_range(vaddr, resource_size(res));
> > > > +#else
> > > > + /*
> > > > + * This architecture does not have memory mapped I/O space,
> > > > + * so this function should never be called.
> > > > + */
> > > > + WARN_ONCE(1, "This architecture does not support memory mapped I/O\n");
> > > > +#endif
> > > > +}
> > > > +
> > > > static void __pci_set_master(struct pci_dev *dev, bool enable)
> > > > {
> > > > u16 old_cmd, cmd;
> > > > diff --git a/include/linux/pci.h b/include/linux/pci.h
> > > > index c28adb4..df1f33d 100644
> > > > --- a/include/linux/pci.h
> > > > +++ b/include/linux/pci.h
> > > > @@ -1168,6 +1168,7 @@ int pci_register_io_range(phys_addr_t addr, resource_size_t size);
> > > > unsigned long pci_address_to_pio(phys_addr_t addr);
> > > > phys_addr_t pci_pio_to_address(unsigned long pio);
> > > > int pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr);
> > > > +void pci_unmap_iospace(struct resource *res);
> > > >
> > > > static inline pci_bus_addr_t pci_bus_address(struct pci_dev *pdev, int bar)
> > > > {
> > > > --
> > > > 1.9.1
> > > >
> > >
> >
>
> --
> ====================
> | I would like to |
> | fix the world, |
> | but they're not |
> | giving me the |
> \ source code! /
> ---------------
> ??\_(???)_/??

Next message: Jassi Brar: "Re: [PATCH] dmaengine: pl330: Fix race in pl330_get_desc()"
Previous message: Frederic Weisbecker: "Re: [PATCH/RFC v3] perf core: Allow setting up max frame stack depth via sysctl"
In reply to: Liviu . Dudau: "Re: [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources."
Next in thread: Bjorn Helgaas: "Re: [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]