Re: [PATCH 05/11] x86, pci, acpi: Move arch-agnostic MMCONFIG (aka ECAM) and ACPI code out of arch/x86/ directory

From: Tomasz Nowicki
Date: Wed Sep 09 2015 - 09:48:43 EST


Hi Lorenzo,

On 08.09.2015 17:07, Lorenzo Pieralisi wrote:
Hi Tomasz,

On Mon, Sep 07, 2015 at 10:59:44AM +0100, Tomasz Nowicki wrote:
On 31.08.2015 13:01, Tomasz Nowicki wrote:
On 08.06.2015 17:14, Lorenzo Pieralisi wrote:
On Mon, Jun 08, 2015 at 03:57:38AM +0100, Hanjun Guo wrote:

[...]

Why can't we make use of the ECAM implementation used by
pci-host-generic
and drivers/pci/access.c?

We had that question when I had posted MMCFG patch set separately,
please see:
https://lkml.org/lkml/2015/3/11/492

Yes, but the real question is, why do we need to have PCI config
space
up and running before a bus struct is even created ? I think the
reason is
the PCI configuration address space format (ACPI 6.0, Table 5-27,
page
108):

"PCI Configuration space addresses must be confined to devices on
PCI Segment Group 0, bus 0. This restriction exists to accommodate
access to fixed hardware prior to PCI bus enumeration".

On HW reduced platforms I do not even think this is required at all,
we have to look into this to avoid code duplication that might well
turn out useless.

This is only for the fixed hardware, which will be not available for
ARM64 (reduced hardware mode), but in Generic Hardware Programming
Model, we using OEM-provided ACPI Machine Language (AML) code to
access
generic hardware registers, this will be available for reduced
hardware
too.

So in ACPI spec, it says: (ACPI 6.0 page 66, last paragraph)

ACPI defines eight address spaces that may be accessed by generic
hardware implementations. These include:
* System I/O space
* System memory space
* PCI configuration space
* Embedded controller space
* System Management Bus (SMBus) space
* CMOS
* PCI BAR Target
* IPMI space

So if any device using the PCI address space for control, such
as a system reset control device, its address space can be reside
in PCI configuration space (who can prevent a OEM do that crazy
thing? :) ), and it should be accessible before the PCI bus is
created.

Us, by changing attitude and questioning features whose usefulness
is questionable. I will look into this and raise the point, I am not
thrilled by the idea of adding another set of PCI accessor functions
and drivers because we have to access a register through PCI before
enumerating the bus (and on arm64 this is totally useless since
we are not meant to support fixed HW anyway). Maybe we can make acpica
code use a "special" stub (ACPI specific, PCI configuration space
address
space has restrictions anyway), I have to review this set in its
entirety to see how to do that (and I would kindly ask you to do
it too, before saying it is not possible to implement it).

I'm willing to do that, actually, if we don't need a mechanism to
access PCI config space before the bus is created, the code can be
simplified a lot.

After more investigation on the spec and the ACPI core code, I'm
still not convinced that accessing to PCI config space before PCI
bus creating is impossible, also there is no enough ARM64 hardware
to prove that too. But I think we can go in this way, reuse the
ECAM implementation by pci-host-generic for now, and implement the PCI
accessor functions before enumerating PCI bus when needed in the
future, does it make sense?

You mean we rewrite the patch to make sure we can use the PCI host
generic
driver with MCFG and we leave the acpica PCI config call empty stubs on
arm64 (as they are now) ?


Hi Bjorn, Rafael,

Lorenzo pointed out very important problem we are having with PCI config
space access for ARM64. Please refer to the above discussion and add
your 2 cents. Can we forget about accessing PCI config space (for
Hardware Reduced profile) before PCI bus creation? If not, do you see a
way to use drivers/pci/access.c accessors here, like acpica change? Any
opinion is very appreciated.


Kindly remainder.

I think (but I am happy to be corrected) that the map_bus() hook
(ie that's why struct pci_bus is required in eg pci_generic_config_write)
is there to ensure that when the generic accessors are called
a) we have a valid bus b) the host controllers implementing it
has been initialized.

I had another look and I noticed you are trying to solve multiple
things at once:

1) ACPICA seems to need PCI config space on bus 0 to be working
before PCI enumerates (ie before we have a root bus), we need to
countercheck on that, but you can't use the generic PCI accessors
for that reasons (ie root bus might not be available, you do not
have a pci_bus struct)
2) the raw_pci_read/write require _generic_ mmio back-ends, since AMD
can't cope with standard x86 read/write{b,w,l}

Overall, it seems to me that we can avoid code duplication by
shuffling your code a bit.

You could modify the generic accessors in drivers/pci/access.c to
use your mmio back-end instead of using plain read/write{b,w,l}
functions (we should check if RobH is ok with that there can be
reasons that prevent this from happening). This would solve the
AMD mmio issue.

By factoring out the code that actually carries out the reads
and writes in the accessors basically you decouple the functions
requiring the struct pci_bus from the ones that does not require it
(ie raw_pci_{read/write}.

The generic MMIO layer belongs in the drivers/pci/access.c file, it has
nothing to do with ECAM.

The mmcfg interface should probably live in pci-acpi.c, I do not think
you need an extra file in there but that's a detail.

Basically the generic accessors would become something like eg:

int pci_generic_config_write(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 val)
{
void __iomem *addr;

addr = bus->ops->map_bus(bus, devfn, where);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;

pci_mmio_write(size, addr + where, value);

return PCIBIOS_SUCCESSFUL;
}

With that in place using raw_pci_write/read or the generic accessors
becomes almost identical, with code requiring the pci_bus to be
created using the generic accessors and ACPICA using the raw version.

I might be missing something, so apologies if that's the case.


Actually, I think you showed me the right direction :) Here are some conclusions/comments/concerns. Please correct me if I am wrong:

1. We need raw_pci_write/read accessors (based on ECAM) for ARM64 too but only up to the point where buses are enumerated. From that point on, we should reuse generic accessors from access.c file, right?

2. For ARM64 ACPI PCI, we can use generic accessors right away, .map_bus would call common code part (pci_dev_base()). The only thing that worry me is fact that MCFG regions are RCU list so it needs rcu_read_lock() for the .map_bus (mcfg lookup) *and* read/write operation.

3. Changing generic accessors to introduce generic MMIO layer (because of AMD issue) like this:
int pci_generic_config_write(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 val)
{
void __iomem *addr;

addr = bus->ops->map_bus(bus, devfn, where);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;

pci_mmio_write(size, addr + where, val);

return PCIBIOS_SUCCESSFUL;
}
would imply using those accessors for x86 ACPI PCI host bridge driver, see arch/x86/pci/common.c

int raw_pci_read(unsigned int domain, unsigned int bus, unsigned int devfn,
int reg, int len, u32 *val)
{
if (domain == 0 && reg < 256 && raw_pci_ops)
return raw_pci_ops->read(domain, bus, devfn, reg, len, val);
if (raw_pci_ext_ops)
return raw_pci_ext_ops->read(domain, bus, devfn, reg, len, val);
return -EINVAL;
}
[...]
static int pci_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
{
return raw_pci_read(pci_domain_nr(bus), bus->number,
devfn, where, size, value);
}
[...]
struct pci_ops pci_root_ops = {
.read = pci_read,
.write = pci_write,
};

Currently, the above code may call lots of different accessors (not necessarily generic accessor friendly :), moreover it possible that x86 may have registered two accessor sets (raw_pci_ops, raw_pci_ext_ops). I am happy to fix that but I would need x86 PCI expert to get know if that is possible at all.

I really appreciate your help.

Thanks,
Tomasz
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/