Re: [PATCH] acpi: pci: don't ignore function ID of bridge device in _PRT

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 7 April 2017 at 14:22, Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx> wrote:
> We currently derive legacy interrupt routing by matching _PRT
> entries on the PCI device only, presumably under the assumption
> that PRT entries always have a value of 0xffff in the function
> field, meaning 'match all functions'.
>
> This no longer holds for modern PCIe topologies, where the
> legacy interrupts for different slots may be wired to different
> functions on the same bridge device. For instance, on AMD Seattle,
> we may have something like
>
> -[0000:00]-+-00.0  Advanced Micro Devices, Inc. [AMD] Device 1a00
>            +-02.0  Advanced Micro Devices, Inc. [AMD] Device 1a01
>            +-02.2-[01]----00.0  Renesas uPD720202 USB 3.0 Host Controller
>            \-02.3-[02]----00.0  Realtek RTL8169 PCIe Gigabit Ethernet
>
> where the _PRT describes the legacy interrupt routing as
>
>     Name (_PRT, Package ()  // _PRT: PCI Routing Table
>     {
>         // slot 1: dev 2 fn 1
>         Package () { 0x20001, 0x0, 0x0, 0x140 },
>         Package () { 0x20001, 0x1, 0x0, 0x141 },
>         Package () { 0x20001, 0x2, 0x0, 0x142 },
>         Package () { 0x20001, 0x3, 0x0, 0x143 },
>
>         // slot 1: dev 2 fn 2
>         Package () { 0x20002, 0x0, 0x0, 0x144 },
>         Package () { 0x20002, 0x1, 0x0, 0x145 },
>         Package () { 0x20002, 0x2, 0x0, 0x146 },
>         Package () { 0x20002, 0x3, 0x0, 0x147 },
>
>         // slot 1: dev 2 fn 3
>         Package () { 0x20003, 0x0, 0x0, 0x148 },
>         Package () { 0x20003, 0x1, 0x0, 0x149 },
>         Package () { 0x20003, 0x2, 0x0, 0x14a },
>         Package () { 0x20003, 0x3, 0x0, 0x14b }
>     }) // _PRT
>
> The current code always matches on the first entry when trying to
> derive the legacy interrupt routing for the USB and Ethernet controllers
> behind bridges 02.2 and 02.3, which results in the wrong entry to
> be selected.
>
> So fix this by matching the function ID to the value in the _PRT entry
> if the _PRT entries are not using 0xffff in the lower word to match
> all functions.
>
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> ---
>  drivers/acpi/pci_irq.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c
> index c576a6fe4ebb..caac4ac94418 100644
> --- a/drivers/acpi/pci_irq.c
> +++ b/drivers/acpi/pci_irq.c
> @@ -160,6 +160,8 @@ static int acpi_pci_irq_check_entry(acpi_handle handle, struct pci_dev *dev,
>         struct acpi_prt_entry *entry;
>
>         if (((prt->address >> 16) & 0xffff) != device ||
> +           ((prt->address & 0xffff) != 0xffff &&
> +            (prt->address & 0xffff) != PCI_FUNC(dev->devfn)) ||
>             prt->pin + 1 != pin)
>                 return -ENODEV;
>
> --
> 2.9.3
>

It has been pointed out to me by Alan that the current wording of the
ACPI spec requires the function field of the address to be 0xffff.

Section 6.2.13 of the ACPI spec has this note

Note: The PCI function number in the Address field of the _PRT
packages must be 0xFFFF, indicating
“any” function number or “all functions”

of which I wonder if it makes sense for bridges as well.
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux IBM ACPI]     [Linux Power Management]     [Linux Kernel]     [Linux Laptop]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Samba]     [Video 4 Linux]     [Device Mapper]     [Linux Resources]

  Powered by Linux