[PATCH v4] PCI: of: Clear 64-bit flag for non-prefetchable memory below 4GB

[Punit Agrawal]

Hi Alex,

Alexandru Elisei <alexandru.elisei at arm.com> writes:

> Hi Punit,
>
> Thank you for working on this!
>
> On 6/15/21 12:04 AM, Punit Agrawal wrote:
>> Alexandru and Qu reported this resource allocation failure on
>> ROCKPro64 v2 and ROCK Pi 4B, both based on the RK3399:
>>
>>   pci_bus 0000:00: root bus resource [mem 0xfa000000-0xfbdfffff 64bit]
>>   pci 0000:00:00.0: PCI bridge to [bus 01]
>>   pci 0000:00:00.0: BAR 14: no space for [mem size 0x00100000]
>>   pci 0000:01:00.0: reg 0x10: [mem 0x00000000-0x00003fff 64bit]
>>
>> "BAR 14" is the PCI bridge's 32-bit non-prefetchable window, and our
>> PCI allocation code isn't smart enough to allocate it in a host
>> bridge window marked as 64-bit, even though this should work fine.
>>
>> A DT host bridge description includes the windows from the CPU
>> address space to the PCI bus space.  On a few architectures
>> (microblaze, powerpc, sparc), the DT may also describe PCI devices
>> themselves, including their BARs.
>>
>> Before 9d57e61bf723 ("of/pci: Add IORESOURCE_MEM_64 to resource
>> flags for 64-bit memory addresses"), of_bus_pci_get_flags() ignored
>> the fact that some DT addresses described 64-bit windows and BARs.
>> That was a problem because the virtio virtual NIC has a 32-bit BAR
>> and a 64-bit BAR, and the driver couldn't distinguish them.
>>
>> 9d57e61bf723 set IORESOURCE_MEM_64 for those 64-bit DT ranges, which
>> fixed the virtio driver.  But it also set IORESOURCE_MEM_64 for host
>> bridge windows, which exposed the fact that the PCI allocator isn't
>> smart enough to put 32-bit resources in those 64-bit windows.
>>
>> Clear IORESOURCE_MEM_64 from host bridge windows since we don't need
>> that information.
>
> I've tested the patch on my rockpro64. Kernel built from tag v5.13-rc6:
>
> [    0.345676] pci 0000:01:00.0: 8.000 Gb/s available PCIe bandwidth, limited by
> 2.5 GT/s PCIe x4 link at 0000:00:00.0 (capable of 31.504 Gb/s with 8.0 GT/s PCIe
> x4 link)
> [    0.359300] pci_bus 0000:01: busn_res: [bus 01-1f] end is updated to 01
> [    0.359343] pci 0000:00:00.0: BAR 14: no space for [mem size 0x00100000]
> [    0.359365] pci 0000:00:00.0: BAR 14: failed to assign [mem size 0x00100000]
> [    0.359387] pci 0000:01:00.0: BAR 0: no space for [mem size 0x00004000 64bit]
> [    0.359407] pci 0000:01:00.0: BAR 0: failed to assign [mem size 0x00004000 64bit]
> [    0.359428] pci 0000:00:00.0: PCI bridge to [bus 01]
> [    0.359862] pcieport 0000:00:00.0: PME: Signaling with IRQ 76
> [    0.360190] pcieport 0000:00:00.0: AER: enabled with IRQ 76
>
> Kernel built from tag v5.13-rc6 with this patch applied:
>
> [    0.345434] pci 0000:01:00.0: 8.000 Gb/s available PCIe bandwidth, limited by
> 2.5 GT/s PCIe x4 link at 0000:00:00.0 (capable of 31.504 Gb/s with 8.0 GT/s PCIe
> x4 link)
> [    0.359081] pci_bus 0000:01: busn_res: [bus 01-1f] end is updated to 01
> [    0.359128] pci 0000:00:00.0: BAR 14: assigned [mem 0xfa000000-0xfa0fffff]
> [    0.359155] pci 0000:01:00.0: BAR 0: assigned [mem 0xfa000000-0xfa003fff 64bit]
> [    0.359217] pci 0000:00:00.0: PCI bridge to [bus 01]
> [    0.359239] pci 0000:00:00.0:   bridge window [mem 0xfa000000-0xfa0fffff]
> [    0.359422] pcieport 0000:00:00.0: enabling device (0000 -> 0002)
> [    0.359687] pcieport 0000:00:00.0: PME: Signaling with IRQ 76
> [    0.360001] pcieport 0000:00:00.0: AER: enabled with IRQ 76
>
> And the NVME on the PCIE expansion card works as expected:
>
> Tested-by: Alexandru Elisei <alexandru.elisei at arm.com>

Thanks a lot for the retest and the detailed logs.

Punit

[...]