2023-06-15 16:20:55

by Timothy Pearson

[permalink] [raw]
Subject: Re: linux-next: Tree for Jun 2 (arch/powerpc/kernel/iommu.c)



----- Original Message -----
> From: "Randy Dunlap" <[email protected]>
> To: "Timothy Pearson" <[email protected]>, "Michael Ellerman" <[email protected]>
> Cc: "Stephen Rothwell" <[email protected]>, "Linux Next Mailing List" <[email protected]>, "linux-kernel"
> <[email protected]>, "linuxppc-dev" <[email protected]>, "Alexey Kardashevskiy" <[email protected]>
> Sent: Thursday, June 15, 2023 11:00:08 AM
> Subject: Re: linux-next: Tree for Jun 2 (arch/powerpc/kernel/iommu.c)

> Hi Timothy,
>
> On 6/3/23 20:57, Timothy Pearson wrote:
>>
>>
>> ----- Original Message -----
>>> From: "Michael Ellerman" <[email protected]>
>>> To: "Randy Dunlap" <[email protected]>, "Stephen Rothwell"
>>> <[email protected]>, "Linux Next Mailing List"
>>> <[email protected]>
>>> Cc: "linux-kernel" <[email protected]>, "linuxppc-dev"
>>> <[email protected]>, "Alexey
>>> Kardashevskiy" <[email protected]>, "Timothy Pearson"
>>> <[email protected]>
>>> Sent: Saturday, June 3, 2023 7:22:51 PM
>>> Subject: Re: linux-next: Tree for Jun 2 (arch/powerpc/kernel/iommu.c)
>>
>>> Randy Dunlap <[email protected]> writes:
>>>> On 6/1/23 21:01, Stephen Rothwell wrote:
>>>>> Hi all,
>>>>>
>>>>> Changes since 20230601:
>>>>>
>>>>
>>>> On powerpc64, a randconfig failed with:
>>>>
>>>> In file included from ../include/linux/list.h:5,
>>>> from ../include/linux/preempt.h:11,
>>>> from ../include/linux/spinlock.h:56,
>>>> from ../include/linux/mmzone.h:8,
>>>> from ../include/linux/gfp.h:7,
>>>> from ../include/linux/slab.h:15,
>>>> from ../arch/powerpc/kernel/iommu.c:15:
>>>> ../arch/powerpc/kernel/iommu.c: In function
>>>> 'spapr_tce_setup_phb_iommus_initcall':
>>>> ../arch/powerpc/kernel/iommu.c:1391:36: error: 'hose_list' undeclared (first use
>>>> in this function); did you mean 'zonelist'?
>>>> 1391 | list_for_each_entry(hose, &hose_list, list_node) {
>>>> | ^~~~~~~~~
>>> ...
>>>
>>> hose_list is in pci-common.c which is built when PCI=y.
>>>
>>> PSERIES and POWERNV force PCI=y.
>>>
>>> But this config has neither:
>>>
>>> # CONFIG_PPC_POWERNV is not set
>>> # CONFIG_PPC_PSERIES is not set
>>> CONFIG_HAVE_PCI=y
>>> # CONFIG_PCI is not set
>>> # CONFIG_COMMON_CLK_RS9_PCIE is not set
>>>
>>>
>>> Probably the spapr_tce code should be wrapped in an #ifdef that is only
>>> enabled when POWERNV || PSERIES is enabled.
>>>
>>> cheers
>>
>> Sounds reasonable, I was going to look into this further over the weekend. I
>> can put together a patch for Monday if that works?
>
> Did you prepare a patch for this? I am still seeing this build error.
>
> thanks.
> --
> ~Randy

Yes, it was sent in to the linuxppc-dev list some weeks ago. Did it not arrive?