On 07/08/2013 05:20 PM, Benjamin Herrenschmidt wrote:
> On Mon, 2013-07-08 at 14:44 +1000, Alexey Kardashevskiy wrote:
>
>> diff --git a/arch/powerpc/platforms/powernv/pci.h
>> b/arch/powerpc/platforms/powernv/pci.h
>> index 25d76c4..7ea82c1 100644
>> --- a/arch/powerpc/platforms/powernv/pci.h
>> +++ b/arch/powerpc/platforms/powernv/pci.h
>> @@ -52,6 +52,7 @@ struct pnv_ioda_pe {
>> int tce32_seg;
>> int tce32_segcount;
>> struct iommu_table tce32_table;
>> + phys_addr_t it_index_rm;
>
> Please ....
>
> The fact that we hijack the it_index field of the iommu table
> for the virtual address is bad enough, but really don't need
> to perpetuate this :-)
>
> Call the field something decent such as "tce_inval_reg_phys"
Yes we can. I just find it veeeeeeery attractive when I can grep "\<it_"
and get all users of iommu_table.
btw is phys_addr_t correct here?
--
Alexey
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
