Re: [PATCH v5 09/11] hugetlb: Introduce generic version of huge_ptep_set_wrprotect

[Alexandre Ghiti]

Hi Michael,

Thanks, I will then remove those two specific implementations and we'll 
use the generic ones.

I send a v6 asap.

Thanks again,

Alex


On 08/03/2018 10:51 AM, Michael Ellerman wrote:
> Hi Alex,
>
> Sorry missed your previous mail.
>
> Alex Ghiti <alex@xxxxxxxx> writes:
> > Ok, I tried every defconfig available:
> >
> > - for the nohash/32, I found that I could use mpc885_ads_defconfig and I
> > activated HUGETLBFS.
> > I removed the definition of huge_ptep_set_wrprotect from
> > nohash/32/pgtable.h, add an #error in
> > include/asm-generic/hugetlb.h right before the generic definition of
> > huge_ptep_set_wrprotect,
> > and fell onto it at compile-time:
> > => I'm pretty confident then that removing the definition of
> > huge_ptep_set_wrprotect does not
> > break anythingin this case.
> Thanks, that sounds good.
>
> > - regardind book3s/32, I did not find any defconfig with
> > CONFIG_PPC_BOOK3S_32, CONFIG_PPC32
> > allowing to enable huge page support (ie CONFIG_SYS_SUPPORTS_HUGETLBFS)
> > => Do you have a defconfig that would allow me to try the same as above ?
> I think you're right, it's dead code AFAICS.
>
> We have:
>
> config PPC_BOOK3S_64
>          ...
> 	select SYS_SUPPORTS_HUGETLBFS
>
> config PPC_FSL_BOOK3E
>          ...
> 	select SYS_SUPPORTS_HUGETLBFS if PHYS_64BIT || PPC64
>
> config PPC_8xx
> 	...
> 	select SYS_SUPPORTS_HUGETLBFS
>
>
> So we can't ever enable HUGETLBFS for Book3S 32.
>
> Presumably the code got copied when we split the headers apart.
>
> So I think you can just ignore that one, and we'll delete it.
>
> cheers