Re: [PATCH v2 2/5] mm: Factor out the pdx compression logic in ma/va converters

[Julien Grall <julien@xxxxxxx>]

Hi Jan,

On 08/08/2023 07:05, Jan Beulich wrote:
> On 07.08.2023 18:26, Alejandro Vallejo wrote:
> > On Mon, Jul 31, 2023 at 05:15:19PM +0200, Jan Beulich wrote:
> > > Especially since the names of the helper functions are longish,
> > > I'm afraid I'm not fully convinced of the transformation.
> > In what sense? If it's just naming style I'm happy to consider other names,
> > but taking compression logic out of non-pdx code is essential to removing
> > compiling it out.
> I understand that, but my dislike for long names of functions doing
> basic transformations remains. I'm afraid I have no good alternative
> suggestion, though, and while it's been a long time, this may also
> have been one reason why I didn't go with helpers back at the time.
>
> Plus of course I remain unconvinced that compiling out really is the
> better option compared to patching out, as was done by my series.
I am with Alejandro here. The altpatching is more complex and would 
likely require arch specific code.
I don't exactly see the benefits of such approach given that are no 
known x86 platform where PDX might be useful. Even for Arm, I am not 
aware of a platform where PDX could be disabled. I agree this could 
change in the future, but this could be revisit if needed.
Cheers,

--
Julien Grall