Re: [PATCH RFC V2 02/17] x86/fpu: Refactor arch_set_user_pkey_access() for PKS support

[Peter Zijlstra]

On Fri, Jul 17, 2020 at 01:52:55PM -0700, Ira Weiny wrote:
> On Fri, Jul 17, 2020 at 10:54:42AM +0200, Peter Zijlstra wrote:

> > Then we at least have a little clue wtf the thing does.. Yes I started
> > with a rename and then got annoyed at the implementation too.
> 
> On the code I think this is fair.  I've also updated the calling function to be
> a bit cleaner as well.
> 
> However, I think the name 'update' is a bit misleading.  Here is the new
> calling code:
> 
> ...
>         pkru = read_pkru();
> 	pkru = update_pkey_reg(pkru, pkey, init_val);
> 	write_pkru(pkru);
> ...
> 
> 
> I think it is odd to have a function called update_pkey_reg() called right
> before a write_pkru().  Can we call this update_pkey_value?  or just 'val'?
> Because write_pkru() actually updates the register.

Fair enough, update_pkey_val() works fine for me.