Re: [PATCH v2 05/18] x86/fixmap: Generalize the GDT fixmap mechanism

From: Andy Lutomirski
Date: Thu Nov 23 2017 - 10:25:07 EST

On Wed, Nov 22, 2017 at 9:32 AM, Borislav Petkov <bp@xxxxxxx> wrote:
> On Wed, Nov 22, 2017 at 09:16:00AM -0800, Andy Lutomirski wrote:
>> Agreed, except that the fixmap enum needs to know
>> sizeof(cpu_entry_area), and I'm really hesitant to add yet another
>> header dependency.
> Perhaps a separate asm/cpuarea.h. asm/cpu.h looks small enough but it
> has hotplug and other misc stuff in there.
> Bah, our header separation needs a serious cleanup. ;-\
>> My general habit is that I like the != 0 here because I'm doing
>> arithmetic rather than thinking of % as some kind of logical operator.
>> I.e. I find it easier to understand the way I wrote it.
> I know, and I can recognize the code you wrote in arch/x86/ just by
> those tests, without looking at git blame output. :-)
> In my case, without the != 0 reads easier.

I'm making the kernel more personal!