[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YbjEvMTaW9+AiBZ8@zn.tnic>
Date: Tue, 14 Dec 2021 17:22:20 +0100
From: Borislav Petkov <bp@...en8.de>
To: Yazen Ghannam <yazen.ghannam@....com>
Cc: linux-edac@...r.kernel.org, linux-kernel@...r.kernel.org,
mchehab@...nel.org, tony.luck@...el.com, james.morse@....com,
rric@...nel.org, Smita.KoralahalliChannabasappa@....com,
william.roche@...cle.com
Subject: Re: [PATCH 4/4] EDAC/amd64: Add DDR5 support and related register
changes
On Mon, Dec 13, 2021 at 05:46:55PM +0000, Yazen Ghannam wrote:
> Yeah, sorry it's not clear. The purpose of the flag is to indicate some minor
> changes that show up with future systems like register offsets changes, etc. I
> didn't want to tie the name to a specific model or core name. I went with DDR5
> as a new feature that shows up with these changes, but they're not directly
> tied to DDR5.
>
> But yes, a system may support DDR5 and DDR4. And this can be detected from the
> hardware.
>
> What do you think about calling the flag "uses_f19h_m10h_offsets" or something
> like that? I was trying to avoid family/model in the name, but the code
> already does this all over. And the convention has been to call something by
> the first family/model where it shows up.
Good question.
So AFAIU, these register offset changes are probably going to propagate
beyond F19M10... In any case, they won't be tied to the family/model
so your flag idea is in the right direction, AFAICT. I'd do something
shorter, though, so that the code accessing it is short'n'sweet:
if (pvt->flags.f19h_regs_ng) - "new generation" regs :-)
or even
if (pvt->flags.zn_new_regs_fmt)
or whatever that's called. The GPU UMC is called UMC_v2 so I guess
if (pvt->flags.zn_regs_v2)
:-)
You get the idea...
With an ample explanation in a comment what that means, ofc.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
Powered by blists - more mailing lists