Re: [PATCH v3] x86/amd: Address AMD erratum #1485

["Roger Pau Monné" <roger.pau@citrix.com>]

On Tue, Oct 17, 2023 at 11:21:47AM +0200, Jan Beulich wrote:
> On 17.10.2023 11:13, Roger Pau Monné wrote:
> > On Tue, Oct 17, 2023 at 08:50:45AM +0100, Andrew Cooper wrote:
> >> On 17/10/2023 8:44 am, Jan Beulich wrote:
> >>> On 13.10.2023 17:38, Alejandro Vallejo wrote:
> >>>> Fix adapted off Linux's mailing list:
> >>>>   https://lore.kernel.org/lkml/D99589F4-BC5D-430B-87B2-72C20370CF57@exactcode.com/T/#u
> >>> Why reference the bug report when there's a proper commit (f454b18e07f5) now?
> >>> Plus in any event a short summary of the erratum would help if put right here
> >>> (without needing to look up any documents or follow any links).
> >>
> >> That is not public information yet.  The erratum number alone is the
> >> best we can do at this juncture.
> >>>> --- a/xen/arch/x86/cpu/amd.c
> >>>> +++ b/xen/arch/x86/cpu/amd.c
> >>>> @@ -1004,6 +1004,28 @@ static void cf_check zen2_disable_c6(void *arg)
> >>>>  	wrmsrl(MSR_AMD_CSTATE_CFG, val & mask);
> >>>>  }
> >>>>  
> >>>> +static void amd_check_erratum_1485(void)
> >>>> +{
> >>>> +	uint64_t val, chickenbit = (1 << 5);
> >>> Linux gives the bit a name. Any reason you don't?
> >>
> >> There are multiple different names depending on where you look, and none
> >> are particularly relevant here.
> > 
> > Could we make chickenbit const static?
> > 
> > I would also use ULL just to be on the safe side, because we then copy
> > this for a different bit and it explodes.
> 
> I guess the way it is resembles what we already have in amd_check_zenbleed().
> Also it's not clear to me why besides "const" you also ask for "static".

Yes, makes no sense to put in .rodata, sorry, just const.

Roger.