RE: [PATCH v3 1/2] cpu/smt: add a notifier to notify the SMT changes

[Yang, WenYou]

[AMD Official Use Only - General]



> -----Original Message-----
> From: Yang, WenYou
> Sent: Wednesday, March 29, 2023 5:43 PM
> To: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Deucher, Alexander <Alexander.Deucher@xxxxxxx>; Koenig, Christian
> <Christian.Koenig@xxxxxxx>; Pan, Xinhui <Xinhui.Pan@xxxxxxx>; Quan, Evan
> <Evan.Quan@xxxxxxx>; Limonciello, Mario <Mario.Limonciello@xxxxxxx>;
> bp@xxxxxxxxx; jpoimboe@xxxxxxxxxx; Phillips, Kim <kim.phillips@xxxxxxx>;
> tglx@xxxxxxxxxxxxx; Yuan, Perry <Perry.Yuan@xxxxxxx>; Liang, Richard qi
> <Richardqi.Liang@xxxxxxx>; Li, Ying <YING.LI@xxxxxxx>; Liu, Kun
> <Kun.Liu2@xxxxxxx>; gpiccoli@xxxxxxxxxx; amd-gfx@xxxxxxxxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx
> Subject: RE: [PATCH v3 1/2] cpu/smt: add a notifier to notify the SMT changes
> 
> [AMD Official Use Only - General]
> 
> 
> 
> > -----Original Message-----
> > From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> > Sent: Wednesday, March 29, 2023 4:50 PM
> > To: Yang, WenYou <WenYou.Yang@xxxxxxx>
> > Cc: Deucher, Alexander <Alexander.Deucher@xxxxxxx>; Koenig, Christian
> > <Christian.Koenig@xxxxxxx>; Pan, Xinhui <Xinhui.Pan@xxxxxxx>; Quan,
> > Evan <Evan.Quan@xxxxxxx>; Limonciello, Mario
> > <Mario.Limonciello@xxxxxxx>; bp@xxxxxxxxx; jpoimboe@xxxxxxxxxx;
> > Phillips, Kim <kim.phillips@xxxxxxx>; tglx@xxxxxxxxxxxxx; Yuan, Perry
> > <Perry.Yuan@xxxxxxx>; Liang, Richard qi <Richardqi.Liang@xxxxxxx>; Li,
> > Ying <YING.LI@xxxxxxx>; Liu, Kun <Kun.Liu2@xxxxxxx>;
> > gpiccoli@xxxxxxxxxx; amd-gfx@xxxxxxxxxxxxxxxxxxxxx;
> > linux-kernel@xxxxxxxxxxxxxxx
> > Subject: Re: [PATCH v3 1/2] cpu/smt: add a notifier to notify the SMT
> > changes
> >
> > Caution: This message originated from an External Source. Use proper
> > caution when opening attachments, clicking links, or responding.
> >
> >
> > On Wed, Mar 29, 2023 at 07:23:29AM +0000, Yang, WenYou wrote:
> > > [AMD Official Use Only - General]
> >
> > ^^^ that has no business being in a public email.
> >
> > > Hi Peter,
> > >
> > > Thank you for your review.
> > >
> > > The purpose of the patch set is to improve the performance when
> > > playing
> > game for some AMD APUs with SMT enabled/disabled.
> > >
> > > When change the SMT state on the fly through " echo on/off >
> > /sys/devices/system/cpu/smt/control", the kernel needs to send a
> > message to notify PMFW to adjust a variable's value, which impacts the
> performance.
> >
> > When top posting I normally ignore the email. When not wrapping email
> > I typically get cranky. You 'win' *3* 'I cannot use email' trophies in a singly try.
> > Surely AMD has a HOWTO somewhere you can read?
> 
> Yes. It is my fault. Sorry.
> 
> >
> > So what do you want to have happen when someone goes and manually
> > offlines all the SMT siblings using
> > /sys/devices/system/cpu/cpu*/online
> > ?
> 
> I don't consider this situation.  Any suggestions will be deeply appreciated.

Hi Peter,

I don't find a good method to handle this situation.
Yes, manually offlining all the SMT sibling will get the same result of SMT disabling on the fly.

Actually, the normal way to enable/disable SMT on the fly is to echo on/off > /sys/device/system/cpu/smt/control

What are your opinions?

Best Regards,
Wenyou

> 
> >
> > I'm thinking that wants the same PMFW (whatever the heck that is)
> > notification change done, right?
> >
> > If the answer is "yes", then your patch does not meet the goals and is
> > inadequate.
> 
> Yes.  Need to improve.
> 
> Thank you.
> Wenyou