[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y2kuixb0RU9BxKls@araj-dh-work>
Date: Mon, 7 Nov 2022 16:12:59 +0000
From: Ashok Raj <ashok.raj@...el.com>
To: Borislav Petkov <bp@...en8.de>
CC: Thomas Gleixner <tglx@...utronix.de>,
LKML Mailing List <linux-kernel@...r.kernel.org>,
X86-kernel <x86@...nel.org>, Tony Luck <tony.luck@...el.com>,
Dave Hansen <dave.hansen@...el.com>,
Arjan van de Ven <arjan.van.de.ven@...el.com>,
Andy Lutomirski <luto@...nel.org>,
"Jacon Jun Pan" <jacob.jun.pan@...el.com>,
Tom Lendacky <thomas.lendacky@....com>,
"Kai Huang" <kai.huang@...el.com>,
Andrew Cooper <andrew.cooper3@...rix.com>,
Ashok Raj <ashok.raj@...el.com>
Subject: Re: [v2 01/13] x86/microcode/intel: Prevent printing updated
microcode rev multiple times
On Sun, Nov 06, 2022 at 02:35:58PM +0100, Borislav Petkov wrote:
> On Thu, Nov 03, 2022 at 05:58:49PM +0000, Ashok Raj wrote:
> > @@ -696,8 +697,7 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
> >
> > csig->rev = c->microcode;
> >
> > - /* No extra locking on prev, races are harmless. */
> > - if (csig->sig != prev.sig || csig->pf != prev.pf || csig->rev != prev.rev) {
> > + if (bsp && csig->rev != prev.rev) {
> > pr_info("sig=0x%x, pf=0x%x, revision=0x%x\n",
> > csig->sig, csig->pf, csig->rev);
>
> And now that we've established that we don't do mixed steppings anymore
> and the microcode revision is the same system-wide, you should simply
> drop this pr_info(), in your next patch you're adding
>
> +static u32 early_old_rev;
>
> That thing should simply be
>
> /* Currently applied microcode revision */
> static u32 microcode_rev;
>
> and you simply update that one each time you update microcode and print
> it as the previous and the new one and then write the new one into this
> var and that's it. Simple.
>
I removed that preparing for the next round. We already have late-loading
capture the previous rev already. So we don't need any new changes.
[ 482.242727] microcode: Reload completed, microcode revision: 0x2b000070 -> 0x2b000081
Only missing is the ucode date, not a big deal missing it.
Powered by blists - more mailing lists