Re: [patch v3 5/7] x86/smp: Cure kexec() vs. mwait_play_dead() breakage

From: Thomas Gleixner <tglx@linutronix.de>
To: Borislav Petkov <bp@alien8.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
	x86@kernel.org, Mario Limonciello <mario.limonciello@amd.com>,
	Tom Lendacky <thomas.lendacky@amd.com>,
	Tony Battersby <tonyb@cybernetics.com>,
	Ashok Raj <ashok.raj@linux.intel.com>,
	Tony Luck <tony.luck@intel.com>,
	Arjan van de Veen <arjan@linux.intel.com>,
	Eric Biederman <ebiederm@xmission.com>,
	Ashok Raj <ashok.raj@intel.com>
Subject: Re: [patch v3 5/7] x86/smp: Cure kexec() vs. mwait_play_dead() breakage
Date: Tue, 20 Jun 2023 14:25:48 +0200	[thread overview]
Message-ID: <87bkha8gmr.ffs@tglx> (raw)
In-Reply-To: <20230620092324.GWZJFwDDVo3TRUI0Ck@fat_crate.local>

On Tue, Jun 20 2023 at 11:23, Borislav Petkov wrote:
> On Thu, Jun 15, 2023 at 10:33:57PM +0200, Thomas Gleixner wrote:
>> TLDR: It's a mess.
>>  	while (1) {
>> @@ -1824,10 +1836,57 @@ static inline void mwait_play_dead(void)
>
> JFYI: that last hunk has some conflicts applying to latest tip/master.
> Might need merge resolving...

Yes, I know.

>> +/*
>> + * Kick all "offline" CPUs out of mwait on kexec(). See comment in
>> + * mwait_play_dead().
>> + */
>> +void smp_kick_mwait_play_dead(void)
>> +{
>> +	u32 newstate = CPUDEAD_MWAIT_KEXEC_HLT;
>
> Do you even need this newstate thing?

Yes, for two reasons:

  1) To explicitely tell the other CPU to go into HLT. MWAIT can resume
     execution due to SMIs or NMIs, so we don't want to go them into HLT
     unconditionally. TLD; .... :)

  2) Two have the state feedback from the other CPU.

>> +
>> +		if (READ_ONCE(md->status) != newstate)
>> +			pr_err("CPU%u is stuck in mwait_play_dead()\n", cpu);
>
> Shouldn't this be a pr_err_once thing so that it doesn't flood the
> console unnecessarily?

Yes, no, do not know :)