Re: [PATCH v6] x86: mce: kexec: switch MCE handler for kexec/kdump

From: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
To: Borislav Petkov <bp@alien8.de>
Cc: "Luck, Tony" <tony.luck@intel.com>,
	Prarit Bhargava <prarit@redhat.com>,
	Vivek Goyal <vgoyal@redhat.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Junichi Nomura <j-nomura@ce.jp.nec.com>,
	Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Subject: Re: [PATCH v6] x86: mce: kexec: switch MCE handler for kexec/kdump
Date: Fri, 6 Mar 2015 09:09:30 +0000	[thread overview]
Message-ID: <20150306090930.GA14982@hori1.linux.bs1.fc.nec.co.jp> (raw)
In-Reply-To: <20150306083421.GD3514@pd.tnic>

On Fri, Mar 06, 2015 at 09:34:21AM +0100, Borislav Petkov wrote:
> On Fri, Mar 06, 2015 at 02:59:13AM +0000, Naoya Horiguchi wrote:
> > From 8890e9976c525a4b480bf5f86008641688de8c11 Mon Sep 17 00:00:00 2001
> > From: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
> > Date: Fri, 6 Mar 2015 11:52:10 +0900
> > Subject: [PATCH v6] x86: mce: kexec: switch MCE handler for kexec/kdump
> > 
> > kexec disables (or "shoots down") all CPUs other than a crashing CPU before
> > entering the 2nd kernel. But the MCE handler is still enabled after that,
> > so if MCE happens and broadcasts over the CPUs after the main thread starts
> > the 2nd kernel (which might not initialize MCE device yet, or might decide
> > not to enable it,) MCE handler runs only on the other CPUs (not on the main
> > thread,) leading to kernel panic with MCE synchronization. The user-visible
> > effect of this bug is kdump failure.
> > 
> > Our standard MCE handler do_machine_check() assumes some about system's
> > status and it's hard to alter it to cover kexec/kdump context, so let's add
> > another kdump-specific one and switch to it.
> > 
> > Note that this problem exists since current MCE handler was implemented in
> > 2.6.32, and recently commit 716079f66eac ("mce: Panic when a core has reached
> > a timeout") made it more visible by changing the default behavior of the
> > synchronization timeout from "ignore" to "panic".
> > 
> > Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
> 
> ...
> 
> > +static void machine_check_under_kdump(struct pt_regs *regs, long error_code)
> > +{
> > +	struct mce m = {};
> > +	char *msg = NULL;
> > +	char *nmsg = NULL;
> > +	int i;
> > +	int worst = 0;
> > +	int severity;
> > +	int ret;
> 
> if you do here
> 
> 	if (mce_cfg.disabled)
> 		return;
> 
> you can use the simple rdmsrl variants and not the _safe() ones with
> exception handling.

I'm not sure why that works, could you elabroate it?

I feel that we need some comment about why it's OK to use rdmsrl_safe()
variant *only* in mce_rdmsrl() (other MCE registers is accessed via rdmsrl(),)
which seems not clear to me from reading current code and git history.

Thanks,
Naoya Horiguchi