[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1161eb84-4ebc-f938-f0ef-82ff6440dbd5@amazon.de>
Date: Thu, 9 Jan 2020 23:30:18 +0100
From: Jan H. Schönherr <jschoenh@...zon.de>
To: "Luck, Tony" <tony.luck@...el.com>,
"Ghannam, Yazen" <Yazen.Ghannam@....com>,
Borislav Petkov <bp@...en8.de>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>, "x86@...nel.org" <x86@...nel.org>
Subject: Re: [PATCH v2 6/6] x86/mce: Dynamically register default MCE handler
On 09/01/2020 22.54, Luck, Tony wrote:
>> It seems to me that the issue is the mcelog notifier counts toward the number
>> of notifiers, so the default notifier doesn't print anything.
>
> If we gave a API to the notifiers to say whether to suppress printing, then the
> dev_mcelog() code could do the suppression only if some process had
> /dev/mcelog open. So if mcelog(8) wasn't running, you'd still see the console
> message.
I briefly looked into that.
There is the issue that mcelog code buffers MCEs unconditionally. And we probably
don't want to deactivate that, so that MCEs during boot can be queried
a bit later via /dev/mcelog.
We would get a bit of duplicate logging, if we let mcelog report "supress
printing" only if there is an actual consumer. (Or if there was a consumer
once, in case there are periodically polling consumers.)
Regards
Jan
--
Amazon Development Center Germany GmbH
Krausenstr. 38
10117 Berlin
Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss
Eingetragen am Amtsgericht Charlottenburg unter HRB 149173 B
Sitz: Berlin
Ust-ID: DE 289 237 879
Powered by blists - more mailing lists