[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1502732561.2042.141.camel@hpe.com>
Date: Mon, 14 Aug 2017 17:52:25 +0000
From: "Kani, Toshimitsu" <toshi.kani@....com>
To: "bp@...en8.de" <bp@...en8.de>
CC: "linux-acpi@...r.kernel.org" <linux-acpi@...r.kernel.org>,
"lenb@...nel.org" <lenb@...nel.org>,
"mchehab@...nel.org" <mchehab@...nel.org>,
"tony.luck@...el.com" <tony.luck@...el.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
"rjw@...ysocki.net" <rjw@...ysocki.net>
Subject: Re: [PATCH v2 4/7] ghes_edac: avoid multiple calls to dmi_walk()
On Mon, 2017-08-14 at 19:05 +0200, Borislav Petkov wrote:
> On Mon, Aug 14, 2017 at 04:48:57PM +0000, Kani, Toshimitsu wrote:
> > Right, but the issue is how [ghes_edac_]report_mem_error() protects
> > from possible concurrent calls from multiple GHES sources when
> > there is only a single mci.
>
> Do you know of an actual firmware reporting multiple errors
> concurrently?
I do not know. We have multiple GHES entries, but they all use SCI.
Since ACPICA uses a single threaded workqueue for notify handlers, they
are serialized among SCIs.
ACPI 6.2 defines multiple notification types in Table 18-383, and
ghes_proc() can be called from ghes_poll_func(), ghes_irq_func(), and
ghes_notify_sci(). So, I think it is safe to operate per an entry
basis.
> GHES v2 even needs to ACK the current error first before it can read
> the next one.
Yes, but this ACK is done per a GHES entry as well.
Thanks,
-Toshi
Powered by blists - more mailing lists