[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <3908561D78D1C84285E8C5FCA982C28F7B37DE1B@ORSMSX110.amr.corp.intel.com>
Date: Wed, 21 Feb 2018 01:05:59 +0000
From: "Luck, Tony" <tony.luck@...el.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
CC: Joe Konno <joe.konno@...ux.intel.com>,
"linux-efi@...r.kernel.org" <linux-efi@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Ard Biesheuvel <ard.biesheuvel@...aro.org>,
Matthew Garrett <matthew.garrett@...ula.com>,
Jeremy Kerr <jk@...abs.org>, Andi Kleen <ak@...ux.intel.com>,
Matthew Garrett <mjg59@...gle.com>,
"Peter Jones" <pjones@...hat.com>,
Andy Lutomirski <luto@...nel.org>,
James Bottomley <james.bottomley@...senpartnership.com>
Subject: RE: [PATCH 1/2] fs/efivarfs: restrict inode permissions
>> EFI[1] stinks. Reading any file in /sys/firmware/efi/efivars/ generates
>> 4 (yes FOUR!) SMIs.
> Is that actualkly the normal implementation?
I don't know if there are other implementations. This is what I see on my
lab system.
> Also, if these are just synchronous SMI's, then don't we just end up
> correctly assigning the CPU load to the reader, and it doesn't
> actually matter? Where's the DoS?
SMI are broadcast to all logical CPUs. The bigger the system the
more the pain from an SMI as code tries to rendezvous them all
(4 socket * 24 core * 2 HyperThreads = 192 CPUs on my system).
Looping reading files from efivarfs doesn't stop the system, but it
does slow to a crawl while it is happening. Not a full blown DoS,
but fairly unpleasant.
-Tony
Powered by blists - more mailing lists