Date: Fri, 12 Jan 2018 11:07:06 -0700 From: Brett Glass <brett@lariat.org> To: Oliver Pinter <oliver.pinter@hardenedbsd.org>, "Zahrir, Abderrahmane" <Abderrahmane.Zahrir@ca.com> Cc: "freebsd-security@freebsd.org" <freebsd-security@freebsd.org> Subject: Re: Response to Meltdown and Spectre Message-ID: <201801121807.LAA16736@mail.lariat.net> In-Reply-To: <CAPQ4ffsL40LsNM1deHLeSQtwAcjszqJC%2BLSd5KiSvncrPiU6jQ@mail.g mail.com> References: <CY1PR01MB124768D9AE4AB4D9CDAB565B8F170@CY1PR01MB1247.prod.exchangelabs.com> <CAPQ4ffsL40LsNM1deHLeSQtwAcjszqJC%2BLSd5KiSvncrPiU6jQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
All: The fix in this patch appears to be unconditional. I do a lot of work with embedded Intel Atom processors, which do not implement speculative execution. (Only one of them even implements limited out-of-order execution.) These systems aren't terribly powerful by today's standards, and would suffer serious performance impacts should the fix be turned on unnecessarily. Will there be automatic detection of 32-bit Atoms (all of which are definitely not susceptible) and 64-bit Atoms (all of which, except for Avoton, are definitely not susceptible; Avoton might have some very limited exposure due to out-of-order execution, but may also not be vulnerable because the OOE is not speculative) to avoid unnecessary performance impacts? --Brett Glass At 05:14 AM 1/12/2018, Oliver Pinter wrote: >The test patch is here: https://reviews.freebsd.org/D13797
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201801121807.LAA16736>