Re: [PATCH 1/2] x86: Remove dynamic NOP selection
From: Thorsten Glaser
Date: Sun Jan 21 2024 - 20:18:56 EST
Steven Rostedt dixit:
>> >This was, incidentally, triggered by looking into a problem report that
>> >something did *not* work on a Geode LX system.
>
>What problem happened?
It turned out to be a compiler issue (GCC thinks i686 means PPro,
not 686-class CPUs, and -fcf-protection causes long NOPs, which
not all 686-class CPUs support, to be inserted). This turned out
to break a large part of Debian stable on OLPCs and other systems,
and the kernel’s changes in nopl handling were tabled as arguments.
https://www.jookia.org/wiki/Nopl has a longer writeup on the nopl
history.
bye,
//mirabilos
--
<igli> exceptions: a truly awful implementation of quite a nice idea.
<igli> just about the worst way you could do something like that, afaic.
<igli> it's like anti-design. <mirabilos> that too… may I quote you on that?
<igli> sure, tho i doubt anyone will listen ;)