<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">On 04/01/18 19:56, Paul Davis wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CAFa_cK=kHDS8jq2BmtVjPnBF4y8FdKC4f8zvvKruXrLXYEQ3vw@mail.gmail.com">
<div dir="ltr">
<div class="gmail_extra"><br>
<div class="gmail_quote">On Thu, Jan 4, 2018 at 2:43 PM, Robin
Gareus <span dir="ltr"><<a
href="mailto:robin@gareus.org" target="_blank"
moz-do-not-send="true">robin@gareus.org</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex"><span
class=""><br>
</span>Depends on your DAW. In Ardour's case there is no
context-switch. All<br>
tracks, Busses and FX run in the same process. So there
won't be any<br>
direct effect.<br>
<br>
Low-latency I/O however (soundcard) will see additional
overhead, as<br>
will JACK with multiple clients (one context switch per
client).<br>
</blockquote>
<div><br>
<div class="h5">is this correct? my impression is that
KPTI makes user->kernel and kernel->user switches
more expensive tha they used to be, but does not change
the cost of a context switch that stays in the same ring
mode.<br>
<div style="font-family:arial,helvetica,sans-serif"
class="gmail_default"><br>
</div>
<div style="font-family:arial,helvetica,sans-serif"
class="gmail_default">so, basically, system calls cost
more; context switches stay at about the same cost.<br>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<br>
<a class="moz-txt-link-freetext" href="https://lwn.net/Articles/741878/">https://lwn.net/Articles/741878/</a> - top-most page table is duplicated
and stripped of kernel pages; duplicate is active when running
userspace.<br>
<br>
</body>
</html>