<br>
<div class="gmail_quote"><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div class="im"><br>
On Tue, Jan 4, 2011 at 22:44, Dave Hylands <<a href="mailto:dhylands@gmail.com">dhylands@gmail.com</a>> wrote:<br>
> That is correct. In some architectures, attempts to use floating point<br>
> from the kernel will work. I've seen some x86 code that uses it.<br>
<br>
<br>
</div>AFAIK, once x86 didn't supported due to floating point related<br>
registers are not correctly (or even doing?) saved and restored during<br>
context switching. So maybe it is fixed now...<br>
<font color="#888888"><br>
--</font><br></blockquote><div><br>I've often wondered about this oft-cited kernel behaviour too, in my naivety. I understand<br>that this must be on a per-arch basis, but does this mean that the kernel doesn't police FP<br>
access at _all_ (perhaps this is what Mohit means too)? Does code like X for example have to access it directly, or does it just use the GPU? What about other user-space code - does it<br>have a separate library and do its own security? Video drivers?<br>
Sorry if these are basic questions, I grepped for float in the kernel but as-yet the associated<br>code looks really arcane to me - if anyone could answer any of these questions generally <br>(if that's possible) that would be very helpful with visualizing the mechanism. <br>
Maybe I'm looking in the wrong place.<br><br>When I started looking at the kernel I imagined this small, neat, concise piece of highly<br>efficient code so I wasn't surprised there was no float (don't laugh - how one learns :-/ ) ... <br>
I suppose any float per-arch 'hacks' (to get a larger word size) would not be worth<br>the overhead of the mode switch and extra code?<br><br>Thanks <br>Julie<br></div></div>