Re: Saving the Floating Point State
Re: Saving the Floating Point State
- Subject: Re: Saving the Floating Point State
- From: Terry Lambert <email@hidden>
- Date: Mon, 26 Apr 2010 10:31:43 -0700
On Apr 26, 2010, at 8:49 AM, Duane Murphy <email@hidden> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Apr 26, 2010, at 8:35 AM, Michael Smith wrote:
On Apr 25, 2010, at 7:02 PM, Quinn wrote:
At 13:40 -0700 23/4/10, Duane Murphy wrote:
Does OS X/Darwin have a function for saving and restoring the
state of the floating pointer registers?
No. The Darwin model is that you should use the floating point
unit at will; the kernel will notice, trap, save the user's FP
state, clear the FP state, and continue running your code. At that
point the kernel knows that the FP unit contains kernel stuff, and
preserve it across kernel preemptions, and restores the user FP
state when the thread returns to user space.
Note that this only applies to the use of FP resources in a user
context, whereas I believe the OP was asking about the use of same
in the kernel.
This is correct. We want to add a process that uses FP resources in
the kernel.
Kernel code must not use FP resources.
Huh? That seems like an onerous restriction. Can someone please
explain this a little better?
There are some useful algorithms that take use floating point
registers. How might I take advantage of these algorithms?
By doing them in user space?
Why does your code need to run in protected mode?
-- Terry
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden