Authored by: Anonymous on Sunday, March 31 2013 @ 04:27 PM EDT |
There's some RAID code that does on x86, as well as (out
of tree) OSS4.
The code that does is wrapped with
kernel_fpu_begin()/kernel_fpu_end()[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Sunday, March 31 2013 @ 04:47 PM EDT |
There's some RAID code that does on x86, as well as (out
of tree) OSS4.
The code that does is wrapped with
kernel_fpu_begin()/kernel_fpu_end()[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, April 01 2013 @ 01:14 PM EDT |
>No, IIRC, the kernel doesn’t do floating point it manages it for user
processes.
Can't speak for whether it DOES floating point, but no way does it manage it for
user processes.
The hardware has registers that do floating point; at most, the kernel may save
user-registers on context swaps (while being utterly agnostic about the
registers' purpose or functionality).[ Reply to This | Parent | # ]
|
|