On Wed, 31 Mar 1999 [EMAIL PROTECTED] wrote:
> Good points. I think we need a better shared memory and understand the
> need for shared memory. I guess I was unclear in my original note. My
> real doubt is whether it is worth it to use mmu tricks to speed up fifo
> operation. We can make fifos use page flipping so that there is no
No, not to speed FIFO. I wanted mmap method in FIFO to be completely separate
from their main functionality. Just not to create special devices only for
mmap. I was thinking about adding char *mbuf, long mbuff_size, int mbuff_cnt
to FIFO structure and copying mmap from mbuff "driver". But there remains
question of bookkeeping/deallocation.
I am leaving for about 10 days, I will continue working on this after.
I have just found the patch Ian Ridley has sent us a year ago. It is in
http://crds.chemie.unibas.ch/PCI-MIO-E/rt-shmem-Ian.Ridley.Apr98.tgz
Regards,
--
Tomek
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
