* SGA155D DMA Burst length question - Memory to S5933
@ 2003-09-25 14:57 Horvath Gyorgy
0 siblings, 0 replies; only message in thread
From: Horvath Gyorgy @ 2003-09-25 14:57 UTC (permalink / raw)
To: Linux Kernel Mailing List
Hi all,
I am on the way developing the driver for SGA155D.
The core driver is OK. I simply copy the firmware.bin file
to /dev/sga155d0, and after the firmware identification -
a proper satellite driver is loaded - actually the POS
(Packet Over SONET).
I can then ifconfig my pos0 - and here weeee goooo.
Now I can DMA well from the card to the PC memory -
even for long-bursts. The data is OK.
I see on the oscilloscope a short transfer at the beginning
(8 DWORDS or so) - a little gap, and the rest go smoothly
in one long piece.
I tried then to DMA from the PC memory to the card.
What I see is: 9 DWORDS, some 9+more gap, 9 DWORDS, some 9 more gap,
etc...
I'd like to see a smooooooooooth transfer - not that 50%.
I have S5933 on-board, and I find some issues elsewhere
about advanced PCI commands when spanning across cache-lines,
but I am not sure that is the issue... (MR/MRL/MRM)
Can S5933 do them?
Any tips/suggestions?
Also, what is the suggested burst size under Linux for
such a NET device like mine? (cca. 2x 140 Mbps, MTU is 4470)
Thanks for in advance:
Gyuri
Gyorgy Horvath, Technical University of Budapest
-------------- Dept. of Telecom. and Telematics
Tel.: +36-1-463-1865, Fax.: +36-1-463-1865
Mail: horvaath@bme-tel.ttt.bme.hu
FTP: ttt-pub.ttt.bme.hu ./income
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2003-09-25 15:01 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-09-25 14:57 SGA155D DMA Burst length question - Memory to S5933 Horvath Gyorgy
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®