Hi guys,
I am back!
Cyrille - i'll test your new patch later this evening.
It looks little more verbose to me, but i think there
is room for optimization. Basically i wish to URBify
all pegasus_*_registers routines. The main problem is
that i can't touch the registers from bottom-havves.
I hope if URBifying work ok this will do the job.
best,
Petkan
"Cyrille Chepelov (home)" wrote:
>
> Hi folks,
>
> This patch is a cleaner version of my hacked pegasus driver. It works
> almost great on my (broken ?) OHCI machine, doesn't crash, and spits out
> solid 200KByte/second on my system (That poor P133 CPU seems to be the
> bootleneck, but it seems OK, since D-Link advertised their toy for P166
> minimum). At least now, NFS, FTP, X, SSH, APT *and* LPRNG are usable
> (LPRNG is really really sensible to the smallest timeouts and lost
> packets, which made the stock driver or my early "failover-with-thread"
> approach unusable).
>
> I'd really be happy to hear about people testing this one, on both host
> interfaces.
>
> Issues still to check :
> * does it work on UHCI as great as unpatched 0.3.13 did ?
> * on OHCI, requires 2.4.0-test1-ac7 *and* my little OHCI fix I've
> sent here yesterday. Also, there are some conditions now (UE) where the
> OHCI controller will disable itself ; basically what's needed now is to
> recover from what the controller says is an unrecoverable error. OHCI
> wizards, any idea ?
> * what happens when there are still timeouts (probably on a
> CPU-challenged machine like mine) and the host controller honours
> urb->timeout (which should keep the NETDEV watchdog silent) ? I can't test
> this code path on current OHCI, and my UHCI machine is too critical to
> allow too much of tinkering with it.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
