On Fri, 2003-11-14 at 15:20, Alan Stern wrote:
> On Fri, 14 Nov 2003, Alex Bennee wrote:
> > Is there some way to pace/slow down the commands at the usb level or
> > should I tweak my transport driver to do that?
>
> Not at the level of usb-storage. Your HCD would have to handle it.
I added this to my submit_urb function:
printk(KERN_ALERT "hci_submit_urb: %ld\n",jiffies);
mdelay(10); /* hack to slow down requests */
Which seemed to do the trick for memory sticks. I'm trying with the
cheap-ass CD-RW at the moment but that seems to have other issues.
> > Unfortunately not an option. I'm getting this working for a production
> > embedded system for which 2.6 is just too new to be an option yet. Once
> > we start shipping version 1 I fully intend to have a go at trying out
> > 2.6.0, it just doesn't currently line up with our release schedules :-)
>
> You could try plugging the device into a standard PC using one of the
> regular HC drivers. If that works okay, it would definitely point to a
> problem with your modified driver.
Well it works under windows but unfortunately my mandrake box has never
recognised whatever USB chipset I have on this thing. However I think
the experiment shows whats causing the problems.
The ISP116x chip generates an interrupt for every USB SOF. I assume if
the rest of the system is fast enough a whole series of commands will
just get written back 2 back. It would be nice to know is anyone else
using the various isp116x driver out there have come across these sorts
of problems. My hacks have been pretty minor (changing from port to
mem-mapped io) but its not exactly the prettiest code I've worked on.
At least I have a workaround for the time being :-)
Is there anyone who claims ownership of the kernel hc_simple code that
the isp116x driver uses?
I'm thinking its about time to clean up the isp116x driver and submit it
to usb-devel/lkml. But that does require resolving some merge issues
with the hc_sl1811 driver thats already there that also uses hc_simple.
--
Alex, homepage: http://www.bennee.com/~alex/
The worst is not so long as we can say "This is the worst."
-- King Lear
-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
