On Fri, Oct 29, 2004, Oliver Neukum <[EMAIL PROTECTED]> wrote:
> Am Freitag, 29. Oktober 2004 20:44 schrieb David Brownell:
> > On Friday 29 October 2004 11:13, Johannes Erdfelt wrote:
> > > >          If usbfs instead grabbed a bunch of non-contiguous
> > > > pages, copied the data into them from user-space, and then sent it (using
> > > > scatter-gather io), then there is no longer any memory pressure problem.
> > > > What's more, there would also be no reason not to increase the maximum
> > > > data size from 16k to something much bigger, say 256k.  Allowing bigger
> > > > buffers should increase performance.
> > > 
> > > Duh, that would be the obvious way to increase performance and
> > > reliability.
> > > 
> > > Not to mention it should be relatively easy to implement.
> > 
> > Yep, and you'd be able to clean up properly after an error in
> > the middle of your N urbs ... the scatterlist code handles
> > such things.
> > 
> > Another tweak would be to remove the limit on transfer size.
> > Use whatever algorithm you want to transfer the first chunk;
> > for high speed, a scatterlist is *really* desirable.  Then
> > repeat for successive chunks ... until error, or all done.
> > (And if you can't get enough buffers to use big chunks, then
> > you could still make progress using smaller chunks...)
> 
> I am afraid there is a problem. In fact, the current situation
> cannot be allowed. There has to be a limit on the number
> of URBs and buffers, or we allow any user to eat up all memory.

I'm ok with a limit and I agree that one is necessary. In reality, you
reach a point of diminishing returns anyway.

We just need to figure out roughly where that point is and set the limit
accordingly.

JE



-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to