Interesting ... it did have an effect. Not entirely unexpected, though I'm still curious which debug-related delays matter.
I've gone back and played with it, and the sure way to make it go bonk is to make emacs/gnus look for new mail, at which point I cause massive I/O to my hard drive (and on a laptop thats sloow).
So you're saying that a high interrupt rate on _another_ device is a key factor in your "high load" problem? Is that hard drive maybe using PIO, or one of the less robust IDE drivers?
Please send along a copy of your "lspci -v" output and your /proc/interrupts. And try disabling ACPI, if you're using it.
play all night with bouncing files around and see if it keels over at some point. The interrupt rate is roughly 9k/sec for the USB controller.
Can you forward your "bounce files around" test script?
Clearly the tests I run aren't producing misbehaviors; maybe yours would cause some on other hardware too.
That interrupt rate is absurdly high, considering ...
Something would seem to be getting measured wrong, or else there's something else causing a lot of your load problems.
And in fact, that "around 1000" IRQ guesstimate seems to have been on the mark on my hardware ... I did some testing during a loop (copying a kernel source tree from one part of a high speed disc to another part of the same disk, so not many "big" writes), on the same host, and averaged over 100 seconds I saw:
- Philips isp1562 (EHCI 0.95) 1250 irq/sec - VIA vt6212 (EHCI 1.0) 2250 irq/sec (*) - VIA vt8235 (EHCI 1.0) 1200 irq/sec
The VT6212 number unfortunately includes shared eth0 IRQs (LAN wasn't idle), so it's nothing more than a ceiling. But it's clearly not in the 9000+ IRQ/sec range. (The VT6212 is an update to the VT6202, which is what you seem to be using.)
Hmm, I guess if the files you're copying are all just a page or two in size, that IRQ rate might be more reasonable. If that's your situation, I'd consider using a bigger irq threshold (it's a module parameter) ... at least try the experiment, since that does seem to be a factor (given the IDE issue you mentioned).
- Dave
------------------------------------------------------- This SF.net email is sponsored by: Etnus, makers of TotalView, The best thread debugger on the planet. Designed with thread debugging features you've never dreamed of, try TotalView 6 free at www.etnus.com. _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
