On Tuesday 01 May 2007, Gene Heskett wrote: >On Monday 30 April 2007, Greg KH wrote: >>On Sat, Apr 28, 2007 at 01:32:51PM -0400, Gene Heskett wrote: >>> Greetings; >>> >>> This is driving me batty (its just a short drive, folks) [horribly word wrapped old log file] >>Are you yanking and plugging this device back in? If not, you have some >>flaky hardware there :( > >In the above instance, I had plugged it in to make sure it was recognized, >then assembled the rest of the cable and plugged it into a different hub, a >2.0 hub. When I was using the pl2303's for that, they won't work unless >plugged into a usb1.1 hub, and I also have a 2.0 hub, which is where the > FTDI adapter is plugged in now, getting rid of one cable plugged into the > front panel breakout which is a 1.1 only hub. 3 years Old. > >>> During a boot to 2.6.21-sd046, it was one way, and now with 2.6.21-CFS-v7 >>> is reversed. I have NDI, nor can I connect mentally, the scheduler diffs >>> to this, it just doesn't grok. >> >>Does 2.6.21 work ok? > >Bare 2.6.21? NDI Greg, that mainline scheduler is such a pain compared to > the 2 I'm playing with I haven't actually tried it. The point being, the > only thing I change in the two 2.6.21 kernel builds is the scheduler patch, > the only patch applied, from Ingo's v7 to Con's SD-0.46, and the two > devices are interchanged after the reboot, so heyu is trying to talk to the > ups, and the upsd is trying to talk to the X10's cm11a. And now with > everything running on FTDI adapters since pl2303's are now non-functional, > I can't even look at the dmesg output and see which is which, I'll just > have to kill both and blindly reconfigure heyu till I can read the clock in > the cm11a, and then give the other one to upsd. Everytime I reboot? Gee, > come on folks, its getting old, and makes a reboot a 30+ minute operation. > :( > >The one thing I haven't done yet is to try putting upsd back on a real > serial port, which would make their daemon happy, but I have a legacy > machine in the basement that needs an rs232 port open here, which was the > reason for buying all these adapters in the first place.
And here I am forced to correct myself, this daemon, which did work very well back around 2.6.18 or so when running through a real serial port, now does exactly the same as I described here previously when moved to a real hardware serial port. I just did it to confirm my claim, which was obviously bogus. So my apologies on that point. Now the question is, since this code hasn't changed since 2002, what new feature of the serial protocol is now eating this things lunch and taking 40% of the cpu with it for company? >And that brings up the other point I was also fussing about, that the >adapters, at one end or the other of this software chain, are not actually >getting me a /dev/ttyUSB# with behavior that's bit for bit identical to what >I'd be getting through a 16550 compatible chipset and /dev/ttyS#. Something >isn't telling upsd that it has all the data (12 bytes I believe) there is to >be fetched in this read cycle and that it can send it on up the pipe and go >back to sleep. As is, its sitting there eating 40% of the cpu by >busy-waiting when its connected to a /dev/ttyUSB#, but behaves as expected >when talking to a real /dev/ttyS#. This is a bug in my book. > >So is the recent failure of the pl2303 cables. They haven't gone bad, they >work just fine on my lappy running FC5 when connecting roadnav to my gps, >again a not very demanding application, a few bytes a second, rx only to the >computer. But they don't seem to work for anything with 2.6.21 final, > having stopped working someplace in the middle of the 2.6.21-rc period. > >>thanks, >> >>greg k-h -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) There is a great discovery still to be made in Literature: that of paying literary men by the quantity they do NOT write. ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
