Andy Green wrote:
But looking closer at your error reports, it is reminding me of something I read a few weeks ago about max packet sizes, or max buffer sizes for different USB <-> ATA chipsets? Some can handle 128KB, some 64KB, something like that. I think it was someone's patch reducing the limit in kernel sources on a ml I was looking at.
Well, it seems like usb-storage still has a way to go even in Arjan's kernels. While things are better for some people, I can reproduce the bug with 100% certainty. Let's hope things improve soon.
Thanks for the strace tip. It was hanging trying to connect to an unresponsive network printer through cups. I deleted this printer from the cups list and that did it, but I'd argue that OO should recover more gracefully from this kind of problem.
Glad I was of some help. Yeah it should be more graceful.
As I mentioned, kprinter was being a victim of the same thing. I'm also seeing now various problem reports about OO, so this is probably going to rise in priority quickly. I don't have time now to trace the problem too deep, but I suspect it's somewhere in cups, since a config cleanup and restart of cups 'unfroze' both kprinter and OOffice.
Display doesn't shut off ------------------------
ACPI prob? Maybe try to fiddle with ACPI=blah on kernel commandline.
I'll try a few options and see what happens.
I get the same behaviour as you on some laptops and not others here. Makes me think it can be (video) BIOS or video driver related as well as ACPI.
I tried a bunch of combinations of apmd/acpi on/off and nothing helped. For now I give up.
In summary:
usb-storage: BROKEN, reproducible. OpenOffice, kprinter hangs: most likely caused by cups problems. Display power management: broken, reported by others also.
I'll definitely wait for a few weeks with this test box before even considering putting FC2 on the rest of our machines.
Again, many thanks for all the help.
Regards,
f.
