© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
Christopher,
Please, don't mark this as Invalid... in subsequent 3.8.0 revisions the behavior has returned. I don't think it was ever truly fixed, I think there was perhaps some performance regression that prevented this bug from occuring. I've notice that this bug conistently occurs at high read IO load, but sometimes wont occur if you're more delicate in your use.
I've also been able to test with mulitple USB3 host controllers now (before I thought perhaps it was a quirk with just the NEC host chip) and I get the same behavior. The new controller is:
Intel Corporation 7 Series/C210 Series Chipset Family USB xHCI Host Controller (rev 04) (prog-if 30 [XHCI])
And I've tested with multiple brands of USB3 card readers and have gotten the same behavior. In all cases, it seems like the power management is powering down the link (if that's even the right terimnology) when it shouldn't.
Hopefully in the next few weeks I can really dig into this in the kernel, toss in extra debugging, etc. Interestingly, removable USB3 HDD seem to work fine. On the surface the only difference I can think of is that these card readers present mulitple USB mass storage devices (most have at least a CF and an SD slot).
Anyway, thanks for your help!