On Sunday 01 February 2009, Qiuping Chen wrote: > It cannot pass some subtests in test #10 of test.sh. That's usually a case of some *particular* fault path not working right ... which you can isolate by taking certain tests out of the loop. The way that test works is by taking N of the M subtests and running them in a loop for T times ... so if you find that it works with N-1 subtests in the loop, but fails with N or more, then the Nth test is causing a problem. That tends to be un-obvious because the faults that get uncovered there are mostly races, which won't reproduce every time. I've certainly had cases where for example recovery after the fault triggered by subtest #10 would work fine ... except that if subtest #11 ran next, it would fail intermittently. As in, sometimes it took a few hours to trigger! > It can pass ping stress test. I think you mentioned that already. Is that "ping -f" in both directions? Which "ping stress" test? > OK to change "Intel USB Client Controller" to "Intel Poulsbo USB > Client Controller". The latter is better. Done. > Should I re-submit the patch with this change? Why don't you hold off on that; I have a few other cleanups in my copy. If there's anything else that's non-cosmetic, I'll do something at that time. - Dave -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
