On Thu, Aug 17, 2006 at 03:43:55PM +0400, Michael Tokarev wrote: > Matthew Wilcox wrote: > > On Thu, Aug 17, 2006 at 02:55:58PM +0400, Michael Tokarev wrote: > [sporadic disk disappearing, no logging] > > > > I'd recommend turning on scsi logging; it might give you a clue about > > which bit of scanning is failing to work properly. > > > > Try booting with scsi_mod.scsi_logging_level = 448 (I think I have that > > number right; 7 shifted left by 6) and then you can compare failing and > > non-failing runs and see if there's any difference. > > It should be the same as > echo $((7<<6)) > /sys/module/scsi_mod/parameters/scsi_logging_level > (which indeed is 448) at runtime, right? (And yes, CONFIG_SCSI_LOGGING > is set to y). That's right. > Heh oh those magic numbers!.. ;) Yeah, but the alternative is an in-kernel named symbol parser ... which we have in some drivers, but boy is it ugly. > Ok, I've turned on the logging on a bunch of machines (using the sysfs > method), let's see what will happen next. Thank you! > > By the way, should kernel pefrorm at least *some* "minimal" logging of > such a serious events by default? Well ok, ok, it's not known yet what > the event really is, so I'm shutting up now, at least for a while.. ;) That's the problem -- if it turns out the event is a reasonable thing to happen for some devices, we'll annoy everyone with those devices. It's hard to please everybody ;-) - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
