On Thu, Jul 09, 2009 at 11:04:05AM -0400, Alan Stern wrote: > Your "USB: allow malformed LANGID descriptors" patch (commit > b7af0bb26899bb47ae16fb41d2296111b0784a56) is causing a regression: > > http://bugzilla.kernel.org/show_bug.cgi?id=13700 > > The usbmon log attached to comment #11 shows how the 5-2 device > doesn't work with this patch. It rejects attempts to read string 0 > (the language ID list) with a STALL, and it crashes when the kernel > attempts to read a string using language 0. > > In fact, 0x0409 (US English) would probably be a better fallback than > 0, since it is so common. But I don't know whether this particular > device would be able to handle that either. Jouni, you should try the > patch below (on vanilla 2.6.30) to see what happens. 0x0409 would be equally ok for me - the intention about my change was purely not to fail hard (i.e., not to report any string at all any more) when the LANGID could not be written. > Maybe changing the default will work, maybe there should be a > NO_STRINGS quirk, or maybe the commit should be reverted. I don't know > which would be best. Reverting it would cause a regression for me, so I'd like to try a different fallback default first and see if it helps for the devices reported. Thanks, Daniel -- 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
