> I installed 9atom downloaded from Erik's site on 5/12 on a > Gigabyte GA-D525TUD > (http://www.gigabyte.com/products/product-page.aspx?pid=3549#ov). > > When I have *nomp=1 in plan9.ini, I am able to log in and > things work OK (except the mouse wheel scroll direction is > opposite to what I would expect), When I remove *nomp=1 from > plan9.ini, 9atom reports four processors, and every keystroke > produces 6 or 8 or more tokens of the character I am trying to > type. If I'm lucky and I tap a key lightly and quickly enough, I > can get a single character, but it's not easy. > > I tried the only other usb keyboard I have and had the same > problem. > > Any suggestions?
9atom's usb is identical to sources, with the exception that i've grabbed cinap's bios handoff code for ehci and enabled all the controllers. this was necessary to get usb to work on some new hardware. i'd recommend doing pull (as glenda) and rebuilding the kernel, first. i do know there is a bug that causes spurious kb repeating. but typically it's not quite as severe, and depends a lot on the keyboard in question. it's interesting that you corellate this to mp interrupts. does the same hapen without *nomp and with *ncpu=1? i haven't done much research on this, so although the usb/kb repeat mechanism at first blush looks like a good place to start, it's not clear to me that this is the issue. - erik
