-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Werner Almesberger wrote: > Andy Green wrote: >> In addition there are hard lockups from Glamo nWAIT, > > BTW, I think you once electrically forced nWAIT high to free the > system from the Glamo's deadly grasp, didn't you ? Could that be > a way for better debugging ?
It lets me reach the crash caused by the AUX key debug code, so I can see where we hung. But the Glamo is dead at that point, it's fundamentally a sign it was broken either by the suspend actions on it or the resume attempt on it. So that's really all that can be squeezed out of forcing it over nWAIT blockage. > E.g., if the same wire went to a currently unused interrupt, e.g., > conveniently accessible EINT3, then it could also activate some > debug dump code at the time nWAIT is forced. Procedure AUX key while stalled and then force nWAIT deasserted runs the dmesg dump / panic code without it (or more accurately given what you wrote, AUX interrupt is used). - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkjnMdAACgkQOjLpvpq7dMqYZQCeLt7NmJp8SLDgDLTn0Q5aclHh zgQAn1iuJaWTlkRgVJ9azvNTr1EY9sST =tGUD -----END PGP SIGNATURE-----
