Hi All, Thanks a lot for great response. All the solutions seems to be working in the testing environment. But a commercial driver can do well in the testing environment but may fail in the customer site due to some other component in costumer's system. In such scenario disabling the driver loading at grub seems to be the easiest way out of the problem.
Thanks & best regards, ajit On Thu, Jan 8, 2009 at 1:21 AM, Andrew Gallatin <[email protected]>wrote: > Garrett D'Amore wrote: > >> That would work. >> >> The only drawback I see is that when the crash occurs, you have to catch >> the system during boot and select the other BE. Since I often test on >> systems remotely (where the console of the system is on the other side of my >> lab), I prefer the "automatic safe reboot" of my approach. (It keeps me >> from running over and trying to find the keyboard for the system I'm >> debugging. :-) >> > > When I'm testing drivers, I don't copy them to /kernel/amd64/drv, > I make a softlink into my (nfs mounted) work area instead. Then > if the crashbox lives up (down?) to its name and crashes, I can easily > rm the driver by the time it comes back up. > > Drew > >
_______________________________________________ driver-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/driver-discuss
