Eric Biederman writes:
>My big concern is with getting the shutdown path setup in a manner
>that works, and gets testing.
Rebooting without traversing the device tree seems to have
essentially worked fine for 2.4.x.
>When booting linux from linux with
>sys_kexec a lot of my problems come back to some device driver not
>getting shutdown.
kmonte and sys_kexec skip the BIOS reset code and therefore
may need to do more elaborate shutdown, but please do not saddle the
normal reboot case with the reliability risk of calling code in each
driver when a user might be rebooting a remote machine precisely
because of a a confused device driver or the potential slow down
(especially since you want an interface where the function that gets
called before reboot may need to do blocking IO). For
kmonte/sys_kexec, this high cost might be necessary, but for the
normal reboot the cost is not worth the benefit.
By way, given the ability to register reboot notifiers in the
device tree, I would be happy to see one registered at the top of the
PCI bus tree (so it would be called last) that would shut off the PCI
bus before reboot, along the lines of what Richard B. Johnson posted.
That would not involve walking a lot of data structures in many
different device drivers and it would be just a few instrutions.
Adam J. Richter __ ______________ 575 Oroville Road
adam@yggdrasil.com \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Oct 23 2002 - 22:00:55 EST