Charles Fultz wrote:
> > BTW, I just discovered that BREAK does not drop to PROM in SILO.
> > Maybe that should be suggested to the SILO developers. If they
> > are doing I/O via PROM code (I suspect likely) I'm surprised it
> > isn't automatically there. It should also accept a special boot
> > name "prom" to go to PROM as well. Then, maybe this can be
> > configured (I haven't looked).
> >
> > I think our discussion regarding serial break is just how it
> > should behave. I think the consensus is that there is not
> > 100% consensus on what it should do, but rather, it should be
> > configurable.
>
> These /proc/sys/sysrq and /proc/sys/stop-a may be the answers to our
> problems. If they are configurable as Phil suggests they are, then
> we're set and our squabling is done.
First, they're really:
/proc/sys/kernel/sysrq
/proc/sys/kernel/stop-a
The current values for me are:
/proc/sys/kernel/sysrq = 1
/proc/sys/kernel/stop-a = 1
I believe /proc/sys/kernel/sysrq turns on the SysRQ feature, which works
on the serial console in Intel, and the video console in both.
I believe /proc/sys/kernel/stop-a is only in the Sparc kernel, and it
turns on Stop+A at the keyboard.
How should I set these so that Stop+A works at the keyboard (which I
think means I need to leave /proc/sys/kernel/stop-a = 1) works and
BREAK does a SysRQ (I'd need to have /proc/sys/kernel/sysrq = 1)
How should I set these so that Stop+A works at the keyboard works
and BREAK goes to PROM from serial? The problem here is that this,
too, needs both set to = 1 if you want SysRQ to work at the keyboard.
So I don't think it is solving the problem yet. The variables are:
SysRQ works at keyboard or not
Stop+A works at keyboard or not
BREAK does PROM or SysRQ
There are at least 8 different combinations and 2 booleans to hold them?
-- ----------------------------------------------------------------- | Phil Howard - KA9WGN | Dallas | http://linuxhomepage.com/ | | phil-nospam@ipal.net | Texas, USA | http://phil.ipal.org/ | -----------------------------------------------------------------
This archive was generated by hypermail 2b30 : Thu Sep 19 2002 - 11:00:02 PDT