I have two machines, both SS-10, quad Ross CPU's, both doing the same
spin_lock_irqsave thing. I had a 4/670MP doing the same thing. None of these
were unstable under 2.0.35 (except that on the 4/670MP it wouldn't work with 4
CPU's, for some reason Linux didn't want to work with a second module
installed). 2.2.x is MUCH faster than 2.0.35 was but it is not stable, it has
gotten better with time, 2.2.0 often didn't even make it into multi-user mode,
now 2.2.10 usually runs a couple of days between crashes.
I also have another machine, SS-10 with the quad Ross's, still running
SunOS 4.1.4, no instability problems.
As I mentioned, the machines are not running terribly hot. Only one has
an internal drive and then only one drive. The other has no internal drives.
The cases are warm, if you open up the CPU modules are hot but not so hot that
you can't hold your fingers on them.
These aren't random OOPs, they get into this spin_lock_irqsave loop.
Also, another strangeness. One of the SS-10's did this and I hit L1-A,
which usually when it does this is ineffective, usually I have to power cycle,
but this time I got the PROM Monitor AND the messages kept printing, as if one
CPU stopped and went into running PROM code but another continued with broken
kernel code.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/