Hi
Following up on i386 PAE issues: the machine works fine with PAE and
SMP disabled (boot -1) when it has 16 GB of RAM. Attempting to add
more memory and reach 32 GB causes the unpleasant UVM crash shown
below. Does it rings a bell for someone:
panic: lock error
cpu0: Begin traceback...
On Fri, Dec 13, 2013 at 01:16:08PM +, Emmanuel Dreyfus wrote:
panic: lock error
Can you try this with a LOCKDEBUG kernel?
Martin
Martin Husemann mar...@duskware.de wrote:
On Fri, Dec 13, 2013 at 01:16:08PM +, Emmanuel Dreyfus wrote:
panic: lock error
Can you try this with a LOCKDEBUG kernel?
Martin
Will not help. This is a trap in x86 pmap, the locking error is just a
side effect, not an issue in UVM. Quite
Hi everybody.
I ve just compiled development branch with Dtrace support but I d like to
know if it supports userland probes ? I tried to produce an object file
with dtrace -G -s ... but tried to find /usr/lib/dtrace/drti.o (which
really does not exist...) so it fails ... is it expected ? Or maybe
To test the impact on other concurrent SCSI commands
It looks like the patch cures my original, real, problem, too. It probably
saved my day (or those of the 250 people with their home dirs on my file
server) or my entire weekend. Without it, I would have lost one RAID
component at 02:05 and a
Mindaugas Rasiukevicius rm...@netbsd.org wrote:
If NetBSD ships PAE,
we ought to fix its bugs, but generally - i386 PAE is inferior to amd64
in practically all aspects, so just do not use it if you can.
The idea of i386 PAE if that if an i386 lacks some RAM, you can just
drop a PAE kernel
