On Thu, Feb 19, 2004 at 01:23:03PM -0800, Matt Thomas wrote:
> At 02:57 AM 2/19/2004, Niels S.Eliasen wrote:
> >Hi flks
> >I'm getting a "pmap_pinit: out of segments" when trying to make a new 
> >build....
> >
> >What can I do to make the build continue ??
> That implies your system has more than 32K vmspaces which is doubtful.
> I wonder if something is corrupting the bitmap.

I'm seeing this too - backtraces from a couple of recent panics follow
(unfortunately, not from a DEBUG/DIAGNOSTIC kernel):

panic: pmap_pinit: out of segments
Begin traceback...
0xd6263d70: at pmap_pinit+12c
0xd6263d90: at pmap_create+34
0xd6263db0: at uvmspace_init+88
0xd6263dd0: at uvmspace_alloc+40
0xd6263e00: at uvmspace_fork+50
0xd6263e30: at uvm_proc_fork+20
0xd6263e50: at fork1+384
0xd6263eb0: at sys_fork+30
0xd6263ed0: at syscall_plain+e0
0xd6263f40: user SC trap #2 by 0x419173ec: srr1=0xd032
            r1=0xffffc8e0 cr=0x20024022 xer=0 ctr=0x419125cc
End traceback...
syncing disks... 9 8 7 7 7 6 5 4 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 giving up
wd0: flush cache command didn't complete
rebooting

panic: pmap_pinit: out of segments
Begin traceback...
0xd72f3c60: at pmap_pinit+12c
0xd72f3c80: at pmap_create+34
0xd72f3ca0: at uvmspace_init+88
0xd72f3cc0: at uvmspace_alloc+40
0xd72f3cf0: at uvmspace_exec+48
0xd72f3d20: at sys_execve+618
0xd72f3ed0: at syscall_plain+e0
0xd72f3f40: user SC trap #59 by 0x41852d18: srr1=0xd032
            r1=0xffffc5d0 cr=0x80000044 xer=0 ctr=0x41894fd0
End traceback...
syncing disks... 6 5 4 3 2 1 done

There is also a thread started by Aymeric Vincent on Jan 1st called
'macppc unstable in -current' (or something) with details of some other
problems with recent macppc kernels.

As a further datapoint (I'm not 100% sure this is related to the panics
I've been having), one of Perl's configure tests segfaults - it's always
the same one, and it's one of the IPC/SHM ones.

Ian.