Re: mlx on sparc?

[Jan-Hinrich Fessel <oskar%fessel.org@localhost>]

Am 02.06.2008 um 22:30 schrieb Andrew Doran:

> On Mon, Jun 02, 2008 at 10:10:55PM +0200, Jan-Hinrich Fessel wrote:
>
> > Now that I was donated a set of mlx boards with newer firmware ;-) I
> > get the following:
> >
> > mlx0 at pci0 dev 2 function 0: Mylex RAID (v3 interface)
> > mlx0: interrupting at ivec 10
> > mlx0: initializing (may take some time)...
> > mlx0: physical drive 2:0 not responding
> > mlx0: physical drive 2:1 not responding
> > mlx0: physical drive 2:2 not responding
> > mlx0: physical drive 2:3 not responding
> > mlx0: logical drive installation aborted
> > mlx0: initialization complete
> > mlx0: DAC960P/PD, 3 channels, firmware 3.51-0-09, 4MB RAM
>
> Seems ok, it needs a configuration (and some disks by the look of it).

;-)
> > mlx1 at pci0 dev 4 function 1: Mylex RAID (v4 interface)
> > mlx1: interrupting at ivec 18
> > mlx1: DAC960PG, 2 channels, firmware 4.06-0-85, 16MB RAM
> > ppb1 at pci0 dev 5 function 0: Digital Equipment DC21154 PCI-PCI
> > Bridge (rev. 0x02)
> > pci2 at ppb1 bus 2
> > Distributed Processing Technology Memory Controller (miscellaneous
> > memory, revision 0x02) at pci2 dev 3 function 0 not configured
> >
> > basically the whole setup looks fishy to me. mlx0 gets initialized,
> > mlx1 not.
>
> mlx1 has probably completed its initialization at power own, so the  
> kernel
> does not need to wait for it.

Okay?
> > iop0 at pci2 dev 4 function 0: I2O adapteriop0: reset rejected,  
> > status
> > 0x0
> > iop0: not responding (reset)
> >
> > And the DPT PM3755U2B controller (which is basically the
> > single channel version of the SUN SRC/P controller) is not accessible
> > (I have created the relevant node entries...)
>
> The I2O code was developed with exactly this board. What dance are  
> the LEDS
> on the board doing before and after the kernel tries to reset the  
> board?

I have to find a way to look at the LEDs, probably shorten the cover  
swithc - the machine will not operate without the cover (probably due  
to thermal considerations).

> > # mlxctl -f /dev/mlx0 -v cstatus
> > DAC960P/PD, 3 channels, firmware 3.51-0-09, 4MB RAM
> > mlxctl: cmd 0x1c failed: Operation not permitted
> > # mlxctl -f /dev/mlx1 -v cstatus
> > DAC960PG, 2 channels, firmware 4.06-0-85, 16MB RAM
> > mlxctl: cmd 0x1c failed: Operation not permitted
> > # mlxctl -f /dev/mlx0 -v config
> > mlxctl: cmd 0x4e failed: Operation not permitted
> > # mlxctl -f /dev/mlx1 -v config
> > mlxctl: cmd 0x4e failed: Operation not permitted
> > # iopctl showstatus
> > iopctl: /dev/iop0: Device not configured
>
> Looks like you need to drop the securelevel to -1, since the kernel  
> won't
> let you talk do disk controllers at securelevel 1 or above.

# mlxctl -v cstatus
DAlignment error: pid=14.1 comm=mlxctl dsfsr=00000000:00800001  
dsfar=ffffffff:ffffba26 isfsr=00000000:00000000 pc=101908
AC960P/PD, 3 channels, fpid 14 (mlxctl), uid 0: exited on signal 10  
(core dumped)
irmware 3.51-0-09, 4MB RAM
[1]   Bus error (core dumped) mlxctl -v cstatus
# mlxctl -f /dev/mlx1 -v cstatus
DAlignment error: pid=15.1 comm=mlxctl dsfsr=00000000:00800001  
dsfar=ffffffff:ffffba06 isfsr=00000000:00000000 pc=101908
AC960PG, 2 channels, firmware 4.06-0-pid 15 (mlxctl), uid 0: exited on  
signal 10 (core dumped)
85, 16MB RAM
[1]   Bus error (core dumped) mlxctl -f /dev/mlx1 -v cstatus
# iopctl showstatus
iopctl: /dev/iop0: Device not configured

gdb isn't that helpful:
# gdb obj/mlxctl mlxctl.core
GNU gdb 6.5
Copyright (C) 2006 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and  
you are
welcome to change it and/or distribute copies of it under certain  
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for  
details.
This GDB was configured as "sparc64--netbsd"...
Reading symbols from /usr/lib/libkvm.so.5...done.
Loaded symbols for /usr/lib/libkvm.so.5
Reading symbols from /usr/lib/libc.so.12...done.
Loaded symbols for /usr/lib/libc.so.12
Reading symbols from /usr/libexec/ld.elf_so...done.
Loaded symbols for /usr/libexec/ld.elf_so
Core was generated by `mlxctl'.
Program terminated with signal 10, Bus error.
#0  0x0000000000101908 in cmd_cstatus ()
(gdb) bt
#0  0x0000000000101908 in cmd_cstatus ()
#1  0x0000000000102dfc in main (argc=3, argv=0xffffffffffffbcb8)
    at /usr/src/usr.sbin/mlxctl/main.c:145
(gdb)


Cheers
        Oskar

Attachment: PGP.sig
Description: This is a digitally signed message part