[parisc-linux] C180 not booting Sarge kernel

[parisc-linux] C180 not booting Sarge kernel

[Chris Jantzen]

FYI, my sarge-based C180 has not been able to boot 2.4.19-32_18 or _22.1;
I have to interact with IPL and fall back to 2.4.18_62.2. It gets as
far as the "...if this is the last message you see..." and I never hear
from it again, even on the network. And I think after a very very long
time, it falls back to firmware. At least, I see DHCP requests coming
out of it after a while. (I only just hooked up serial console to
interact with IPL after getting no response. No I don't boot it up too
often, it's a "side project".)

Also, "uname -a" segfaults.

I'll try unplugging the GSC and EISA cards plugged into it tomorrow.
I'll also try letting it just sit there under supervision and see what
really happens.

-- 
chris jantzen kb7rnl =->         __O
Insert witty comment here.     _`\<,_
http://www.maybe.net/         (*)/ (*)

Re: [parisc-linux] C180 not booting Sarge kernel

[Carlos O'Donell]

> FYI, my sarge-based C180 has not been able to boot 2.4.19-32_18 or _22.1;
> I have to interact with IPL and fall back to 2.4.18_62.2. It gets as
> far as the "...if this is the last message you see..." and I never hear
> from it again, even on the network. And I think after a very very long
> time, it falls back to firmware. At least, I see DHCP requests coming
> out of it after a while. (I only just hooked up serial console to
> interact with IPL after getting no response. No I don't boot it up too
> often, it's a "side project".)

HP hardware is occasionally quirky :}

> Also, "uname -a" segfaults.

That bug got squashed in unstable. You'll see the fix soon.
 
> I'll try unplugging the GSC and EISA cards plugged into it tomorrow.
> I'll also try letting it just sit there under supervision and see what
> really happens.

Good idea :)
If it doesn't work, get back to the list.

c.

Re: [parisc-linux] C180 not booting Sarge kernel

[Tom]

On Thu, 21 Nov 2002, Chris Jantzen wrote:
> FYI, my sarge-based C180 has not been able to boot 2.4.19-32_18 or _22.1;
> I have to interact with IPL and fall back to 2.4.18_62.2. It gets as
> far as the "...if this is the last message you see..." and I never hear
> from it again, even on the network. And I think after a very very long
> time, it falls back to firmware. At least, I see DHCP requests coming
> out of it after a while. (I only just hooked up serial console to
> interact with IPL after getting no response. No I don't boot it up too
> often, it's a "side project".)

This is the "bus-walk" issue I asked about a few weeks back. Apparently
the system is looking through every possible connection, looking for
peripherals, and doesn't continue until it finishes. My C180 sits for ~20
minutes at that point, then continues and runs normally. Boots normally on
2.4.18-pa59, happens on all the 2.4.19 kernels (pa20, pa24 for sure, as I
tried both of those). Not sure about 2.4.20-rc1-pa6, but it did boot (I
wasn't paying attention - powered up before dinner, after dinner
everything was waiting for me <g>)

"They" swear this has a fix in 2.5, but I just let it take it's time and
all seems happy...

Re: [parisc-linux] C180 not booting Sarge kernel

[Chris Jantzen]

On Thu, Nov 21, 2002 at 08:57:54AM -0500, Carlos O'Donell wrote:
> > FYI, my sarge-based C180 has not been able to boot 2.4.19-32_18 or _22.1;
> > I have to interact with IPL and fall back to 2.4.18_62.2. It gets as
> > far as the "...if this is the last message you see..." and I never hear
> > from it again, even on the network. And I think after a very very long
> > time, it falls back to firmware. At least, I see DHCP requests coming
> > out of it after a while. (I only just hooked up serial console to
> > interact with IPL after getting no response. No I don't boot it up too
> > often, it's a "side project".)
> 
> HP hardware is occasionally quirky :}
> 
> > Also, "uname -a" segfaults.
> 
> That bug got squashed in unstable. You'll see the fix soon.
>  
> > I'll try unplugging the GSC and EISA cards plugged into it tomorrow.
> > I'll also try letting it just sit there under supervision and see what
> > really happens.
> 
> Good idea :)
> If it doesn't work, get back to the list.

Well, after leaving 22.1 running, it does appear to come up after a
while. About 20 minutes. I am rather certain 18 never booted. Taking all
the expansion cards out does not make any difference.

After doing this, it occurred to me that maybe this was the bus walk
issue that I had ignored on the mailing list from a week ago. :-)

On Thu, Nov 21, 2002 at 08:34:38AM -0900, Tom wrote:
> "They" swear this has a fix in 2.5, but I just let it take it's time and
> all seems happy...

Well, we'll have to keep on them, eh? :-)

While I have somebody's attention: I have some EISA 3c597's that I'd
like to use in this machine. I carefully compile the kernel module with
the same parameters and install it and all I ever get is:

ryoko:~# modprobe 3c59x
3c59x: Donald Becker and others. www.scyld.com/network/vortex.html
3c59x: 3Com EISA 3c590 Vortex 10Mbps at 0x1000. Vers LK1.1.16

ccio-dma.c:638: Assertion dev failed!
Kernel panic: dev

Message from syslogd@ryoko at  Thu Nov 21 09:47:32 2002 ...
ryoko kernel: Kernel panic: dev

I don't get any kind of "HMPC" or register dump or anything else. Just
sits there and stops responding.

Is it ever possible that these cards would work in PA-RISC? They
certainly fit in the slot (;-)), and don't have any x86 code on them. Or
do I need to buckle down and buy a replacement backplane and go PCI?

-- 
chris jantzen kb7rnl =->         __O
Insert witty comment here.     _`\<,_
http://www.maybe.net/         (*)/ (*)

Re: [parisc-linux] C180 not booting Sarge kernel

[Matthew Wilcox]

On Thu, Nov 21, 2002 at 02:09:30PM -0800, Chris Jantzen wrote:
> While I have somebody's attention: I have some EISA 3c597's that I'd
> like to use in this machine. I carefully compile the kernel module with
> the same parameters and install it and all I ever get is:
> 
> ryoko:~# modprobe 3c59x
> 3c59x: Donald Becker and others. www.scyld.com/network/vortex.html
> 3c59x: 3Com EISA 3c590 Vortex 10Mbps at 0x1000. Vers LK1.1.16
> 
> ccio-dma.c:638: Assertion dev failed!
> Kernel panic: dev
> 
> Message from syslogd@ryoko at  Thu Nov 21 09:47:32 2002 ...
> ryoko kernel: Kernel panic: dev

Yeah, so somebody needs to write the code that allows EISA devices to
do DMA back to the host ... I have the specs here, maybe I should write
them up and somebody else can write the code?

-- 
Revolutions do not require corporate support.