Re: [parisc-linux] Couple of questions

Re: [parisc-linux] Couple of questions

[davisg]

Greetings,

Corne' Beerse wrote:
>
> As I read this, the TOC button can have the next implementations:
> - hardware reset, as the button on intel platforms
> - three finger salute (<ctrl><alt><del>) as on intel platforms
> - Boot prom prompt (if available, like Sun's <Stop><A>)
> - Kernel core dump and [stop|halt|reboot|continue|shutdown]
> - What does HPUX with it?
> a combination of the above.

HP-UX provides an OS_TOC vector which is invoked by the platform firmware's
PDCE_TOC handler. When control is transferred to the OS_TOC handler, the OS
proceeds to perform a core dump. Then reboots (resets) the system.

>
> The main reason for this button is a kernel debugging aid. Not that I'm a
> kernel hacker, I follow this group out of interest.

Correct! TOC is used as a diagnostic tool when debugging systems which appear
to be hung and are no longer responding, i.e. TOC is a high priority interrupt
(although it's not technically classified as an interrupt - rather, it is a
special form of reset which preserves machine state) which ideally invokes
a core dump then reboots (resets) the system. The core can then be reviewed
after reboot to determine the potential cause of the hang condition.

>
> BTW, Where can I find this TOC button on my 9000/712/60 machine? is it the
> white button on the front? If so, please give some software control to it.
> In the past, I frequently pushed my keyboard against it, rebooting the
> machine while typing......
>

There is no TOC button available on the Model 712. A special debug card is
used by HP hardware and OS developers for TOC among other things on this
platform.

BTW, the white button on the front left side of the Model 712 is a soft
power switch. Upon initial application of power, the switch operates as
a hard power switch which immediately affects the power supply state (ON/OFF).
After the OS (HP-UX) has booted and initialized, a bit buried in the Lasi
power control register can be flipped by the OS to enable the soft power
feature which, with proper setup of the hardware and inclusion of the
necessary OS hooks, invokes an interrupt when depressed thus allowing
a clean shutdown of the system, i.e. the soft power switch interrupt
handler invokes 'shutdown -h now' to perform a clean unmount of all
file systems, etc,. Note that not all systems which use Lasi utilize
this feature - some don't implement any form of soft power feature,
while others use a different method entirely. But in the case of the
Model 712...


Regards,
Geoge G. Davis
Software Engineer,  Celestica Chelmsford Design Center
20 Alpha Road
Chelmsford, MA 01824

Re: [parisc-linux] Couple of questions

[Philippe Benard]

davisg@Celestica.com wrote:
> 

> 
> There is no TOC button available on the Model 712. A special debug card is
> used by HP hardware and OS developers for TOC among other things on this
> platform.

Not sure about it, but I think that pushing the white button 3 time, i.e from
ON state going OFF/ON/OFF (may be more don't remember) does trig a TOC, and
this is High prio, i.e for dump capable OS, this mean post mortem dump
analysis due to a hang possible (i.e what I like :-)

Cheers,
Phi

--
mailto:phi@hpfrcu81.france.hp.com
WTEC Project. Kernel debugging tools

Re: [parisc-linux] Couple of questions

[Corne Beerse]

davisg@Celestica.com wrote:
> 
> Greetings,
> 
> Corne' Beerse wrote:
> >
> > As I read this, the TOC button can have the next implementations:
> > - hardware reset, as the button on intel platforms
> > - three finger salute (<ctrl><alt><del>) as on intel platforms
> > - Boot prom prompt (if available, like Sun's <Stop><A>)
> > - Kernel core dump and [stop|halt|reboot|continue|shutdown]
> > - What does HPUX with it?
> > a combination of the above.
> 
> HP-UX provides an OS_TOC vector which is invoked by the platform firmware's
> PDCE_TOC handler. When control is transferred to the OS_TOC handler, the OS
> proceeds to perform a core dump. Then reboots (resets) the system.
> 
> >
> > The main reason for this button is a kernel debugging aid. Not that I'm a
> > kernel hacker, I follow this group out of interest.
> 
> Correct! TOC is used as a diagnostic tool when debugging systems which appear
> to be hung and are no longer responding, i.e. TOC is a high priority interrupt
> (although it's not technically classified as an interrupt - rather, it is a
> special form of reset which preserves machine state) which ideally invokes
> a core dump then reboots (resets) the system. The core can then be reviewed
> after reboot to determine the potential cause of the hang condition.
> 
> >
> > BTW, Where can I find this TOC button on my 9000/712/60 machine? is it the
> > white button on the front? If so, please give some software control to it.
> > In the past, I frequently pushed my keyboard against it, rebooting the
> > machine while typing......
> >
> 
> There is no TOC button available on the Model 712. A special debug card is
> used by HP hardware and OS developers for TOC among other things on this
> platform.
> 
> BTW, the white button on the front left side of the Model 712 is a soft
> power switch. Upon initial application of power, the switch operates as
> a hard power switch which immediately affects the power supply state (ON/OFF).
> After the OS (HP-UX) has booted and initialized, a bit buried in the Lasi
> power control register can be flipped by the OS to enable the soft power
> feature which, with proper setup of the hardware and inclusion of the
> necessary OS hooks, invokes an interrupt when depressed thus allowing
> a clean shutdown of the system, i.e. the soft power switch interrupt
> handler invokes 'shutdown -h now' to perform a clean unmount of all
> file systems, etc,. Note that not all systems which use Lasi utilize
> this feature - some don't implement any form of soft power feature,
> while others use a different method entirely. But in the case of the
> Model 712...

To my experience, in power-off state it is the power-on. I think it is a
pushbutton which activates a hold-relais to keep the power on. In power-on
mode it invokes `shutdown -h now` as you describe. This is on a machine
running HPUX 10.20, same experience (as I recal by head) on 10.10, 10.01 and
9.x.

For the ease of use in the future, my wich is to control the behavoure from
linux. I think of an entry in the init.tab file, similar to the
<ctrl><alt><delete> entry which is in there for intel platforms. For Sun
sparc machines, there is a similar switch. How is that handled in sparc
linux?

CB

btw, if I reply to messages received trouth parisc-linux@thepuffingroup.com,
it replies only to the originator, not to the puffin list. Other maillists
generate a Reply-to: header line with the maillist address. Can this be
added to the puffin headers too? Or are there reasons not to do it?

-- 
Everything should be as simple as possible but not simpler - A. Einstein
Corne' Beerse					| Alcatel Telecom Nederland
mailto:beerse@ats.nld.alcatel.nl		| Postbus 3292
talkto:+31(70)3079108 faxto:+31(70)3079191	| NL-2280 GG  Rijswijk

Re: [parisc-linux] Couple of questions

[willy]

On Wed, Mar 29, 2000 at 11:02:20AM +0200, Corne Beerse wrote:
> btw, if I reply to messages received trouth parisc-linux@thepuffingroup.com,
> it replies only to the originator, not to the puffin list. Other maillists
> generate a Reply-to: header line with the maillist address. Can this be
> added to the puffin headers too? Or are there reasons not to do it?

The current behaviour is correct.  If you want to reply to the person
sending the message privately, you reply; if you want to reply to the
list, you reply to all (possibly trimming the cc list as appropriate).
With the reply-to header, there is no way to accomplish this.  I'm sure
I've seen more than one `Considered Harmful' document on this subject.

Re: [parisc-linux] Couple of questions

[Peter A. Peterson II]

Quoting willy@thepuffingroup.com:
> On Wed, Mar 29, 2000 at 11:02:20AM +0200, Corne Beerse wrote:
> > btw, if I reply to messages received trouth parisc-linux@thepuffingroup.com,
> > it replies only to the originator, not to the puffin list. Other maillists
> > generate a Reply-to: header line with the maillist address. Can this be
> > added to the puffin headers too? Or are there reasons not to do it?
> 
> The current behaviour is correct.  If you want to reply to the person
> sending the message privately, you reply; if you want to reply to the
> list, you reply to all (possibly trimming the cc list as appropriate).
> With the reply-to header, there is no way to accomplish this.  I'm sure
> I've seen more than one `Considered Harmful' document on this subject.

'Reply-to Munging' is, I believe, the proper term for 'reply'
being header-forced to operate as 'reply-to-all'. 

The Goodness of reply-to and reply-to-all being separate will become
clear with time, when after some period here, you go back to another
list and accidentally send a extremely personal message to *everyone*.

Speaking as one who has sent such messages,

Peter

PS: I just got an HP-9000 E45... soon to be a Linux box!

Re: [parisc-linux] Couple of questions

[willy]

On Wed, Mar 29, 2000 at 01:10:46PM -0600, Peter A. Peterson II wrote:
> PS: I just got an HP-9000 E45... soon to be a Linux box!

let's hope so :-)  It's got lots of devices in it which are not top
of our list to get HP to relelase docs on, nor support in any way,
but eventually it should be supportable.

Re: [parisc-linux] Couple of questions

[davisg]

Greetings,

Phi wrote:
> >
> > There is no TOC button available on the Model 712. A special debug card is
> > used by HP hardware and OS developers for TOC among other things on this
> > platform.
>
> Not sure about it, but I think that pushing the white button 3 time, i.e from
> ON state going OFF/ON/OFF (may be more don't remember) does trig a TOC, and
> this is High prio, i.e for dump capable OS, this mean post mortem dump
> analysis due to a hang possible (i.e what I like :-)
>

Ooops, I stand corrected, the power switch (used in this manner while running
HP-UX) can indeed be
used to trigger a TOC. However, I still assert that the power switch itself
cannot be directly used to
generate a TOC. Refer to the "lasi_ers.ps" on or about page 78. There is a
diagram which describes
the power switch control circuit. As described in this diagram and elsewhere in
this document, the power
switch generates a power fail interrupt to the CPU only. The hardware itself
cannot directly trigger a TOC
on the Model 712. But, apparently, you can cause a TOC on the Model 712 (while
running HP-UX) by
toggling the power switch a few times. I've tried, it works, I'll be damned, I
didn't know that.

Thanks for the insight!    :  )



Regards,
Geoge G. Davis
Software Engineer,  Celestica Chelmsford Design Center
20 Alpha Road
Chelmsford, MA 01824

Re: [parisc-linux] Couple of questions

[davisg]

Greetings,

Philipp Rumpf wrote:
> > Transfer-Of-Control button/switch merely appears to result in a system
> > reset, its primary use is to recover from system hang conditions while
> > preserving the machine state for later analysis, i.e. where and why was
> > the sytem hung.
> >
> > So, in summary, although TOC on the surface appears to merely 'reset' the
> > system, it is capable of much more provided the OS specifies an OS_TOC
> > entry point for recovery and debug of machine state at any point during
> > system run time.
>
> Actually, that's exactly what reset does on (most) x86 boxes.  Remember
> getting 286s out of protected mode by resetting them ?
>

Oh yeah - I vaguely recall this. Painful, ugly and slow.

>
> I'm not convinced having customized TOC / HPMC handlers for Linux/PA-RISC
> is a good idea yet.

Agreed!  It's a lot of effort to add support for this. Looking back over the
parisc-linux archives, the alternative
is to use the platform firmware 'pim [hpmc|toc]' command to retrieve the machine
state at the time of crash/hang
and use this information to manually glark the vmlinux image to see where you
were at the time. Although there
is limited information to understand the overall machine context at time of
crash/hang, there's generally enough
information available from PIM to find out where you were (assuming you were
even in the kernel at the time).

BTW, there is a reference in the parisc-linux archives to trap handlers in the
Utah code, Mach I presume, which
may be useful as a model upon which to base HPMC/TOC handlers in Linux for
PA-RISC. Maybe someday...

Regards,
Geoge G. Davis
Software Engineer,  Celestica Chelmsford Design Center
20 Alpha Road
Chelmsford, MA 01824

Re: [parisc-linux] Couple of questions

[davisg]

Greetings,

Just a slight clarification on what TOC does. Although on the surface the
Transfer-Of-Control
button/switch merely appears to result in a system reset, its primary use is to
recover from
system hang conditions while preserving the machine state for later analysis,
i.e. where and
why was the sytem hung.

The TOC switch invokes the platform firmware PDCE_TOC entry point to save
machine state
to non-valitile storage (PIM TOC) after which the firmware checks MEM_TOC to
determine if
the operating system has identified an entry point to be called for recovery. If
an OS_TOC entry
point has been specified (by the OS), the platform firmware transfers control to
the specified
OS procedure. Typically, the OS_TOC entry point calls PDC_PIM to retrieve the
machine
state for display of a tomb stone marker, syncs/flushes all disk buffers,
performs a core dump
and then resets the system (not necessarily in the order specified here). One
could invoke a
debugger at any point in here to have a look around to see what was going on at
the time of
TOC.

The above information is all contained in the "PA-RISC x.x Firmware Architecture
Reference
Specification" documents available at
http://www.thepuffinegroup.com/parisc/documentation.html.

So, in summary, although TOC on the surface appears to merely 'reset' the
system, it is capable
of much more provided the OS specifies an OS_TOC entry point for recovery and
debug of
machine state at any point during system run time.


Regards,
Geoge G. Davis
Software Engineer,  Celestica Chelmsford Design Center
20 Alpha Road
Chelmsford, MA 01824





Philipp Rumpf <prumpf@puffin.external.hp.com> on 03/26/2000 01:02:15 PM

To:   Bas Valkema <b_valkema@hotmail.com>, parisc-linux@thepuffingroup.com
cc:    (bcc: George G Davis/CCH/Celestica)

Subject:  Re: [parisc-linux] Couple of questions



> 1 - how does a 715/50 boot over a network? Probably not bootp...

rboot.  there's an RPM for it on puffin.external.hp.com.

> 3 - What does the TOC button do????

Reset, usually.

---------------------------------------------------------------------------
To unsubscribe: send e-mail to parisc-linux-request@thepuffingroup.com with
`unsubscribe' as the subject.

Re: [parisc-linux] Couple of questions

[Philipp Rumpf]

> Just a slight clarification on what TOC does. Although on the surface the
> Transfer-Of-Control button/switch merely appears to result in a system
> reset, its primary use is to recover from system hang conditions while
> preserving the machine state for later analysis, i.e. where and why was
> the sytem hung.
> 
> So, in summary, although TOC on the surface appears to merely 'reset' the
> system, it is capable of much more provided the OS specifies an OS_TOC
> entry point for recovery and debug of machine state at any point during
> system run time.

Actually, that's exactly what reset does on (most) x86 boxes.  Remember
getting 286s out of protected mode by resetting them ?

I'm not convinced having customized TOC / HPMC handlers for Linux/PA-RISC
is a good idea yet.

	Philipp Rumpf

Re: [parisc-linux] Couple of questions

[Corne Beerse]

Philipp Rumpf wrote:
> 
> > Just a slight clarification on what TOC does. Although on the surface the
> > Transfer-Of-Control button/switch merely appears to result in a system
> > reset, its primary use is to recover from system hang conditions while
> > preserving the machine state for later analysis, i.e. where and why was
> > the sytem hung.
> >
> > So, in summary, although TOC on the surface appears to merely 'reset' the
> > system, it is capable of much more provided the OS specifies an OS_TOC
> > entry point for recovery and debug of machine state at any point during
> > system run time.
> 
> Actually, that's exactly what reset does on (most) x86 boxes.  Remember
> getting 286s out of protected mode by resetting them ?
> 
> I'm not convinced having customized TOC / HPMC handlers for Linux/PA-RISC
> is a good idea yet.
> 
As I read this, the TOC button can have the next implementations:
- hardware reset, as the button on intel platforms
- three finger salute (<ctrl><alt><del>) as on intel platforms
- Boot prom prompt (if available, like Sun's <Stop><A>)
- Kernel core dump and [stop|halt|reboot|continue|shutdown]
- What does HPUX with it?
a combination of the above.

The main reason for this button is a kernel debugging aid. Not that I'm a
kernel hacker, I follow this group out of interest.

BTW, Where can I find this TOC button on my 9000/712/60 machine? is it the
white button on the front? If so, please give some software control to it.
In the past, I frequently pushed my keyboard against it, rebooting the
machine while typing......

CB

-- 
Everything should be as simple as possible but not simpler - A. Einstein
Corne' Beerse					| Alcatel Telecom Nederland
mailto:beerse@ats.nld.alcatel.nl		| Postbus 3292
talkto:+31(70)3079108 faxto:+31(70)3079191	| NL-2280 GG  Rijswijk

[parisc-linux] Couple of questions

[Bas Valkema]

[-- Attachment #1: Type: text/plain, Size: 540 bytes --]

Hi,

a few questions:

1 - how does a 715/50 boot over a network? Probably not bootp...
2 - what is the maximum size of memory(module) (8 slots) supported in the 715/50? And what does Linux support?
3 - What does the TOC button do????
4 - Linux reports SCSI-Clock of 50MHz, this is probably wrong, right?
5 - I am willing to write a fb-driver, but can't find any specs of the graphics adapter..
6 - Is someone (thinking of) porting ld.so??
7 - Is it possible to use an ext2fs partition as root?

phew that's it.

Bas Valkema

[-- Attachment #2: Type: text/html, Size: 1361 bytes --]

Re: [parisc-linux] Couple of questions

[Philipp Rumpf]

> 1 - how does a 715/50 boot over a network? Probably not bootp...

rboot.  there's an RPM for it on puffin.external.hp.com.

> 3 - What does the TOC button do????

Reset, usually.

Re: [parisc-linux] Couple of questions

[willy]

On Sun, Mar 26, 2000 at 02:17:01PM +0200, Bas Valkema wrote:
> 1 - how does a 715/50 boot over a network? Probably not bootp...

rbootd.  There's an RPM and a tgz on puffin.external.hp.com in
/pub/parisc/binaries/

> 2 - what is the maximum size of memory(module) (8 slots) supported in the 715/50? And what does Linux support?

not sure.  You need to add them in pairs.  It's documented in the Hardware
Compatibility List that's available from the documentation page.

> 3 - What does the TOC button do????

Transfer Of Control.  It resets the machine.

> 4 - Linux reports SCSI-Clock of 50MHz, this is probably wrong, right?

According to the documentation we have, that is correct.

> 5 - I am willing to write a fb-driver, but can't find any specs of the graphics adapter..

we're working on getting those specs released.  steve shack is already
working on it though.

> 6 - Is someone (thinking of) porting ld.so??

yes, it's part of glibc.

> 7 - Is it possible to use an ext2fs partition as root?

The SCSI driver isn't working yet.  If you want to add support for the
NCR53c700 which is in the 715/50, then you will be very popular :-)