Updated Hercules 3.02 RPMs available

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Updated Hercules 3.02 RPMs available

Jay Maynard
6 months later than they should have been available, I've finally fixed the
problem with the Hercules RPM version not being able to find the device
libraries and made updated RPMs available. This only affects the versions of
3.02 on Linux installed from binary RPM; all other versions are not affected
by this update.

You can download the updated RPM file from the Hercules web site at
http://www.conmicro.cx/hercules/hercules-3.02-2.i686.rpm . There's a
corresponding source RPM, as well. This version will isntall directly over
previous versions. If you've changed your configuration files to point to
the library directory, you can remove those changes after isntalling this
version.


Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

Re: Updated Hercules 3.02 RPMs available

Alessandro Brezzi
Thank you Jay.
I was able to build my x86-64 package. It work. efore I've tried with the
SRPM from freshrpms, hich is quite different in the spec file from your.
 It's possible to "measure" the raw MIPS, just to see if and optimized
version of hercules for the amd-64 arch work better and how much better?
 Regards
 Alessandro
 2005/5/24, Jay Maynard <[hidden email]>:

>
> 6 months later than they should have been available, I've finally fixed
> the
> problem with the Hercules RPM version not being able to find the device
> libraries and made updated RPMs available. This only affects the versions
> of
> 3.02 on Linux installed from binary RPM; all other versions are not
> affected
> by this update.
>
> You can download the updated RPM file from the Hercules web site at
> http://www.conmicro.cx/hercules/hercules-3.02-2.i686.rpm . There's a
> corresponding source RPM, as well. This version will isntall directly over
> previous versions. If you've changed your configuration files to point to
> the library directory, you can remove those changes after isntalling this
> version.
>
>
> Community email addresses:
> Post message: [hidden email]
> Subscribe: [hidden email]
> Unsubscribe: [hidden email]
> List owner: [hidden email]
>
> Files and archives at:
> http://groups.yahoo.com/group/hercules-390
>
> Get the latest version of Hercules from:
> http://www.conmicro.cx/hercules 
>
>
> ------------------------------
> *Yahoo! Groups Links*
>
>    - To visit your group on the web, go to:
>    http://groups.yahoo.com/group/hercules-390/
>    - To unsubscribe from this group, send an email to:
>    [hidden email]<[hidden email]?subject=Unsubscribe>
>    - Your use of Yahoo! Groups is subject to the Yahoo! Terms of
>    Service <http://docs.yahoo.com/info/terms/>.
>
>


--
Alessandro Brezzi


[Non-text portions of this message have been removed]



Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

RE: Updated Hercules 3.02 RPMs available

"Fish" (David B. Trout)
 
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Alessandro wrote:
<snip>
>  [...] It's possible to "measure" the raw MIPS, just to see
> if and optimized version of hercules for the amd-64 arch work
> better and how much better?

There is a "maxrates" command which will display the highest observed
MIPS and SIOS rate during a user-specified time interval, but it is
currently only available for the Cygwin/Windows version of Herc when
used with my HercGUI.

The upcoming 3.03 version of Hercules (we're still working on it
folks!) will support 'maxrates' for all platforms.

Note that, as stated, it will only show the MAXIMUM *observed* MIPS /
SIOS rate (for the given intervals) -- which, as we all know, is not
an accurate indicator of true performance. (But hey, at least it's
*something!*)

I've been toying with the idea of maybe calculating an *average*
MIPS/SIOS rate as well, but haven't "gotten a round tuit" yet. :)

FYI: Hercules does support a special diagnose instruction (code
X'0F08') which will return the Hercules "instruction count" value
into your specified operand-1 register, so you always have the option
of writing your own program to calculate a rough MIPS rate yourself.
<shrug>

HTH

- --
"Fish" (David B. Trout) - fish(at)infidels.org
Fight Spam! Join CAUCE! <http://www.cauce.org/>
(Any HTML email received will be deleted unread)

-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4

iQA/AwUBQpQynEj11/TE7j4qEQLP+gCghVr+gxfDP1fy9db60WQVXBFNa9YAn2Kp
7gVNH6iBeOv6xjMdSE7UtvuC
=m05t
-----END PGP SIGNATURE-----




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

Question on Speedometer and MIPS on Console Interface

Norman Hollander
Is the CPU Utilization, the MIPS and SIOs a real number?
Where do they come from?  Can they be recorded at some
interval when the OS is busy?

I'm sure this was discussed here somewhere. But I'd like to
determine a way to convert between GHz of an Intel system to
MIPs.  




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

Re: Question on Speedometer and MIPS on Console Interface

John-85
--- In [hidden email], "Norman Hollander" <norman-
hollander@h...> wrote:
> Is the CPU Utilization, the MIPS and SIOs a real number?
> Where do they come from?  Can they be recorded at some
> interval when the OS is busy?
>
> I'm sure this was discussed here somewhere. But I'd like to
> determine a way to convert between GHz of an Intel system to
> MIPs.

"Use the source, Luke!" <GRIN>

The MIPS and SIOs displayed are tracked by the Hercules emulator.
There is a counter in "cpu.c" and "sie.c" which counts the number of
instructions executed so far. It is used to determine the MIPS (but I
can seem to find the exact location).  The same process is used to
count the SIOs or SSCHs.

Note that the MIPS is actually MIPS and the value displayed depends
on the instructions used and the "wait time". For instance, if the
guest OS is in a "hard wait", then the MIPS are 0. On my Intel 2.4
Ghz, under MVS 3.8j, with a program running which is basically BR R15
where R15 points to the BR instruction (hard CPU loop using a branch
register), I get about 30 MIPS (IIRC). Now, if I change that to do a
B instead of a BR, my MIPS go down (don't remember). If I put an
really involved instruction (like MVCL) in a hard loop, my MIPS
really go down.

Therefore you will NEVER get a ?? Ghz == ?? MIPS. Also, the number
will change depending on other factors, such as other processes
running, taking host CPU time away from Hercules. Maybe even paging.
And remember that the I/O processing is being done in a separate
thread, so a heavy I/O load may decrease the MIPS due to host CPU
contention. This doesn't occur on a "real" machine because the I/O is
done by a separate, dedicated, I/O processor (the SAP - Service
Assist Processor).




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

Re: Question on Speedometer and MIPS on Console Interface

Greg Smith-14
In reply to this post by Norman Hollander
On Wed, 2005-05-25 at 13:22 -0700, Norman Hollander wrote:
> Is the CPU Utilization, the MIPS and SIOs a real number?
> Where do they come from?  Can they be recorded at some
> interval when the OS is busy?
>
> I'm sure this was discussed here somewhere. But I'd like to
> determine a way to convert between GHz of an Intel system to
> MIPs

This is a tricky question.  Whatever answer you get will have some
margin of error.

Each cpu (represented by the REGS structure) has an instcount field.
This is updated once per every 8 executed instructions.  This keeps us
from incrementing it for each instruction.  Actually, we execute 3
instructions, add 8 to the instcount, then execute 5 more instructions,
then check for pending interrupt.  Since a program interrupt can
potentially occur for any particular instruction, the value is not an
exact number.  However, tests have shown that this spot is the best
place to get closest to the actual number.

However, what an `executed instruction' is is debatable.  I think Jan
mentioned one time that a z900 SS instruction could be executed over a
dozen times due to page related program checks.

Anyway, the `timer pop' thread, which runs at a higher priority than the
cpu threads, consolidates each cpu's instcount into sysblk.instcount.
The MIPS value shows the delta between this sysblk.instcount and the
previous sysblk.instcount divided by the last time.  The same is true
for the SIOs value.

I would guesstimate that we execute 200 intel instructions per emulated
instruction.  But that depends on the instruction mix, I/O rate, how
many times we have to call DAT, etc.  Ivan has a benchmarking page but
I've lost the url.  And then, each intel instruction is going to take so
many integral clock ticks depending on how well the pipeline is doing
and how successful the l1 and l2 cache performs.

If I had to guess, I'd guess perhaps 8-10 Mips per Ghz for a distributed
workload.  I have had my dual amd 2800+ machine sustain 50 Mips while
building a package under s390 linux.

Greg



Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

RE: Question on Speedometer and MIPS on ConsoleInterface

Norman Hollander
I can buy that description.  I was trying to use HERC as a means for a
Server Consolidation to Linux on z/Series presentation.  Everyone usually
asks: "How many standalone servers can I fit onto an IFL engine?"  Of course
the answer is: "it depends".  But from a ballpark measure, I was looking for
something you indicated-  1 GHz is roughly the same as 10 MIPS.  Your
mileage will definitely vary.  I know the Workload Characteristics will play
heavily into any comparison.  But is a good place to start.  

Many Thanx-

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On
Behalf Of Greg Smith
Sent: Wednesday, May 25, 2005 SYSN 2:43 PM
To: [hidden email]
Subject: Re: [hercules-390] Question on Speedometer and MIPS on
ConsoleInterface

On Wed, 2005-05-25 at 13:22 -0700, Norman Hollander wrote:
> Is the CPU Utilization, the MIPS and SIOs a real number?
> Where do they come from?  Can they be recorded at some
> interval when the OS is busy?
>
> I'm sure this was discussed here somewhere. But I'd like to
> determine a way to convert between GHz of an Intel system to
> MIPs

This is a tricky question.  Whatever answer you get will have some
margin of error.

Each cpu (represented by the REGS structure) has an instcount field.
This is updated once per every 8 executed instructions.  This keeps us
from incrementing it for each instruction.  Actually, we execute 3
instructions, add 8 to the instcount, then execute 5 more instructions,
then check for pending interrupt.  Since a program interrupt can
potentially occur for any particular instruction, the value is not an
exact number.  However, tests have shown that this spot is the best
place to get closest to the actual number.

However, what an `executed instruction' is is debatable.  I think Jan
mentioned one time that a z900 SS instruction could be executed over a
dozen times due to page related program checks.

Anyway, the `timer pop' thread, which runs at a higher priority than the
cpu threads, consolidates each cpu's instcount into sysblk.instcount.
The MIPS value shows the delta between this sysblk.instcount and the
previous sysblk.instcount divided by the last time.  The same is true
for the SIOs value.

I would guesstimate that we execute 200 intel instructions per emulated
instruction.  But that depends on the instruction mix, I/O rate, how
many times we have to call DAT, etc.  Ivan has a benchmarking page but
I've lost the url.  And then, each intel instruction is going to take so
many integral clock ticks depending on how well the pipeline is doing
and how successful the l1 and l2 cache performs.

If I had to guess, I'd guess perhaps 8-10 Mips per Ghz for a distributed
workload.  I have had my dual amd 2800+ machine sustain 50 Mips while
building a package under s390 linux.

Greg



Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links



 






Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

RE: Updated Hercules 3.02 RPMs available

Gerhard Postpischil
In reply to this post by Jay Maynard
--- "Fish" (David B. Trout) - fish(at)infidels.org wrote:
I've been toying with the idea of maybe calculating an *average*
MIPS/SIOS rate as well, but haven't "gotten a round tuit" yet. :)

FYI: Hercules does support a special diagnose instruction (code
X'0F08') which will return the Hercules "instruction count" value
into your specified operand-1 register, so you always have the option
of writing your own program to calculate a rough MIPS rate yourself.
<shrug>

I love it. Any more secret goodies?

1) The format and use of DIAGnose used to be model dependent. Also use under VM
differs from hardware definitions. For those programmers not cognizant of
Diagnose, write this as a DC X'83x00F08' (where "x" is the output register)
in-line with your processing code (in my case I put an (E)SPIE around it)

2) When I tried it, I got a high byte of x'0D', which leads me to believe the
value could overflow during normal use.  

If this will be formally supported and documented, it might be advisable to
make the result a double-word (e.g., even/odd register).
I'd also like to see extensions or additional subcodes. At a minimum, it might
be nice to get the start time/date corresponding to the zero instruction count,
perhaps the version/release of Hercules; a platform indication (Win XP, Tiger,
linux flavor, etc.); number of CPUs; and .....?

For really fancy stuff, it would be interesting to have a formal interface for
sending host system commands, and getting responses.

Gerhard Postpischil
Bradford, VT


Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

RE: Updated Hercules 3.02 RPMs available

"Fish" (David B. Trout)
 
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

[hidden email] wrote:

> --- "Fish" (David B. Trout) - fish(at)infidels.org wrote:
> > I've been toying with the idea of maybe calculating an
> > *average* MIPS/SIOS rate as well, but haven't "gotten a
> > round tuit" yet. :)
> >
> > FYI: Hercules does support a special diagnose instruction
> > (code X'0F08') which will return the Hercules "instruction
> > count" value into your specified operand-1 register, so you
> > always have the option of writing your own program to
> > calculate a rough MIPS rate yourself. <shrug>
>
> I love it. Any more secret goodies?

Nope.  No secrets at all with open source.  ;-)


diagnose.c:


#ifdef FEATURE_HERCULES_DIAGCALLS
    case 0xF00:
   
/*---------------------------------------------------------------*/
    /* Diagnose F00: Hercules normal mode                          
*/
   
/*---------------------------------------------------------------*/
        sysblk.inststep = 0;
        SET_IC_TRACE;
        break;

    case 0xF04:
   
/*---------------------------------------------------------------*/
    /* Diagnose F04: Hercules single step mode                      
*/
   
/*---------------------------------------------------------------*/
        sysblk.inststep = 1;
        SET_IC_TRACE;
        break;

    case 0xF08:
   
/*---------------------------------------------------------------*/
    /* Diagnose F08: Hercules get instruction counter              
*/
   
/*---------------------------------------------------------------*/
        regs->GR_L(r1) = (U32)regs->instcount;
        break;

    case 0xF0C:
   
/*---------------------------------------------------------------*/
    /* Diagnose F0C: Set/reset bad frame indicator                  
*/
   
/*---------------------------------------------------------------*/
        /* Load 4K block address from R2 register */
        n = regs->GR_L(r2) & ADDRESS_MAXWRAP(regs);

        /* Convert real address to absolute address */
        n = APPLY_PREFIXING (n, regs->PX);

        /* Addressing exception if block is outside main storage */
        if ( n > regs->mainlim )
        {
            ARCH_DEP(program_interrupt) (regs,
PGM_ADDRESSING_EXCEPTION);
            break;
        }

        /* Update the storage key from R1 register bit 31 */
        STORAGE_KEY(n, regs) &= ~(STORKEY_BADFRM);
        STORAGE_KEY(n, regs) |= regs->GR_L(r1) & STORKEY_BADFRM;

        break;

    case 0xF10:
   
/*---------------------------------------------------------------*/
    /* Diagnose F10: Hercules CPU stop                              
*/
   
/*---------------------------------------------------------------*/
        regs->cpustate = CPUSTATE_STOPPING;
        ON_IC_INTERRUPT(regs);
        break;

#if defined(OPTION_DYNAMIC_LOAD)
    case 0xF14:
   
/*---------------------------------------------------------------*/
    /* Diagnose F14: Hercules DLL interface                        
*/
   
/*---------------------------------------------------------------*/
        ARCH_DEP(diagf14_call) (r1, r2, regs);
        break;
#endif /*defined(OPTION_DYNAMIC_LOAD)*/

#if !defined(NO_SIGABEND_HANDLER)
    /* The following diagnose calls cause a exigent (non-repressible)
       machine check, and are used for test purposes only *JJ */
    case 0xFE8:
   
/*---------------------------------------------------------------*/
    /* Diagnose FE8: Simulate Illegal Instruction                  
*/
   
/*---------------------------------------------------------------*/
        raise(SIGILL);
        break;

    case 0xFEC:
   
/*---------------------------------------------------------------*/
    /* Diagnose FEC: Simulate Floating Point Exception              
*/
   
/*---------------------------------------------------------------*/
        raise(SIGFPE);
        break;

    case 0xFF0:
   
/*---------------------------------------------------------------*/
    /* Diagnose FF0: Simulate Segment Violation                    
*/
   
/*---------------------------------------------------------------*/
        raise(SIGSEGV);
        break;

    case 0xFF4:
   
/*---------------------------------------------------------------*/
    /* Diagnose FF4: Simulate BUS Error                            
*/
   
/*---------------------------------------------------------------*/
        raise(SIGBUS);
        break;

    case 0xFF8:
   
/*---------------------------------------------------------------*/
    /* Diagnose FF8: Simulate Loop                                  
*/
   
/*---------------------------------------------------------------*/
        while(1);
        break;

    case 0xFFC:
   
/*---------------------------------------------------------------*/
    /* Diagnose FFC: Simulate Wait                                  
*/
   
/*---------------------------------------------------------------*/
        SLEEP(300);
        break;
#endif /*!defined(NO_SIGABEND_HANDLER)*/

#endif /*FEATURE_HERCULES_DIAGCALLS*/


> 1) The format and use of DIAGnose used to be model
> dependent. Also use under VM differs from hardware
> definitions. For those programmers not cognizant of
> Diagnose, write this as a DC X'83x00F08' (where "x"
> is the output register) in-line with your processing
> code (in my case I put an (E)SPIE around it)

DIAGnose instructions are privileged instructions, so yes, you have
to be in supervisor state (PSW) in order to execute them, otherwise
you'll get a program-check interruption ("illegal instruction"
variety).


> 2) When I tried it, I got a high byte of x'0D', which leads me
> to believe the value could overflow during normal use.

Correct. The instruction count value is maintained internally as a
64-bit counter (and thus won't likely overflow unless Herc is left
running full out for quite a long time), but we cast it to a 32-bit
value (thus causing the high-order 32-bits to get truncated/lost)
whenever we return it to the guest via the DIAG X'F08' interface, so
you need to take that into consideration whenever you use it: -- the
value that gets returned on a subsequent call could easily be
*smaller* than the value returned by a previous call.

You also need to be aware that Herc maintains the instruction/sios
counts *separately* for each CPU in your configuration too, so if
you're running on a multiple-cpu configuration and your guest o/s is
multi-cpu aware (and thus starts and actively uses the other cpu(s)
in your configuration), there's no guarantee that the value returned
from any given DIAG X'F08' call will be for the same CPU that was
returned from a previous call unless your guest o/s supports the
ability for a process/thread to specify which specific processor the
given process/thread should always be scheduled on (i.e. some sort of
"processor affinity" type function)) and your code is designed to
make use of that particular guest o/s feature/function.


> If this will be formally supported and documented,

Well, it's been in Herc since forever (several version anyway (I
think)). It just hasn't been formally documented, that's all. <shrug>
There's probably a LOT of Herc that hasn't been formally/officially
documented yet (you know how much programmers [*don't!*] like to
document their code after all!). I guess we all (we developers) keep
*hoping* that one (or more!) of our many thousands of users [of this
COMPLETELY FREE product of "ours"] will offer to "help pay for it" by
contributing whatever time/effort is involved in helping to
"spiff-up" our currently rather lame and incomplete documentation
(both *end-user* documentation AS WELL AS *internal-development*
documentation too).

But unfortunately so far no one has. :(

Thus we simply continue with our development efforts and avoid
documenting what we do like the plague.[1] :)


> it might be advisable to make the result a double-word (e.g.,
> even/odd register).

Makes sense to me.

You have our permission to go ahead and do it and offer the results
(in the form of a patch) to the rest of the community. ;-)

(Hey! This *IS* "open source" after all! It's a *community effort!*
There's nothing stopping *anyone* from contributing!)


> I'd also like to see extensions or additional subcodes.

See further/just above. :)


> At a minimum, it might be nice to get the start time/date
> corresponding to the zero instruction count, perhaps the
> version/release of Hercules; a platform indication (Win
> XP, Tiger, linux flavor, etc.); number of CPUs; and .....?

Once again, you have our explicit permission to write such code and
offer it back to the community as "payment" for using Herc. :)

(We can't after all be expected to do *everything*. We *try*, but
it's simply too much for any one person or team or people to do.
We're *always* looking for competent help.)


> For really fancy stuff, it would be interesting to have a formal
> interface for sending host system commands, and getting responses.

There already is: it's called diagnose code X'008'. :)

See the Hercules configuration file documentation for the "DIAG8CMD"
configuration file statement:

  http://www.conmicro.cx/hercules/hercconf.html#DIAG8CMD

and pay *particular* attention to the documented "Caution": enabling
DIAG 8 support can have *serious* security implications due to its
ability to issue *any* supported Hercules command -- *including* the
'sh' ("shell") command! (which allows you to issue commands to the
*HOST SYSTEM* (e.g. Windows/Linux/Apple/etc)).

Thus with diagnose 8, it's possible to issue Windows/Linux HOST
commands (via Herc's "sh" command), such as 'cd / && rm -rf *.*'
(which I most definitely wouldn't recommend doing!).

The key word in all this I guess is:  ** BE CAREFUL! **  (duh!)  :)

Hercules's diag 8 for the most part behaves identically to IBM's
existing/documented[2] DIAGNOSE Code X'08' (Virtual Console
Function).




Anyway, I guess I've rambled on long enough for this reply. If you
have any more questions, feel free to examine the code for yourself
or ask someone else. I'm unfortunately still kinda busy right now and
only responded 'cause I needed a brief few-minutes break. I
unfortunately gotta get back to work[3] now.




- ------------------------------------
[1] As one of our developers likes to say, whenever he gets the
feeling to write doc, he simply lies down until the feeling passes.
;-)     (or something like that; you get the idea...)

[2] See: SC24-5760-03: "CP Programming Services".

[3] I'm still unemployed (coming up on **>> 7 YEARS <<** now!), but
my paying client (I'm doing Herc contracting work[4] on the side)
needs me to do some work for him, and since I VERY MUCH NEED THE
MONEY (especially since Laura lost her job too recently and so now
there's only one person in the family that's officially working (my
daughter who's only making slightly above minimum wage)), I need to
get back to "taking care of business" so to speak so we can pay our
rent and buy food. :(

(It sucks being poor)

- --
"Fish" (David B. Trout) - fish(at)infidels.org
Fight Spam! Join CAUCE! <http://www.cauce.org/>
(Any HTML email received will be deleted unread)

-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4

iQA/AwUBQpbSR0j11/TE7j4qEQIK9wCfVXvlI++Ow4Zow1NJeMy0abpnblsAn14O
LYyzxJrFCOoNnM7SG0emXfSx
=1/Wt
-----END PGP SIGNATURE-----




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Reply | Threaded
Open this post in threaded view
|

Documentation effort (was Re: Updated Hercules 3.02 RPMs available)

Roger Bowler
--- In [hidden email], "Fish" <fish@i...> wrote:
> I guess we all (we developers) keep > *hoping* that one (or more!)
> of our many thousands of users [of this COMPLETELY FREE product
> of "ours"] will offer to "help pay for it" by contributing
> whatever time/effort is involved in helping to "spiff-up" our
> currently rather lame and incomplete documentation (both *end-
> user* documentation AS WELL AS *internal-development*
> documentation too).
>
> But unfortunately so far no one has. :(

Don't forget there are actually quite a few sites with Hercules
documentation. One that springs to mind is Axel Schwarzer's site
which has some howto documentation plus a superb pocket reference
card all indexed by a delightful 3279-style screen coded in HTML:

http://pweb.uunet.de/schwarzer.d/projects/hercules/

Regards,
Roger Bowler




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply | Threaded
Open this post in threaded view
|

RE: Documentation effort (was Re: Updated Hercules 3.02 RPMs available)

"Fish" (David B. Trout)
 
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Roger Bowler wrote:

> --- In [hidden email], "Fish" <fish@i...> wrote:
> > I guess we all (we developers) keep > *hoping* that one (or
> > more!) of our many thousands of users [of this COMPLETELY FREE
> > product of "ours"] will offer to "help pay for it" by
> > contributing
> > whatever time/effort is involved in helping to "spiff-up" our
> > currently rather lame and incomplete documentation (both *end-
> > user* documentation AS WELL AS *internal-development*
> > documentation too).
> >
> > But unfortunately so far no one has. :(
>
> Don't forget there are actually quite a few sites with Hercules
> documentation.

Quite true, quite true...

It'd be nice though if we could somehow gather them all together into
once place and/or add the best of them to the official distribution
(with the author's permission of course!).

Just a thought...  :)

> One that springs to mind is Axel Schwarzer's site
> which has some howto documentation plus a superb pocket
> reference card all indexed by a delightful 3279-style
> screen coded in HTML:
>
> http://pweb.uunet.de/schwarzer.d/projects/hercules/

WOW! COOL!

That *is* nice!  (Thanks Axel!)

(and thank YOU Roger for mentioning it!)

<me: immediately downloads Herc ref cards>

- --
"Fish" (David B. Trout) - fish(at)infidels.org
Fight Spam! Join CAUCE! <http://www.cauce.org/>
(Any HTML email received will be deleted unread)

-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4

iQA/AwUBQpb6wEj11/TE7j4qEQJ93QCeJVpF3+pdr6wJ1fzSHTRMVKCR+VUAoLpK
DV5nNN2LLZoUPXbRUmGJEhmH
=hqfw
-----END PGP SIGNATURE-----




Community email addresses:
  Post message: [hidden email]
  Subscribe:    [hidden email]
  Unsubscribe:  [hidden email]
  List owner:   [hidden email]

Files and archives at:
  http://groups.yahoo.com/group/hercules-390

Get the latest version of Hercules from:
  http://www.conmicro.cx/hercules 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/hercules-390/

<*> To unsubscribe from this group, send an email to:
    [hidden email]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/