On Thu, 22 May 2014, Johnny Billquist wrote:
On 2014-05-22 12:11, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
boot [1,54]rsx11m.sys
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
After disabling parity this is what it does consistently.
@004620/000776
Do I have a bad memory board? It's far, far too late to return this one...
Is that some output already at power on before anything is booted? That would mean some selftest failed.
Nope. The only output I get at power on is 173000 (halt switch in).
I commented earlier on the other error you saw, which definitely is a case of parity error...
Yeah.
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On Thu, 22 May 2014, Johnny Billquist wrote:
I've never seen a board with the ability to disable parity...
It's a clearpoint board. Technically it's just disabling the parity circuitry CSR.
The "***EXEC PARITY ERROR STOP***" message is when you get a parity error in the memory where the kernel resides.
RSX can deal with memory errors in other parts of the memory, by just making that memory unavailable for programs to use. But it cannot do this with the parts that the kernel use, so then it panics instead.
Ahh. Damn. I'd be okay if the only other memory board I had spare wasn't PMI.
There's that 16K board...I could adjust the memory offsets.
First bugcheck is essentially an illegal instruction in kernel mode.
(Error code=110 is illegal instruction, while facility=300 means the executive.)
Ahh.
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On 2014-05-22 14:06, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
boot [1,54]rsx11m.sys
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
After disabling parity this is what it does consistently.
@004620/000776
Do I have a bad memory board? It's far, far too late to return this
one...
With parity disabled and RSTS/E:
https://ghostbin.com/paste/ksd3c a bad ~50K? I can disable everything
above 512kByte which is...i forget how many words.
512Kbytes is 256Kword. :-)
But it's not clear that the address are in Kwords...
Johnny
On 2014-05-22 12:11, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
boot [1,54]rsx11m.sys
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
After disabling parity this is what it does consistently.
@004620/000776
Do I have a bad memory board? It's far, far too late to return this one...
Is that some output already at power on before anything is booted? That would mean some selftest failed.
I commented earlier on the other error you saw, which definitely is a case of parity error...
Johnny
On 2014-05-22 11:43, Cory Smelosky wrote:
On Thu, 22 May 2014, Johnny Billquist wrote:
On 2014-05-22 11:26, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
What the hell?
boot [1,54]RSX11M.SYS
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
boot [1,54]rsx11m.sys
***EXEC PARITY ERROR STOP***
025746
@025746/000310
Uh. This doesn't make a whole lot of sense...
That sounds like a memory parity error to me...
I was just skimming through the memory board's manual for its parity
settings to see if perhaps it was disabled. Or enabled.
Doesn't look like there's a jumper for that. :(
Perhaps I have the model without parity...
I've never seen a board with the ability to disable parity...
The "***EXEC PARITY ERROR STOP***" message is when you get a parity error in the memory where the kernel resides.
RSX can deal with memory errors in other parts of the memory, by just making that memory unavailable for programs to use. But it cannot do this with the parts that the kernel use, so then it panics instead.
First bugcheck is essentially an illegal instruction in kernel mode.
(Error code=110 is illegal instruction, while facility=300 means the executive.)
Johnny
On Thu, 22 May 2014, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
boot [1,54]rsx11m.sys
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
After disabling parity this is what it does consistently.
@004620/000776
Do I have a bad memory board? It's far, far too late to return this one...
With parity disabled and RSTS/E:
https://ghostbin.com/paste/ksd3c a bad ~50K? I can disable everything above 512kByte which is...i forget how many words.
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On Thu, 22 May 2014, Cory Smelosky wrote:
boot [1,54]rsx11m.sys
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
After disabling parity this is what it does consistently.
@004620/000776
Do I have a bad memory board? It's far, far too late to return this one...
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Great, now you are set. Sorry for not answering quickly enough!
To answer your question; it depends on the configuration.
With DECnet Phase IV a cluster member should be a routing node. With DECnet Phase V you should have a router in the same LAN (and Area, of course).
Kari
On 22.5.2014 12:25, Jordi Guillaumes i Pons wrote:
El 22/05/2014, a les 11.08, Jordi Guillaumes i Pons <jg at jordi.guillaumes.name> va escriure:
El 22/05/2014, a les 9.27, Kari Uusim ki <uusimaki at exdecfinland.org> va escriure:
For the DECnet cluster alias you need (at least) one routing node (L1) or a router.
Does any of the cluster nodes have routing enabled?
I hqve a L1 router and an area router in the same LAN. Is that enough, or should I enable routing in one of the cluster members?
Answering to myself: yes, enabling the boot member as L1 router fixed the thing and now the cluster alias is visible and accessible. Thanks!
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
.
On Thu, 22 May 2014, Johnny Billquist wrote:
On 2014-05-22 11:26, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
What the hell?
boot [1,54]RSX11M.SYS
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
boot [1,54]rsx11m.sys
***EXEC PARITY ERROR STOP***
025746
@025746/000310
Uh. This doesn't make a whole lot of sense...
That sounds like a memory parity error to me...
I was just skimming through the memory board's manual for its parity settings to see if perhaps it was disabled. Or enabled.
Doesn't look like there's a jumper for that. :(
Perhaps I have the model without parity...
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On 2014-05-22 11:26, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
What the hell?
boot [1,54]RSX11M.SYS
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
boot [1,54]rsx11m.sys
***EXEC PARITY ERROR STOP***
025746
@025746/000310
Uh. This doesn't make a whole lot of sense...
That sounds like a memory parity error to me...
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On Thu, 22 May 2014, Cory Smelosky wrote:
On Thu, 22 May 2014, Cory Smelosky wrote:
What the hell?
boot [1,54]RSX11M.SYS
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
boot [1,54]rsx11m.sys
***EXEC PARITY ERROR STOP***
025746
@025746/000310
Uh. This doesn't make a whole lot of sense...
During initial boot of baseline:
SET /CRASHDEV=MM0:
00:01:39 Task "LOAT0 " terminated
Reserved inst execution
R0=140134
R1=000000
R2=000000
R3=000000
R4=001100
R5=060630
SP=120400
PC=140162
PS=170010
SET -- Crash device MM000: has been successfully loaded
SYSGEN log:
https://ghostbin.com/paste/z2z28
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On 2014-05-22 11:21, Cory Smelosky wrote:
Okay, I'm not finding a whole lot of documentation on this...
Can ZRQCH0 be used to format disks when the CSR isn't the default? Does
it prompt for them during formatting? Will I have to shove my SCSI
controller to a different address temporarily?
I have a worksheet detailing how to patch ZRQCH0 to format unsupported
geometries so upon getting XXDP booted I can FINALLY use one of my
drives...which will finally allow me to boot RT-11 on the 11/03 from an
old CMI drive...if it even still works. ;)
Unless I remember wrong, the tool asks for the CSR.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On Thu, 22 May 2014, Cory Smelosky wrote:
What the hell?
boot [1,54]RSX11M.SYS
SYSTEM FAULT DETECTED AT PC=042714 FACILITY=000300 ERROR CODE=000110
CRASH -- CRASH DRIVER NOT LOADED
boot [1,54]rsx11m.sys
***EXEC PARITY ERROR STOP***
025746
@025746/000310
Uh. This doesn't make a whole lot of sense...
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
El 22/05/2014, a les 11.08, Jordi Guillaumes i Pons <jg at jordi.guillaumes.name> va escriure:
El 22/05/2014, a les 9.27, Kari Uusim ki <uusimaki at exdecfinland.org> va escriure:
For the DECnet cluster alias you need (at least) one routing node (L1) or a router.
Does any of the cluster nodes have routing enabled?
I hqve a L1 router and an area router in the same LAN. Is that enough, or should I enable routing in one of the cluster members?
Answering to myself: yes, enabling the boot member as L1 router fixed the thing and now the cluster alias is visible and accessible. Thanks!
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
Okay, I'm not finding a whole lot of documentation on this...
Can ZRQCH0 be used to format disks when the CSR isn't the default? Does it prompt for them during formatting? Will I have to shove my SCSI controller to a different address temporarily?
I have a worksheet detailing how to patch ZRQCH0 to format unsupported geometries so upon getting XXDP booted I can FINALLY use one of my drives...which will finally allow me to boot RT-11 on the 11/03 from an old CMI drive...if it even still works. ;)
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
El 22/05/2014, a les 9.27, Kari Uusim ki <uusimaki at exdecfinland.org> va escriure:
For the DECnet cluster alias you need (at least) one routing node (L1) or a router.
Does any of the cluster nodes have routing enabled?
I hqve a L1 router and an area router in the same LAN. Is that enough, or should I enable routing in one of the cluster members?
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
On Thu, 22 May 2014, Johnny Billquist wrote:
It's in a BA123 so I currently have a MicroPDPVAX 11/23 II
FYI: PDP-11s were also sold in BA123 configurations...
I wasn't aware PDP-11s were sold in BA123s! I had wondered if they were though.
But nice. So now the system is running as it should.
Yup. Save for the one fan that seems to be failing.
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
For the DECnet cluster alias you need (at least) one routing node (L1) or a router.
Does any of the cluster nodes have routing enabled?
Kari
On 22.5.2014 9:46, Jordi Guillaumes i Pons wrote:
El 22/05/2014, a les 8.29, Kari Uusim ki <uusimaki at exdecfinland.org> va escriure:
I guess you have a DECnet router involved, but just asking.
Not really. All my machines (virtual and real) are in the same LAN, and they appear to be all adjacent (excepting the simulated KS10, which connects to one of the virtual VAXen via a KDP/DUP <-> DMC link).
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
.
El 22/05/2014, a les 8.29, Kari Uusim ki <uusimaki at exdecfinland.org> va escriure:
I guess you have a DECnet router involved, but just asking.
Not really. All my machines (virtual and real) are in the same LAN, and they appear to be all adjacent (excepting the simulated KS10, which connects to one of the virtual VAXen via a KDP/DUP <-> DMC link).
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
On 2014-05-22 05:15, Cory Smelosky wrote:
Everything I know about the Qbus tells me my bus shouldn't be
contiguous...but...
Processor Type: 11/23-PLUS Memory Size: 512. Kw
Options:
Extended Instruction Set (EIS)
Extended (22-Bit) Addressing
Parity Memory
Name Vector CSR Unit Type Remark
DUA 154 172150
0 RA82
1 RA82
2 RA82
3 RA82
MUB 330 160404
YLA 060 177560
YLB 300 176500
YLC 310 176510
YLD 320 176520
It's in a BA123 so I currently have a MicroPDPVAX 11/23 II
FYI: PDP-11s were also sold in BA123 configurations...
But nice. So now the system is running as it should.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
I guess you have a DECnet router involved, but just asking.
Kari
On 22.5.2014 0:26, Jordi Guillaumes i Pons wrote:
Hello,
I have built a simh based LAVC (just because it is cool), and I have a problem. The booting/sharing/clustering works as usual (that is, awesomly), but I can't make the decnet cluster alias work. I have defined it and if I do a show exec char I see it is defined. If I do a SET HOST to the cluster alias _FROM_ any node in the cluster, it also works (as SET HOST 0), but the address does not propagate thru the LAN, so it is not visible in SHOW NET/OLD and it is regarded as unreachable from any other node.
I have some real iron here, and there the cluster alias works as intended.
I wonder if it is a simh-related issue. I don't really know how does this thing work internally. Since DECNET has no ARP resolution I guess the MAC address corresponding to the DECNET address is somehow added to the NIC so it "listens" to both MAC addresses (the one corresponding to the node and the cluster one). Perhaps SIMH implementation does not allow this magic. If that was the case I'd consider it something to be fixed.
Any idea if I am missing something stupid before reporting this as a bug?
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
.
On Wed, 21 May 2014, Dave McGuire wrote:
:-)
Only problem is: cooling for the HDD and teh fact it's 4.3 instead of 4.6.
On 05/21/2014 11:15 PM, Cory Smelosky wrote:
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
:-)
On 05/21/2014 11:15 PM, Cory Smelosky wrote:
Everything I know about the Qbus tells me my bus shouldn't be
contiguous...but...
Processor Type: 11/23-PLUS Memory Size: 512. Kw
Options:
Extended Instruction Set (EIS)
Extended (22-Bit) Addressing
Parity Memory
Name Vector CSR Unit Type Remark
DUA 154 172150
0 RA82
1 RA82
2 RA82
3 RA82
MUB 330 160404
YLA 060 177560
YLB 300 176500
YLC 310 176510
YLD 320 176520
It's in a BA123 so I currently have a MicroPDPVAX 11/23 II
--
Dave McGuire, AK4HZ
New Kensington, PA
Everything I know about the Qbus tells me my bus shouldn't be contiguous...but...
Processor Type: 11/23-PLUS Memory Size: 512. Kw
Options:
Extended Instruction Set (EIS)
Extended (22-Bit) Addressing
Parity Memory
Name Vector CSR Unit Type Remark
DUA 154 172150
0 RA82
1 RA82
2 RA82
3 RA82
MUB 330 160404
YLA 060 177560
YLB 300 176500
YLC 310 176510
YLD 320 176520
It's in a BA123 so I currently have a MicroPDPVAX 11/23 II
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Hello,
I have built a simh based LAVC (just because it is cool), and I have a problem. The booting/sharing/clustering works as usual (that is, awesomly), but I can't make the decnet cluster alias work. I have defined it and if I do a show exec char I see it is defined. If I do a SET HOST to the cluster alias _FROM_ any node in the cluster, it also works (as SET HOST 0), but the address does not propagate thru the LAN, so it is not visible in SHOW NET/OLD and it is regarded as unreachable from any other node.
I have some real iron here, and there the cluster alias works as intended.
I wonder if it is a simh-related issue. I don't really know how does this thing work internally. Since DECNET has no ARP resolution I guess the MAC address corresponding to the DECNET address is somehow added to the NIC so it "listens" to both MAC addresses (the one corresponding to the node and the cluster one). Perhaps SIMH implementation does not allow this magic. If that was the case I'd consider it something to be fixed.
Any idea if I am missing something stupid before reporting this as a bug?
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
