Conference mvblab::alphaserver_4100

    Very interesting!  A 2-cpu 2GB cacheless system may be somewhat rare.
    (One would imagine that most people with the $$ to buy 2GB would also
    probably buy cached CPUs).  So it's possible that you might be building
    a configuration that has less testing than other configurations.  
    
    The assertion that genvmunix from V3.2G works on your new disk but not
    on the old disk is interesting.  Could you do an ls -l on both of them
    just to make sure that there aren't any surprises?
    
    If they really are the same, you may need the help of the Unix hardware
    support team.  You could file a QAR addressed to their attention (set
    host gorge, username QAR_INTERNAL), or you could visit their home page
    and look for clues
    
           http://www.zk3.dec.com/uhs/

    I have forwarded your note off to the UNIX support group. A question
    for you in the meantime, however. Are you sure you haven't moved your
    KZPSA or other SCSI adapters at any time during your boot experiments?
    In other words, only the memory config has changed?
    
    The UNIX boot audit trail shows CAM errors on the KZPSA and NCR810 SCSI
    adapters on PCI hose 1 (slots 4, 5) and ends up not configuring them.
    
    BC

    Bill Arvidson will be getting you a patch for this known problem.
    Looks like it was fixed in the V4.0A and B stream.
    
    BC
    
    
From:	LANDO::WASTED::freiss "Rich Freiss USG  07-Mar-1997 1320"  7-MAR-1997 13:20:15.75
To:	arvidson@DEC:.zko.wasted
CC:	lando::cummins, freiss@DEC:.zko.wasted, ROMTSS::ZAGARIA,
	dutile@DEC:.zko.wasted
Subj:	Patch NEEDED ASAP for a Rawdhie Customer for V3.2G and TCR.

Bill,

	I've attached a known problem that was reported in the Rawhide
notes file.  This is a known problem that we sent you a change for
V3.2G. I've checked the support pool and the code is not submited.
The code only needs to be put in the V3.2G support pool. V4.0A has the
fix.

Can you please get a patch to the customer or the CSE ASAP?

I find it very interesting that this problem has not been reported
prior to this on other TCR V3.2G large memory installations.

Regards,
Rich Freiss


------- Forwarded Message

Return-Path:  dutile 
Delivery-Date:  Fri, 07 Mar 97 11:42:01 -0500
Return-Path:  dutile
Received:  from localhost by wasted.zk3.dec.com; (5.65v3.2/1.1.8.2/18Feb95-1123
AM)
	  	id AA22090; Fri, 7 Mar 1997 11:41:59 -0500
Message-Id:  <[email protected]>
To:  freiss
Subject:  The v3.2g patch for 4100+TCR installed
Date:  Fri, 07 Mar 97 11:41:59 -0500
From:  "Don Dutile, UNIX Systems Hardware Support" <dutile>
X-Mts:  smtp


------- Forwarded Message

To: arvidson
Cc: dutile
Subject: Rawhide patch-fix for >1GB and TCR installed
Date: Wed, 28 Aug 96 17:05:42 -0400
From: "Don Dutile, UNIX Systems Hardware Support" <dutile>


 Bill,
	There is a bug in the v3.2g support of RAWHIDE 
 such that when > 1GB of main memory is on the system, and
 TCR is installed, the kernel will not boot (it installs & builds,
 but won't boot).

 The fix for the bug can be found at:
	wasted:~dutile/rawhide/mcpcia.c.tcr_1gigplus

 
 This file is ready to be bci'd into a sandbox backed by
 v32g(supportos).  So, all one has to do is
	workon -sb your_sb_to_fix
	cd kernel/io/dec/pci
	bco mcpcia.c
	cp wasted:~dutile/rawhide/mcpcia.c.tcr_1gigplus mcpcia.c
	bco mcpcia.c

 then you can srequest, bsubmit, build, etc. the above, and
 mkpatch it into a fix-kit.

 We don't have to back port the fix to v3.2f since TCR isn't
 supported on V3.2f.  If you want to, you could do the above
 to v3.2f, and make two patch kits, one for v3.2f & v3.2g, since
 the above file has not been modified in v32fsupportos, i.e,
 mcpcia.c is the same in v3.2f, v3.2fsupportos & v3.2g.  
 One would do the v3.2f patch as a covering-the-bases in case
 someone were to set the /etc/sysconfigtab variable basic-dma-window-size
 to less than 1GB on a 1GB+ RAWHIDE/AS4100.

 The above fix has been applied and tested on the v4.0a kit, and was
 verified on a v3.2g system (both types of AS4100's reported a QAR
 on this problem).

 If you need or want more info, ping me.... Don


------- End of Forwarded Message



Return-Path: [email protected] 
Delivery-Date: Fri, 07 Mar 97 09:11:38 -0500
Return-Path: [email protected]
Received: from DECnet-Mail11.wasted.zk3.dec.com by wasted.zk3.dec.com; (5.65v3.2/1.1.8.2/18Feb95-1123AM)
	id AA04578; Fri, 7 Mar 1997 09:11:18 -0500
Date: Fri, 7 Mar 1997 09:11:18 -0500
Message-Id: <[email protected]>
Mime-Version: 1.0
From: [email protected] (Bill Cummins, PKO3-2/Q21, 223-4641)
To: [email protected]
Cc: CUMMINS
Subject: Anything jump out at you with respect to this customer log? Thanks, BC
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

      <<< MVBLAB::SYS$SYSDEVICE:[NOTES$LIBRARY]ALPHASERVER_4100.NOTE;1 >>>
                             -< AlphaServer 4100 >-
================================================================================
Note 522.0                       Memory upgrade                        2 replies
ROMTSS::ZAGARIA                                     348 lines   6-MAR-1997 04:49
--------------------------------------------------------------------------------
   
   Hi, all
   
   I have two AlphaServer 4100 with D.U. 3.2g that are working in 
   an existing TruCluster configuration . 
       
   We have upgraded both systems with a CPU board B3001-CA and
   2 x 512MB Edo memory boards (B3030-FA).
   So now we have 2 CPU B3001-ca and 2GB of memory.

   At boot time we detected a crash as listed below. The crash
   occurred with both vmunix and genvmunix .
  
   If we downgrade only the memory to 1 GB the systems work fine.
   The memory boards are 4 x 512 Edo and their locations on the 
   backplane are 0L-0H and 1L-1H , the memory test env. is set "full".


   Another hit : we reinstalled Digital Unix 3.2g with 2CPu and 2GB 
                 memory on a new spare disk and boot performed fine.
                 ( vmunix and genvmunix )

   At this point my asks are:
   
   Why old genvmunix doesn't boot correctly as the new one ?
                  
   How can I integrate new memory boards with the original software
   installation ?  

   Regards Antonio

**************************************************************************
    
 SROM V1.1 on cpu0
 SROM V1.1 on cpu1
XSROM V3.0 on cpu1
XSROM V3.0 on cpu0
mem_pair0 - 1024 MB 
mem_pair1 - 1024 MB 
20..20..21..21..23..
please wait 48 seconds for T24 to complete
24..24..
Memory testing complete on cpu0
Memory testing complete on cpu1
starting console on CPU 0
sizing memory
  0   1024 MB EDO
  1   1024 MB EDO
starting console on CPU 1
probing IOD1 hose 1 
  bus 0 slot 1 - NCR 53C810
  bus 0 slot 2 - DEC PCI FDDI
  bus 0 slot 3 - DECchip 21040-AA
  bus 0 slot 4 - DEC KZPSA
  bus 0 slot 5 - NCR 53C810
probing IOD0 hose 0 
  bus 0 slot 1 - PCEB
    probing EISA Bridge, bus 1
  bus 0 slot 2 - DEC PCI MC
  bus 0 slot 3 - S3 Trio64/Trio32
  bus 0 slot 4 - DEC KZPSA
  bus 0 slot 5 - DEC KZPSA
configuring I/O adapters...
  ncr0, hose 1, bus 0, slot 1
  pfi0, hose 1, bus 0, slot 2
  tulip0, hose 1, bus 0, slot 3
  kzpsa0, hose 1, bus 0, slot 4
  ncr1, hose 1, bus 0, slot 5
  floppy0, hose 0, bus 1, slot 0
  kzpsa1, hose 0, bus 0, slot 4
  kzpsa2, hose 0, bus 0, slot 5
  mc0, hose 0, bus 0, slot 2
System temperature is 26 degrees C
AlphaServer 4100 Console V3.0-10, 19-NOV-1996 13:57:07
P00>>>sho mem*
memory_test         	full            
P00>>>sho fru
                           Digital Equipment Corporation
                                 AlphaServer 4100

 Console V3.0-10  OpenVMS PALcode V1.19-2, Digital UNIX PALcode V1.21-14

 Module                Part #    Type     Rev    Name      Serial #
 System Motherboard    23803-01  0        0000   mthrbrd0  AY63807378
 Memory 1024 MB EDO    N/A       0        0000   mem0      N/A       
 Memory 1024 MB EDO    N/A       0        0000   mem1      N/A       
 CPU (Uncached)        B3001-CA  0        0001   cpu0      KA624PNHYD
 CPU (Uncached)        B3001-CA  0        0001   cpu1      KA621PEHVA
 Bridge (IOD0/IOD1)    B3040-AA  600      0032   iod0/iod1 AY64201654
 PCI Motherboard       B3050-AA  8        0002   saddle0   KA626PVDBB

 Bus 0  iod0 (PCI0)
 Slot   Option Name              Type     Rev    Name
 1      PCEB                     4828086  0005   pceb0    
 2      DEC PCI MC               181011   000B   mc0      
 3      S3 Trio64/Trio32         88115333 0000   vga0     
 4      DEC KZPSA                81011    0000   kzpsa1   
 5      DEC KZPSA                81011    0000   kzpsa2   

 Bus 1  pceb0 (EISA Bridge connected to iod0, slot 1)
 Slot   Option Name              Type     Rev    Name

 Bus 0  iod1 (PCI1)
 Slot   Option Name              Type     Rev    Name
 1      NCR 53C810               11000    0002   ncr0     
 2      DEC PCI FDDI             f1011    0000   pfi0     
 3      DECchip 21040-AA         21011    0024   tulip0   
 4      DEC KZPSA                81011    0000   kzpsa0   
 5      NCR 53C810               11000    0002   ncr1     
P00>>>sho device
polling ncr0 (NCR 53C810) slot 1, bus 0 PCI, hose 1   SCSI Bus ID 7
dka500.5.0.1.1     DKa500                   RRD45  0436
polling kzpsa0 (DEC KZPSA) slot 4, bus 0 PCI, hose 1   TPwr 1 Fast 1 Bus ID 7
kzpsa0.7.0.4.1     dkb     TPwr 1 Fast 1 Bus ID 7   N01  A10    
jkb0.0.0.4.1       JKb0                     TL810  1.20
mkb200.2.0.4.1     MKb200                    TZ88  CC34
mkb300.3.0.4.1     MKb300                    TZ88  CC34
mkb400.4.0.4.1     MKb400                    TZ88  CC34
mkb500.5.0.4.1     MKb500                    TZ88  CC34
polling ncr1 (NCR 53C810) slot 5, bus 0 PCI, hose 1   SCSI Bus ID 7
mkc0.0.0.5.1       MKc0                     EXABYTE EXB-85058SQANXR1  07J0
mkc300.3.0.5.1     MKc300                   TLZ09  0165
dkc400.4.0.5.1     DKc400                   RZ28D  0010
dkc500.5.0.5.1     DKc500                   RZ29B  0016
dkc600.6.0.5.1     DKc600                   RZ29B  0016
polling floppy0 (FLOPPY) PCEB - XBUS hose 0   
dva0.0.0.1000.0    DVA0                      RX23
polling kzpsa1 (DEC KZPSA) slot 4, bus 0 PCI, hose 0   TPwr 1 Fast 1 Bus ID 7
kzpsa1.7.0.4.0     dkd     TPwr 1 Fast 1 Bus ID 7   P01  A10    
dkd100.1.0.4.0     DKd100                   HSZ50  V50Z
dkd101.1.0.4.0     DKd101                   HSZ50  V50Z
dkd200.2.0.4.0     DKd200                   HSZ50  V50Z
dkd300.3.0.4.0     DKd300                   HSZ50  V50Z
dkd301.3.0.4.0     DKd301                   HSZ50  V50Z
dkd302.3.0.4.0     DKd302                   HSZ50  V50Z
dkd400.4.0.4.0     DKd400                   HSZ50  V50Z
jkd607.6.0.4.0     JKd607                   DIGITAL  ffff
polling kzpsa2 (DEC KZPSA) slot 5, bus 0 PCI, hose 0   TPwr 1 Fast 1 Bus ID 7
kzpsa2.7.0.5.0     dke     TPwr 1 Fast 1 Bus ID 7   P01  A10    
dke100.1.0.5.0     DKe100                   HSZ50  V50Z
dke101.1.0.5.0     DKe101                   HSZ50  V50Z
dke200.2.0.5.0     DKe200                   HSZ50  V50Z
dke201.2.0.5.0     DKe201                   HSZ50  V50Z
dke300.3.0.5.0     DKe300                   HSZ50  V50Z
dke400.4.0.5.0     DKe400                   HSZ50  V50Z
jke607.6.0.5.0     JKe607                   DIGITAL  ffff
polling pfi0 (DEC PCI FDDI) slot 2, bus 0 PCI, hose 1 
fwa0.0.0.2.1: 00-00-F8-40-0A-09
polling tulip0 (DECchip 21040-AA) slot 3, bus 0 PCI, hose 1 
ewa0.0.0.3.1       00-00-F8-22-82-90


P00>>>boot -fl i
(boot dkc600.6.0.5.1 -flags i)
Building FRU table
block 0 of dkc600.6.0.5.1 is a valid boot block
reading 32 blocks from dkc600.6.0.5.1
bootstrap code read in
base = 200000, image_start = 0, image_bytes = 4000
initializing HWRPB at 2000
initializing page table at 1f2000
initializing machine state
setting affinity to the primary CPU
jumping to bootstrap code

OSF boot - Mon Jul 24 21:56:39 EDT 1995

Enter [kernel_name] [option_1 ... option_n]: genvmunix
Loading genvmunix ...
Current PAL Revision <0x4000200010113>
Switching to OSF PALcode Succeeded
New PAL Revision <0x4000e00020115>
Loading into KSEG Address Space

Sizes:
text = 5293968
data = 1350768
bss  = 2798416
Starting at 0xfffffc00002394b0

Alpha boot: available memory from 0x2c08000 to 0x7fff6000
Digital UNIX V3.2G (Rev. 62); Thu Jul 18 22:21:57 EDT 1996
physical memory = 2048.00 megabytes.
available memory = 2003.92 megabytes.
using 7856 buffers containing 61.37 megabytes of memory
Master cpu at slot 0.
Firmware revision: 3.0
PALcode: Digital-UNIX/OSF version 1.21
AlphaServer 4100 5/300 0MB
pci1 at mcbus0 slot 5
psiop0 at pci1 slot 1
Loading SIOP: script c0001900, reg 4121300, data c000d838
scsi0 at psiop0 slot 0
rz5 at scsi0 bus 0 target 5 lun 0 (DEC     RRD45   (C) DEC  0436)
fta0 DEC DEFPA FDDI Module, Hardware Revision 0
fta0 at pci1 slot 2
fta0: DMA Available.
fta0: DEC DEFPA (PDQ) FDDI Interface, Hardware address: 00-00-F8-40-0A-09
fta0: Firmware rev: 2.46
tu0: DECchip 21040-AA: Revision: 2.4
tu0 at pci1 slot 3
tu0: DEC TULIP Ethernet Interface, hardware address: 00-00-F8-22-82-90
tu0: console mode: selecting 10BaseT (UTP) port: half duplex: no link
pza0 at pci1 slot 4
asr = 0x200
cam_logger: CAM_ERROR packet
cam_logger: bus 1 
spo_adap_reinit
Adapter State couldn't be set
ccfg_simattach(): SIM Attach Failed, sim_init() reported FAILURE - Retry.
pza0 not probed 
pza in slot 4 not configured.
psiop1 at pci1 slot 5
Loading SIOP: script c0013900, reg 4121000, data c0027c38
cam_logger: CAM_ERROR packet
cam_logger: bus 2 
psiop_pci_script_cntl
SCRIPTS did not start
ccfg_simattach(): SIM Attach Failed, sim_init() reported FAILURE - Retry.
psiop1 not probed 
psiop in slot 5 not configured.
gpc0 at eisa0
pci0 at mcbus0 slot 4
eisa0 at pci0
ace0 at eisa0
ace1 at eisa0
lp0 at eisa0
fdi0 at eisa0
fd0 at fdi0 unit 0
vga0 at pci0 slot 3
PCXAL keyboard, language English (American)
 1024x768 (S3TRIO  )
pza0 at pci0 slot 4
asr = 0x200
cam_logger: CAM_ERROR packet
cam_logger: bus 1 
spo_adap_reinittrap: invalid memory read access from kernel mode
    faulting virtual address:     0x00000000000000b0
    pc of faulting instruction:   0xfffffc0000529db0
    ra contents at time of fault: 0xfffffc0000529d34
    sp contents at time of fault: 0xffffffffb66f2da8

panic (cpu 0): kernel memory fault

DUMP: No primary swap, no explicit dumpdev.
	  Nowhere to put header, giving up.halted CPU 0

halt code = 5
HALT instruction executed
PC = fffffc0000454ad0
P00>>>
P00>>>
AlphaServer 4100 Console V3.0-10, 19-NOV-1996 13:57:07
P00>>>sho *
arc_enable          	OFF             
auto_action         	HALT            
boot_dev            	dkc600.6.0.5.1 dkc500.5.0.5.1
boot_file           	                
boot_osflags        	a               
boot_reset          	OFF             
bootdef_dev         	dkc600.6.0.5.1 dkc500.5.0.5.1
booted_dev          	                
booted_file         	                
booted_osflags      	                
cda0                	dka500.5.0.1.1  
char_set            	0
com1_baud           	9600            
com1_flow           	SOFTWARE        
com1_modem          	OFF             
com2_baud           	9600            
com2_flow           	SOFTWARE        
com2_modem          	OFF             
console             	serial          
cpu_enabled         	f
d_group             	field           
d_omit              	                
d_passes            	1
d_runtime           	0
d_verbose           	0
dump_dev            	                
enable_audit        	ON              
ewa0_loop_count     	3e8
ewa0_loop_inc       	a
ewa0_loop_patt      	ffffffff
ewa0_loop_size      	2e
ewa0_lp_msg_node    	1
ewa0_mode           	Twisted-Pair    
exdep_data          	5555555555555555
exdep_location      	0               
exdep_size          	0
exdep_space         	pmem            
exdep_type          	3
fru_table           	ON              
full_powerup_diags  	ON              
fwa0_loop_count     	3e8
fwa0_loop_inc       	a
fwa0_loop_patt      	ffffffff
fwa0_loop_size      	2e
fwa0_lp_msg_node    	1
graphics_background 	4               
graphics_foreground 	7               
graphics_page       	0               
graphics_switch     	-1              
graphics_sync       	0               
graphics_type       	VIDEO           
kbd_hardware_type   	LK411           
kzpsa0_fast         	1
kzpsa0_host_id      	7
kzpsa0_termpwr      	1
kzpsa1_fast         	1
kzpsa1_host_id      	7
kzpsa1_termpwr      	1
kzpsa2_fast         	1
kzpsa2_host_id      	7
kzpsa2_termpwr      	1
language            	36
language_name       	English (American)
license             	MU              
memory_test         	full            
ocp_text            	BITTP1          
os_type             	UNIX            
pal                 	OpenVMS PALcode V1.19-2, Digital UNIX PALcode V1.21-14
pci_arbmode         	Round-Robin     
pci_parity          	ON              
pci_req64           	                
pka0_disconnect     	1
pka0_fast           	1
pka0_host_id        	7
pkc0_disconnect     	1
pkc0_fast           	1
pkc0_host_id        	7
prompt              	>>>             
rcm_answer          	
rcm_dialout         	
rcm_init            	
reset_boot_arg0     	0               
reset_boot_arg1     	                
reset_boot_arg2     	                
sys_model_num       	4100            
sys_serial_num      	AY64912663      
sys_type            	RACKMOUNT       
tga_sync_green      	0
tt_allow_login      	1
tta0_page           	0               
tta0_type           	VIDEO           
tty_dev             	0               
version             	V3.0-10, 19-NOV-1996 13:57:07
P00>>>
================================================================================
Note 522.1                       Memory upgrade                           1 of 2
PERFOM::HENNING                                      15 lines   6-MAR-1997 07:13
                              -< Interesting... >-
--------------------------------------------------------------------------------
    Very interesting!  A 2-cpu 2GB cacheless system may be somewhat rare.
    (One would imagine that most people with the $$ to buy 2GB would also
    probably buy cached CPUs).  So it's possible that you might be building
    a configuration that has less testing than other configurations.  
    
    The assertion that genvmunix from V3.2G works on your new disk but not
    on the old disk is interesting.  Could you do an ls -l on both of them
    just to make sure that there aren't any surprises?
    
    If they really are the same, you may need the help of the Unix hardware
    support team.  You could file a QAR addressed to their attention (set
    host gorge, username QAR_INTERNAL), or you could visit their home page
    and look for clues
    
           http://www.zk3.dec.com/uhs/
================================================================================
Note 522.2                       Memory upgrade                           2 of 2
HARMNY::CUMMINS                                       9 lines   6-MAR-1997 13:54
                -< No SCSI adapters moved, removed, or added? >-
--------------------------------------------------------------------------------
    I have forwarded your note off to the UNIX support group. A question
    for you in the meantime, however. Are you sure you haven't moved your
    KZPSA or other SCSI adapters at any time during your boot experiments?
    In other words, only the memory config has changed?
    
    The UNIX boot audit trail shows CAM errors on the KZPSA and NCR810 SCSI
    adapters on PCI hose 1 (slots 4, 5) and ends up not configuring them.
    
    BC

------- End of Forwarded Message