Conference 7.286::fddi

1215.0. "DEFQA crashes 4000-400 ?" by ZUR01::PMAIER (recycled hardware engineer) Thu Jan 27 1994 05:00

SDA> sho crash
    
Version of system: VAX/VMS VERSION V5.5-2H4

VAXcluster node: VAXQ, a VAX 4000-400

CPU bugcheck codes:
	CPU 00 -- SSRVEXCEPT, Unexpected system service exception

Process currently executing on this CPU: CLUSTER_SERVER
    
Current image file: $1$DIA0:[SYS0.SYSCOMMON.][SYSEXE]CSP.EXE;1
    
Current IPL: 0  (decimal)

General registers:

	R0  = 00000009   R1  = 80002398   R2  = 000063E9   R3  = 09AB0E8F
	R4  = 806B4490   R5  = 00000000   R6  = 00040A00   R7  = 00000003
	R8  = 00000000   R9  = 7FFC0480   R10 = 7FFC053E   R11 = 7FFC050A
	AP  = 7FFE776C   FP  = 7FFE7754   SP  = 7FFE7754   PC  = 8000239E
	PSL = 00000000


Current operating stack (KERNEL):

		7FFE774C  8000239E	EXE$EXCPTN+00006
		7FFE7750  00000000	

	 SP =>  7FFE7754  00000000	
		7FFE7758  00000000	
		7FFE775C  7FF85368	
		7FFE7760  7FFE77A8	CTL$GL_KSTKBAS+005A8
		7FFE7764  80000014	EXE$QIOW_3+00004
		7FFE7768  802742C4	EXCEPTION+024C4
		7FFE776C  00000002	
		7FFE7770  7FFE7790	CTL$GL_KSTKBAS+00590
		7FFE7774  7FFE7778	CTL$GL_KSTKBAS+00578
		7FFE7778  00000004	
		7FFE777C  7FFE77E4	CTL$GL_KSTKBAS+005E4
		7FFE7780  FFFFFFFD	
		7FFE7784  070063E9	
		7FFE7788  8942AA6F	
		7FFE778C  000008F8	BUG$_SYSAPLERR
		7FFE7790  00000005	
		7FFE7794  0000000C	
		7FFE7798  00000001	
		7FFE779C  8942AA6F	
		7FFE77A0  00000827	BUG$_SECURITY2+00007
		7FFE77A4  08C00001	
		7FFE77A8  00000000	

SDA> exam/instr 827-20;20

BUG$_SECURITY2+00007:   MOVC3   R0,@1E(R2),@1E(R1)[R1]

Extract from CSP.LIS listing


	  MOVL    CSP$GL_CURCTX, R6			     
	  MOVQ    12(R6), R1				      
	  MOVL    26(R2), SIZE				      
	  ADDL3   30(R1), SIZE, R3			      
	  CMPZV   #0, #16, 8(R1), R3			      
	  BGEQU   1$					      
	  MOVZWL  8(R1), SIZE				      
	  SUBL2   30(R1), SIZE				      
	  MOVL    #3, COMMAND				      
	  BRB     3$					      
1$:	  TSTL    26(R2)				      
	  BEQL    2$					      
 	  MOVL    #5, COMMAND				      
 	  BRB     3$					      
2$:	  MOVL    #4, COMMAND				      
3$:	  MOVC3   SIZE, @30(R2), @30(R1)[R1]		      


I have verified that R6 has CSP$GL_CURCTX as contents.

SDA> exam R6
R6:	     00040A00

SDA> exam @r6+C;4          (C is Decimal 12)
00000000 00000000 00040B40 806C6F60  `ol.@.......................     00040A0C

R1 is 806C6F60 after the MOVQ 12(R6), R1
R2 is 00040B40 after the MOVQ 12(R6), R1


SDA> exam 806C6F60+1E     (1E is Decimal 30)
806C6f7E:	893E9F2F

MOVC3   SIZE, @30(R2), @30(R1)[R1] 

fails with 893E9F2F + 40A0C = 8942AA6F --> @30(R2)  
On the stack R2 and R1 are turned around. This is correct. Microcode
writes the results of the MOVC3 instruction in this registers. 

Both addresses in the MOVC3 instruction are wrong:

@30(R2)     R2 = 40B40   30(R2) = 40B5E =  8942AA6F

@30(R1)[R1]  R1 = 806C6F60  30(R1) = 893E9F2F 

 		      
One of the problem is the value of 806C6F60 in location
CSP$GL_CURCTX+C

The search of the memory for this address ends up with two consecutive
VCRP data structures :

VCRP    806C4610  5280
			 004C018E 004E14A0 806C0830 806C1CD0 �.l.0.l...N...L.
			 00000000 AAAA0659 000400AA 00000000 ....�...Y.��....
			 806C4780 01AD0016 00000000 8058F43A :�X.......�..Gl.
			 .
			 .
			 20202020 20202020 00000000 00000000 ........        
			 20202020 003B0120 00001020 806C6F60 <---------------- 
			 20202020 20202020 20202020 20202020                 
			 20202020 20202020 20202020 20202020                 
			 20202020 20202020 20202020 20202020                 
			 20202020 20202020 20202020 20202020                 
VCRP    806C5AB0  5280
			 0030016E 004E14A0 807239D0 806C0830 0.l.�9r...N.n.0.
			 806C0000 AAAA0406 000400AA 00000000 ....�.....��..l.
			 806C5C00 02470001 00000000 8058F43A :�X.......G..\l.
			 .
			 .
			 .
			 20202020 20202030 00000000 00000000 ........0       
			 00000000 003B0120 00000380 806C6F60 <------------------
			 00000000 00000000 00000000 00000000 ................
			 00000000 00000000 00000000 00000000 ................
			 00000000 00000000 00000000 00000000 ................


Removing both FDDI interfaces (FQA and FQB) "fixed" this problem 


Question: Why are this VCRP's 5280 bytes long ?
    
    Regards Peter
    

    can somebody give me the listings  for the FQDRIVER ?
    
    I could not find any single listing for VMS5.5-2H .
    
    Thanks    Peter