Conference vaxaxp::alphanotes

    Without the full output from all the boot messages there's no way
    to know what that system space address belongs to, and no way
    to tell you what is wrong.
    
    d.

Try booting with flags of 0,30000 to obtain more information about loaded
images.


							-Paul

Can the customer downgrade to the V4.6 console?

I made a change in the V4.7 console which moved the page tables built by the
boot command from low memory to high memory.  Apparently this is interacting
with a change that OpenVMS made between V7.0 and V7.1 in how PHYSICAL_MEMORY is
treated.  I will work on this problem with OpenVMS and report back here when it
is resolved.

stephen

    
    Thank you for the replies so far. I will suggest the customer try
    downgrading the console. This customer has been doing further testing. 
    He sees this problem on an Alpha 2100 4/233 and a 2100 4/200, but
    not on an Alphastation 250 4/266 or a 4100. 
    
      Here are the results of boot -flags 0,30000 dkb100.  The information
    prior the the message "jumping to bootstrap code" and the lines beginning
    with "BOOT QIO:VA = " have been omitted to make the large amount of data 
    a little smaller.  This boot was attempted with PHYSICAL_MEMORY = 103.
    
    P000>>> b -fl 0,30000 dkb100
    
            <Several lines omitted>
    
    
    jumping to bootstrap code
    %APB-I-APBVER, Alpha primary bootstrap, version X6C7
    Initializing Xdelta...
    Initial breakpoint not taken...
    Checking PAL code revision...
    Creating APBs memory pool...
    Initializing TIMEDWAIT constants...
    %APB-I-BOOTDEV, determining bootdevice type
    Initializing the system root specification...
    %APB-I-BOOTDRIV, Selecting boot driver
    %APB-I-BOOTFILE, Selecting boot file
    %APB-I-BOOTVOL, Bootvol, mounting boot volume
    
            < 4 lines omitted beginning with BOOT QIO:VA= >
    
    %APB-I-SYSROOT, System root is [SYS0.]
    
            < 10 lines omitted beginning with BOOT QIO:VA= >
    
    %APB-I-LOADFILE, Loading [sys0.syscommon.sysexe]sysboot.exe;1
    
            < 2 lines omitted beginning with BOOT QIO:VA= >
    
    %APB-I-SECBOOT, Transferring to secondary bootstrap
    %SYSBOOT-I-WLCM, Welcome to Alpha secondary bootstrap.
    Initializing the boot SCB
    Initializing Xdelta
    %SYSBOOT-I-LOADPARAM, Loading parameter file alphavmssys.par
            < 2 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading [SYS0.SYSEXE]ALPHAVMSSYS.PAR;19
    Creating the PFN memory map
    %SYSBOOT-I-CREATSPACE, Creating page table space
    %SYSBOOT-I-SYSPGTBL, Creating the system page tables
    %SYSBOOT-I-ALLOCPGS, Allocating loader huge pages
    Allocating physical memory for execlet code huge page
    Allocating physical memory for data huge page
    Allocating physical memory for VMS exec data huge page
    Allocating physical memory for resident image code huge page
    Allocating physical memory for VMS S2 exec data huge page
    Mapping execlet code huge page
    Mapping resident image code huge page
    Mapping VMS exec data huge page
    Allocating nonpaged pool expansion region
    Mapping execlet data huge page
    Mapping VMS S2 exec data huge page
    %SYSBOOT-I-LOADIMGS, Loading the base system images
    
            < 5 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYS$PUBLIC_VECTORS.EXE;1
    
            < 9 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYS$BASE_IMAGE.EXE;1
    
            < 4 lines omitted beginning with BOOT QIO:VA= >
    
    Fixing up sysboot references to the base images
    %SYSBOOT-I-INITDATA, Initialize MMG/loader/misc. system data cells
    Creating the balance set slots
    Creating then SYSPHD
    swp$gl_sysphd = 80C00000
    mmg$gq_syswsl = 80C00448
    Creating pool
    Allocating PAGED pool
    mmg$gl_pagedyn = 828E8000
    Allocating NONPAGED pool
    mmg$gl_npagedyn - 80C18000
    mmg$gl_npagnext = 810FA000
    Allocating error log buffers
    exe$gl_scb = 82CD0000
    Creating the primary CPU database
    Remapping file cache in system space
    Disconnecting the boot driver
    Remapping the memory descriptors into system space
    %SYSBOOT-I-REMAP. Remap HWRPB/SWRPB
    SWRPB = 80C186C0
    HWRPB = 82D1C000
    Moving the boot IO vector into system space
    Initializing the boot driver
    Initializing the primary CPU db and kstack
    Allocating the system kernel stack
    Kernel stack pointer/base = 82F24000
    %SYSBOOT-I-ALLOCPCB, Allocate boot control block
    
            < 2 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-BITMAP, Opening bitmap.sys
    
            < 1 line omitted beginning with BOOT QIO:VA= >
    
    Creating the PFN database
    pfn$pq_database = FFFFFFFE00000000
    %SYSBOOT-I-LOADEXEC, Loading boot driver execlets
    
            < 1 line omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYS$CNBTDRIVER.EXE;1
    
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    Loading execlets
    
            <1 line omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading [SYS0.SYSCOMMON.SYS$LDR]SYS$OPDRIVER.EXE;1
    
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYSTEM_PRIMITIVES_MIN.EXE;1
    
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYSTEM_SYNCHRONIZATION_MIN.EXE;1
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading [SYS0.SYSCOMMON.SYS$LDR]ERRORLOG.EXE;1
    
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading
    [SYS0.SYSCOMMON.SYS$LDR]SYS$CPU_ROUTINES_0902.EXE;1
    
            < 7 lines omitted beginning with BOOT QIO:VA= >
    
    %SYSBOOT-I-LOADFILE, Loading [SYS0.SYSCOMMON.SYS$LDR]EXEC_INIT.EXE;1
    
            < 5 lines omitted beginning with BOOT QIO:VA= >
    
    Copying the LDR$GQ_HPDESC to pool
    %SYSBOOT-I-INITARG, Init boot param arglist for EXEC_INIT
    Copying boot port allocls list
    Disconnecting the boot driver
    %IOC$DISC_BOOT-I-ENTER_ROUTINE, Entered reoutine.
    %IOC$DISC_BOOT-I-DISC_PBD, Disconnecting from primary boot device
    %IOC$DISC_BOOT-I-EXIT-ROUTINE, Exiting routine.
    Copying the system parameters to base image.
    Swap boot process context to system process.
    %SYSBOOT-I-EXECINIT,Transferring control to EXEC_INIT
    %EXECINIT-I-PERCPU, Initializing per CPU database
    %EXECINIT-I-INITSS, Calling initialization routines
    %LOADER-I-INIT, Initializing SYS$CNBTDRIVER.EXE
    %LOADER-I-INIT, Initializing SYS$OPDRIVER.EXE
    %LOADER-I-INIT, Initializing SYSTEM_PRIMITIVES_MIN.EXE
    %LOADER-I-INIT, Initializing SYSTEM_SYNCHRONIZATION_MIN.EXE
    %LOADER-I-INIT, Initializing ERRORLOG.EXE
    %LOADER-I-INIT, Initializing SYS$CPU_ROUTINES.0902.EXE
    
    
            OpenVMS(TM) Alpha Operating System, Version V7.1
    
    %EXECINIT-I-FREEBS, freeing bootstrap space PFNs
    %EXECINIT-I-PFNDB, building PFN database
    
         THE CRASH OCCURS HERE
    

The customer's observations seem correct.  I only made the change to move the
page tables to high memory on the 2000, 2100, 2100A, 1000, and 1000A.  I did not
change the 250 or the 4100.

stephen

    re: .4
    
    Thanks, but it was the BOOT QIO VA messages that I needed! That's the
    only way to determine where the execlets are loaded. But, Stephen has
    already found the failure in EXEC_INIT.

    The change in the console has exposed this problem, so hopefully
    downgrading it will be an acceptable workaround until I can get
    a fix out.

    
    Thank you. That definately seem to be the problem. My customer did not 
    have SRM V4.6-201, but did have V4.5-51, so he tried this, and the system 
    booted fine with a lowered PHYSICAL_MEMORY. Both of these versions show
    as supporting OpenVMS V7.0, so is it ok to use this with V7.1 in the
    interim? Should I IPMT This, or just watch for a reply here? 
    
    Peter  

    If you IPMT the problem it will be easier for me to get the fix out, 
    so yes, please do. I will also post a note when the fix is ready.
    
    d.