isp driver - inquiry changed

[Rajesh S. Ghanekar]

Matthew Jacob wrote:

>
>
>     (da0:isp0:0:2:1): inquiry changed
>
>
> Did you transcribe by hand? The closest would be:
>
>     "Inquiry data has changed"
>
> This just means that  the the EVA sent back a command with  "Inquiry 
> Data Changed"
>
>     (da0:isp0:0:2:1): . CDB: 2a 0 8 9f 6d 25 0 0 4 0
>
>     Fatal trap 12: page fault while in kernel mode
>     mp_lock = 01000002; cpuid = 1; lapic.id <http://lapic.id> = 06000000
>     fault virtual address   = 0x60
>     fault code              = supervisor read, page not present
>     instruction pointer     = 0x8:0xc01aec1b
>     stack pointer           = 0x10:0xf4f41d5c
>     frame pointer           = 0x10:0xf4f41d74
>     code segment            = base 0x0, limit 0xfffff, type 0x1b
>                             = DPL 0, pres 1, def32 1, gran 1
>     processor eflags        = interrupt enabled, resume, IOPL = 0
>     processor eflags        = interrupt enabled, resume, IOPL = 0
>     current process         = 10750 (IOtest-static)
>     interrupt mask          = none <- SMP: XXX
>     trap number             = 12
>     panic: page fault
>     mp_lock = 01000002; cpuid = 1; lapic.id <http://lapic.id> = 06000000
>     boot() called on cpu#1
>
>
>
> w/o a traceback I can't really say what happened here with certainty. 
> Given the other probe messages somehow we seem to be re-scanning due 
> to the other system going down.
>
> Try a 'boot verbose' and see whether on the surviving system that the 
> disks are being seen as going away by the ISP driver.
>
> What's the connection topology, btw?
>

Its a simple topology with all the hosts presented to all the available 
LUNs on EVA.
I am sorry that I said it is FreeBSD-4.10, it is FreeBSD-4.1. This logs 
the message
as "inquiry changed".

                                        
--------------------------------------------
                                       |    HP 
EVA-8k                               |
                                       |    (cntrl 
1)                                      |
                                        
--------------------------------------------
                                                              |
                                                              |
            
--------------------------------------------------------------------------
            |                 FC 
Switch                                                          |
            
|                                                                                            
|
             
-------------------------------------------------------------------------
                            
|                                                 |
                            
|                                                 |
                     Machine 1                                    Machine 2



Backtrace is as follows:
-----------------------------

(kgdb) #0  0xc01a44af in boot ()
#1  0xc01a4c0f in panic ()
#2  0xc02849f0 in trap_fatal ()
#3  0xc028466d in trap_pfault ()
#4  0xc02841df in trap ()
#5  0xc01aec1b in dscheck ()
#6  0xc01ae94a in diskstrategy ()
#7  0xc01a049c in physio ()
#8  0xc01e1ee6 in spec_read ()
#9  0xc0238160 in ufsspec_read ()
#10 0xc0238a5d in ufs_vnoperatespec ()
#11 0xc01d91fc in vn_read ()
#12 0xc01b31cd in dofileread ()
#13 0xc01b2fd6 in read ()
#14 0xc0284e86 in syscall2 ()
#15 0xc027066b in Xint0x80_syscall ()
cannot read proc at 0

On the surviving system (the machine which rebooted initially) can see
all the LUNs after it comes up on reboot.

Thanks,
Rajesh