OpenSolaris

Printable Version Enter a New Search
Bug ID 6763994
Synopsis Kernel panic after disconnecting usb disk possibly while disk is unmounting
State 10-Fix Delivered (Fix available in build)
Category:Subcategory kernel:scsa
Keywords BOP
Responsible Engineer Larry Liu
Reported Against snv_105 , s10u4_12b
Duplicate Of
Introduced In solaris_nevada
Commit to Fix snv_106
Fixed In snv_106
Release Fixed solaris_nevada(snv_106)
Related Bugs 6520112 , 6531441 , 6766288 , 6770105 , 6811936
Submit Date 27-October-2008
Last Update Date 17-July-2009
Description
During a dull conference call I unmounted my external usb disk via the desktop and disconected the drive. Moments later my workstation, an Ultra 24, paniced. From the messages in syslog just prior to the panic I assume I prematurely disconected the disk.

> ::status
debugging crash dump vmcore.0 (64-bit) from faramir
operating system: 5.11 snv_100 (i86pc)
panic message:
BAD TRAP: type=e (#pf Page fault) rp=ffffff0010fb3ad0 addr=198 occurred in modul
e "genunix" due to a NULL pointer dereference
dump content: kernel pages only

and from ::msgbuf we see

USB 2.0 device (usb59f,c41) operating at hi speed (USB 2.x) on USB 2.0 root hub:
 storage@3, scsa2usb1 at bus address 3
        LaCie  Hard Drive USB 152D203380B6
scsa2usb1 is /pci@0,0/pci108e,5351@1d,7/storage@3
/pci@0,0/pci108e,5351@1d,7/storage@3 (scsa2usb1) online
sd4 at scsa2usb1: target 0 lun 0
sd4 is /pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0
/pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4) online
WARNING: /pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4):
        Command failed to complete...Device is gone

WARNING: /pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4):
        SYNCHRONIZE CACHE command failed (5)

WARNING: /pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4):
        Command failed to complete...Device is gone

/pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4) offline
/pci@0,0/pci108e,5351@1d,7/storage@3/disk@0,0 (sd4) offline
/pci@0,0/pci108e,5351@1d,7/storage@3 (scsa2usb1) offline

^Mpanic[cpu3]/thread=ffffff0010fb3c80:
BAD TRAP: type=e (#pf Page fault) rp=ffffff0010fb3ad0 addr=198 occurred in modul
e "genunix" due to a NULL pointer dereference

Looking in /var/adm/messages* there is at least one occurance of these messages that isn't followed by a kernel panic. Between these two events I updated the system from snv_99 to snv_100. 

Attached are unix.0, vmcore.0, messages (which includes the panic) and messages.0 and messages.3 that appears to include the same warning messages from the driver but didn't panic.
Work Around
N/A
Comments
N/A