Since the problem appears with PCMCIA cards as well, this must
be a problem in the SCI and/or FS layers.  I hope someone that knows
more about it will have a look.

    -- Lennart

Steven M. Bellovin wrote:

>In message <20030721142858.29BA714401@tiamat.goathill.org>, MLH writes:
>  
>
>>Michal 'hramrach' Suchanek wrote:
>>
>>    
>>
>>>On Sat, Jul 19, 2003 at 12:59:27PM +0200, Lennart Augustsson wrote:
>>>      
>>>
>>>>Can you get into the debugger when it happens?
>>>>
>>>>   -- Lennart
>>>>        
>>>>
>>>The first time I encountered this I got an uvm fault and was dropped to
>>>the debugger. However, the command that was supposed to give a stack trace
>>>caused one more (iirc the same) fault and sync did not finish, no coredump
>>>after reboot.
>>>      
>>>
>>FWIW, if I remove my flash reader before unmounting it on all of
>>the NetBSD hosts I use it on, that host completely locks up. No
>>coredump, no debugger, no remote login, it doesn't even respond to
>>a ping at that point. It has been this way since this reader was
>>first enabled (last fall - post 1.6 -current).
>>
>>    
>>
>
>The same is true of pcmcia cards, including ones that are removed while 
>the machine is suspend.  *That* is a bad misfeature.
>
>		--Steve Bellovin, http://www.research.att.com/~smb
>
>
>  
>