[caiman-discuss] Code review for bug 434

[Moinak Ghosh]

Moinak Ghosh wrote:
> Dave Miner wrote:
>   
>> Moinak Ghosh wrote:
>>     
>>> Dave Miner wrote:
>>>       
>>>> Anil Gulecha wrote:
>>>>  
>>>>         
>>>>> The new code review is at 
>>>>> http://cr.opensolaris.org/~anilg/001_fsroot_v3/
>>>>>
>>>>> *Signature checking has been added (searches for solaris.zlib)
>>>>> *the second lofi hard code has been changed.. but I'm not sure on how
>>>>> to proceed for the first. It is being sesnt to the background and the
>>>>> output thsu cannot be caught using ``. Any ideas on this?
>>>>>     
>>>>>           
>>>> Do we really need to do it this way anymore?  I'm curious whether we 
>>>> still need to call devfsadm specifically to get lofi initialized, 
>>>> but even if we do, seems like we could re-order so that lofi 
>>>> initialized first and then run the lofiadm in the foreground.
>>>>   
>>>>         
>>>    Not going to work since devfsadmd must be running
>>>    (see my previous email). That starts when the
>>>    system/device/local service is initialized later on in boot.
>>>
>>>       
>> OK, I thought the number of changes in devfs since you originally 
>> wrote this might have made a difference.  But why wouldn't doing the 
>> devfsadm -i lofi before the first lofiadm work, though?
>>     
>
>    It will not be useful since devfsadm -i lofi is a one-time action;
>    does not start the daemon. It just checks the current device tree,
>    creates links if needed and comes out.
>   

   One thing that can probably be done is to run the daemon: 
/usr/lib/devfsadm/devfsadmd
   perform the lofi actions and then kill it. Should simplify the code 
if it works.

Regards,
Moinak.