[arch-releng] isohybrid -offset 1, can't mount partitions inside live env

[Gerardo Exequiel Pozzi]

On 03/15/2010 10:25 PM, Sven-Hendrik Haase wrote:
> On 15.03.2010 16:41, Gerardo Exequiel Pozzi wrote:
>    
>> Hello,
>>
>> There is another issue with offset 1. This does not happen with offset 0.
>>
>> My scenario:
>> * 2G USB key
>> * isohybrid -offset 1
>> * another primary partition created (also is the same issue happens with a logical part instead)
>>
>>
>> # mount /dev/sdc2 /home/hola
>> mount: /dev/sdc2 already mounted or /home/hola busy
>>
>> ** strace shows:
>> mount("/dev/sdc2", "/home/hola/", "ext2", MS_MGC_VAL, NULL) = -1 EBUSY (Device or resource busy)
>>
>>
>> # mkfs.ext2 /dev/sdc2
>> mke2fs 1.41.10 (10-Feb-2009)
>> /dev/sdc2 is apparently in use by the system; will not make a filesystem here!
>>
>> ** strace shows:
>> open("/dev/sdc2", O_RDONLY|O_EXCL)      = -1 EBUSY (Device or resource busy)
>>
>>
>> My conclusion is, isohybrid mode is a great hack!. But of course, carries some inevitable issues.
>>
>> The big dichotomy:
>> * offset 1:
>>   + allows some partitions programs works without issues. (at least I know about parted issue)
>>   - Can not use extra partitions created in the medium.
>>   - Can not use /dev/disk/by-label/ skip archiso_early.
>>
>> * offset 0:
>>   - Cause issues with some partitions programs.
>>   + Can use extra partitions.
>>   + Can use "by-label" (under condition that all partitions must have a LABEL)
>>
>>
>> Good luck!
>>
>>
>>      
> Ok, so let's use the upstream recommended offset 0 for now as well. By
> the way, I was actually able to use extra partitions and boot from them
> with offset 1 but my results do not appear to be consistent.
>    
Okey.
Note that archboot use the default mode.
> They are currently deciding on a more proper way for isohybrid on the
> syslinux mailing list to work while being integrated in the iso
> generation process itself so as not to be a hack anymore. Until then,
> we'll have to think about something ourselves.
>    
Good.
> So use offset 0 instead of 1 for now?
>    
Yes. I agree.

I will push the patch that removes archiso_early on my git branch to be 
merged in master ;)

Thanks for the feedback.

-- 
Gerardo Exequiel Pozzi ( djgera )
http://www.djgera.com.ar
KeyID: 0x1B8C330D
Key fingerprint = 0CAA D5D4 CD85 4434 A219  76ED 39AB 221B 1B8C 330D