Darren Dunham wrote:
>>On Sun, 15 Apr 2007, Martin Rosenau wrote:
> 
> [...]
> 
>>No objection to the idea, just we have to keep in
>>mind:
>>
>>- we don't want to sacrifice native UFS performance
>>- we don't want to make changes to UFS that make it
>>incompatible to old
>>   existing filesystems
> 
> 
> Would the result have to be a modified UFS driver?  Especially given the 
> problems mentioned of the conditionals, how about separating it out?  It 
> wouldn't be as useful, but could there be a 'ufsx86' for SPARC and a 
> 'ufssparc' for x86?  
> 
> Both could assume the correct byte-swap instead of a conditional.  Uglier 
> than a unified driver, but the functionality would be appreciated if a unifed 
> one will never be written.
> 

Just as a point of reference, NetBSD has had support for different endianess in 
FFS for a while (never tested actually sharing disks or familiar with how it 
works) and it doesn't seem to cause a large performance overhead (at least for 
the native endianess case).

See 'options FFS_EI':
http://netbsd.gw.com/cgi-bin/man-cgi?options++NetBSD-current

Since the code is probably radically different I am not sure how helpful it 
will be, but it is out there.




_______________________________________________
ufs-discuss mailing list
[email protected]

Reply via email to