On Thu, Aug 15, 2013 at 01:41:15PM -0500, Eric Sandeen wrote:
|
| You brought this topic up on an xfs community call months ago,
| as I recall. I thought we had put it to rest, because we didn't hear
| from you again until now.
|
| Am I misremembering?
The issue I brought up before was with one of two aspects of v5
superblock filesystems.
- The fact that CRCs cannot be disabled in a v5 superblock filesystem.
I capitulated on this issue on the grounds that v5 superblock
filesystems are experimental only, and I have the opportunity to
provide the code myself to disable CRCs in a v5 superblock filesystem.
- v5 superblock filesystem minimum inode size is 512 bytes. Dave
pointed out in a followup e-mail that v5 superblock inodes have
bloated so that they no longer fit in 256 bytes.
http://oss.sgi.com/archives/xfs/2013-06/msg00726.html
Yes, I have had recent issues with how Dave is implementing some of
his features. My recent complaint is unrelated to the previous two.
The dirent filetype feature is not dependent neither on v5 superblock
nor on v3 directory block for technical reasons, only by preference.
I do no sit around looking for excuses to throw barriers in Dave's way.
My only motivation is to speak up when I feel a decision has been made
which is not beneficial to the XFS community.
--
Geoffrey Wehrman 651-683-5496 gwehrman@xxxxxxx
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
