[PATCH 1/4] xfs: fix use of agi_newino in finobt lookup

Christoph Hellwig hch at infradead.org
Fri Sep 26 04:47:48 CDT 2014


On Thu, Sep 25, 2014 at 09:37:40AM +1000, Dave Chinner wrote:
> Sparse warns that we are passing the big-endian valueo f agi_newino
> to the initial btree lookup function when trying to find a new
> inode. This is wrong - we need to pass the host order value, not the
> disk order value. This will adversely affect the next inode
> allocated, but given that the free inode btree is usually much
> smaller than the allocated inode btree it is much less likely to be
> a performance issue if we start the search in the wrong place.

Looks good.  I'm rather surprised this wasn't noticed earlier.

Reviewed-by: Christoph Hellwig <hch at lst.de>



More information about the xfs mailing list