[Cluster-devel] [GFS2 PATCH] GFS2: Introduce new gfs2_log_header_v2

[Bob Peterson]

Hi,

----- Original Message -----
| On 01/12/17 19:34, Bob Peterson wrote:
| > Hi,
| >
| > This patch adds a new structure called gfs2_log_header_v2 which is
| > used to store expanded fields into previously unused areas of the
| > log headers. Some of these are used for debug purposes so we can
| > backtrack when problems occur. Others are reserved for future
| > expansion.
| >
| > This is based on a prototype patch from Steve Whitehouse.
(snip)
| Where has setting the CRC gone?  That appears to be missing from this patch.

Well, I'm not convinced we really need another crc in this first patch.
The original part of the log header already has a crc32 field.
The new fields are all just for debugging until we get to the patches
that take advantage of the local statfs and quota fields (rather than
separate inodes), in which case keeping a crc may be a good idea to
ensure their integrity. But if (or when) we do that, we should also add
code to check its integrity during recovery. Setting the crc just seemed
out of the scope of this patch, but I guess it's not that big of a deal;
I have no problem adding it back in on the assumption that we'll
eventually get to that point.

(snip)
| > +	__be32 lh_pad;
| Why is there padding here? There is nothing following it.

Yeah, I fully expected to expand the structure with some more
fields. So I was going to use it, but didn't quite get there. So it was
kind of just a reminder to myself: I wanted to preempt potential
compiler-padding during future expansion.

Bob