On 5/14/15 8:39 AM, Brian Foster wrote:
> On Tue, May 12, 2015 at 02:30:15PM -0500, Eric Sandeen wrote:
>> This adds a new superblock field, sb_meta_uuid. This allows us to
>> change the use-visible UUID on crc-enabled filesytems from userspace
>> if desired, by copying the existing UUID to the new location for
>> metadata comparisons. If this is done, an incompat flag must be
>> set to prevent older filesystems from mounting the filesystem, but
>> the original UUID can be restored, and the incompat flag removed,
>> with a new xfs_db / xfs_admin UUID command, "restore."
>>
>> Much of this patch mirrors the kernel patch in simply renaming
>> the field used for metadata uuid comparison; other bits:
>>
>> * Teach xfs_db to print the new meta_uuid field
>> * Allow xfs_db to generate a new UUID for CRC-enabled filesystems
>> * Allow xfs_db to revert to the original UUID and clear the flag
>> * Fix up xfs_copy to work with CRC-enabled filesystems
>> * Update the xfs_admin manpage to show the UUID "restore" command
>>
>> Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx>
>> ---
>>
>> This version makes the UUID "restore" command silent if there
>> is nothing to do.
>>
>> Still open to bike-shedding on the "restore" name. Maybe "revert?"
>>
>
> Restore seems fine to me...
>
>> diff --git a/copy/xfs_copy.c b/copy/xfs_copy.c
>> index 279527c..a58eee5 100644
>> --- a/copy/xfs_copy.c
>> +++ b/copy/xfs_copy.c
>> @@ -466,6 +466,34 @@ write_wbuf(void)
>> sighold(SIGCHLD);
>> }
>>
>> +void
>> +sb_update_uuid(
>> + xfs_sb_t *sb,
>> + ag_header_t *ag_hdr,
>> + thread_args *tcarg)
>> +{
>> + /*
>> + * If this filesystem has CRCs, the original UUID is stamped into
>> + * all metadata. We need to copy the original UUID into the meta_uuid
>> + * slot and set the incompat flag if that hasn't already been done.
>> + */
>> + if (xfs_sb_version_hascrc(sb) && !xfs_sb_version_hasmetauuid(sb)) {
>> + __be32 feat;
>> +
>> + feat = be32_to_cpu(ag_hdr->xfs_sb->sb_features_incompat);
>> + feat |= XFS_SB_FEAT_INCOMPAT_META_UUID;
>> + ag_hdr->xfs_sb->sb_features_incompat = cpu_to_be32(feat);
>> + platform_uuid_copy(&ag_hdr->xfs_sb->sb_meta_uuid,
>> + &ag_hdr->xfs_sb->sb_uuid);
>> + }
>> +
>> + platform_uuid_copy(&ag_hdr->xfs_sb->sb_uuid, &tcarg->uuid);
>> +
>> + /* We changed the UUID, so update the superblock CRC if needed */
>> + if (xfs_sb_version_hascrc(sb))
>> + xfs_update_cksum((char *)&ag_hdr->xfs_sb, sb->sb_sectsize,
>> + XFS_SB_CRC_OFF);
> xfs_copy doesn't work for me (crc=1,finobt=1)...
>
> # ./copy/xfs_copy /dev/vdb1 /dev/vdb2
> 0% ... 10% ... 20% ... 30% ... 40% ... 50% ... 60% ... 70% ... 80%
> ... 90% ... 100%
>
> All copies completed.
> # mount /dev/vdb2 /mnt/
> mount: mount /dev/vdb2 on /mnt failed: Structure needs cleaning
> # dmesg | tail
> ...
> [ 4145.609403] XFS (vdb2): Metadata CRC error detected at
> xfs_sb_read_verify+0x112/0x170 [xfs], block 0xffffffffffffffff
> [ 4145.611224] XFS (vdb2): Unmount and run xfs_repair
> [ 4145.611945] XFS (vdb2): First 64 bytes of corrupted metadata buffer:
> [ 4145.612873] ffff8800d8def000: 58 46 53 42 00 00 10 00 00 00 00 00 00 25 40
> 00 XFSB.........%@.
> [ 4145.615047] ffff8800d8def010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> 00 ................
> [ 4145.616359] ffff8800d8def020: 73 68 6d 76 a1 6d 43 96 8f 35 cb 98 2c ff fd
> 8d shmv.mC..5..,...
> [ 4145.617926] ffff8800d8def030: 00 00 00 00 00 20 00 05 00 00 00 00 00 00 00
> 60 ..... .........`
> [ 4145.619267] XFS (vdb2): SB validate failed with error -74.
>
> The same thing occurs with -d. It looks like the crc might not have
> updated..?
yeargh,
- xfs_update_cksum((char *)&ag_hdr->xfs_sb, sb->sb_sectsize,
+ xfs_update_cksum((char *)ag_hdr->xfs_sb, sb->sb_sectsize,
will fix up, test harder, and resend :(
-Eric
|