Re: Re: fr1-2.16 patch for 2.6.10 kernel

I've now compiled 2.6.8.1 SMP and have run experiments. I  get no oops:

   143  16:43   sudo mdadm -A /dev/md0 /dev/hda20 /dev/hda21
   146  16:43   sudo raidsetfaulty /dev/md0 /dev/hda21
   147  16:43   sudo raidhotadd /dev/md0 /dev/hda21
   153  16:44   sudo raidsetfaulty /dev/md0 /dev/hda21
   154  16:44   sudo raidhotadd /dev/md0 /dev/hda21
   157  16:44   sudo raidsetfaulty /dev/md0 /dev/hda21
   158  16:44   sudo raidhotadd /dev/md0 /dev/hda21
   162  16:45   sudo raidsetfaulty /dev/md0 /dev/hda21
   163  16:45   sudo raidhotadd /dev/md0 /dev/hda21
   164  16:45   sudo raidsetfaulty /dev/md0 /dev/hda21
   165  16:45   sudo raidhotadd /dev/md0 /dev/hda21
   170  16:45   sudo raidsetfaulty /dev/md0 /dev/hda21
   171  16:45   sudo raidhotadd /dev/md0 /dev/hda21

I tried setfaulty before the resync had finished a couple of times.

% cat /proc/mdstat 
Personalities : [raid1] 
md0 : active raid1 hda21[1] hda20[0]
      7936 blocks [2/2] [UU]
      
unused devices: <none>

The partitions are only 8MB each.

I think I saw the SB recognition issue on alternate experiments, but
I wasn't looking specifically for it. 

I am compiling with gcc 2.95.4. SMP, PREEMPT, etc.

md: trying to hot-add unknown-block(3,21) to md0 ... 
md0: repair of faulty disk 300015!
md: trying to remove unknown-block(3,21) from md0 ... 
md0: notifying dev 300015 it is no longer in array
md: unbind<hda21>
md: export_rdev(hda21)
md 1205: unlock_rdev nulls bdev
md 1197: lock_rdev restores bdev 300015
md0: repairing old mirror component 300015 (disk 51 >= bitmap 0)
md0: WARNING: hda21 appears to be on the same physical disk as hda20.  True
     protection against single-disk failure might be compromised.
md: bind<hda21>
md0: notifying dev 300015 it is now in array
md0: set repair bit to 1 on superblock cff48800
md0: md_check_recovery 2: repair bit 100 on sb cff48800 preserved
RAID1 conf printout:
 --- wd:1 rd:2
 disk 0, wo:0, o:1, dev:hda20
 disk 1, wo:1, o:1, dev:hda21
md: syncing RAID array md0 (sb cff48800)
md: minimum _guaranteed_ reconstruction speed: 1000 KB/sec/disc.
md: using maximum available idle IO bandwith (but not more than 200000
KB/sec) for reconstruction.
md: using 128k window, over a total of 7936 blocks.
md0: repair bit set on sb so retained bitmap cff68d20
raid1: skipped clean sectors 0-15871
md: md0: sync done.
RAID1 conf printout:
 --- wd:2 rd:2
 disk 0, wo:0, o:1, dev:hda20
 disk 1, wo:0, o:1, dev:hda21

Peter