Re: fr1-2.16 patch for 2.6.10 kernel

denis bonnenfant <[email protected]> wrote:
> P> Uh - my tests "wait" until reconstruction is complete to do a setfaulty
> > again, because I'm not fast enough to hit the keys again before it
> > completes!
> > 
> > What does the "even if" in your sentence mean? Do you mean "only if"?
> > (seulement si on n'espere pas ...).
> 
> en francais, "meme si" ! est ce que tu prefere que l'on continue in french ?

Non - je ne suis pas plus habitue a parler en francais, et je me suis
confondu entre le francais et l'espagnol dans cette phrase.  Tous les
langues d'origine latine occupent le meme reseau neuronal dans mon
cerveau!


> I added some printks in hot_add  :
> 
> hotadd
> 
> md0: md_check_recovery 2: cleared recovery bits on sb f7fc8480
> md: 3830
> md: 3838
> md: 3846
> md: 3838
> md: 3846
> md 3875: ** mddev=f7fc8480 rdev=f7ff5980
> md 3882: ** mddev=f7fc8480 rdev=f7ff5980
> md 3875: ** mddev=f7fc8480 rdev=f7ff7dc0
> md 3882: ** mddev=f7fc8480 rdev=f7ff7dc0
> md 3875: ** mddev=f7fc8480 rdev=f7fc8490
> md 3897: ** mddev=f7fc8480 rdev=f7fc8490
> 
> why there is 3 rdevs here ?

?? I don't know, without the code.

> md: 3863
> md: unlock at line 3891
> md 2361: ** mddev=f7fc8480
> 
> very beginning of hot_add_disk
> 
> md: trying to hot-add unknown-block(8,1) to md0 ...
> md 2387: ** mddev=f7fc8480 rdev=f7ff5980
> md0: repair of faulty disk 800001!
> md: trying to remove unknown-block(8,1) from md0 ...
> md0: notifying dev 800001 it is no longer in array
> md: unbind<sda1>
> md: export_rdev(sda1)
> md 2421: ** rdev=f7ff5940
> 
> it has changed ? we are here :
> 
>         if (mddev->persistent) {
>                 rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
>                         printk(KERN_INFO "md %d: ** rdev=%p\n", __LINE__, rdev);
>         }
>         else
>                 rdev->sb_offset =
> 
> 
> md0: repairing old mirror component 800001 (disk 125 >= bitmap 0)
> md0: WARNING: sda1 appears to be on the same physical disk as sda2. True
>       protection against single-disk failure might be compromised.
> md: bind<sda1>
> md0: notifying dev 800001 it is now in array
> md0: set repair bit to 1 on superblock f7fc8480
> md0: md_check_recovery 2: cleared recovery bits on sb f7fc8480
> md0: md_check_recovery 2: restored repair bit on sb f7fc8480
> md: 3830
> md: 3838
> md: 3846
> md: 3838
> md: 3846
> md 3875: ** mddev=f7fc8480 rdev=f7ff5940
> md 3882: ** mddev=f7fc8480 rdev=f7ff5940
> md 3884: ** mddev=f7fc8480 rdev=f7ff5940
> md 3885: ** mddev->pers=f88a8b60
> md 3886: ** mddev->pers->hot_add_disk=f88a5388
> Unable to handle kernel NULL pointer dereference at virtual address 00000004
>   printing eip:

So it still looks like death in hot_add_disk. Did you answer my
question about the address of hot_add_disk, as seen in /proc/kallsysms?

> f88a53d7
> 
> I'm going to add some more printks...

One at the start of hot_add_disk. You might be interested in the
__FUNCTION__ macro! It's a string.

Peter