Re: + dm-raid1-bitops-bug.patch added to -mm tree
From: Heiko Carstens
Date: Mon Mar 17 2008 - 11:46:16 EST
> Subject: dm raid1: bitops bug
> From: Heiko Carstens <heiko.carstens@xxxxxxxxxx>
>
> Commit 8f0205b798f926e2745de5fdebf0a8605c621de6
> "dm raid1: handle recovery failures" introduces a test_bit on an
> integer. test_bit is only defined to do something useful on arrays
> of long. Remove the test_bit code and use hand coded test bit
> instead.
>
> Cc: Jonathan Brassow <jbrassow@xxxxxxxxxx>
> Cc: Alasdair G Kergon <agk@xxxxxxxxxx>
> Signed-off-by: Heiko Carstens <heiko.carstens@xxxxxxxxxx>
> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> ---
>
> drivers/md/dm-raid1.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff -puN drivers/md/dm-raid1.c~dm-raid1-bitops-bug drivers/md/dm-raid1.c
> --- a/drivers/md/dm-raid1.c~dm-raid1-bitops-bug
> +++ a/drivers/md/dm-raid1.c
> @@ -758,7 +758,8 @@ static void recovery_complete(int read_e
> {
> struct region *reg = (struct region *)context;
> struct mirror_set *ms = reg->rh->ms;
> - int m, bit = 0;
> + unsigned int mask = 1;
> + int m;
>
> if (read_err) {
> /* Read error means the failure of default mirror. */
> @@ -776,10 +777,10 @@ static void recovery_complete(int read_e
> for (m = 0; m < ms->nr_mirrors; m++) {
> if (&ms->mirror[m] == get_default_mirror(ms))
> continue;
> - if (test_bit(bit, &write_err))
> + if (write_err & mask)
> fail_mirror(ms->mirror + m,
> DM_RAID1_SYNC_ERROR);
> - bit++;
> + mask <<= 1;
> }
> }
Could we have this patch or something similar in 2.6.25, please?
It's broken doing a test_bit() on an int since the result is random
whenever sizeof(int) != sizeof(long).
It might also lead to exceptions if an architecture requires a
sizeof(long) alignment for test_bit().
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/