> On 24 Jan 2019, at 01.55, Zhoujie Wu <[email protected]> wrote: > > The write pointer of the bad block could be 0 or undefined, ignore > the checking of the bad block wp for pblk_line_wp_is_unbalanced to > avoid fake warning. > > Signed-off-by: Zhoujie Wu <[email protected]> > --- > drivers/lightnvm/pblk-recovery.c | 12 +++++++++--- > 1 file changed, 9 insertions(+), 3 deletions(-) > > diff --git a/drivers/lightnvm/pblk-recovery.c > b/drivers/lightnvm/pblk-recovery.c > index 5ee20da..e1fba98 100644 > --- a/drivers/lightnvm/pblk-recovery.c > +++ b/drivers/lightnvm/pblk-recovery.c > @@ -312,21 +312,27 @@ static int pblk_line_wp_is_unbalanced(struct pblk *pblk, > struct nvm_chk_meta *chunk; > struct ppa_addr ppa; > u64 line_wp; > - int pos, i; > + int pos, i, v_blk; >
v_blk is a bit confusing as it looks like the virtual block abstraction
we use in liblightnvm. In other places we call this simply 'bit'.
> - rlun = &pblk->luns[0];
> + v_blk = find_first_zero_bit(line->blk_bitmap, lm->blk_per_line);
> + if (v_blk > lm->blk_per_line)
> + return 0;
> + rlun = &pblk->luns[v_blk];
> ppa = rlun->bppa;
> pos = pblk_ppa_to_pos(geo, ppa);
> chunk = &line->chks[pos];
>
> line_wp = chunk->wp;
>
> - for (i = 1; i < lm->blk_per_line; i++) {
> + for (i = v_blk + 1; i < lm->blk_per_line; i++) {
> rlun = &pblk->luns[i];
> ppa = rlun->bppa;
> pos = pblk_ppa_to_pos(geo, ppa);
> chunk = &line->chks[pos];
>
> + if (test_bit(pos, line->blk_bitmap))
> + continue;
> +
Instead, why not simply check for chunk->state in the loop? This is in
the end the ground truth for a bad chunk.
> if (chunk->wp > line_wp)
> return 1;
> else if (chunk->wp < line_wp)
> --
> 1.9.1
Javier
signature.asc
Description: Message signed with OpenPGP
