| Message ID | 20230623172525.2513235-1-linan666@huaweicloud.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | md/raid1: prioritize adding disk to 'removed' mirror | expand |
Hi, 在 2023/06/24 1:25, linan666@huaweicloud.com 写道: > From: Li Nan <linan122@huawei.com> > > New disk should be added to "removed" position first instead of to be a > replacement. Commit 6090368abcb4 ("md/raid10: prioritize adding disk to > 'removed' mirror") has fixed this issue for raid10. This patch fixes raid1. This commit message "This patch ..." shound use imperative mood, other than that, this patch LGTM. Reviewed-by: Yu Kuai <yukuai3@huawei.com> > > Signed-off-by: Li Nan <linan122@huawei.com> > --- > drivers/md/raid1.c | 26 +++++++++++++++----------- > 1 file changed, 15 insertions(+), 11 deletions(-) > > diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c > index 68a9e2d9985b..320bede4bfab 100644 > --- a/drivers/md/raid1.c > +++ b/drivers/md/raid1.c > @@ -1782,7 +1782,7 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) > { > struct r1conf *conf = mddev->private; > int err = -EEXIST; > - int mirror = 0; > + int mirror = 0, repl_slot = -1; > struct raid1_info *p; > int first = 0; > int last = conf->raid_disks - 1; > @@ -1825,17 +1825,21 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) > break; > } > if (test_bit(WantReplacement, &p->rdev->flags) && > - p[conf->raid_disks].rdev == NULL) { > - /* Add this device as a replacement */ > - clear_bit(In_sync, &rdev->flags); > - set_bit(Replacement, &rdev->flags); > - rdev->raid_disk = mirror; > - err = 0; > - conf->fullsync = 1; > - rcu_assign_pointer(p[conf->raid_disks].rdev, rdev); > - break; > - } > + p[conf->raid_disks].rdev == NULL && repl_slot < 0) > + repl_slot = mirror; > } > + > + if (err && repl_slot >= 0) { > + /* Add this device as a replacement */ > + p = conf->mirrors + repl_slot; > + clear_bit(In_sync, &rdev->flags); > + set_bit(Replacement, &rdev->flags); > + rdev->raid_disk = repl_slot; > + err = 0; > + conf->fullsync = 1; > + rcu_assign_pointer(p[conf->raid_disks].rdev, rdev); > + } > + > print_conf(conf); > return err; > } >
在 2023/6/25 10:06, Yu Kuai 写道: > Hi, > > 在 2023/06/24 1:25, linan666@huaweicloud.com 写道: >> From: Li Nan <linan122@huawei.com> >> >> New disk should be added to "removed" position first instead of to be a >> replacement. Commit 6090368abcb4 ("md/raid10: prioritize adding disk to >> 'removed' mirror") has fixed this issue for raid10. This patch fixes >> raid1. > > This commit message "This patch ..." shound use imperative mood, other > than that, this patch LGTM. > > Reviewed-by: Yu Kuai <yukuai3@huawei.com> OK, I will change it. Thanks for your review.
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 68a9e2d9985b..320bede4bfab 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -1782,7 +1782,7 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) { struct r1conf *conf = mddev->private; int err = -EEXIST; - int mirror = 0; + int mirror = 0, repl_slot = -1; struct raid1_info *p; int first = 0; int last = conf->raid_disks - 1; @@ -1825,17 +1825,21 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) break; } if (test_bit(WantReplacement, &p->rdev->flags) && - p[conf->raid_disks].rdev == NULL) { - /* Add this device as a replacement */ - clear_bit(In_sync, &rdev->flags); - set_bit(Replacement, &rdev->flags); - rdev->raid_disk = mirror; - err = 0; - conf->fullsync = 1; - rcu_assign_pointer(p[conf->raid_disks].rdev, rdev); - break; - } + p[conf->raid_disks].rdev == NULL && repl_slot < 0) + repl_slot = mirror; } + + if (err && repl_slot >= 0) { + /* Add this device as a replacement */ + p = conf->mirrors + repl_slot; + clear_bit(In_sync, &rdev->flags); + set_bit(Replacement, &rdev->flags); + rdev->raid_disk = repl_slot; + err = 0; + conf->fullsync = 1; + rcu_assign_pointer(p[conf->raid_disks].rdev, rdev); + } + print_conf(conf); return err; }