md/raid5: move some more common code into handle_stripe
authorNeilBrown <neilb@suse.de>
Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)
committerNeilBrown <neilb@suse.de>
Wed, 27 Jul 2011 01:00:36 +0000 (11:00 +1000)
The RAID6 version of this code is usable for RAID5 providing:
  - we test "conf->max_degraded" rather than "2" as appropriate
  - we make sure s->failed_num[1] is meaningful (and not '-1')
    when s->failed > 1

The 'return 1' must become 'goto finish' in the new location.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
drivers/md/raid5.c

index cf60b15b4e3a3179537c3912f12357f90dea6886..63acc51e8406e380ff303f8290cfe0d10edf0378 100644 (file)
@@ -2968,63 +2968,14 @@ static int handle_stripe5(struct stripe_head *sh, struct stripe_head_state *s)
                if (test_bit(R5_ReadError, &dev->flags))
                        clear_bit(R5_Insync, &dev->flags);
                if (!test_bit(R5_Insync, &dev->flags)) {
+                       if (s->failed < 2)
+                               s->failed_num[s->failed] = i;
                        s->failed++;
-                       s->failed_num[0] = i;
                }
        }
        spin_unlock_irq(&conf->device_lock);
        rcu_read_unlock();
 
-       if (unlikely(s->blocked_rdev)) {
-               if (s->syncing || s->expanding || s->expanded ||
-                   s->to_write || s->written) {
-                       set_bit(STRIPE_HANDLE, &sh->state);
-                       return 1;
-               }
-               /* There is nothing for the blocked_rdev to block */
-               rdev_dec_pending(s->blocked_rdev, conf->mddev);
-               s->blocked_rdev = NULL;
-       }
-
-       if (s->to_fill && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) {
-               set_bit(STRIPE_OP_BIOFILL, &s->ops_request);
-               set_bit(STRIPE_BIOFILL_RUN, &sh->state);
-       }
-
-       pr_debug("locked=%d uptodate=%d to_read=%d"
-               " to_write=%d failed=%d failed_num=%d\n",
-               s->locked, s->uptodate, s->to_read, s->to_write,
-               s->failed, s->failed_num[0]);
-       /* check if the array has lost two devices and, if so, some requests might
-        * need to be failed
-        */
-       if (s->failed > 1 && s->to_read+s->to_write+s->written)
-               handle_failed_stripe(conf, sh, s, disks, &s->return_bi);
-       if (s->failed > 1 && s->syncing) {
-               md_done_sync(conf->mddev, STRIPE_SECTORS,0);
-               clear_bit(STRIPE_SYNCING, &sh->state);
-               s->syncing = 0;
-       }
-
-       /* might be able to return some write requests if the parity block
-        * is safe, or on a failed drive
-        */
-       dev = &sh->dev[sh->pd_idx];
-       if (s->written &&
-           ((test_bit(R5_Insync, &dev->flags) &&
-             !test_bit(R5_LOCKED, &dev->flags) &&
-             test_bit(R5_UPTODATE, &dev->flags)) ||
-            (s->failed == 1 && s->failed_num[0] == sh->pd_idx)))
-               handle_stripe_clean_event(conf, sh, disks, &s->return_bi);
-
-       /* Now we might consider reading some blocks, either to check/generate
-        * parity, or to satisfy requests
-        * or to load a block that is being partially written.
-        */
-       if (s->to_read || s->non_overwrite ||
-           (s->syncing && (s->uptodate + s->compute < disks)) || s->expanding)
-               handle_stripe_fill(sh, s, disks);
-
        return 0;
 }
 
@@ -3032,8 +2983,8 @@ static int handle_stripe6(struct stripe_head *sh, struct stripe_head_state *s)
 {
        raid5_conf_t *conf = sh->raid_conf;
        int disks = sh->disks;
-       int i, pd_idx = sh->pd_idx, qd_idx = sh->qd_idx;
-       struct r5dev *dev, *pdev, *qdev;
+       struct r5dev *dev;
+       int i;
 
        /* Now to look around and see what can be done */
 
@@ -3107,65 +3058,6 @@ static int handle_stripe6(struct stripe_head *sh, struct stripe_head_state *s)
        spin_unlock_irq(&conf->device_lock);
        rcu_read_unlock();
 
-       if (unlikely(s->blocked_rdev)) {
-               if (s->syncing || s->expanding || s->expanded ||
-                   s->to_write || s->written) {
-                       set_bit(STRIPE_HANDLE, &sh->state);
-                       return 1;
-               }
-               /* There is nothing for the blocked_rdev to block */
-               rdev_dec_pending(s->blocked_rdev, conf->mddev);
-               s->blocked_rdev = NULL;
-       }
-
-       if (s->to_fill && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) {
-               set_bit(STRIPE_OP_BIOFILL, &s->ops_request);
-               set_bit(STRIPE_BIOFILL_RUN, &sh->state);
-       }
-
-       pr_debug("locked=%d uptodate=%d to_read=%d"
-              " to_write=%d failed=%d failed_num=%d,%d\n",
-              s->locked, s->uptodate, s->to_read, s->to_write, s->failed,
-              s->failed_num[0], s->failed_num[1]);
-       /* check if the array has lost >2 devices and, if so, some requests
-        * might need to be failed
-        */
-       if (s->failed > 2 && s->to_read+s->to_write+s->written)
-               handle_failed_stripe(conf, sh, s, disks, &s->return_bi);
-       if (s->failed > 2 && s->syncing) {
-               md_done_sync(conf->mddev, STRIPE_SECTORS,0);
-               clear_bit(STRIPE_SYNCING, &sh->state);
-               s->syncing = 0;
-       }
-
-       /*
-        * might be able to return some write requests if the parity blocks
-        * are safe, or on a failed drive
-        */
-       pdev = &sh->dev[pd_idx];
-       s->p_failed = (s->failed >= 1 && s->failed_num[0] == pd_idx)
-               || (s->failed >= 2 && s->failed_num[1] == pd_idx);
-       qdev = &sh->dev[qd_idx];
-       s->q_failed = (s->failed >= 1 && s->failed_num[0] == qd_idx)
-               || (s->failed >= 2 && s->failed_num[1] == qd_idx);
-
-       if (s->written &&
-           (s->p_failed || ((test_bit(R5_Insync, &pdev->flags)
-                            && !test_bit(R5_LOCKED, &pdev->flags)
-                            && test_bit(R5_UPTODATE, &pdev->flags)))) &&
-           (s->q_failed || ((test_bit(R5_Insync, &qdev->flags)
-                            && !test_bit(R5_LOCKED, &qdev->flags)
-                            && test_bit(R5_UPTODATE, &qdev->flags)))))
-               handle_stripe_clean_event(conf, sh, disks, &s->return_bi);
-
-       /* Now we might consider reading some blocks, either to check/generate
-        * parity, or to satisfy requests
-        * or to load a block that is being partially written.
-        */
-       if (s->to_read || s->non_overwrite || (s->to_write && s->failed) ||
-           (s->syncing && (s->uptodate + s->compute < disks)) || s->expanding)
-               handle_stripe_fill(sh, s, disks);
-
        return 0;
 }
 
@@ -3173,10 +3065,11 @@ static void handle_stripe(struct stripe_head *sh)
 {
        struct stripe_head_state s;
        int done;
+       raid5_conf_t *conf = sh->raid_conf;
        int i;
        int prexor;
        int disks = sh->disks;
-       raid5_conf_t *conf = sh->raid_conf;
+       struct r5dev *pdev, *qdev;
 
        clear_bit(STRIPE_HANDLE, &sh->state);
        if (test_and_set_bit(STRIPE_ACTIVE, &sh->state)) {
@@ -3213,6 +3106,67 @@ static void handle_stripe(struct stripe_head *sh)
        if (done)
                goto finish;
 
+       if (unlikely(s.blocked_rdev)) {
+               if (s.syncing || s.expanding || s.expanded ||
+                   s.to_write || s.written) {
+                       set_bit(STRIPE_HANDLE, &sh->state);
+                       goto finish;
+               }
+               /* There is nothing for the blocked_rdev to block */
+               rdev_dec_pending(s.blocked_rdev, conf->mddev);
+               s.blocked_rdev = NULL;
+       }
+
+       if (s.to_fill && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) {
+               set_bit(STRIPE_OP_BIOFILL, &s.ops_request);
+               set_bit(STRIPE_BIOFILL_RUN, &sh->state);
+       }
+
+       pr_debug("locked=%d uptodate=%d to_read=%d"
+              " to_write=%d failed=%d failed_num=%d,%d\n",
+              s.locked, s.uptodate, s.to_read, s.to_write, s.failed,
+              s.failed_num[0], s.failed_num[1]);
+       /* check if the array has lost more than max_degraded devices and,
+        * if so, some requests might need to be failed.
+        */
+       if (s.failed > conf->max_degraded && s.to_read+s.to_write+s.written)
+               handle_failed_stripe(conf, sh, &s, disks, &s.return_bi);
+       if (s.failed > conf->max_degraded && s.syncing) {
+               md_done_sync(conf->mddev, STRIPE_SECTORS, 0);
+               clear_bit(STRIPE_SYNCING, &sh->state);
+               s.syncing = 0;
+       }
+
+       /*
+        * might be able to return some write requests if the parity blocks
+        * are safe, or on a failed drive
+        */
+       pdev = &sh->dev[sh->pd_idx];
+       s.p_failed = (s.failed >= 1 && s.failed_num[0] == sh->pd_idx)
+               || (s.failed >= 2 && s.failed_num[1] == sh->pd_idx);
+       qdev = &sh->dev[sh->qd_idx];
+       s.q_failed = (s.failed >= 1 && s.failed_num[0] == sh->qd_idx)
+               || (s.failed >= 2 && s.failed_num[1] == sh->qd_idx)
+               || conf->level < 6;
+
+       if (s.written &&
+           (s.p_failed || ((test_bit(R5_Insync, &pdev->flags)
+                            && !test_bit(R5_LOCKED, &pdev->flags)
+                            && test_bit(R5_UPTODATE, &pdev->flags)))) &&
+           (s.q_failed || ((test_bit(R5_Insync, &qdev->flags)
+                            && !test_bit(R5_LOCKED, &qdev->flags)
+                            && test_bit(R5_UPTODATE, &qdev->flags)))))
+               handle_stripe_clean_event(conf, sh, disks, &s.return_bi);
+
+       /* Now we might consider reading some blocks, either to check/generate
+        * parity, or to satisfy requests
+        * or to load a block that is being partially written.
+        */
+       if (s.to_read || s.non_overwrite
+           || (conf->level == 6 && s.to_write && s.failed)
+           || (s.syncing && (s.uptodate + s.compute < disks)) || s.expanding)
+               handle_stripe_fill(sh, &s, disks);
+
        /* Now we check to see if any write operations have recently
         * completed
         */