Merge branch 'linux-next' of git://git.infradead.org/~dedekind/ubi-2.6

* 'linux-next' of git://git.infradead.org/~dedekind/ubi-2.6:
  UBI: fix checkpatch.pl warnings
  UBI: simplify PEB protection code
  UBI: prepare for protection tree improvements
  UBI: return -ENOMEM upon failing vmalloc
  UBI: document UBI ioctls
  UBI: handle write errors in WL worker
  UBI: fix error path
  UBI: some code re-structuring
  UBI: fix deadlock
  UBI: fix warnings when debugging is enabled

7 files changed, 301 insertions, 330 deletions

diff --git a/drivers/mtd/ubi/build.c b/drivers/mtd/ubi/build.c
index c7630a228310..ba0bd3d5775b 100644
--- a/drivers/mtd/ubi/build.c
+++ b/drivers/mtd/ubi/build.c
@@ -815,19 +815,20 @@ int ubi_attach_mtd_dev(struct mtd_info *mtd, int ubi_num, int vid_hdr_offset)
         if (err)
                 goto out_free;
 
+        err = -ENOMEM;
         ubi->peb_buf1 = vmalloc(ubi->peb_size);
         if (!ubi->peb_buf1)
                 goto out_free;
 
         ubi->peb_buf2 = vmalloc(ubi->peb_size);
         if (!ubi->peb_buf2)
-                 goto out_free;
+                goto out_free;
 
 #ifdef CONFIG_MTD_UBI_DEBUG
         mutex_init(&ubi->dbg_buf_mutex);
         ubi->dbg_peb_buf = vmalloc(ubi->peb_size);
         if (!ubi->dbg_peb_buf)
-                 goto out_free;
+                goto out_free;
 #endif
 
         err = attach_by_scanning(ubi);
diff --git a/drivers/mtd/ubi/cdev.c b/drivers/mtd/ubi/cdev.c
index b30a0b83d7f1..98cf31ed0814 100644
--- a/drivers/mtd/ubi/cdev.c
+++ b/drivers/mtd/ubi/cdev.c
@@ -721,7 +721,8 @@ static int rename_volumes(struct ubi_device *ubi,
                  * It seems we need to remove volume with name @re->new_name,
                  * if it exists.
                  */
-                desc = ubi_open_volume_nm(ubi->ubi_num, re->new_name, UBI_EXCLUSIVE);
+                desc = ubi_open_volume_nm(ubi->ubi_num, re->new_name,
+                                          UBI_EXCLUSIVE);
                 if (IS_ERR(desc)) {
                         err = PTR_ERR(desc);
                         if (err == -ENODEV)
diff --git a/drivers/mtd/ubi/debug.h b/drivers/mtd/ubi/debug.h
index 78e914d23ece..13777e5beac9 100644
--- a/drivers/mtd/ubi/debug.h
+++ b/drivers/mtd/ubi/debug.h
@@ -27,11 +27,11 @@
 #define dbg_err(fmt, ...) ubi_err(fmt, ##__VA_ARGS__)
 
 #define ubi_assert(expr)  do {                                               \
-        if (unlikely(!(expr))) {                                             \
-                printk(KERN_CRIT "UBI assert failed in %s at %u (pid %d)\n", \
-                       __func__, __LINE__, current->pid);                    \
-                ubi_dbg_dump_stack();                                        \
-        }                                                                    \
+        if (unlikely(!(expr))) {                                             \
+                printk(KERN_CRIT "UBI assert failed in %s at %u (pid %d)\n", \
+                       __func__, __LINE__, current->pid);                    \
+                ubi_dbg_dump_stack();                                        \
+        }                                                                    \
 } while (0)
 
 #define dbg_msg(fmt, ...)                                    \
diff --git a/drivers/mtd/ubi/eba.c b/drivers/mtd/ubi/eba.c
index d8966bae0e0b..048a606cebde 100644
--- a/drivers/mtd/ubi/eba.c
+++ b/drivers/mtd/ubi/eba.c
@@ -504,12 +504,9 @@ static int recover_peb(struct ubi_device *ubi, int pnum, int vol_id, int lnum,
         if (!vid_hdr)
                 return -ENOMEM;
 
-        mutex_lock(&ubi->buf_mutex);
-
 retry:
         new_pnum = ubi_wl_get_peb(ubi, UBI_UNKNOWN);
         if (new_pnum < 0) {
-                mutex_unlock(&ubi->buf_mutex);
                 ubi_free_vid_hdr(ubi, vid_hdr);
                 return new_pnum;
         }
@@ -529,20 +526,23 @@ retry:
                 goto write_error;
 
         data_size = offset + len;
+        mutex_lock(&ubi->buf_mutex);
         memset(ubi->peb_buf1 + offset, 0xFF, len);
 
         /* Read everything before the area where the write failure happened */
         if (offset > 0) {
                 err = ubi_io_read_data(ubi, ubi->peb_buf1, pnum, 0, offset);
                 if (err && err != UBI_IO_BITFLIPS)
-                        goto out_put;
+                        goto out_unlock;
         }
 
         memcpy(ubi->peb_buf1 + offset, buf, len);
 
         err = ubi_io_write_data(ubi, ubi->peb_buf1, new_pnum, 0, data_size);
-        if (err)
+        if (err) {
+                mutex_unlock(&ubi->buf_mutex);
                 goto write_error;
+        }
 
         mutex_unlock(&ubi->buf_mutex);
         ubi_free_vid_hdr(ubi, vid_hdr);
@@ -553,8 +553,9 @@ retry:
         ubi_msg("data was successfully recovered");
         return 0;
 
-out_put:
+out_unlock:
         mutex_unlock(&ubi->buf_mutex);
+out_put:
         ubi_wl_put_peb(ubi, new_pnum, 1);
         ubi_free_vid_hdr(ubi, vid_hdr);
         return err;
@@ -567,7 +568,6 @@ write_error:
         ubi_warn("failed to write to PEB %d", new_pnum);
         ubi_wl_put_peb(ubi, new_pnum, 1);
         if (++tries > UBI_IO_RETRIES) {
-                mutex_unlock(&ubi->buf_mutex);
                 ubi_free_vid_hdr(ubi, vid_hdr);
                 return err;
         }
@@ -949,10 +949,14 @@ write_error:
  * This function copies logical eraseblock from physical eraseblock @from to
  * physical eraseblock @to. The @vid_hdr buffer may be changed by this
  * function. Returns:
- *   o %0  in case of success;
- *   o %1 if the operation was canceled and should be tried later (e.g.,
- *     because a bit-flip was detected at the target PEB);
- *   o %2 if the volume is being deleted and this LEB should not be moved.
+ *   o %0 in case of success;
+ *   o %1 if the operation was canceled because the volume is being deleted
+ *        or because the PEB was put meanwhile;
+ *   o %2 if the operation was canceled because there was a write error to the
+ *        target PEB;
+ *   o %-EAGAIN if the operation was canceled because a bit-flip was detected
+ *     in the target PEB;
+ *   o a negative error code in case of failure.
  */
 int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
                      struct ubi_vid_hdr *vid_hdr)
@@ -978,7 +982,7 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
         /*
          * Note, we may race with volume deletion, which means that the volume
          * this logical eraseblock belongs to might be being deleted. Since the
-         * volume deletion unmaps all the volume's logical eraseblocks, it will
+         * volume deletion un-maps all the volume's logical eraseblocks, it will
          * be locked in 'ubi_wl_put_peb()' and wait for the WL worker to finish.
          */
         vol = ubi->volumes[idx];
@@ -986,7 +990,7 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
                 /* No need to do further work, cancel */
                 dbg_eba("volume %d is being removed, cancel", vol_id);
                 spin_unlock(&ubi->volumes_lock);
-                return 2;
+                return 1;
         }
         spin_unlock(&ubi->volumes_lock);
 
@@ -1023,7 +1027,7 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
 
         /*
          * OK, now the LEB is locked and we can safely start moving it. Since
-         * this function utilizes thie @ubi->peb1_buf buffer which is shared
+         * this function utilizes the @ubi->peb1_buf buffer which is shared
          * with some other functions, so lock the buffer by taking the
          * @ubi->buf_mutex.
          */
@@ -1068,8 +1072,11 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
         vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi));
 
         err = ubi_io_write_vid_hdr(ubi, to, vid_hdr);
-        if (err)
+        if (err) {
+                if (err == -EIO)
+                        err = 2;
                 goto out_unlock_buf;
+        }
 
         cond_resched();
 
@@ -1079,14 +1086,17 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
                 if (err != UBI_IO_BITFLIPS)
                         ubi_warn("cannot read VID header back from PEB %d", to);
                 else
-                        err = 1;
+                        err = -EAGAIN;
                 goto out_unlock_buf;
         }
 
         if (data_size > 0) {
                 err = ubi_io_write_data(ubi, ubi->peb_buf1, to, 0, aldata_size);
-                if (err)
+                if (err) {
+                        if (err == -EIO)
+                                err = 2;
                         goto out_unlock_buf;
+                }
 
                 cond_resched();
 
@@ -1101,15 +1111,16 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to,
                                 ubi_warn("cannot read data back from PEB %d",
                                          to);
                         else
-                                err = 1;
+                                err = -EAGAIN;
                         goto out_unlock_buf;
                 }
 
                 cond_resched();
 
                 if (memcmp(ubi->peb_buf1, ubi->peb_buf2, aldata_size)) {
-                        ubi_warn("read data back from PEB %d - it is different",
-                                 to);
+                        ubi_warn("read data back from PEB %d and it is "
+                                 "different", to);
+                        err = -EINVAL;
                         goto out_unlock_buf;
                 }
         }
diff --git a/drivers/mtd/ubi/io.c b/drivers/mtd/ubi/io.c
index 2fb64be44f1b..a74118c05745 100644
--- a/drivers/mtd/ubi/io.c
+++ b/drivers/mtd/ubi/io.c
@@ -637,8 +637,6 @@ int ubi_io_read_ec_hdr(struct ubi_device *ubi, int pnum,
 
         dbg_io("read EC header from PEB %d", pnum);
         ubi_assert(pnum >= 0 && pnum < ubi->peb_count);
-        if (UBI_IO_DEBUG)
-                verbose = 1;
 
         err = ubi_io_read(ubi, ec_hdr, pnum, 0, UBI_EC_HDR_SIZE);
         if (err) {
@@ -685,6 +683,9 @@ int ubi_io_read_ec_hdr(struct ubi_device *ubi, int pnum,
                         if (verbose)
                                 ubi_warn("no EC header found at PEB %d, "
                                          "only 0xFF bytes", pnum);
+                        else if (UBI_IO_DEBUG)
+                                dbg_msg("no EC header found at PEB %d, "
+                                        "only 0xFF bytes", pnum);
                         return UBI_IO_PEB_EMPTY;
                 }
 
@@ -696,7 +697,9 @@ int ubi_io_read_ec_hdr(struct ubi_device *ubi, int pnum,
                         ubi_warn("bad magic number at PEB %d: %08x instead of "
                                  "%08x", pnum, magic, UBI_EC_HDR_MAGIC);
                         ubi_dbg_dump_ec_hdr(ec_hdr);
-                }
+                } else if (UBI_IO_DEBUG)
+                        dbg_msg("bad magic number at PEB %d: %08x instead of "
+                                "%08x", pnum, magic, UBI_EC_HDR_MAGIC);
                 return UBI_IO_BAD_EC_HDR;
         }
 
@@ -708,7 +711,9 @@ int ubi_io_read_ec_hdr(struct ubi_device *ubi, int pnum,
                         ubi_warn("bad EC header CRC at PEB %d, calculated "
                                  "%#08x, read %#08x", pnum, crc, hdr_crc);
                         ubi_dbg_dump_ec_hdr(ec_hdr);
-                }
+                } else if (UBI_IO_DEBUG)
+                        dbg_msg("bad EC header CRC at PEB %d, calculated "
+                                "%#08x, read %#08x", pnum, crc, hdr_crc);
                 return UBI_IO_BAD_EC_HDR;
         }
 
@@ -912,8 +917,6 @@ int ubi_io_read_vid_hdr(struct ubi_device *ubi, int pnum,
 
         dbg_io("read VID header from PEB %d", pnum);
         ubi_assert(pnum >= 0 &&  pnum < ubi->peb_count);
-        if (UBI_IO_DEBUG)
-                verbose = 1;
 
         p = (char *)vid_hdr - ubi->vid_hdr_shift;
         err = ubi_io_read(ubi, p, pnum, ubi->vid_hdr_aloffset,
@@ -960,6 +963,9 @@ int ubi_io_read_vid_hdr(struct ubi_device *ubi, int pnum,
                         if (verbose)
                                 ubi_warn("no VID header found at PEB %d, "
                                          "only 0xFF bytes", pnum);
+                        else if (UBI_IO_DEBUG)
+                                dbg_msg("no VID header found at PEB %d, "
+                                        "only 0xFF bytes", pnum);
                         return UBI_IO_PEB_FREE;
                 }
 
@@ -971,7 +977,9 @@ int ubi_io_read_vid_hdr(struct ubi_device *ubi, int pnum,
                         ubi_warn("bad magic number at PEB %d: %08x instead of "
                                  "%08x", pnum, magic, UBI_VID_HDR_MAGIC);
                         ubi_dbg_dump_vid_hdr(vid_hdr);
-                }
+                } else if (UBI_IO_DEBUG)
+                        dbg_msg("bad magic number at PEB %d: %08x instead of "
+                                "%08x", pnum, magic, UBI_VID_HDR_MAGIC);
                 return UBI_IO_BAD_VID_HDR;
         }
 
@@ -983,7 +991,9 @@ int ubi_io_read_vid_hdr(struct ubi_device *ubi, int pnum,
                         ubi_warn("bad CRC at PEB %d, calculated %#08x, "
                                  "read %#08x", pnum, crc, hdr_crc);
                         ubi_dbg_dump_vid_hdr(vid_hdr);
-                }
+                } else if (UBI_IO_DEBUG)
+                        dbg_msg("bad CRC at PEB %d, calculated %#08x, "
+                                "read %#08x", pnum, crc, hdr_crc);
                 return UBI_IO_BAD_VID_HDR;
         }
 
@@ -1024,7 +1034,7 @@ int ubi_io_write_vid_hdr(struct ubi_device *ubi, int pnum,
 
         err = paranoid_check_peb_ec_hdr(ubi, pnum);
         if (err)
-                return err > 0 ? -EINVAL: err;
+                return err > 0 ? -EINVAL : err;
 
         vid_hdr->magic = cpu_to_be32(UBI_VID_HDR_MAGIC);
         vid_hdr->version = UBI_VERSION;
diff --git a/drivers/mtd/ubi/ubi.h b/drivers/mtd/ubi/ubi.h
index 1c3fa18c26a7..4a8ec485c91d 100644
--- a/drivers/mtd/ubi/ubi.h
+++ b/drivers/mtd/ubi/ubi.h
@@ -74,6 +74,13 @@
 #define UBI_IO_RETRIES 3
 
 /*
+ * Length of the protection queue. The length is effectively equivalent to the
+ * number of (global) erase cycles PEBs are protected from the wear-leveling
+ * worker.
+ */
+#define UBI_PROT_QUEUE_LEN 10
+
+/*
  * Error codes returned by the I/O sub-system.
  *
  * UBI_IO_PEB_EMPTY: the physical eraseblock is empty, i.e. it contains only
@@ -95,7 +102,8 @@ enum {
 
 /**
  * struct ubi_wl_entry - wear-leveling entry.
- * @rb: link in the corresponding RB-tree
+ * @u.rb: link in the corresponding (free/used) RB-tree
+ * @u.list: link in the protection queue
  * @ec: erase counter
  * @pnum: physical eraseblock number
  *
@@ -104,7 +112,10 @@ enum {
  * RB-trees. See WL sub-system for details.
  */
 struct ubi_wl_entry {
-        struct rb_node rb;
+        union {
+                struct rb_node rb;
+                struct list_head list;
+        } u;
         int ec;
         int pnum;
 };
@@ -288,7 +299,7 @@ struct ubi_wl_entry;
  * @beb_rsvd_level: normal level of PEBs reserved for bad PEB handling
  *
  * @autoresize_vol_id: ID of the volume which has to be auto-resized at the end
- *                     of UBI ititializetion
+ *                     of UBI initialization
  * @vtbl_slots: how many slots are available in the volume table
  * @vtbl_size: size of the volume table in bytes
  * @vtbl: in-RAM volume table copy
@@ -306,18 +317,17 @@ struct ubi_wl_entry;
  * @used: RB-tree of used physical eraseblocks
  * @free: RB-tree of free physical eraseblocks
  * @scrub: RB-tree of physical eraseblocks which need scrubbing
- * @prot: protection trees
- * @prot.pnum: protection tree indexed by physical eraseblock numbers
- * @prot.aec: protection tree indexed by absolute erase counter value
- * @wl_lock: protects the @used, @free, @prot, @lookuptbl, @abs_ec, @move_from,
- *           @move_to, @move_to_put @erase_pending, @wl_scheduled, and @works
- *           fields
+ * @pq: protection queue (contain physical eraseblocks which are temporarily
+ *      protected from the wear-leveling worker)
+ * @pq_head: protection queue head
+ * @wl_lock: protects the @used, @free, @pq, @pq_head, @lookuptbl, @move_from,
+ *           @move_to, @move_to_put @erase_pending, @wl_scheduled and @works
+ *           fields
  * @move_mutex: serializes eraseblock moves
- * @work_sem: sycnhronizes the WL worker with use tasks
+ * @work_sem: synchronizes the WL worker with use tasks
  * @wl_scheduled: non-zero if the wear-leveling was scheduled
  * @lookuptbl: a table to quickly find a &struct ubi_wl_entry object for any
  *             physical eraseblock
- * @abs_ec: absolute erase counter
  * @move_from: physical eraseblock from where the data is being moved
  * @move_to: physical eraseblock where the data is being moved to
  * @move_to_put: if the "to" PEB was put
@@ -351,11 +361,11 @@ struct ubi_wl_entry;
  *
  * @peb_buf1: a buffer of PEB size used for different purposes
  * @peb_buf2: another buffer of PEB size used for different purposes
- * @buf_mutex: proptects @peb_buf1 and @peb_buf2
+ * @buf_mutex: protects @peb_buf1 and @peb_buf2
  * @ckvol_mutex: serializes static volume checking when opening
- * @mult_mutex: serializes operations on multiple volumes, like re-nameing
+ * @mult_mutex: serializes operations on multiple volumes, like re-naming
  * @dbg_peb_buf: buffer of PEB size used for debugging
- * @dbg_buf_mutex: proptects @dbg_peb_buf
+ * @dbg_buf_mutex: protects @dbg_peb_buf
  */
 struct ubi_device {
         struct cdev cdev;
@@ -392,16 +402,13 @@ struct ubi_device {
         struct rb_root used;
         struct rb_root free;
         struct rb_root scrub;
-        struct {
-                struct rb_root pnum;
-                struct rb_root aec;
-        } prot;
+        struct list_head pq[UBI_PROT_QUEUE_LEN];
+        int pq_head;
         spinlock_t wl_lock;
         struct mutex move_mutex;
         struct rw_semaphore work_sem;
         int wl_scheduled;
         struct ubi_wl_entry **lookuptbl;
-        unsigned long long abs_ec;
         struct ubi_wl_entry *move_from;
         struct ubi_wl_entry *move_to;
         int move_to_put;
diff --git a/drivers/mtd/ubi/wl.c b/drivers/mtd/ubi/wl.c
index dcb6dac1dc54..14901cb82c18 100644
--- a/drivers/mtd/ubi/wl.c
+++ b/drivers/mtd/ubi/wl.c
@@ -22,7 +22,7 @@
  * UBI wear-leveling sub-system.
  *
  * This sub-system is responsible for wear-leveling. It works in terms of
- * physical* eraseblocks and erase counters and knows nothing about logical
+ * physical eraseblocks and erase counters and knows nothing about logical
  * eraseblocks, volumes, etc. From this sub-system's perspective all physical
  * eraseblocks are of two types - used and free. Used physical eraseblocks are
  * those that were "get" by the 'ubi_wl_get_peb()' function, and free physical
@@ -55,8 +55,39 @@
  *
  * As it was said, for the UBI sub-system all physical eraseblocks are either
  * "free" or "used". Free eraseblock are kept in the @wl->free RB-tree, while
- * used eraseblocks are kept in a set of different RB-trees: @wl->used,
- * @wl->prot.pnum, @wl->prot.aec, and @wl->scrub.
+ * used eraseblocks are kept in @wl->used or @wl->scrub RB-trees, or
+ * (temporarily) in the @wl->pq queue.
+ *
+ * When the WL sub-system returns a physical eraseblock, the physical
+ * eraseblock is protected from being moved for some "time". For this reason,
+ * the physical eraseblock is not directly moved from the @wl->free tree to the
+ * @wl->used tree. There is a protection queue in between where this
+ * physical eraseblock is temporarily stored (@wl->pq).
+ *
+ * All this protection stuff is needed because:
+ *  o we don't want to move physical eraseblocks just after we have given them
+ *    to the user; instead, we first want to let users fill them up with data;
+ *
+ *  o there is a chance that the user will put the physical eraseblock very
+ *    soon, so it makes sense not to move it for some time, but wait; this is
+ *    especially important in case of "short term" physical eraseblocks.
+ *
+ * Physical eraseblocks stay protected only for limited time. But the "time" is
+ * measured in erase cycles in this case. This is implemented with help of the
+ * protection queue. Eraseblocks are put to the tail of this queue when they
+ * are returned by the 'ubi_wl_get_peb()', and eraseblocks are removed from the
+ * head of the queue on each erase operation (for any eraseblock). So the
+ * length of the queue defines how may (global) erase cycles PEBs are protected.
+ *
+ * To put it differently, each physical eraseblock has 2 main states: free and
+ * used. The former state corresponds to the @wl->free tree. The latter state
+ * is split up on several sub-states:
+ * o the WL movement is allowed (@wl->used tree);
+ * o the WL movement is temporarily prohibited (@wl->pq queue);
+ * o scrubbing is needed (@wl->scrub tree).
+ *
+ * Depending on the sub-state, wear-leveling entries of the used physical
+ * eraseblocks may be kept in one of those structures.
  *
  * Note, in this implementation, we keep a small in-RAM object for each physical
  * eraseblock. This is surely not a scalable solution. But it appears to be good
@@ -70,9 +101,6 @@
  * target PEB, we pick a PEB with the highest EC if our PEB is "old" and we
  * pick target PEB with an average EC if our PEB is not very "old". This is a
  * room for future re-works of the WL sub-system.
- *
- * Note: the stuff with protection trees looks too complex and is difficult to
- * understand. Should be fixed.
  */
 
 #include <linux/slab.h>
@@ -85,14 +113,6 @@
 #define WL_RESERVED_PEBS 1
 
 /*
- * How many erase cycles are short term, unknown, and long term physical
- * eraseblocks protected.
- */
-#define ST_PROTECTION 16
-#define U_PROTECTION  10
-#define LT_PROTECTION 4
-
-/*
  * Maximum difference between two erase counters. If this threshold is
  * exceeded, the WL sub-system starts moving data from used physical
  * eraseblocks with low erase counter to free physical eraseblocks with high
@@ -120,64 +140,9 @@
 #define WL_MAX_FAILURES 32
 
 /**
- * struct ubi_wl_prot_entry - PEB protection entry.
- * @rb_pnum: link in the @wl->prot.pnum RB-tree
- * @rb_aec: link in the @wl->prot.aec RB-tree
- * @abs_ec: the absolute erase counter value when the protection ends
- * @e: the wear-leveling entry of the physical eraseblock under protection
- *
- * When the WL sub-system returns a physical eraseblock, the physical
- * eraseblock is protected from being moved for some "time". For this reason,
- * the physical eraseblock is not directly moved from the @wl->free tree to the
- * @wl->used tree. There is one more tree in between where this physical
- * eraseblock is temporarily stored (@wl->prot).
- *
- * All this protection stuff is needed because:
- *  o we don't want to move physical eraseblocks just after we have given them
- *    to the user; instead, we first want to let users fill them up with data;
- *
- *  o there is a chance that the user will put the physical eraseblock very
- *    soon, so it makes sense not to move it for some time, but wait; this is
- *    especially important in case of "short term" physical eraseblocks.
- *
- * Physical eraseblocks stay protected only for limited time. But the "time" is
- * measured in erase cycles in this case. This is implemented with help of the
- * absolute erase counter (@wl->abs_ec). When it reaches certain value, the
- * physical eraseblocks are moved from the protection trees (@wl->prot.*) to
- * the @wl->used tree.
- *
- * Protected physical eraseblocks are searched by physical eraseblock number
- * (when they are put) and by the absolute erase counter (to check if it is
- * time to move them to the @wl->used tree). So there are actually 2 RB-trees
- * storing the protected physical eraseblocks: @wl->prot.pnum and
- * @wl->prot.aec. They are referred to as the "protection" trees. The
- * first one is indexed by the physical eraseblock number. The second one is
- * indexed by the absolute erase counter. Both trees store
- * &struct ubi_wl_prot_entry objects.
- *
- * Each physical eraseblock has 2 main states: free and used. The former state
- * corresponds to the @wl->free tree. The latter state is split up on several
- * sub-states:
- * o the WL movement is allowed (@wl->used tree);
- * o the WL movement is temporarily prohibited (@wl->prot.pnum and
- * @wl->prot.aec trees);
- * o scrubbing is needed (@wl->scrub tree).
- *
- * Depending on the sub-state, wear-leveling entries of the used physical
- * eraseblocks may be kept in one of those trees.
- */
-struct ubi_wl_prot_entry {
-        struct rb_node rb_pnum;
-        struct rb_node rb_aec;
-        unsigned long long abs_ec;
-        struct ubi_wl_entry *e;
-};
-
-/**
  * struct ubi_work - UBI work description data structure.
  * @list: a link in the list of pending works
  * @func: worker function
- * @priv: private data of the worker function
  * @e: physical eraseblock to erase
  * @torture: if the physical eraseblock has to be tortured
  *
@@ -198,9 +163,11 @@ struct ubi_work {
 static int paranoid_check_ec(struct ubi_device *ubi, int pnum, int ec);
 static int paranoid_check_in_wl_tree(struct ubi_wl_entry *e,
                                      struct rb_root *root);
+static int paranoid_check_in_pq(struct ubi_device *ubi, struct ubi_wl_entry *e);
 #else
 #define paranoid_check_ec(ubi, pnum, ec) 0
 #define paranoid_check_in_wl_tree(e, root)
+#define paranoid_check_in_pq(ubi, e) 0
 #endif
 
 /**
@@ -220,7 +187,7 @@ static void wl_tree_add(struct ubi_wl_entry *e, struct rb_root *root)
                 struct ubi_wl_entry *e1;
 
                 parent = *p;
-                e1 = rb_entry(parent, struct ubi_wl_entry, rb);
+                e1 = rb_entry(parent, struct ubi_wl_entry, u.rb);
 
                 if (e->ec < e1->ec)
                         p = &(*p)->rb_left;
@@ -235,8 +202,8 @@ static void wl_tree_add(struct ubi_wl_entry *e, struct rb_root *root)
                 }
         }
 
-        rb_link_node(&e->rb, parent, p);
-        rb_insert_color(&e->rb, root);
+        rb_link_node(&e->u.rb, parent, p);
+        rb_insert_color(&e->u.rb, root);
 }
 
 /**
@@ -331,7 +298,7 @@ static int in_wl_tree(struct ubi_wl_entry *e, struct rb_root *root)
         while (p) {
                 struct ubi_wl_entry *e1;
 
-                e1 = rb_entry(p, struct ubi_wl_entry, rb);
+                e1 = rb_entry(p, struct ubi_wl_entry, u.rb);
 
                 if (e->pnum == e1->pnum) {
                         ubi_assert(e == e1);
@@ -355,50 +322,24 @@ static int in_wl_tree(struct ubi_wl_entry *e, struct rb_root *root)
 }
 
 /**
- * prot_tree_add - add physical eraseblock to protection trees.
+ * prot_queue_add - add physical eraseblock to the protection queue.
  * @ubi: UBI device description object
  * @e: the physical eraseblock to add
- * @pe: protection entry object to use
- * @abs_ec: absolute erase counter value when this physical eraseblock has
- * to be removed from the protection trees.
  *
- * @wl->lock has to be locked.
+ * This function adds @e to the tail of the protection queue @ubi->pq, where
+ * @e will stay for %UBI_PROT_QUEUE_LEN erase operations and will be
+ * temporarily protected from the wear-leveling worker. Note, @wl->lock has to
+ * be locked.
  */
-static void prot_tree_add(struct ubi_device *ubi, struct ubi_wl_entry *e,
-                          struct ubi_wl_prot_entry *pe, int abs_ec)
+static void prot_queue_add(struct ubi_device *ubi, struct ubi_wl_entry *e)
 {
-        struct rb_node **p, *parent = NULL;
-        struct ubi_wl_prot_entry *pe1;
-
-        pe->e = e;
-        pe->abs_ec = ubi->abs_ec + abs_ec;
-
-        p = &ubi->prot.pnum.rb_node;
-        while (*p) {
-                parent = *p;
-                pe1 = rb_entry(parent, struct ubi_wl_prot_entry, rb_pnum);
-
-                if (e->pnum < pe1->e->pnum)
-                        p = &(*p)->rb_left;
-                else
-                        p = &(*p)->rb_right;
-        }
-        rb_link_node(&pe->rb_pnum, parent, p);
-        rb_insert_color(&pe->rb_pnum, &ubi->prot.pnum);
-
-        p = &ubi->prot.aec.rb_node;
-        parent = NULL;
-        while (*p) {
-                parent = *p;
-                pe1 = rb_entry(parent, struct ubi_wl_prot_entry, rb_aec);
+        int pq_tail = ubi->pq_head - 1;
 
-                if (pe->abs_ec < pe1->abs_ec)
-                        p = &(*p)->rb_left;
-                else
-                        p = &(*p)->rb_right;
-        }
-        rb_link_node(&pe->rb_aec, parent, p);
-        rb_insert_color(&pe->rb_aec, &ubi->prot.aec);
+        if (pq_tail < 0)
+                pq_tail = UBI_PROT_QUEUE_LEN - 1;
+        ubi_assert(pq_tail >= 0 && pq_tail < UBI_PROT_QUEUE_LEN);
+        list_add_tail(&e->u.list, &ubi->pq[pq_tail]);
+        dbg_wl("added PEB %d EC %d to the protection queue", e->pnum, e->ec);
 }
 
 /**
@@ -414,14 +355,14 @@ static struct ubi_wl_entry *find_wl_entry(struct rb_root *root, int max)
         struct rb_node *p;
         struct ubi_wl_entry *e;
 
-        e = rb_entry(rb_first(root), struct ubi_wl_entry, rb);
+        e = rb_entry(rb_first(root), struct ubi_wl_entry, u.rb);
         max += e->ec;
 
         p = root->rb_node;
         while (p) {
                 struct ubi_wl_entry *e1;
 
-                e1 = rb_entry(p, struct ubi_wl_entry, rb);
+                e1 = rb_entry(p, struct ubi_wl_entry, u.rb);
                 if (e1->ec >= max)
                         p = p->rb_left;
                 else {
@@ -443,17 +384,12 @@ static struct ubi_wl_entry *find_wl_entry(struct rb_root *root, int max)
  */
 int ubi_wl_get_peb(struct ubi_device *ubi, int dtype)
 {
-        int err, protect, medium_ec;
+        int err, medium_ec;
         struct ubi_wl_entry *e, *first, *last;
-        struct ubi_wl_prot_entry *pe;
 
         ubi_assert(dtype == UBI_LONGTERM || dtype == UBI_SHORTTERM ||
                    dtype == UBI_UNKNOWN);
 
-        pe = kmalloc(sizeof(struct ubi_wl_prot_entry), GFP_NOFS);
-        if (!pe)
-                return -ENOMEM;
-
 retry:
         spin_lock(&ubi->wl_lock);
         if (!ubi->free.rb_node) {
@@ -461,16 +397,13 @@ retry:
                         ubi_assert(list_empty(&ubi->works));
                         ubi_err("no free eraseblocks");
                         spin_unlock(&ubi->wl_lock);
-                        kfree(pe);
                         return -ENOSPC;
                 }
                 spin_unlock(&ubi->wl_lock);
 
                 err = produce_free_peb(ubi);
-                if (err < 0) {
-                        kfree(pe);
+                if (err < 0)
                         return err;
-                }
                 goto retry;
         }
 
@@ -483,7 +416,6 @@ retry:
                  * %WL_FREE_MAX_DIFF.
                  */
                 e = find_wl_entry(&ubi->free, WL_FREE_MAX_DIFF);
-                protect = LT_PROTECTION;
                 break;
         case UBI_UNKNOWN:
                 /*
@@ -492,81 +424,63 @@ retry:
                  * eraseblock with erase counter greater or equivalent than the
                  * lowest erase counter plus %WL_FREE_MAX_DIFF.
                  */
-                first = rb_entry(rb_first(&ubi->free), struct ubi_wl_entry, rb);
-                last = rb_entry(rb_last(&ubi->free), struct ubi_wl_entry, rb);
+                first = rb_entry(rb_first(&ubi->free), struct ubi_wl_entry,
+                                        u.rb);
+                last = rb_entry(rb_last(&ubi->free), struct ubi_wl_entry, u.rb);
 
                 if (last->ec - first->ec < WL_FREE_MAX_DIFF)
                         e = rb_entry(ubi->free.rb_node,
-                                        struct ubi_wl_entry, rb);
+                                        struct ubi_wl_entry, u.rb);
                 else {
                         medium_ec = (first->ec + WL_FREE_MAX_DIFF)/2;
                         e = find_wl_entry(&ubi->free, medium_ec);
                 }
-                protect = U_PROTECTION;
                 break;
         case UBI_SHORTTERM:
                 /*
                  * For short term data we pick a physical eraseblock with the
                  * lowest erase counter as we expect it will be erased soon.
                  */
-                e = rb_entry(rb_first(&ubi->free), struct ubi_wl_entry, rb);
-                protect = ST_PROTECTION;
+                e = rb_entry(rb_first(&ubi->free), struct ubi_wl_entry, u.rb);
                 break;
         default:
-                protect = 0;
-                e = NULL;
                 BUG();
         }
 
+        paranoid_check_in_wl_tree(e, &ubi->free);
+
         /*
-         * Move the physical eraseblock to the protection trees where it will
+         * Move the physical eraseblock to the protection queue where it will
          * be protected from being moved for some time.
          */
-        paranoid_check_in_wl_tree(e, &ubi->free);
-        rb_erase(&e->rb, &ubi->free);
-        prot_tree_add(ubi, e, pe, protect);
-
-        dbg_wl("PEB %d EC %d, protection %d", e->pnum, e->ec, protect);
+        rb_erase(&e->u.rb, &ubi->free);
+        dbg_wl("PEB %d EC %d", e->pnum, e->ec);
+        prot_queue_add(ubi, e);
         spin_unlock(&ubi->wl_lock);
-
         return e->pnum;
 }
 
 /**
- * prot_tree_del - remove a physical eraseblock from the protection trees
+ * prot_queue_del - remove a physical eraseblock from the protection queue.
  * @ubi: UBI device description object
  * @pnum: the physical eraseblock to remove
  *
- * This function returns PEB @pnum from the protection trees and returns zero
- * in case of success and %-ENODEV if the PEB was not found in the protection
- * trees.
+ * This function deletes PEB @pnum from the protection queue and returns zero
+ * in case of success and %-ENODEV if the PEB was not found.
  */
-static int prot_tree_del(struct ubi_device *ubi, int pnum)
+static int prot_queue_del(struct ubi_device *ubi, int pnum)
 {
-        struct rb_node *p;
-        struct ubi_wl_prot_entry *pe = NULL;
-
-        p = ubi->prot.pnum.rb_node;
-        while (p) {
-
-                pe = rb_entry(p, struct ubi_wl_prot_entry, rb_pnum);
-
-                if (pnum == pe->e->pnum)
-                        goto found;
+        struct ubi_wl_entry *e;
 
-                if (pnum < pe->e->pnum)
-                        p = p->rb_left;
-                else
-                        p = p->rb_right;
-        }
+        e = ubi->lookuptbl[pnum];
+        if (!e)
+                return -ENODEV;
 
-        return -ENODEV;
+        if (paranoid_check_in_pq(ubi, e))
+                return -ENODEV;
 
-found:
-        ubi_assert(pe->e->pnum == pnum);
-        rb_erase(&pe->rb_aec, &ubi->prot.aec);
-        rb_erase(&pe->rb_pnum, &ubi->prot.pnum);
-        kfree(pe);
+        list_del(&e->u.list);
+        dbg_wl("deleted PEB %d from the protection queue", e->pnum);
         return 0;
 }
 
@@ -632,47 +546,47 @@ out_free:
 }
 
 /**
- * check_protection_over - check if it is time to stop protecting some PEBs.
+ * serve_prot_queue - check if it is time to stop protecting PEBs.
  * @ubi: UBI device description object
  *
- * This function is called after each erase operation, when the absolute erase
- * counter is incremented, to check if some physical eraseblock  have not to be
- * protected any longer. These physical eraseblocks are moved from the
- * protection trees to the used tree.
+ * This function is called after each erase operation and removes PEBs from the
+ * tail of the protection queue. These PEBs have been protected for long enough
+ * and should be moved to the used tree.
  */
-static void check_protection_over(struct ubi_device *ubi)
+static void serve_prot_queue(struct ubi_device *ubi)
 {
-        struct ubi_wl_prot_entry *pe;
+        struct ubi_wl_entry *e, *tmp;
+        int count;
 
         /*
          * There may be several protected physical eraseblock to remove,
          * process them all.
          */
-        while (1) {
-                spin_lock(&ubi->wl_lock);
-                if (!ubi->prot.aec.rb_node) {
-                        spin_unlock(&ubi->wl_lock);
-                        break;
-                }
-
-                pe = rb_entry(rb_first(&ubi->prot.aec),
-                              struct ubi_wl_prot_entry, rb_aec);
+repeat:
+        count = 0;
+        spin_lock(&ubi->wl_lock);
+        list_for_each_entry_safe(e, tmp, &ubi->pq[ubi->pq_head], u.list) {
+                dbg_wl("PEB %d EC %d protection over, move to used tree",
+                        e->pnum, e->ec);
 
-                if (pe->abs_ec > ubi->abs_ec) {
+                list_del(&e->u.list);
+                wl_tree_add(e, &ubi->used);
+                if (count++ > 32) {
+                        /*
+                         * Let's be nice and avoid holding the spinlock for
+                         * too long.
+                         */
                         spin_unlock(&ubi->wl_lock);
-                        break;
+                        cond_resched();
+                        goto repeat;
                 }
-
-                dbg_wl("PEB %d protection over, abs_ec %llu, PEB abs_ec %llu",
-                       pe->e->pnum, ubi->abs_ec, pe->abs_ec);
-                rb_erase(&pe->rb_aec, &ubi->prot.aec);
-                rb_erase(&pe->rb_pnum, &ubi->prot.pnum);
-                wl_tree_add(pe->e, &ubi->used);
-                spin_unlock(&ubi->wl_lock);
-
-                kfree(pe);
-                cond_resched();
         }
+
+        ubi->pq_head += 1;
+        if (ubi->pq_head == UBI_PROT_QUEUE_LEN)
+                ubi->pq_head = 0;
+        ubi_assert(ubi->pq_head >= 0 && ubi->pq_head < UBI_PROT_QUEUE_LEN);
+        spin_unlock(&ubi->wl_lock);
 }
 
 /**
@@ -680,8 +594,8 @@ static void check_protection_over(struct ubi_device *ubi)
  * @ubi: UBI device description object
  * @wrk: the work to schedule
  *
- * This function enqueues a work defined by @wrk to the tail of the pending
- * works list.
+ * This function adds a work defined by @wrk to the tail of the pending works
+ * list.
  */
 static void schedule_ubi_work(struct ubi_device *ubi, struct ubi_work *wrk)
 {
@@ -739,13 +653,11 @@ static int schedule_erase(struct ubi_device *ubi, struct ubi_wl_entry *e,
 static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
                                 int cancel)
 {
-        int err, put = 0, scrubbing = 0, protect = 0;
-        struct ubi_wl_prot_entry *uninitialized_var(pe);
+        int err, scrubbing = 0, torture = 0;
         struct ubi_wl_entry *e1, *e2;
         struct ubi_vid_hdr *vid_hdr;
 
         kfree(wrk);
-
         if (cancel)
                 return 0;
 
@@ -781,7 +693,7 @@ static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
                  * highly worn-out free physical eraseblock. If the erase
                  * counters differ much enough, start wear-leveling.
                  */
-                e1 = rb_entry(rb_first(&ubi->used), struct ubi_wl_entry, rb);
+                e1 = rb_entry(rb_first(&ubi->used), struct ubi_wl_entry, u.rb);
                 e2 = find_wl_entry(&ubi->free, WL_FREE_MAX_DIFF);
 
                 if (!(e2->ec - e1->ec >= UBI_WL_THRESHOLD)) {
@@ -790,21 +702,21 @@ static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
                         goto out_cancel;
                 }
                 paranoid_check_in_wl_tree(e1, &ubi->used);
-                rb_erase(&e1->rb, &ubi->used);
+                rb_erase(&e1->u.rb, &ubi->used);
                 dbg_wl("move PEB %d EC %d to PEB %d EC %d",
                        e1->pnum, e1->ec, e2->pnum, e2->ec);
         } else {
                 /* Perform scrubbing */
                 scrubbing = 1;
-                e1 = rb_entry(rb_first(&ubi->scrub), struct ubi_wl_entry, rb);
+                e1 = rb_entry(rb_first(&ubi->scrub), struct ubi_wl_entry, u.rb);
                 e2 = find_wl_entry(&ubi->free, WL_FREE_MAX_DIFF);
                 paranoid_check_in_wl_tree(e1, &ubi->scrub);
-                rb_erase(&e1->rb, &ubi->scrub);
+                rb_erase(&e1->u.rb, &ubi->scrub);
                 dbg_wl("scrub PEB %d to PEB %d", e1->pnum, e2->pnum);
         }
 
         paranoid_check_in_wl_tree(e2, &ubi->free);
-        rb_erase(&e2->rb, &ubi->free);
+        rb_erase(&e2->u.rb, &ubi->free);
         ubi->move_from = e1;
         ubi->move_to = e2;
         spin_unlock(&ubi->wl_lock);
@@ -844,46 +756,67 @@ static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
 
         err = ubi_eba_copy_leb(ubi, e1->pnum, e2->pnum, vid_hdr);
         if (err) {
-
+                if (err == -EAGAIN)
+                        goto out_not_moved;
                 if (err < 0)
                         goto out_error;
-                if (err == 1)
+                if (err == 2) {
+                        /* Target PEB write error, torture it */
+                        torture = 1;
                         goto out_not_moved;
+                }
 
                 /*
-                 * For some reason the LEB was not moved - it might be because
-                 * the volume is being deleted. We should prevent this PEB from
-                 * being selected for wear-levelling movement for some "time",
-                 * so put it to the protection tree.
+                 * The LEB has not been moved because the volume is being
+                 * deleted or the PEB has been put meanwhile. We should prevent
+                 * this PEB from being selected for wear-leveling movement
+                 * again, so put it to the protection queue.
                  */
 
-                dbg_wl("cancelled moving PEB %d", e1->pnum);
-                pe = kmalloc(sizeof(struct ubi_wl_prot_entry), GFP_NOFS);
-                if (!pe) {
-                        err = -ENOMEM;
-                        goto out_error;
-                }
+                dbg_wl("canceled moving PEB %d", e1->pnum);
+                ubi_assert(err == 1);
+
+                ubi_free_vid_hdr(ubi, vid_hdr);
+                vid_hdr = NULL;
+
+                spin_lock(&ubi->wl_lock);
+                prot_queue_add(ubi, e1);
+                ubi_assert(!ubi->move_to_put);
+                ubi->move_from = ubi->move_to = NULL;
+                ubi->wl_scheduled = 0;
+                spin_unlock(&ubi->wl_lock);
 
-                protect = 1;
+                e1 = NULL;
+                err = schedule_erase(ubi, e2, 0);
+                if (err)
+                        goto out_error;
+                mutex_unlock(&ubi->move_mutex);
+                return 0;
         }
 
+        /* The PEB has been successfully moved */
         ubi_free_vid_hdr(ubi, vid_hdr);
-        if (scrubbing && !protect)
+        vid_hdr = NULL;
+        if (scrubbing)
                 ubi_msg("scrubbed PEB %d, data moved to PEB %d",
                         e1->pnum, e2->pnum);
 
         spin_lock(&ubi->wl_lock);
-        if (protect)
-                prot_tree_add(ubi, e1, pe, protect);
-        if (!ubi->move_to_put)
+        if (!ubi->move_to_put) {
                 wl_tree_add(e2, &ubi->used);
-        else
-                put = 1;
+                e2 = NULL;
+        }
         ubi->move_from = ubi->move_to = NULL;
         ubi->move_to_put = ubi->wl_scheduled = 0;
         spin_unlock(&ubi->wl_lock);
 
-        if (put) {
+        err = schedule_erase(ubi, e1, 0);
+        if (err) {
+                e1 = NULL;
+                goto out_error;
+        }
+
+        if (e2) {
                 /*
                  * Well, the target PEB was put meanwhile, schedule it for
                  * erasure.
@@ -894,13 +827,6 @@ static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
                         goto out_error;
         }
 
-        if (!protect) {
-                err = schedule_erase(ubi, e1, 0);
-                if (err)
-                        goto out_error;
-        }
-
-
         dbg_wl("done");
         mutex_unlock(&ubi->move_mutex);
         return 0;
@@ -908,20 +834,24 @@ static int wear_leveling_worker(struct ubi_device *ubi, struct ubi_work *wrk,
         /*
          * For some reasons the LEB was not moved, might be an error, might be
          * something else. @e1 was not changed, so return it back. @e2 might
-         * be changed, schedule it for erasure.
+         * have been changed, schedule it for erasure.
          */
 out_not_moved:
+        dbg_wl("canceled moving PEB %d", e1->pnum);
         ubi_free_vid_hdr(ubi, vid_hdr);
+        vid_hdr = NULL;
         spin_lock(&ubi->wl_lock);
         if (scrubbing)
                 wl_tree_add(e1, &ubi->scrub);
         else
                 wl_tree_add(e1, &ubi->used);
+        ubi_assert(!ubi->move_to_put);
         ubi->move_from = ubi->move_to = NULL;
-        ubi->move_to_put = ubi->wl_scheduled = 0;
+        ubi->wl_scheduled = 0;
         spin_unlock(&ubi->wl_lock);
 
-        err = schedule_erase(ubi, e2, 0);
+        e1 = NULL;
+        err = schedule_erase(ubi, e2, torture);
         if (err)
                 goto out_error;
 
@@ -938,8 +868,10 @@ out_error:
         ubi->move_to_put = ubi->wl_scheduled = 0;
         spin_unlock(&ubi->wl_lock);
 
-        kmem_cache_free(ubi_wl_entry_slab, e1);
-        kmem_cache_free(ubi_wl_entry_slab, e2);
+        if (e1)
+                kmem_cache_free(ubi_wl_entry_slab, e1);
+        if (e2)
+                kmem_cache_free(ubi_wl_entry_slab, e2);
         ubi_ro_mode(ubi);
 
         mutex_unlock(&ubi->move_mutex);
@@ -988,7 +920,7 @@ static int ensure_wear_leveling(struct ubi_device *ubi)
                  * erase counter of free physical eraseblocks is greater then
                  * %UBI_WL_THRESHOLD.
                  */
-                e1 = rb_entry(rb_first(&ubi->used), struct ubi_wl_entry, rb);
+                e1 = rb_entry(rb_first(&ubi->used), struct ubi_wl_entry, u.rb);
                 e2 = find_wl_entry(&ubi->free, WL_FREE_MAX_DIFF);
 
                 if (!(e2->ec - e1->ec >= UBI_WL_THRESHOLD))
@@ -1050,7 +982,6 @@ static int erase_worker(struct ubi_device *ubi, struct ubi_work *wl_wrk,
                 kfree(wl_wrk);
 
                 spin_lock(&ubi->wl_lock);
-                ubi->abs_ec += 1;
                 wl_tree_add(e, &ubi->free);
                 spin_unlock(&ubi->wl_lock);
 
@@ -1058,7 +989,7 @@ static int erase_worker(struct ubi_device *ubi, struct ubi_work *wl_wrk,
                  * One more erase operation has happened, take care about
                  * protected physical eraseblocks.
                  */
-                check_protection_over(ubi);
+                serve_prot_queue(ubi);
 
                 /* And take care about wear-leveling */
                 err = ensure_wear_leveling(ubi);
@@ -1190,12 +1121,12 @@ retry:
         } else {
                 if (in_wl_tree(e, &ubi->used)) {
                         paranoid_check_in_wl_tree(e, &ubi->used);
-                        rb_erase(&e->rb, &ubi->used);
+                        rb_erase(&e->u.rb, &ubi->used);
                 } else if (in_wl_tree(e, &ubi->scrub)) {
                         paranoid_check_in_wl_tree(e, &ubi->scrub);
-                        rb_erase(&e->rb, &ubi->scrub);
+                        rb_erase(&e->u.rb, &ubi->scrub);
                 } else {
-                        err = prot_tree_del(ubi, e->pnum);
+                        err = prot_queue_del(ubi, e->pnum);
                         if (err) {
                                 ubi_err("PEB %d not found", pnum);
                                 ubi_ro_mode(ubi);
@@ -1255,11 +1186,11 @@ retry:
 
         if (in_wl_tree(e, &ubi->used)) {
                 paranoid_check_in_wl_tree(e, &ubi->used);
-                rb_erase(&e->rb, &ubi->used);
+                rb_erase(&e->u.rb, &ubi->used);
         } else {
                 int err;
 
-                err = prot_tree_del(ubi, e->pnum);
+                err = prot_queue_del(ubi, e->pnum);
                 if (err) {
                         ubi_err("PEB %d not found", pnum);
                         ubi_ro_mode(ubi);
@@ -1290,7 +1221,7 @@ int ubi_wl_flush(struct ubi_device *ubi)
         int err;
 
         /*
-         * Erase while the pending works queue is not empty, but not more then
+         * Erase while the pending works queue is not empty, but not more than
          * the number of currently pending works.
          */
         dbg_wl("flush (%d pending works)", ubi->works_count);
@@ -1308,7 +1239,7 @@ int ubi_wl_flush(struct ubi_device *ubi)
         up_write(&ubi->work_sem);
 
         /*
-         * And in case last was the WL worker and it cancelled the LEB
+         * And in case last was the WL worker and it canceled the LEB
          * movement, flush again.
          */
         while (ubi->works_count) {
@@ -1337,11 +1268,11 @@ static void tree_destroy(struct rb_root *root)
                 else if (rb->rb_right)
                         rb = rb->rb_right;
                 else {
-                        e = rb_entry(rb, struct ubi_wl_entry, rb);
+                        e = rb_entry(rb, struct ubi_wl_entry, u.rb);
 
                         rb = rb_parent(rb);
                         if (rb) {
-                                if (rb->rb_left == &e->rb)
+                                if (rb->rb_left == &e->u.rb)
                                         rb->rb_left = NULL;
                                 else
                                         rb->rb_right = NULL;
@@ -1436,15 +1367,13 @@ static void cancel_pending(struct ubi_device *ubi)
  */
 int ubi_wl_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si)
 {
-        int err;
+        int err, i;
         struct rb_node *rb1, *rb2;
         struct ubi_scan_volume *sv;
         struct ubi_scan_leb *seb, *tmp;
         struct ubi_wl_entry *e;
 
-
         ubi->used = ubi->free = ubi->scrub = RB_ROOT;
-        ubi->prot.pnum = ubi->prot.aec = RB_ROOT;
         spin_lock_init(&ubi->wl_lock);
         mutex_init(&ubi->move_mutex);
         init_rwsem(&ubi->work_sem);
@@ -1458,6 +1387,10 @@ int ubi_wl_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si)
         if (!ubi->lookuptbl)
                 return err;
 
+        for (i = 0; i < UBI_PROT_QUEUE_LEN; i++)
+                INIT_LIST_HEAD(&ubi->pq[i]);
+        ubi->pq_head = 0;
+
         list_for_each_entry_safe(seb, tmp, &si->erase, u.list) {
                 cond_resched();
 
@@ -1552,33 +1485,18 @@ out_free:
 }
 
 /**
- * protection_trees_destroy - destroy the protection RB-trees.
+ * protection_queue_destroy - destroy the protection queue.
  * @ubi: UBI device description object
  */
-static void protection_trees_destroy(struct ubi_device *ubi)
+static void protection_queue_destroy(struct ubi_device *ubi)
 {
-        struct rb_node *rb;
-        struct ubi_wl_prot_entry *pe;
-
-        rb = ubi->prot.aec.rb_node;
-        while (rb) {
-                if (rb->rb_left)
-                        rb = rb->rb_left;
-                else if (rb->rb_right)
-                        rb = rb->rb_right;
-                else {
-                        pe = rb_entry(rb, struct ubi_wl_prot_entry, rb_aec);
-
-                        rb = rb_parent(rb);
-                        if (rb) {
-                                if (rb->rb_left == &pe->rb_aec)
-                                        rb->rb_left = NULL;
-                                else
-                                        rb->rb_right = NULL;
-                        }
+        int i;
+        struct ubi_wl_entry *e, *tmp;
 
-                        kmem_cache_free(ubi_wl_entry_slab, pe->e);
-                        kfree(pe);
+        for (i = 0; i < UBI_PROT_QUEUE_LEN; ++i) {
+                list_for_each_entry_safe(e, tmp, &ubi->pq[i], u.list) {
+                        list_del(&e->u.list);
+                        kmem_cache_free(ubi_wl_entry_slab, e);
                 }
         }
 }
@@ -1591,7 +1509,7 @@ void ubi_wl_close(struct ubi_device *ubi)
 {
         dbg_wl("close the WL sub-system");
         cancel_pending(ubi);
-        protection_trees_destroy(ubi);
+        protection_queue_destroy(ubi);
         tree_destroy(&ubi->used);
         tree_destroy(&ubi->free);
         tree_destroy(&ubi->scrub);
@@ -1661,4 +1579,27 @@ static int paranoid_check_in_wl_tree(struct ubi_wl_entry *e,
         return 1;
 }
 
+/**
+ * paranoid_check_in_pq - check if wear-leveling entry is in the protection
+ *                        queue.
+ * @ubi: UBI device description object
+ * @e: the wear-leveling entry to check
+ *
+ * This function returns zero if @e is in @ubi->pq and %1 if it is not.
+ */
+static int paranoid_check_in_pq(struct ubi_device *ubi, struct ubi_wl_entry *e)
+{
+        struct ubi_wl_entry *p;
+        int i;
+
+        for (i = 0; i < UBI_PROT_QUEUE_LEN; ++i)
+                list_for_each_entry(p, &ubi->pq[i], u.list)
+                        if (p == e)
+                                return 0;
+
+        ubi_err("paranoid check failed for PEB %d, EC %d, Protect queue",
+                e->pnum, e->ec);
+        ubi_dbg_dump_stack();
+        return 1;
+}
 #endif /* CONFIG_MTD_UBI_DEBUG_PARANOID */