diff options
Diffstat (limited to 'drivers/mtd/ubi/eba.c')
-rw-r--r-- | drivers/mtd/ubi/eba.c | 474 |
1 files changed, 315 insertions, 159 deletions
diff --git a/drivers/mtd/ubi/eba.c b/drivers/mtd/ubi/eba.c index 7d27edaee2a..3c2a7e69e18 100644 --- a/drivers/mtd/ubi/eba.c +++ b/drivers/mtd/ubi/eba.c @@ -7,20 +7,20 @@ */ /* - * The UBI Eraseblock Association (EBA) unit. + * The UBI Eraseblock Association (EBA) sub-system. * - * This unit is responsible for I/O to/from logical eraseblock. + * This sub-system is responsible for I/O to/from logical eraseblock. * * Although in this implementation the EBA table is fully kept and managed in * RAM, which assumes poor scalability, it might be (partially) maintained on * flash in future implementations. * - * The EBA unit implements per-logical eraseblock locking. Before accessing a - * logical eraseblock it is locked for reading or writing. The per-logical - * eraseblock locking is implemented by means of the lock tree. The lock tree - * is an RB-tree which refers all the currently locked logical eraseblocks. The - * lock tree elements are &struct ubi_ltree_entry objects. They are indexed by - * (@vol_id, @lnum) pairs. + * The EBA sub-system implements per-logical eraseblock locking. Before + * accessing a logical eraseblock it is locked for reading or writing. The + * per-logical eraseblock locking is implemented by means of the lock tree. The + * lock tree is an RB-tree which refers all the currently locked logical + * eraseblocks. The lock tree elements are &struct ubi_ltree_entry objects. + * They are indexed by (@vol_id, @lnum) pairs. * * EBA also maintains the global sequence counter which is incremented each * time a logical eraseblock is mapped to a physical eraseblock and it is @@ -29,13 +29,15 @@ * 64 bits is enough to never overflow. */ -#ifdef UBI_LINUX +#define __UBOOT__ +#ifndef __UBOOT__ #include <linux/slab.h> #include <linux/crc32.h> -#include <linux/err.h> +#else +#include <ubi_uboot.h> #endif -#include <ubi_uboot.h> +#include <linux/err.h> #include "ubi.h" /* Number of physical eraseblocks reserved for atomic LEB change operation */ @@ -49,7 +51,7 @@ * global sequence counter value. It also increases the global sequence * counter. */ -static unsigned long long next_sqnum(struct ubi_device *ubi) +unsigned long long ubi_next_sqnum(struct ubi_device *ubi) { unsigned long long sqnum; @@ -181,9 +183,7 @@ static struct ubi_ltree_entry *ltree_add_entry(struct ubi_device *ubi, le->users += 1; spin_unlock(&ubi->ltree_lock); - if (le_free) - kfree(le_free); - + kfree(le_free); return le; } @@ -215,22 +215,18 @@ static int leb_read_lock(struct ubi_device *ubi, int vol_id, int lnum) */ static void leb_read_unlock(struct ubi_device *ubi, int vol_id, int lnum) { - int _free = 0; struct ubi_ltree_entry *le; spin_lock(&ubi->ltree_lock); le = ltree_lookup(ubi, vol_id, lnum); le->users -= 1; ubi_assert(le->users >= 0); + up_read(&le->mutex); if (le->users == 0) { rb_erase(&le->rb, &ubi->ltree); - _free = 1; + kfree(le); } spin_unlock(&ubi->ltree_lock); - - up_read(&le->mutex); - if (_free) - kfree(le); } /** @@ -266,7 +262,6 @@ static int leb_write_lock(struct ubi_device *ubi, int vol_id, int lnum) */ static int leb_write_trylock(struct ubi_device *ubi, int vol_id, int lnum) { - int _free; struct ubi_ltree_entry *le; le = ltree_add_entry(ubi, vol_id, lnum); @@ -281,12 +276,9 @@ static int leb_write_trylock(struct ubi_device *ubi, int vol_id, int lnum) ubi_assert(le->users >= 0); if (le->users == 0) { rb_erase(&le->rb, &ubi->ltree); - _free = 1; - } else - _free = 0; - spin_unlock(&ubi->ltree_lock); - if (_free) kfree(le); + } + spin_unlock(&ubi->ltree_lock); return 1; } @@ -299,23 +291,18 @@ static int leb_write_trylock(struct ubi_device *ubi, int vol_id, int lnum) */ static void leb_write_unlock(struct ubi_device *ubi, int vol_id, int lnum) { - int _free; struct ubi_ltree_entry *le; spin_lock(&ubi->ltree_lock); le = ltree_lookup(ubi, vol_id, lnum); le->users -= 1; ubi_assert(le->users >= 0); + up_write(&le->mutex); if (le->users == 0) { rb_erase(&le->rb, &ubi->ltree); - _free = 1; - } else - _free = 0; - spin_unlock(&ubi->ltree_lock); - - up_write(&le->mutex); - if (_free) kfree(le); + } + spin_unlock(&ubi->ltree_lock); } /** @@ -347,8 +334,10 @@ int ubi_eba_unmap_leb(struct ubi_device *ubi, struct ubi_volume *vol, dbg_eba("erase LEB %d:%d, PEB %d", vol_id, lnum, pnum); + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = UBI_LEB_UNMAPPED; - err = ubi_wl_put_peb(ubi, pnum, 0); + up_read(&ubi->fm_sem); + err = ubi_wl_put_peb(ubi, vol_id, lnum, pnum, 0); out_unlock: leb_write_unlock(ubi, vol_id, lnum); @@ -425,9 +414,10 @@ retry: * may try to recover data. FIXME: but this is * not implemented. */ - if (err == UBI_IO_BAD_VID_HDR) { - ubi_warn("bad VID header at PEB %d, LEB" - "%d:%d", pnum, vol_id, lnum); + if (err == UBI_IO_BAD_HDR_EBADMSG || + err == UBI_IO_BAD_HDR) { + ubi_warn("corrupted VID header at PEB %d, LEB %d:%d", + pnum, vol_id, lnum); err = -EBADMSG; } else ubi_ro_mode(ubi); @@ -508,16 +498,12 @@ static int recover_peb(struct ubi_device *ubi, int pnum, int vol_id, int lnum, struct ubi_vid_hdr *vid_hdr; vid_hdr = ubi_zalloc_vid_hdr(ubi, GFP_NOFS); - if (!vid_hdr) { + if (!vid_hdr) return -ENOMEM; - } - - mutex_lock(&ubi->buf_mutex); retry: - new_pnum = ubi_wl_get_peb(ubi, UBI_UNKNOWN); + new_pnum = ubi_wl_get_peb(ubi); if (new_pnum < 0) { - mutex_unlock(&ubi->buf_mutex); ubi_free_vid_hdr(ubi, vid_hdr); return new_pnum; } @@ -531,39 +517,45 @@ retry: goto out_put; } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); err = ubi_io_write_vid_hdr(ubi, new_pnum, vid_hdr); if (err) goto write_error; data_size = offset + len; - memset(ubi->peb_buf1 + offset, 0xFF, len); + mutex_lock(&ubi->buf_mutex); + memset(ubi->peb_buf + 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); + err = ubi_io_read_data(ubi, ubi->peb_buf, pnum, 0, offset); if (err && err != UBI_IO_BITFLIPS) - goto out_put; + goto out_unlock; } - memcpy(ubi->peb_buf1 + offset, buf, len); + memcpy(ubi->peb_buf + offset, buf, len); - err = ubi_io_write_data(ubi, ubi->peb_buf1, new_pnum, 0, data_size); - if (err) + err = ubi_io_write_data(ubi, ubi->peb_buf, new_pnum, 0, data_size); + if (err) { + mutex_unlock(&ubi->buf_mutex); goto write_error; + } mutex_unlock(&ubi->buf_mutex); ubi_free_vid_hdr(ubi, vid_hdr); + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = new_pnum; - ubi_wl_put_peb(ubi, pnum, 1); + up_read(&ubi->fm_sem); + ubi_wl_put_peb(ubi, vol_id, lnum, pnum, 1); ubi_msg("data was successfully recovered"); return 0; -out_put: +out_unlock: mutex_unlock(&ubi->buf_mutex); - ubi_wl_put_peb(ubi, new_pnum, 1); +out_put: + ubi_wl_put_peb(ubi, vol_id, lnum, new_pnum, 1); ubi_free_vid_hdr(ubi, vid_hdr); return err; @@ -573,9 +565,8 @@ write_error: * get another one. */ ubi_warn("failed to write to PEB %d", new_pnum); - ubi_wl_put_peb(ubi, new_pnum, 1); + ubi_wl_put_peb(ubi, vol_id, lnum, new_pnum, 1); if (++tries > UBI_IO_RETRIES) { - mutex_unlock(&ubi->buf_mutex); ubi_free_vid_hdr(ubi, vid_hdr); return err; } @@ -591,7 +582,6 @@ write_error: * @buf: the data to write * @offset: offset within the logical eraseblock where to write * @len: how many bytes to write - * @dtype: data type * * This function writes data to logical eraseblock @lnum of a dynamic volume * @vol. Returns zero in case of success and a negative error code in case @@ -599,7 +589,7 @@ write_error: * written to the flash media, but may be some garbage. */ int ubi_eba_write_leb(struct ubi_device *ubi, struct ubi_volume *vol, int lnum, - const void *buf, int offset, int len, int dtype) + const void *buf, int offset, int len) { int err, pnum, tries = 0, vol_id = vol->vol_id; struct ubi_vid_hdr *vid_hdr; @@ -640,14 +630,14 @@ int ubi_eba_write_leb(struct ubi_device *ubi, struct ubi_volume *vol, int lnum, } vid_hdr->vol_type = UBI_VID_DYNAMIC; - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); vid_hdr->vol_id = cpu_to_be32(vol_id); vid_hdr->lnum = cpu_to_be32(lnum); vid_hdr->compat = ubi_get_compat(ubi, vol_id); vid_hdr->data_pad = cpu_to_be32(vol->data_pad); retry: - pnum = ubi_wl_get_peb(ubi, dtype); + pnum = ubi_wl_get_peb(ubi); if (pnum < 0) { ubi_free_vid_hdr(ubi, vid_hdr); leb_write_unlock(ubi, vol_id, lnum); @@ -667,14 +657,15 @@ retry: if (len) { err = ubi_io_write_data(ubi, buf, pnum, offset, len); if (err) { - ubi_warn("failed to write %d bytes at offset %d of " - "LEB %d:%d, PEB %d", len, offset, vol_id, - lnum, pnum); + ubi_warn("failed to write %d bytes at offset %d of LEB %d:%d, PEB %d", + len, offset, vol_id, lnum, pnum); goto write_error; } } + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = pnum; + up_read(&ubi->fm_sem); leb_write_unlock(ubi, vol_id, lnum); ubi_free_vid_hdr(ubi, vid_hdr); @@ -693,7 +684,7 @@ write_error: * eraseblock, so just put it and request a new one. We assume that if * this physical eraseblock went bad, the erase code will handle that. */ - err = ubi_wl_put_peb(ubi, pnum, 1); + err = ubi_wl_put_peb(ubi, vol_id, lnum, pnum, 1); if (err || ++tries > UBI_IO_RETRIES) { ubi_ro_mode(ubi); leb_write_unlock(ubi, vol_id, lnum); @@ -701,7 +692,7 @@ write_error: return err; } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); ubi_msg("try another PEB"); goto retry; } @@ -713,7 +704,6 @@ write_error: * @lnum: logical eraseblock number * @buf: data to write * @len: how many bytes to write - * @dtype: data type * @used_ebs: how many logical eraseblocks will this volume contain * * This function writes data to logical eraseblock @lnum of static volume @@ -725,13 +715,12 @@ write_error: * to the real data size, although the @buf buffer has to contain the * alignment. In all other cases, @len has to be aligned. * - * It is prohibited to write more then once to logical eraseblocks of static + * It is prohibited to write more than once to logical eraseblocks of static * volumes. This function returns zero in case of success and a negative error * code in case of failure. */ int ubi_eba_write_leb_st(struct ubi_device *ubi, struct ubi_volume *vol, - int lnum, const void *buf, int len, int dtype, - int used_ebs) + int lnum, const void *buf, int len, int used_ebs) { int err, pnum, tries = 0, data_size = len, vol_id = vol->vol_id; struct ubi_vid_hdr *vid_hdr; @@ -756,7 +745,7 @@ int ubi_eba_write_leb_st(struct ubi_device *ubi, struct ubi_volume *vol, return err; } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); vid_hdr->vol_id = cpu_to_be32(vol_id); vid_hdr->lnum = cpu_to_be32(lnum); vid_hdr->compat = ubi_get_compat(ubi, vol_id); @@ -769,7 +758,7 @@ int ubi_eba_write_leb_st(struct ubi_device *ubi, struct ubi_volume *vol, vid_hdr->data_crc = cpu_to_be32(crc); retry: - pnum = ubi_wl_get_peb(ubi, dtype); + pnum = ubi_wl_get_peb(ubi); if (pnum < 0) { ubi_free_vid_hdr(ubi, vid_hdr); leb_write_unlock(ubi, vol_id, lnum); @@ -794,7 +783,9 @@ retry: } ubi_assert(vol->eba_tbl[lnum] < 0); + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = pnum; + up_read(&ubi->fm_sem); leb_write_unlock(ubi, vol_id, lnum); ubi_free_vid_hdr(ubi, vid_hdr); @@ -813,7 +804,7 @@ write_error: return err; } - err = ubi_wl_put_peb(ubi, pnum, 1); + err = ubi_wl_put_peb(ubi, vol_id, lnum, pnum, 1); if (err || ++tries > UBI_IO_RETRIES) { ubi_ro_mode(ubi); leb_write_unlock(ubi, vol_id, lnum); @@ -821,7 +812,7 @@ write_error: return err; } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); ubi_msg("try another PEB"); goto retry; } @@ -833,7 +824,6 @@ write_error: * @lnum: logical eraseblock number * @buf: data to write * @len: how many bytes to write - * @dtype: data type * * This function changes the contents of a logical eraseblock atomically. @buf * has to contain new logical eraseblock data, and @len - the length of the @@ -845,7 +835,7 @@ write_error: * LEB change may be done at a time. This is ensured by @ubi->alc_mutex. */ int ubi_eba_atomic_leb_change(struct ubi_device *ubi, struct ubi_volume *vol, - int lnum, const void *buf, int len, int dtype) + int lnum, const void *buf, int len) { int err, pnum, tries = 0, vol_id = vol->vol_id; struct ubi_vid_hdr *vid_hdr; @@ -862,7 +852,7 @@ int ubi_eba_atomic_leb_change(struct ubi_device *ubi, struct ubi_volume *vol, err = ubi_eba_unmap_leb(ubi, vol, lnum); if (err) return err; - return ubi_eba_write_leb(ubi, vol, lnum, NULL, 0, 0, dtype); + return ubi_eba_write_leb(ubi, vol, lnum, NULL, 0, 0); } vid_hdr = ubi_zalloc_vid_hdr(ubi, GFP_NOFS); @@ -874,7 +864,7 @@ int ubi_eba_atomic_leb_change(struct ubi_device *ubi, struct ubi_volume *vol, if (err) goto out_mutex; - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); vid_hdr->vol_id = cpu_to_be32(vol_id); vid_hdr->lnum = cpu_to_be32(lnum); vid_hdr->compat = ubi_get_compat(ubi, vol_id); @@ -887,7 +877,7 @@ int ubi_eba_atomic_leb_change(struct ubi_device *ubi, struct ubi_volume *vol, vid_hdr->data_crc = cpu_to_be32(crc); retry: - pnum = ubi_wl_get_peb(ubi, dtype); + pnum = ubi_wl_get_peb(ubi); if (pnum < 0) { err = pnum; goto out_leb_unlock; @@ -911,12 +901,14 @@ retry: } if (vol->eba_tbl[lnum] >= 0) { - err = ubi_wl_put_peb(ubi, vol->eba_tbl[lnum], 1); + err = ubi_wl_put_peb(ubi, vol_id, lnum, vol->eba_tbl[lnum], 0); if (err) goto out_leb_unlock; } + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = pnum; + up_read(&ubi->fm_sem); out_leb_unlock: leb_write_unlock(ubi, vol_id, lnum); @@ -936,18 +928,45 @@ write_error: goto out_leb_unlock; } - err = ubi_wl_put_peb(ubi, pnum, 1); + err = ubi_wl_put_peb(ubi, vol_id, lnum, pnum, 1); if (err || ++tries > UBI_IO_RETRIES) { ubi_ro_mode(ubi); goto out_leb_unlock; } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); ubi_msg("try another PEB"); goto retry; } /** + * is_error_sane - check whether a read error is sane. + * @err: code of the error happened during reading + * + * This is a helper function for 'ubi_eba_copy_leb()' which is called when we + * cannot read data from the target PEB (an error @err happened). If the error + * code is sane, then we treat this error as non-fatal. Otherwise the error is + * fatal and UBI will be switched to R/O mode later. + * + * The idea is that we try not to switch to R/O mode if the read error is + * something which suggests there was a real read problem. E.g., %-EIO. Or a + * memory allocation failed (-%ENOMEM). Otherwise, it is safer to switch to R/O + * mode, simply because we do not know what happened at the MTD level, and we + * cannot handle this. E.g., the underlying driver may have become crazy, and + * it is safer to switch to R/O mode to preserve the data. + * + * And bear in mind, this is about reading from the target PEB, i.e. the PEB + * which we have just written. + */ +static int is_error_sane(int err) +{ + if (err == -EIO || err == -ENOMEM || err == UBI_IO_BAD_HDR || + err == UBI_IO_BAD_HDR_EBADMSG || err == -ETIMEDOUT) + return 0; + return 1; +} + +/** * ubi_eba_copy_leb - copy logical eraseblock. * @ubi: UBI device description object * @from: physical eraseblock number from where to copy @@ -957,10 +976,9 @@ 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 %MOVE_CANCEL_RACE, %MOVE_TARGET_WR_ERR, %MOVE_TARGET_BITFLIPS, etc; + * 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) @@ -972,7 +990,7 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, vol_id = be32_to_cpu(vid_hdr->vol_id); lnum = be32_to_cpu(vid_hdr->lnum); - dbg_eba("copy LEB %d:%d, PEB %d to PEB %d", vol_id, lnum, from, to); + dbg_wl("copy LEB %d:%d, PEB %d to PEB %d", vol_id, lnum, from, to); if (vid_hdr->vol_type == UBI_VID_STATIC) { data_size = be32_to_cpu(vid_hdr->data_size); @@ -986,17 +1004,16 @@ 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]; + spin_unlock(&ubi->volumes_lock); if (!vol) { /* No need to do further work, cancel */ - dbg_eba("volume %d is being removed, cancel", vol_id); - spin_unlock(&ubi->volumes_lock); - return 2; + dbg_wl("volume %d is being removed, cancel", vol_id); + return MOVE_CANCEL_RACE; } - spin_unlock(&ubi->volumes_lock); /* * We do not want anybody to write to this logical eraseblock while we @@ -1008,12 +1025,15 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, * (@from). This task locks the LEB and goes sleep in the * 'ubi_wl_put_peb()' function on the @ubi->move_mutex. In turn, we are * holding @ubi->move_mutex and go sleep on the LEB lock. So, if the - * LEB is already locked, we just do not move it and return %1. + * LEB is already locked, we just do not move it and return + * %MOVE_RETRY. Note, we do not return %MOVE_CANCEL_RACE here because + * we do not know the reasons of the contention - it may be just a + * normal I/O on this LEB, so we want to re-try. */ err = leb_write_trylock(ubi, vol_id, lnum); if (err) { - dbg_eba("contention on LEB %d:%d, cancel", vol_id, lnum); - return err; + dbg_wl("contention on LEB %d:%d, cancel", vol_id, lnum); + return MOVE_RETRY; } /* @@ -1022,30 +1042,30 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, * cancel it. */ if (vol->eba_tbl[lnum] != from) { - dbg_eba("LEB %d:%d is no longer mapped to PEB %d, mapped to " - "PEB %d, cancel", vol_id, lnum, from, - vol->eba_tbl[lnum]); - err = 1; + dbg_wl("LEB %d:%d is no longer mapped to PEB %d, mapped to PEB %d, cancel", + vol_id, lnum, from, vol->eba_tbl[lnum]); + err = MOVE_CANCEL_RACE; goto out_unlock_leb; } /* - * OK, now the LEB is locked and we can safely start moving iy. Since - * this function utilizes thie @ubi->peb1_buf buffer which is shared - * with some other functions, so lock the buffer by taking the + * OK, now the LEB is locked and we can safely start moving it. Since + * this function utilizes the @ubi->peb_buf buffer which is shared + * with some other functions - we lock the buffer by taking the * @ubi->buf_mutex. */ mutex_lock(&ubi->buf_mutex); - dbg_eba("read %d bytes of data", aldata_size); - err = ubi_io_read_data(ubi, ubi->peb_buf1, from, 0, aldata_size); + dbg_wl("read %d bytes of data", aldata_size); + err = ubi_io_read_data(ubi, ubi->peb_buf, from, 0, aldata_size); if (err && err != UBI_IO_BITFLIPS) { ubi_warn("error %d while reading data from PEB %d", err, from); + err = MOVE_SOURCE_RD_ERR; goto out_unlock_buf; } /* - * Now we have got to calculate how much data we have to to copy. In + * Now we have got to calculate how much data we have to copy. In * case of a static volume it is fairly easy - the VID header contains * the data size. In case of a dynamic volume it is more difficult - we * have to read the contents, cut 0xFF bytes from the end and copy only @@ -1056,14 +1076,14 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, */ if (vid_hdr->vol_type == UBI_VID_DYNAMIC) aldata_size = data_size = - ubi_calc_data_len(ubi, ubi->peb_buf1, data_size); + ubi_calc_data_len(ubi, ubi->peb_buf, data_size); cond_resched(); - crc = crc32(UBI_CRC32_INIT, ubi->peb_buf1, data_size); + crc = crc32(UBI_CRC32_INIT, ubi->peb_buf, data_size); cond_resched(); /* - * It may turn out to me that the whole @from physical eraseblock + * It may turn out to be that the whole @from physical eraseblock * contains only 0xFF bytes. Then we have to only write the VID header * and do not write any data. This also means we should not set * @vid_hdr->copy_flag, @vid_hdr->data_size, and @vid_hdr->data_crc. @@ -1073,28 +1093,37 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, vid_hdr->data_size = cpu_to_be32(data_size); vid_hdr->data_crc = cpu_to_be32(crc); } - vid_hdr->sqnum = cpu_to_be64(next_sqnum(ubi)); + vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi)); err = ubi_io_write_vid_hdr(ubi, to, vid_hdr); - if (err) + if (err) { + if (err == -EIO) + err = MOVE_TARGET_WR_ERR; goto out_unlock_buf; + } cond_resched(); /* Read the VID header back and check if it was written correctly */ err = ubi_io_read_vid_hdr(ubi, to, vid_hdr, 1); if (err) { - if (err != UBI_IO_BITFLIPS) - ubi_warn("cannot read VID header back from PEB %d", to); - else - err = 1; + if (err != UBI_IO_BITFLIPS) { + ubi_warn("error %d while reading VID header back from PEB %d", + err, to); + if (is_error_sane(err)) + err = MOVE_TARGET_RD_ERR; + } else + err = MOVE_TARGET_BITFLIPS; goto out_unlock_buf; } if (data_size > 0) { - err = ubi_io_write_data(ubi, ubi->peb_buf1, to, 0, aldata_size); - if (err) + err = ubi_io_write_data(ubi, ubi->peb_buf, to, 0, aldata_size); + if (err) { + if (err == -EIO) + err = MOVE_TARGET_WR_ERR; goto out_unlock_buf; + } cond_resched(); @@ -1102,28 +1131,33 @@ int ubi_eba_copy_leb(struct ubi_device *ubi, int from, int to, * We've written the data and are going to read it back to make * sure it was written correctly. */ - - err = ubi_io_read_data(ubi, ubi->peb_buf2, to, 0, aldata_size); + memset(ubi->peb_buf, 0xFF, aldata_size); + err = ubi_io_read_data(ubi, ubi->peb_buf, to, 0, aldata_size); if (err) { - if (err != UBI_IO_BITFLIPS) - ubi_warn("cannot read data back from PEB %d", - to); - else - err = 1; + if (err != UBI_IO_BITFLIPS) { + ubi_warn("error %d while reading data back from PEB %d", + err, to); + if (is_error_sane(err)) + err = MOVE_TARGET_RD_ERR; + } else + err = MOVE_TARGET_BITFLIPS; 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", + if (crc != crc32(UBI_CRC32_INIT, ubi->peb_buf, data_size)) { + ubi_warn("read data back from PEB %d and it is different", to); + err = -EINVAL; goto out_unlock_buf; } } ubi_assert(vol->eba_tbl[lnum] == from); + down_read(&ubi->fm_sem); vol->eba_tbl[lnum] = to; + up_read(&ubi->fm_sem); out_unlock_buf: mutex_unlock(&ubi->buf_mutex); @@ -1133,28 +1167,165 @@ out_unlock_leb: } /** - * ubi_eba_init_scan - initialize the EBA unit using scanning information. + * print_rsvd_warning - warn about not having enough reserved PEBs. * @ubi: UBI device description object - * @si: scanning information + * + * This is a helper function for 'ubi_eba_init()' which is called when UBI + * cannot reserve enough PEBs for bad block handling. This function makes a + * decision whether we have to print a warning or not. The algorithm is as + * follows: + * o if this is a new UBI image, then just print the warning + * o if this is an UBI image which has already been used for some time, print + * a warning only if we can reserve less than 10% of the expected amount of + * the reserved PEB. + * + * The idea is that when UBI is used, PEBs become bad, and the reserved pool + * of PEBs becomes smaller, which is normal and we do not want to scare users + * with a warning every time they attach the MTD device. This was an issue + * reported by real users. + */ +static void print_rsvd_warning(struct ubi_device *ubi, + struct ubi_attach_info *ai) +{ + /* + * The 1 << 18 (256KiB) number is picked randomly, just a reasonably + * large number to distinguish between newly flashed and used images. + */ + if (ai->max_sqnum > (1 << 18)) { + int min = ubi->beb_rsvd_level / 10; + + if (!min) + min = 1; + if (ubi->beb_rsvd_pebs > min) + return; + } + + ubi_warn("cannot reserve enough PEBs for bad PEB handling, reserved %d, need %d", + ubi->beb_rsvd_pebs, ubi->beb_rsvd_level); + if (ubi->corr_peb_count) + ubi_warn("%d PEBs are corrupted and not used", + ubi->corr_peb_count); +} + +/** + * self_check_eba - run a self check on the EBA table constructed by fastmap. + * @ubi: UBI device description object + * @ai_fastmap: UBI attach info object created by fastmap + * @ai_scan: UBI attach info object created by scanning + * + * Returns < 0 in case of an internal error, 0 otherwise. + * If a bad EBA table entry was found it will be printed out and + * ubi_assert() triggers. + */ +int self_check_eba(struct ubi_device *ubi, struct ubi_attach_info *ai_fastmap, + struct ubi_attach_info *ai_scan) +{ + int i, j, num_volumes, ret = 0; + int **scan_eba, **fm_eba; + struct ubi_ainf_volume *av; + struct ubi_volume *vol; + struct ubi_ainf_peb *aeb; + struct rb_node *rb; + + num_volumes = ubi->vtbl_slots + UBI_INT_VOL_COUNT; + + scan_eba = kmalloc(sizeof(*scan_eba) * num_volumes, GFP_KERNEL); + if (!scan_eba) + return -ENOMEM; + + fm_eba = kmalloc(sizeof(*fm_eba) * num_volumes, GFP_KERNEL); + if (!fm_eba) { + kfree(scan_eba); + return -ENOMEM; + } + + for (i = 0; i < num_volumes; i++) { + vol = ubi->volumes[i]; + if (!vol) + continue; + + scan_eba[i] = kmalloc(vol->reserved_pebs * sizeof(**scan_eba), + GFP_KERNEL); + if (!scan_eba[i]) { + ret = -ENOMEM; + goto out_free; + } + + fm_eba[i] = kmalloc(vol->reserved_pebs * sizeof(**fm_eba), + GFP_KERNEL); + if (!fm_eba[i]) { + ret = -ENOMEM; + goto out_free; + } + + for (j = 0; j < vol->reserved_pebs; j++) + scan_eba[i][j] = fm_eba[i][j] = UBI_LEB_UNMAPPED; + + av = ubi_find_av(ai_scan, idx2vol_id(ubi, i)); + if (!av) + continue; + + ubi_rb_for_each_entry(rb, aeb, &av->root, u.rb) + scan_eba[i][aeb->lnum] = aeb->pnum; + + av = ubi_find_av(ai_fastmap, idx2vol_id(ubi, i)); + if (!av) + continue; + + ubi_rb_for_each_entry(rb, aeb, &av->root, u.rb) + fm_eba[i][aeb->lnum] = aeb->pnum; + + for (j = 0; j < vol->reserved_pebs; j++) { + if (scan_eba[i][j] != fm_eba[i][j]) { + if (scan_eba[i][j] == UBI_LEB_UNMAPPED || + fm_eba[i][j] == UBI_LEB_UNMAPPED) + continue; + + ubi_err("LEB:%i:%i is PEB:%i instead of %i!", + vol->vol_id, i, fm_eba[i][j], + scan_eba[i][j]); + ubi_assert(0); + } + } + } + +out_free: + for (i = 0; i < num_volumes; i++) { + if (!ubi->volumes[i]) + continue; + + kfree(scan_eba[i]); + kfree(fm_eba[i]); + } + + kfree(scan_eba); + kfree(fm_eba); + return ret; +} + +/** + * ubi_eba_init - initialize the EBA sub-system using attaching information. + * @ubi: UBI device description object + * @ai: attaching information * * This function returns zero in case of success and a negative error code in * case of failure. */ -int ubi_eba_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si) +int ubi_eba_init(struct ubi_device *ubi, struct ubi_attach_info *ai) { int i, j, err, num_volumes; - struct ubi_scan_volume *sv; + struct ubi_ainf_volume *av; struct ubi_volume *vol; - struct ubi_scan_leb *seb; + struct ubi_ainf_peb *aeb; struct rb_node *rb; - dbg_eba("initialize EBA unit"); + dbg_eba("initialize EBA sub-system"); spin_lock_init(&ubi->ltree_lock); mutex_init(&ubi->alc_mutex); ubi->ltree = RB_ROOT; - ubi->global_sqnum = si->max_sqnum + 1; + ubi->global_sqnum = ai->max_sqnum + 1; num_volumes = ubi->vtbl_slots + UBI_INT_VOL_COUNT; for (i = 0; i < num_volumes; i++) { @@ -1174,24 +1345,27 @@ int ubi_eba_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si) for (j = 0; j < vol->reserved_pebs; j++) vol->eba_tbl[j] = UBI_LEB_UNMAPPED; - sv = ubi_scan_find_sv(si, idx2vol_id(ubi, i)); - if (!sv) + av = ubi_find_av(ai, idx2vol_id(ubi, i)); + if (!av) continue; - ubi_rb_for_each_entry(rb, seb, &sv->root, u.rb) { - if (seb->lnum >= vol->reserved_pebs) + ubi_rb_for_each_entry(rb, aeb, &av->root, u.rb) { + if (aeb->lnum >= vol->reserved_pebs) /* * This may happen in case of an unclean reboot * during re-size. */ - ubi_scan_move_to_list(sv, seb, &si->erase); - vol->eba_tbl[seb->lnum] = seb->pnum; + ubi_move_aeb_to_list(av, aeb, &ai->erase); + vol->eba_tbl[aeb->lnum] = aeb->pnum; } } if (ubi->avail_pebs < EBA_RESERVED_PEBS) { ubi_err("no enough physical eraseblocks (%d, need %d)", ubi->avail_pebs, EBA_RESERVED_PEBS); + if (ubi->corr_peb_count) + ubi_err("%d PEBs are corrupted and not used", + ubi->corr_peb_count); err = -ENOSPC; goto out_free; } @@ -1204,9 +1378,7 @@ int ubi_eba_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si) if (ubi->avail_pebs < ubi->beb_rsvd_level) { /* No enough free physical eraseblocks */ ubi->beb_rsvd_pebs = ubi->avail_pebs; - ubi_warn("cannot reserve enough PEBs for bad PEB " - "handling, reserved %d, need %d", - ubi->beb_rsvd_pebs, ubi->beb_rsvd_level); + print_rsvd_warning(ubi, ai); } else ubi->beb_rsvd_pebs = ubi->beb_rsvd_level; @@ -1214,7 +1386,7 @@ int ubi_eba_init_scan(struct ubi_device *ubi, struct ubi_scan_info *si) ubi->rsvd_pebs += ubi->beb_rsvd_pebs; } - dbg_eba("EBA unit is initialized"); + dbg_eba("EBA sub-system is initialized"); return 0; out_free: @@ -1222,23 +1394,7 @@ out_free: if (!ubi->volumes[i]) continue; kfree(ubi->volumes[i]->eba_tbl); + ubi->volumes[i]->eba_tbl = NULL; } return err; } - -/** - * ubi_eba_close - close EBA unit. - * @ubi: UBI device description object - */ -void ubi_eba_close(const struct ubi_device *ubi) -{ - int i, num_volumes = ubi->vtbl_slots + UBI_INT_VOL_COUNT; - - dbg_eba("close EBA unit"); - - for (i = 0; i < num_volumes; i++) { - if (!ubi->volumes[i]) - continue; - kfree(ubi->volumes[i]->eba_tbl); - } -} |