mtd: nand: misc BBT fixes
Subsystems (2)
cover mtd: nand: misc BBT fixes
A few assorted changes. Some of these are geared toward refactoring the BBT code for being reused for SPI NAND.
If possible, I'd really like to get some testing for the diskonchip changes. It's an old driver, but it appears as if some still use it. Perhaps Alexander (CC'd) can report?
Brian Norris (5):
mtd: nand_bbt: drop unnecessary header
mtd: diskonchip: don't call nand_scan_bbt() directly
mtd: nand_bbt: make nand_scan_bbt() static
mtd: nand_bbt: unify/fix error handling in nand_scan_bbt()
mtd: nand_bbt: fix theoretical integer overflow in BBT write
drivers/mtd/nand/diskonchip.c | 27 ++++++++++++++++----------- drivers/mtd/nand/nand_bbt.c | 24 ++++++++++++++---------- include/linux/mtd/nand.h | 1 - 3 files changed, 30 insertions(+), 22 deletions(-)
Signature · 1 line
2.1.0
PATCH 1/5 mtd: nand_bbt: drop unnecessary header
--- drivers/mtd/nand/nand_bbt.c | 1 - 1 file changed, 1 deletion(-)
drivers/mtd/nand/nand_bbt.c +0 -1
@@ -64,7 +64,6 @@#include <linux/mtd/mtd.h> #include <linux/mtd/bbm.h> #include <linux/mtd/nand.h> -#include <linux/mtd/nand_ecc.h> #include <linux/bitops.h> #include <linux/delay.h> #include <linux/vmalloc.h>
Signature · 1 line
2.1.0
PATCH 2/5 mtd: diskonchip: don't call nand_scan_bbt() directly
In reply to Brian NorrisThe diskonchip driver almost uses the default nand_base hooks as-is, except that it provides custom on-flash BBT descriptors and avoids using factory-marked bad blockers.
So let's refactor the BBT initialization code into a private 'late_init' hook which handles all the private details. Note the usage of NAND_SKIP_BBTSCAN, which allows us to defer the BBT scan until we've prepared everything.
--- drivers/mtd/nand/diskonchip.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-)
drivers/mtd/nand/diskonchip.c +16 -11
@@ -69,6 +69,9 @@ struct doc_priv {int mh0_page;int mh1_page;struct mtd_info *nextdoc; + +/* Handle the last stage of initialization (BBT scan, partitioning) */ +int (*late_init)(struct mtd_info *mtd); };/* This is the syndrome computed by the HW ecc generator upon reading an empty
@@ -1294,10 +1297,10 @@ static int __init nftl_scan_bbt(struct mtd_info *mtd) this->bbt_md =NULL ; } -/* It's safe to set bd=NULL below because NAND_BBT_CREATE is not set. -At least as nand_bbt.c is currently written. */ -if ((ret =nand_scan_bbt (mtd,NULL ))) + ret = this->scan_bbt (mtd); +if (ret)return ret; +mtd_device_register (mtd,NULL ,0 );if (!no_autopart)mtd_device_register (mtd, parts, numparts);
@@ -1344,10 +1347,10 @@ static int __init inftl_scan_bbt(struct mtd_info *mtd) this->bbt_md->pattern ="TBB_SYSM" ; } -/* It's safe to set bd=NULL below because NAND_BBT_CREATE is not set. -At least as nand_bbt.c is currently written. */ -if ((ret =nand_scan_bbt (mtd,NULL ))) + ret = this->scan_bbt (mtd); +if (ret)return ret; +memset ((char *)parts,0 ,sizeof (parts)); numparts =inftl_partscan (mtd, parts);/* At least for now, require the INFTL Media Header. We could probably
@@ -1369,7 +1372,7 @@ static inline int __init doc2000_init(struct mtd_info *mtd) this->read_byte = doc2000_read_byte; this->write_buf = doc2000_writebuf; this->read_buf = doc2000_readbuf; - this->scan_bbt = nftl_scan_bbt; + doc->late_init = nftl_scan_bbt; doc->CDSNControl =CDSN_CTRL_FLASH_IO |CDSN_CTRL_ECC_IO ;doc2000_count_chips (mtd);
@@ -1396,13 +1399,13 @@ static inline int __init doc2001_init(struct mtd_info *mtd)can have multiple chips. */ doc2000_count_chips (mtd); mtd->name ="DiskOnChip 2000 (INFTL Model)" ; - this->scan_bbt = inftl_scan_bbt; + doc->late_init = inftl_scan_bbt;return (4 * doc->chips_per_floor); }else {/* Bog-standard Millennium */ doc->chips_per_floor =1 ; mtd->name ="DiskOnChip Millennium" ; - this->scan_bbt = nftl_scan_bbt; + doc->late_init = nftl_scan_bbt;return 1 ; } }
@@ -1415,7 +1418,7 @@ static inline int __init doc2001plus_init(struct mtd_info *mtd) this->read_byte = doc2001plus_read_byte; this->write_buf = doc2001plus_writebuf; this->read_buf = doc2001plus_readbuf; - this->scan_bbt = inftl_scan_bbt; + doc->late_init = inftl_scan_bbt; this->cmd_ctrl =NULL ; this->select_chip = doc2001plus_select_chip; this->cmdfunc = doc2001plus_command;
@@ -1591,6 +1594,8 @@ static int __init doc_probe(unsigned long physadr) nand->ecc.bytes =6 ; nand->ecc.strength =2 ; nand->bbt_options =NAND_BBT_USE_FLASH ; +/* Skip the automatic BBT scan so we can run it manually */ + nand->options |=NAND_SKIP_BBTSCAN ; doc->physadr = physadr; doc->virtadr = virtadr;
@@ -1608,7 +1613,7 @@ static int __init doc_probe(unsigned long physadr)else numchips =doc2001_init (mtd); -if ((ret =nand_scan (mtd, numchips))) { +if ((ret =nand_scan (mtd, numchips)) || (ret = doc->late_init (mtd))) {/* DBB note: i believe nand_release is necessary here, as buffers may have been allocated in nand_base. Check with Thomas. FIX ME! */
Signature · 1 line
2.1.0
PATCH 3/5 mtd: nand_bbt: make nand_scan_bbt() static
In reply to Brian NorrisThis implementation detail is no longer needed outside of nand_bbt.c.
--- drivers/mtd/nand/nand_bbt.c | 2 +- include/linux/mtd/nand.h | 1 - 2 files changed, 1 insertion(+), 2 deletions(-)
drivers/mtd/nand/nand_bbt.c +1 -1
@@ -1074,7 +1074,7 @@ static void verify_bbt_descr(struct mtd_info *mtd, struct nand_bbt_descr *bd)* The bad block table memory is allocated here. It must be freed by calling * the nand_free_bbt function. */ -int nand_scan_bbt (struct mtd_info *mtd,struct nand_bbt_descr *bd) +static int nand_scan_bbt (struct mtd_info *mtd,struct nand_bbt_descr *bd) {struct nand_chip *this = mtd->priv;int len, res =0 ;
include/linux/mtd/nand.h +0 -1
@@ -833,7 +833,6 @@ struct nand_manufacturers {extern struct nand_flash_dev nand_flash_ids[];extern struct nand_manufacturers nand_manuf_ids[]; -extern int nand_scan_bbt (struct mtd_info *mtd,struct nand_bbt_descr *bd);extern int nand_default_bbt (struct mtd_info *mtd);extern int nand_markbad_bbt (struct mtd_info *mtd,loff_t offs);extern int nand_isreserved_bbt (struct mtd_info *mtd,loff_t offs);
Signature · 1 line
2.1.0
PATCH 4/5 mtd: nand_bbt: unify/fix error handling in nand_scan_bbt()
In reply to Brian NorrisDon't leak this->bbt, and return early if check_create() fails. It helps to have a single error path to avoid these problems.
--- drivers/mtd/nand/nand_bbt.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-)
drivers/mtd/nand/nand_bbt.c +12 -7
@@ -1077,7 +1077,7 @@ static void verify_bbt_descr(struct mtd_info *mtd, struct nand_bbt_descr *bd)static int nand_scan_bbt (struct mtd_info *mtd,struct nand_bbt_descr *bd) {struct nand_chip *this = mtd->priv; -int len, res =0 ; +int len, res;uint8_t *buf;struct nand_bbt_descr *td = this->bbt_td;struct nand_bbt_descr *md = this->bbt_md;
@@ -1098,10 +1098,9 @@ static int nand_scan_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd)if (!td) {if ((res =nand_memory_bbt (mtd, bd))) {pr_err ("nand_bbt: can't scan flash and build the RAM-based BBT\n" ); -kfree (this->bbt); - this->bbt =NULL ; + goto err; } -return res; +return 0 ; }verify_bbt_descr (mtd, td);verify_bbt_descr (mtd, md);
@@ -1111,9 +1110,8 @@ static int nand_scan_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd) len += (len >> this->page_shift) * mtd->oobsize; buf =vmalloc (len);if (!buf) { -kfree (this->bbt); - this->bbt =NULL ; -return -ENOMEM ; + res = -ENOMEM ; + goto err; }/* Is the bbt at a given page? */
@@ -1125,6 +1123,8 @@ static int nand_scan_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd) } res =check_create (mtd, buf, bd); +if (res) + goto err;/* Prevent the bbt regions from erasing / writing */ mark_bbt_region (mtd, td);
@@ -1132,6 +1132,11 @@ static int nand_scan_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd)mark_bbt_region (mtd, md);vfree (buf); +return 0 ; + +err: +kfree (this->bbt); + this->bbt =NULL ;return res; }
Signature · 1 line
2.1.0
PATCH 5/5 mtd: nand_bbt: fix theoretical integer overflow in BBT write
In reply to Brian NorrisThis statement was written with a cast-to-loff_t to be sure to have a full 64-bit mask. However, we don't account for the fact that '1 << this->bbt_erase_shift' might already overflow.
This will not be a problem in practice, since eraseblocks should never be anywhere near 4GiB. But we can do this for completeness, and quiet Coverity in the meantime. CID #1226806.
--- drivers/mtd/nand/nand_bbt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mtd/nand/nand_bbt.c +1 -1
@@ -719,7 +719,7 @@ static int write_bbt(struct mtd_info *mtd, uint8_t *buf,/* Must we save the block contents? */ if (td->options &NAND_BBT_SAVECONTENT ) {/* Make it block aligned */ - to &= ~((loff_t)((1 << this->bbt_erase_shift) -1 )); + to &= ~(((loff_t )1 << this->bbt_erase_shift) -1 ); len =1 << this->bbt_erase_shift; res =mtd_read (mtd, to, len, &retlen, buf);if (res <0 ) {
Signature · 1 line
2.1.0
Quoted 6 lines · Brian Norris
A few assorted changes. Some of these are geared toward refactoring the BBT code for being reused for SPI NAND.
If possible, I'd really like to get some testing for the diskonchip changes. It's an old driver, but it appears as if some still use it. Perhaps Alexander (CC'd) can report?
Pushed all 5 to l2-mtd.git, since the only comments I got were reposts of the patches...
Brian