Skip to contentSkip to conversation

mtd: CONFIG_NAND_LAZY_SCAN option support

PATCH New Ilya Yanok 3 participants 4 messages Jun 26, 2008 – Jul 6, 2008

Subsystems (2)
Memory technology devices (MTD) NAND flash
lore
Share & tools
lore

Ilya Yanok

author
#0 lore raw

PATCH 1/1 mtd: CONFIG_NAND_LAZY_SCAN option support

With this option set, the nand_init() doesn't scan NAND for bad blocks.
This allows one to avoid boot delays caused by the scanning procedure for the boards which do not need NAND for booting.

To work with NAND having this option set, one should explicitely call
nand_lazy_scan_finish() before using the NAND chip. This function automatically
called at the first access to NAND chip from the U-Boot command line.

Signed-off-by Ilya Yanok <[email protected]>
Signed-off-by Yuri Tikhonov <[email protected]>
---
 common/cmd_nand.c            |    2 ++
 drivers/mtd/nand/nand_base.c |    5 +++++
 include/linux/mtd/nand.h     |    4 ++++
 include/nand.h               |   14 ++++++++++++++
 4 files changed, 25 insertions(+), 0 deletions(-)
common/cmd_nand.c +0 -0
index 37eb41b..6f5d13d 100644
@@ -236,6 +236,8 @@ int do_nand(cmd_tbl_t * cmdtp, int flag, int argc,

char *argv[])

     }
     nand = &nand_info[nand_curr_device];

+ nand_lazy_scan_finish(nand);
+

     if (strcmp(cmd, "bad") == 0) {
         printf("\nDevice %d bad blocks:\n", nand_curr_device);
         for (off = 0; off < nand->size; off += nand->erasesize)
drivers/mtd/nand/nand_base.c +5 -0
index 740d3fc..f3ee705 100644
@@ -2652,8 +2652,13 @@ int nand_scan (struct mtd_info *mtd, int maxchips)
 #if 0
     mtd->owner = THIS_MODULE;
 #endif
+#ifdef CONFIG_NAND_LAZY_SCAN
+    this->options &= ~NAND_BBT_SCANNED;
+    return 0;
+#else
     /* Build bad block table */
     return this->scan_bbt (mtd);
+#endif
 }
 
 /**
include/linux/mtd/nand.h +4 -0
index 4cc4a7d..8a6d69e 100644
@@ -450,6 +450,10 @@ struct nand_bbt_descr {
 #define NAND_BBT_SAVECONTENT    0x00002000
 /* Search good / bad pattern on the first and the second page */
 #define NAND_BBT_SCAN2NDPAGE    0x00004000
+#ifdef CONFIG_NAND_LAZY_SCAN
+/* bbt is already read */
+#define NAND_BBT_SCANNED    0x80000000
+#endif
 
 /* The maximum number of blocks to scan for a bbt */
 #define NAND_BBT_SCAN_MAXBLOCKS 4
include/nand.h +0 -0
index 247d346..6d3e91a 100644
@@ -122,4 +122,18 @@ int nand_get_lock_status(nand_info_t *meminfo,

ulong offset);
void board_nand_select_device(struct nand_chip *nand, int chip); #endif

+#ifdef CONFIG_NAND_LAZY_SCAN
+static inline void nand_lazy_scan_finish(nand_info_t *info) +{
+ struct nand_chip *chip = info->priv;
+
+ if (!(chip->options & NAND_BBT_SCANNED)) {
+ chip->scan_bbt(info);
+ chip->options |= NAND_BBT_SCANNED;
+ }
+}
+#else
+#define nand_lazy_scan_finish(info) do {} while(0)
+#endif
+
#endif

Signature · 1 line
1.5.4.2
#1 lore raw

PATCH 1/1 mtd: CONFIG_NAND_LAZY_SCAN option support

common/cmd_nand.c int do_nand(cmd_tbl_t * cmdtp, int flag, int argc, · line 241

Quoted 11 lines · Ilya Yanok
common/cmd_nand.c +2 -0
index 37eb41b..6f5d13d 100644
@@ -236,6 +236,8 @@ int do_nand(cmd_tbl_t * cmdtp, int flag, int argc,
 char *argv[])
     }
     nand = &nand_info[nand_curr_device];
 
+    nand_lazy_scan_finish(nand);
+

There are other entry points that need to be covered (e.g. do_nandboot, do_onenand, env_nand, jffs2, etc). Probably better to put the call in nand_block_isbad().

Quoted 5 lines
drivers/mtd/nand/nand_base.c +0 -0
index 740d3fc..f3ee705 100644

drivers/mtd/nand/nand_base.c int nand_scan (struct mtd_info *mtd, int maxchips) · line 2661

Quoted 10 lines
drivers/mtd/nand/nand_base.c +5 -0
@@ -2652,8 +2652,13 @@ int nand_scan (struct mtd_info *mtd, int maxchips)
 #if 0
     mtd->owner = THIS_MODULE;
 #endif
+#ifdef CONFIG_NAND_LAZY_SCAN
+    this->options &= ~NAND_BBT_SCANNED;
+    return 0;
+#else
     /* Build bad block table */
     return this->scan_bbt (mtd);
+#endif

Is there any reason not to enable this unconditionally?

Also, this patch is whitespace-mangled.

-Scott

author
#2 lore raw

PATCH 1/1 mtd: CONFIG_NAND_LAZY_SCAN support (2nd rev)

Quoted 9 lines · Scott Wood
Quoted 9 lines
common/cmd_nand.c +0 -0
index 37eb41b..6f5d13d 100644
@@ -236,6 +236,8 @@ int do_nand(cmd_tbl_t * cmdtp, int flag, int argc,
 char *argv[])
     }
     nand = &nand_info[nand_curr_device];
Quoted 6 lines

common/cmd_nand.c int do_nand(cmd_tbl_t * cmdtp, int flag, int argc, · line 241

Quoted 2 lines
common/cmd_nand.c +2 -0
+    nand_lazy_scan_finish(nand);
+

There are other entry points that need to be covered (e.g. do_nandboot, do_onenand, env_nand, jffs2, etc). Probably better to put the call in nand_block_isbad().

You're right. I've added scanning in nand_block_checkbad.

Quoted 1 line

Is there any reason not to enable this unconditionally?

Don't really know. I've used config option just to provide old behavior.

--
Ilya

#3 lore raw

PATCH 1/1 mtd: CONFIG_NAND_LAZY_SCAN support (2nd rev)

Quoted 4 lines · In message <[email protected]> you wrote:
Quoted 2 lines

Is there any reason not to enable this unconditionally?

Don't really know. I've used config option just to provide old behavior.

I vote for adding this unconditionally, too.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
O Staat!   Wie tief dir alle Besten fluchen!  Du bist kein Ziel.  Der
Mensch mu? weiter suchen.                     - Christian Morgenstern