[U-Boot] [PATCH 1/1] jffs2: Fix zero sector_size when not using CONFIG_JFFS2_CMDLINE

[U-Boot] [PATCH 1/1] jffs2: Fix zero sector_size when not using CONFIG_JFFS2_CMDLINE

[Tomasz Figa]

This patch fixes a bug (?) introduced after inclusion of the new JFFS2 code.

When not using CONFIG_JFFS2_CMDLINE, the code in cmd_jffs2.c doesn't fill in 
part->sector_size (keeping it as 0), but a correct value is needed by the code 
in jffs2_1pass.c. This causes all JFFS2 accesses to be in the same place of the 
memory, what obviously means impossibility to use the JFFS2 partition.

This problem is fixed in this patch by including sector size calculation in 
non-CONFIG_JFFS2_CMDLINE mtdparts_init variant.

Signed-off-by: Tomasz Figa <tomasz.figa_at_gmail.com>
---
diff --git a/common/cmd_jffs2.c b/common/cmd_jffs2.c
index 7866c80..fe0f8d7 100644
--- a/common/cmd_jffs2.c
+++ b/common/cmd_jffs2.c
@@ -1778,6 +1778,96 @@ int mtdparts_init(void)
  */
 
 /**
+ * Calculate sector size.
+ *
+ * @return sector size
+ */
+static inline u32 get_part_sector_size_nand(struct mtdids *id)
+{
+#if defined(CONFIG_JFFS2_NAND) && defined(CONFIG_CMD_NAND)
+#if defined(CONFIG_NAND_LEGACY)
+	extern struct nand_chip nand_dev_desc[CONFIG_SYS_MAX_NAND_DEVICE];
+
+	return nand_dev_desc[id->num].erasesize;
+#else
+	nand_info_t *nand;
+
+	nand = &nand_info[id->num];
+
+	return nand->erasesize;
+#endif
+#else
+	BUG();
+	return 0;
+#endif
+}
+
+static inline u32 get_part_sector_size_nor(struct mtdids *id, struct 
part_info *part)
+{
+#if defined(CONFIG_CMD_FLASH)
+	extern flash_info_t flash_info[];
+	
+	u32 end_phys, start_phys, sector_size = 0, size = 0;
+	int i;
+	flash_info_t *flash;
+	
+	flash = &flash_info[id->num];
+
+	start_phys = flash->start[0] + part->offset;
+	end_phys = start_phys + part->size;
+	
+	for (i = 0; i < flash->sector_count; i++) {
+		if (flash->start[i] >= end_phys)
+			break;
+		
+		if (flash->start[i] >= start_phys) {
+			if (i == flash->sector_count - 1) {
+				size = flash->start[0] + flash->size - flash->start[i];
+			} else {
+				size = flash->start[i+1] - flash->start[i];
+			}
+			
+			if (sector_size < size)
+				sector_size = size;
+		}
+	}
+
+	return sector_size;
+#else
+	BUG();
+	return 0;
+#endif
+}
+
+static inline u32 get_part_sector_size_onenand(void)
+{
+#if defined(CONFIG_CMD_ONENAND)
+	struct mtd_info *mtd;
+
+	mtd = &onenand_mtd;
+
+	return mtd->erasesize;
+#else
+	BUG();
+	return 0;
+#endif
+}
+
+static inline u32 get_part_sector_size(struct mtdids *id, struct part_info 
*part)
+{
+	if (id->type == MTD_DEV_TYPE_NAND)
+		return get_part_sector_size_nand(id);
+	else if (id->type == MTD_DEV_TYPE_NOR)
+		return get_part_sector_size_nor(id, part);
+	else if (id->type == MTD_DEV_TYPE_ONENAND)
+		return get_part_sector_size_onenand();
+	else
+		DEBUGF("Error: Unknown device type.\n");
+		
+	return 0;
+}
+
+/**
  * Parse and initialize global mtdids mapping and create global
  * device/partition list.
  *
@@ -1846,6 +1936,8 @@ int mtdparts_init(void)
 		part->offset = 0x00000000;
 #endif
 
+		part->sector_size = get_part_sector_size(id, part);
+
 		part->dev = current_dev;
 		INIT_LIST_HEAD(&part->link);
 

[U-Boot] [PATCH 1/1] jffs2: Fix zero sector_size when not using CONFIG_JFFS2_CMDLINE

[Wolfgang Denk]

Dear Tomasz Figa,

In message <200812301835.57435.tomasz.figa@gmail.com> you wrote:
> This patch fixes a bug (?) introduced after inclusion of the new JFFS2 code.
> 
> When not using CONFIG_JFFS2_CMDLINE, the code in cmd_jffs2.c doesn't fill in 
> part->sector_size (keeping it as 0), but a correct value is needed by the code 
> in jffs2_1pass.c. This causes all JFFS2 accesses to be in the same place of the 
> memory, what obviously means impossibility to use the JFFS2 partition.
> 
> This problem is fixed in this patch by including sector size calculation in 
> non-CONFIG_JFFS2_CMDLINE mtdparts_init variant.
> 
> Signed-off-by: Tomasz Figa <tomasz.figa_at_gmail.com>

Applied, thanks, but please be a bit more careful next time; this
patch had a lot of trailing whitespace issues, and some wrapped long
lines.


Best regards,

Wolfgang Denk

-- 
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
To program is to be.