Patchwork [v3,5/6] UBI: add ioctl for max_beb_per1024

login
register
mail settings
Submitter Richard Genoud
Date Aug. 20, 2012, 4 p.m.
Message ID <1345478416-23900-6-git-send-email-richard.genoud@gmail.com>
Download mbox | patch
Permalink /patch/178869/
State New
Headers show

Comments

Richard Genoud - Aug. 20, 2012, 4 p.m.
This patch provides the possibility to adjust the "maximum expected number of
bad blocks per 1024 blocks" (max_beb_per1024) for each mtd device from
UBI_IOCATT ioctl.

The majority of NAND devices have their max_beb_per1024 equal to 20, but
sometimes it's more.
We already could adjust that via a kernel parameter, now we can also use
UBI_IOCATT ioctl:
struct ubi_attach_req {
	__s32 ubi_num;
	__s32 mtd_num;
	__s32 vid_hdr_offset;
	__u16 max_beb_per1024;
	__s8 padding[10];
};

Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
---
 drivers/mtd/ubi/Kconfig |    3 ++-
 drivers/mtd/ubi/build.c |    2 ++
 drivers/mtd/ubi/cdev.c  |    2 +-
 include/mtd/ubi-user.h  |   19 ++++++++++++++++++-
 4 files changed, 23 insertions(+), 3 deletions(-)
Artem Bityutskiy - Aug. 21, 2012, 7:34 p.m.
On Mon, 2012-08-20 at 18:00 +0200, Richard Genoud wrote:
> This patch provides the possibility to adjust the "maximum expected number of
> bad blocks per 1024 blocks" (max_beb_per1024) for each mtd device from
> UBI_IOCATT ioctl.

Thanks, pushed to linux-ubi.git with minor amendments.

>  struct ubi_attach_req {
>  	__s32 ubi_num;
>  	__s32 mtd_num;
>  	__s32 vid_hdr_offset;
> -	__s8 padding[12];
> +	__u16 max_beb_per1024;
> +	__s8 padding[10];
>  };

I've made that to be _s16, for consistency. AFAIR, the reasons I used
signed types all over the place (unless the really needed an unsigned
type) are:

1. Easy to write functions with return a positive resulting value of
something (e.g., pnum) in case of success and a negative error code in
case of failure.
2. Easy to check for sane value by just comparing to 0.

But these are not very strong arguments, and I read somewhere that some
CPUs handle unsigned integers a bit faster, not sure what architectures
was that about.

Thanks!

Patch

diff --git a/drivers/mtd/ubi/Kconfig b/drivers/mtd/ubi/Kconfig
index 37e070c..9406d26 100644
--- a/drivers/mtd/ubi/Kconfig
+++ b/drivers/mtd/ubi/Kconfig
@@ -51,7 +51,8 @@  config MTD_UBI_BEB_LIMIT
 	  MTD partitions of the same size, UBI will reserve 40 eraseblocks when
 	  attaching a partition.
 
-	  This option can be overridden by the kernel parameter ubi.mtd.
+	  This option can be overridden by the kernel parameter ubi.mtd and the
+	  ioctl UBI_IOCATT.
 
 	  Leave the default value if unsure.
 
diff --git a/drivers/mtd/ubi/build.c b/drivers/mtd/ubi/build.c
index 5b987cb..a690d86 100644
--- a/drivers/mtd/ubi/build.c
+++ b/drivers/mtd/ubi/build.c
@@ -865,6 +865,8 @@  int ubi_attach_mtd_dev(struct mtd_info *mtd, int ubi_num,
 
 	/*
 	 * Use the default value if max_beb_per1024 isn't provided.
+	 * This way, we are keeping the same behaviour between the UBI_IOCATT
+	 * ioctl and the module parameter.
 	 */
 	if (!max_beb_per1024)
 		max_beb_per1024 = CONFIG_MTD_UBI_BEB_LIMIT;
diff --git a/drivers/mtd/ubi/cdev.c b/drivers/mtd/ubi/cdev.c
index 619f914..7885dc0 100644
--- a/drivers/mtd/ubi/cdev.c
+++ b/drivers/mtd/ubi/cdev.c
@@ -1011,7 +1011,7 @@  static long ctrl_cdev_ioctl(struct file *file, unsigned int cmd,
 		 */
 		mutex_lock(&ubi_devices_mutex);
 		err = ubi_attach_mtd_dev(mtd, req.ubi_num, req.vid_hdr_offset,
-					 CONFIG_MTD_UBI_BEB_LIMIT);
+					 req.max_beb_per1024);
 		mutex_unlock(&ubi_devices_mutex);
 		if (err < 0)
 			put_mtd_device(mtd);
diff --git a/include/mtd/ubi-user.h b/include/mtd/ubi-user.h
index 8787349..2a763ae 100644
--- a/include/mtd/ubi-user.h
+++ b/include/mtd/ubi-user.h
@@ -222,6 +222,7 @@  enum {
  * @ubi_num: UBI device number to create
  * @mtd_num: MTD device number to attach
  * @vid_hdr_offset: VID header offset (use defaults if %0)
+ * @max_beb_per1024: Maximum expected bad eraseblocks per 1024 eraseblocks
  * @padding: reserved for future, not used, has to be zeroed
  *
  * This data structure is used to specify MTD device UBI has to attach and the
@@ -245,12 +246,28 @@  enum {
  * be 2KiB-64 bytes = 1984. Note, that this position is not even 512-bytes
  * aligned, which is OK, as UBI is clever enough to realize this is 4th
  * sub-page of the first page and add needed padding.
+ *
+ * The @max_beb_per1024 is the maximum bad eraseblocks UBI expects on the ubi
+ * device per 1024 eraseblocks.
+ * This value is often given in an other form in the NAND datasheet (min NVB
+ * i.e. minimal number of valid blocks). The maximum expected bad eraseblocks
+ * per 1024 is then:
+ *   1024 * (1 - MinNVB / MaxNVB)
+ * Which gives 20 for most NAND devices.
+ * This limit is used in order to derive amount of eraseblock UBI reserves for
+ * handling new bad blocks.
+ * If the device has more bad eraseblocks than this limit, UBI does not reserve
+ * any physical eraseblocks for new bad eraseblocks, but attempts to use
+ * available eraseblocks (if any).
+ * The accepted range is 0-768. If 0 is given, the default kernel value will be
+ * used for compatibility.
  */
 struct ubi_attach_req {
 	__s32 ubi_num;
 	__s32 mtd_num;
 	__s32 vid_hdr_offset;
-	__s8 padding[12];
+	__u16 max_beb_per1024;
+	__s8 padding[10];
 };
 
 /**