[PATCH] [MTD] NAND: fix race in nand_base

[PATCH] [MTD] NAND: fix race in nand_base

[Artem Bityutskiy]

From: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Date: Mon, 23 Jul 2007 16:06:50 +0300
Subject: [PATCH] [MTD] NAND: fix race in nand_base

When we mark block bad we have to get chip because this involves
writing to the page's OOB. We hit this bug in UBI - we observed
random obscure crashes when it marks block bad from the background
thread and there is some parallel task which utilizes flash.

Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
---
 drivers/mtd/nand/nand_base.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
index 7e68203..3391b05 100644
--- a/drivers/mtd/nand/nand_base.c
+++ b/drivers/mtd/nand/nand_base.c
@@ -348,6 +348,8 @@ static int nand_default_block_markbad(struct mtd_info *mtd, loff_t ofs)
 	uint8_t buf[2] = { 0, 0 };
 	int block, ret;
 
+	nand_get_device(chip, mtd, FL_READING);
+
 	/* Get block number */
 	block = (int)(ofs >> chip->bbt_erase_shift);
 	if (chip->bbt)
@@ -370,6 +372,8 @@ static int nand_default_block_markbad(struct mtd_info *mtd, loff_t ofs)
 	}
 	if (!ret)
 		mtd->ecc_stats.badblocks++;
+
+	nand_release_device(mtd);
 	return ret;
 }
 
-- 
1.5.0.6


-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: [PATCH] [MTD] NAND: fix race in nand_base

[Artem Bityutskiy]

On Mon, 2007-07-23 at 14:17 +0300, Artem Bityutskiy wrote:
> diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
> index 7e68203..3391b05 100644
> --- a/drivers/mtd/nand/nand_base.c
> +++ b/drivers/mtd/nand/nand_base.c
> @@ -348,6 +348,8 @@ static int nand_default_block_markbad(struct mtd_info *mtd, loff_t ofs)
>  	uint8_t buf[2] = { 0, 0 };
>  	int block, ret;
>  
> +	nand_get_device(chip, mtd, FL_READING);
Err, this should be FL_WRITING I guess.

> +
>  	/* Get block number */
>  	block = (int)(ofs >> chip->bbt_erase_shift);
>  	if (chip->bbt)
> @@ -370,6 +372,8 @@ static int nand_default_block_markbad(struct mtd_info *mtd, loff_t ofs)
>  	}
>  	if (!ret)
>  		mtd->ecc_stats.badblocks++;
> +
> +	nand_release_device(mtd);
>  	return ret;
>  }

There is still the problem with nand_update_bbt() which in turn calls
mtd->read() which will deadlock. I have not found how BBT is protected.
Any suggestions/pointers?

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)