diff mbox

[BUG] Nand support broken with v2.6.36-rc1

Message ID 4C6B024A.7040505@broadcom.com
State Accepted
Commit 065a1ed8de85583888b3d4f22c64b534a1fbdaaa
Headers show

Commit Message

Brian Norris Aug. 17, 2010, 9:42 p.m. UTC 
On 08/17/2010 01:05 PM, Michael Guntsche wrote:
> On 17 Aug 10 11:49, Brian Norris wrote:
>> First, have you ever used any of the mtdutils? In particular,
>> running the command "mtdinfo -a" and sending the output is helpful
>> if you have the utility installed on your board.
> hmm mtdinfo tries to open /sys/class/mtd/mtd0/dev which das not exist
> the device is working ok as block device on the other hand so let's try
> the next thing.

I'm not an expert on the workings of mtdutils, so I don't know
why your device does not have the necessary sysfs entries. Perhaps
weird hardware features or a strange, incorrect driver (I can't work
on your board specific driver for you). I suppose it's OK to ignore
this problem for the moment.

> Output booting with a patched .36-rc1
> 
> [    0.279217] rbppc_nand_probe: MikroTik RouterBOARD 600 series NAND driver, version 0.0.2
> [    0.287535] ID byte 0: 0xad
> [    0.290373] ID byte 1: 0x76
> [    0.293185] ID byte 2: 0xad
> [    0.295985] ID byte 3: 0x76
> [    0.298798] ID byte 4: 0xad
> [    0.301610] ID byte 5: 0x76
> [    0.304423] ID byte 6: 0xad
> [    0.307223] ID byte 7: 0x76
> [    0.310046] NAND device: Manufacturer ID: 0xad, Chip ID: 0x76 (Hynix NAND 64MiB 3,3V 8-bit)
> 
> Hope this helps...
> 

Honestly, that doesn't really help :) I guess the device is old
enough it does not have an extended ID. In that case, I will need
the part number to be able to diagnose for sure. Can you find the
physical chip on the board and give me whatever labeling is on it?

In place of that, though, you can just try this patch on 2.6.36-rc1.
I believe it should satisfy the intention of my previous (faulty)
commit while reverting the regression behavior. If this works OK, I
will submit it to be included in the mainline kernel.

Thanks for taking the time to debug this.

Brian

---
 drivers/mtd/nand/nand_base.c |   10 +++-------
 1 files changed, 3 insertions(+), 7 deletions(-)

Comments

Michael Guntsche Aug. 18, 2010, 5:53 a.m. UTC | #1 
On 17 Aug 10 14:42, Brian Norris wrote:
> In place of that, though, you can just try this patch on 2.6.36-rc1.
> I believe it should satisfy the intention of my previous (faulty)
> commit while reverting the regression behavior. If this works OK, I
> will submit it to be included in the mainline kernel.

Hi Brian,

Applying this patch fixes the problem for me. I removed my workaround
and just tried your patch and was able to mount both mtdblock devices.

Thank you for fixing this,
Michael Guntsche
diff mbox

Patch

diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
index a3c7473..a22ed7b 100644
--- a/drivers/mtd/nand/nand_base.c
+++ b/drivers/mtd/nand/nand_base.c
@@ -2934,14 +2934,10 @@  static struct nand_flash_dev *nand_get_flash_type(struct mtd_info *mtd,
 		chip->chip_shift = ffs((unsigned)(chip->chipsize >> 32)) + 32 - 1;
 
 	/* Set the bad block position */
-	if (!(busw & NAND_BUSWIDTH_16) && (*maf_id == NAND_MFR_STMICRO ||
-				(*maf_id == NAND_MFR_SAMSUNG &&
-				 mtd->writesize == 512) ||
-				*maf_id == NAND_MFR_AMD))
-		chip->badblockpos = NAND_SMALL_BADBLOCK_POS;
-	else
+	if (mtd->writesize > 512 || (busw & NAND_BUSWIDTH_16))
 		chip->badblockpos = NAND_LARGE_BADBLOCK_POS;
-
+	else
+		chip->badblockpos = NAND_SMALL_BADBLOCK_POS;
 
 	/* Get chip options, preserve non chip based options */
 	chip->options &= ~NAND_CHIPOPTIONS_MSK;