Hi Brian,

On Fri, 4 Mar 2016 18:26:21 -0800
Brian Norris [off-list ref] wrote:

On Fri, Feb 26, 2016 at 01:57:24AM +0100, Boris Brezillon wrote:

quoted

Use the mtd_set_ecclayout() helper instead of directly assigning the
mtd->ecclayout field.

Signed-off-by: Boris Brezillon <redacted>
---
 drivers/mtd/nand/nand_base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
index 17504f2..5093a3c 100644
--- a/drivers/mtd/nand/nand_base.c
+++ b/drivers/mtd/nand/nand_base.c

@@ -4288,7 +4288,7 @@ int nand_scan_tail(struct mtd_info *mtd)
 		ecc->write_oob_raw = ecc->write_oob;
 
 	/* propagate ecc info to mtd_info */
-	mtd->ecclayout = ecc->layout;
+	mtd_set_ecclayout(mtd, ecc->layout);

I'm having trouble applying this one. For the life of me, I can't figure
out where you got this context from. This block only appears much later
in nand_scan_tail()...

Patch 7 has moved this section upper in the function to avoid problems
when calculating the number of available/free OOB bytes.

Do you think you could post a git tree with your intended changes? I may
just try to pull something in like that instead.

Yep, it's there [1]. Note that this branch contains the two fixes I
talked about with Harvey and Stephan. I also made a few changes to use
ecc->total instead of calculating (ecc->steps * ecc->bytes).

BTW, I'm not sure the OMAP refactorings are going to come in time, but I
was planning to pull those directly from the TI folks (i.e., they won't
be rebased on l2-mtd.git), since there's some intermingling of platform
changes there. I think I can fix the conflicts fine, but FYI.

Okay, then I'll let you deal with those conflicts. I can check your
conflict resolution if you're unsure.

Thanks,

Boris


-- 
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com