On 01/23/2015 12:59 PM, Gregory CLEMENT wrote:

On 23/01/2015 16:41, Maxime Ripard wrote:

quoted

The NDDB register holds the data that are needed by the read and write
commands.

However, during a read PIO access, the datasheet specifies that after each 32
bits read in that register, when BCH is enabled, we have to make sure that the
RDDREQ bit is set in the NDSR register.

This fixes an issue that was seen on the Armada 385, and presumably other mvebu
SoCs, when a read on a newly erased page would end up in the driver reporting a
timeout from the NAND.

Cc: <redacted>

It would help the stable maintainer if you could indicate since which commit or
kernel release this fix should be applied.

This is a fix for the BCH support, namely commit 43bcfd2bb24a
"mtd: nand: pxa3xx: Add driver-specific ECC BCH support". The commit was merged
in v3.14.

However, this patch won't apply directly there. It will apply on commit
fa543bef72d6 "mtd: nand: pxa3xx: Add a read/write buffers markers"; which
was also merged in v3.14.

Therefore, I guess it's OK to say

Cc: <redacted> # v3.14.x

quoted

Signed-off-by: Maxime Ripard <redacted>
---
 drivers/mtd/nand/pxa3xx_nand.c | 36 ++++++++++++++++++++++++++++++------
 1 file changed, 30 insertions(+), 6 deletions(-)

diff --git a/drivers/mtd/nand/pxa3xx_nand.c b/drivers/mtd/nand/pxa3xx_nand.c
index 96b0b1d27df1..320c2ab14d4e 100644
--- a/drivers/mtd/nand/pxa3xx_nand.c
+++ b/drivers/mtd/nand/pxa3xx_nand.c

@@ -480,6 +480,30 @@ static void disable_int(struct pxa3xx_nand_info *info, uint32_t int_mask)
 	nand_writel(info, NDCR, ndcr | int_mask);
 }
 
+static void drain_fifo(struct pxa3xx_nand_info *info,
+		       void *data,
+		       int len)

^^
You don't need to split that line, it seems to fit 80 characters as is.

quoted

+{
+	u32 *dst = (u32 *)data;
+
+	if (info->ecc_bch) {
+		while (len--) {
+			*dst++ = nand_readl(info, NDDB);
+
+			/*
+			 * According to the datasheet, when reading
+			 * from NDDB with BCH enabled, after each 32
+			 * bits reads, we have to make sure that the
+			 * NDSR.RDDREQ bit is set
+			 */
+			while (!(nand_readl(info, NDSR) & NDSR_RDDREQ))
+				cpu_relax();

Are we sure that we won't be blocked here?
If not, what about adding a timeout?

Definitely. I think we shouldn't have an infinite loop, no matter what the hw specs say.
-- 
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com