On Tuesday, October 28, 2014 at 11:45:08 PM, Scott Wood wrote:
On Sat, 2014-10-11 at 18:42 +0200, Marek Vasut wrote:
- /* sometimes people do not think about using the ECC, so check
* to see if we have an 0xff,0xff,0xff read ECC and then ignore* the error, on the assumption that this is an un-eccd page.*/Eww. I suppose I won't argue too loudly if Linux is doing the same thing, but what if it's a corrupted blank page, or the ECC just happened to turn out as all 0xff? It seems like there should at least be a warning the first time this happens, and ideally it should be configurable.
if (read_ecc[0] == 0xff && read_ecc[1] == 0xff && read_ecc[2] == 0xff
/*&& info->platform->ignore_unset_ecc*/)return 0;
So it looks like it is configurable in Linux, but you've commented it out here.
@@ -221,6 +298,8 @@ int board_nand_init(struct nand_chip *nand)
nand->dev_ready = s3c24x0_dev_ready;
- nand->chip_delay = 50;
I'm not sure how this is related to the changes described in the changelog...
Can you collect the MTD patches which are applicable at least and drop this one?
Best regards, Marek Vasut