Wolfgang Denk wrote:

Dear Scott,

In message 1263145126-23165-1-git-send-email-liwenhao.g@gmail.com Li Wenha wrote:

...

Signed-off-by: Li Wenhao liwenhao.g@gmail.com

common/cmd_nand.c | 21 +++++++++++++++++++++ 1 files changed, 21 insertions(+), 0 deletions(-)

diff --git a/common/cmd_nand.c b/common/cmd_nand.c index 075a8af..38c6480 100644 --- a/common/cmd_nand.c +++ b/common/cmd_nand.c @@ -390,6 +390,27 @@ int do_nand(cmd_tbl_t * cmdtp, int flag, int argc, char *argv[]) ret = nand->read_oob(nand, off, &ops); else ret = nand->write_oob(nand, off, &ops);

• } else if (!strcmp(s, ".yaffs2") && !read) {
  

• 	mtd_oob_ops_t ops = {
  

• 		.mode = MTD_OOB_AUTO,
  

• 		.len = 2048,	/* page size */
  

• 		.ooblen = 64,	/* spare size */
  

• 	};
  

What about 512 or 4096 byte pages?

...

• 	ulong page = 0;
  

• 	ulong block_size = ops.len + ops.ooblen;
  

• 	while (page * block_size < size) {
  

What if "size" is not a multiple of "block_size"? Should not read past the end of the input array, and should warn the user.

...

• 		ret = nand->write_oob(nand, 
  

• 				      off + page * ops.len,
  

• 				      &ops);
  

• 		if (ret) break;
  

The break should go on its own line, and you should tell the user about the failure.

...

• 		page++;
  

• 	}
  

  } else { printf("Unknown nand command suffix '%s'.\n", s); return 1;

-- 1.6.3.3

What is the status of this patch? Is it in your queue?

No, I didn't notice it -- the subject doesn't mention NAND, and it presumably wasn't CCed to me.

Li Wenhao, could you send a new patch addressing the above issues, plus updated documentation?

-Scott