[PATCH] UBI: add ubi_err() to report the failure of leb read

[Richard Weinberger]

Am 16.12.2014 um 08:54 schrieb hujianyang:
> If an error occur while reading from PEBs, for example, an ECC error,
> ubi_io_read() will print some error messages. But it's not enough for
> debugging. These messages don't show the mapping info for a read from
> UBIFS layer.
> 
> Although UBIFS will soon print its error messages after catching the
> return value from UBI layer,  multi-path reading will confuse the
> relationship between LEBs and PEBs showed by these messages.
> 
> This patch adds an ubi_err() to report reading errors in the function
> ubi_eba_read_leb(). The mapping info of LEB and PEB is showed by
> this error message.
> 
> Signed-off-by: hujianyang <hujianyang at huawei.com>
> ---
>  drivers/mtd/ubi/eba.c |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/mtd/ubi/eba.c b/drivers/mtd/ubi/eba.c
> index b698534..b4e69e1 100644
> --- a/drivers/mtd/ubi/eba.c
> +++ b/drivers/mtd/ubi/eba.c
> @@ -477,6 +477,8 @@ out_free:
>  	ubi_free_vid_hdr(ubi, vid_hdr);
>  out_unlock:
>  	leb_read_unlock(ubi, vol_id, lnum);
> +	ubi_err(ubi, "err %d while reading %d bytes from offset %d of LEB %d:%d, PEB %d",
> +		err, len, offset, vol_id, lnum, pnum);

This label will also be reached if we run out of memory.
Please make sure that the new ubi_err() can only be reached in case of an MTD error.
Also make sure that the function prints only one message and not two.

Thanks,
//richard