Re: [PATCH v2] cifs: small underflow in cnvrtDosUnixTm()

[walter harms <wharms@bfs.de>]

Am 28.04.2017 14:51, schrieb Dan Carpenter:
> January is month 1.  There is no zero-th month.  If someone passes a
> zero month then it means we read from one space before the start of the
> total_days_of_prev_months[] array.
> 
> We may as well also be strict about days as well.
> 
> Fixes: 1bd5bbcb6531 ("[CIFS] Legacy time handling for Win9x and OS/2 part 1")
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> v2:  Be strict about days as well.  My first patch was less intrusive
> because it only prevented the out of bounds access.  I have no idea how
> common it is to pass in an illegal day but, hopefully, not very common.
> 
> diff --git a/fs/cifs/netmisc.c b/fs/cifs/netmisc.c
> index abae6dd2c6b9..4b2726ee4fad 100644
> --- a/fs/cifs/netmisc.c
> +++ b/fs/cifs/netmisc.c
> @@ -980,10 +980,10 @@ struct timespec cnvrtDosUnixTm(__le16 le_date, __le16 le_time, int offset)
>  		cifs_dbg(VFS, "illegal hours %d\n", st->Hours);
>  	days = sd->Day;
>  	month = sd->Month;
> -	if ((days > 31) || (month > 12)) {
> +	if (days < 1 || days > 31 || month < 1 || month > 12) {
>  		cifs_dbg(VFS, "illegal date, month %d day: %d\n", month, days);
> -		if (month > 12)
> -			month = 12;
> +		days = (days < 1) ? 1 : ((days <= 31) ? days : 31);
> +		month = (month < 1) ? 1 : ((month <= 12) ? month : 12);
>  	}
>  	month -= 1;
>  	days += total_days_of_prev_months[month];

The mixing in now a bit unfortunate ... why not simply

	if (days < 1 || days > 31 || month < 1 || month > 12)
		cifs_dbg(VFS, "illegal date, month %d day: %d\n", month, days);

	month = (month < 1) ? 1 : ((month <= 12) ? month : 12);
	days = (days < 1) ? 1 : ((days <= 31) ? days : 31);

	
hope that helps,
re,
 wh


> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>