linux-kernel - Re: [PATCH v2 tracing 3/6] tracing: support IPv6 filter predicates

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230609171227.56dc45b6@gandalf.local.home>
Date:   Fri, 9 Jun 2023 17:12:27 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     Alan Maguire <alan.maguire@...cle.com>
Cc:     mhiramat@...nel.org, corbet@....net, shuah@...nel.org,
        linux-trace-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
        linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 tracing 3/6] tracing: support IPv6 filter predicates


BTW, the subjects for the tracing subsystem should always start with a
capital letter.

  "tracing: Support IPv6 filter predicates"

But that's not why I'm replying here.

On Fri, 28 Apr 2023 16:34:46 +0100
Alan Maguire <alan.maguire@...cle.com> wrote:

> Support '==' and '!=' predicates for IPv6 addresses;
> for example
> 
>  cd /sys/kernel/debug/tracing/events/tcp/tcp_receive_reset
>  echo "saddr_v6 == ::1" > filter
> 
>  or equivalently
> 
>  echo "saddr_v6 == 0:0:0:0:0:0:0:1" > filter
> 
> Signed-off-by: Alan Maguire <alan.maguire@...cle.com>
> ---
>  kernel/trace/trace_events_filter.c | 73 ++++++++++++++++++++++++++++++
>  1 file changed, 73 insertions(+)
> 
> diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c
> index d8e08d3c3594..e2521574f3c4 100644
> --- a/kernel/trace/trace_events_filter.c
> +++ b/kernel/trace/trace_events_filter.c
> @@ -1665,6 +1665,79 @@ static int parse_pred(const char *str, void *data,
>  		if (pred->op == OP_NE)
>  			pred->not = 1;
>  
> +	} else if (field->size == 16 &&
> +		   (str[i] == ':' ||
> +		    (isalnum(str[i]) && tolower(str[i + 1]) != 'x'))) {
> +		u8 j, gap_size, gap = 0, gap_count = 0, index = 0;
> +		u16 tmp_v6addr[8] = {};
> +		u16 v6addr[8] = {};
> +
> +		/* For IPv6 addresses, only '==' or '!=' are supported. */
> +		if (pred->op != OP_EQ && pred->op != OP_NE) {
> +			parse_error(pe, FILT_ERR_ILLEGAL_FIELD_OP, pos + i);
> +			goto err_free;
> +		}
> +		/* Store the u16s in the address string consecutively in
> +		 * tmp_v6addr while tracking the presence of a "::" (if any)
> +		 * in the IPv6 address string; we will use its location
> +		 * to determine how many u16s it represents (the gap_size
> +		 * below).  Only one "::" is allowed in an IPv6 address
> +		 * string.
> +		 */
> +		while (isalnum(str[i]) || str[i] == ':') {
> +			switch (str[i]) {
> +			case ':':
> +				i++;
> +				/* mark "::" index by setting gap */
> +				if (str[i] == ':') {
> +					gap = index;
> +					gap_count++;
> +					i++;
> +				}
> +				if (gap_count > 1) {
> +					parse_error(pe, FILT_ERR_ILLEGAL_FIELD_OP,
> +						    pos + s);
> +					goto err_free;
> +				}
> +				break;
> +			default:
> +				if (sscanf(&str[i], "%hx", &tmp_v6addr[index]) != 1) {
> +					parse_error(pe, FILT_ERR_ILLEGAL_FIELD_OP,
> +						    pos + s);
> +					goto err_free;
> +				}
> +				index++;
> +				while (isalnum(str[i]))
> +					i++;
> +				break;
> +			}
> +		}

There appears to be no limit to the above loop. I panic'd my machine with:

 # echo 'saddr_v6 == 0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef:0123:4567:89ab:cdef' > /sys/kernel/tracing/events/sock/inet_sk_error_report/filter

-- Steve

> +		/* The gap_size here represents the number of u16s the "::"
> +		 * represents; for ::1 the gap size is 7, for feed::face
> +		 * it is 6, etc.
> +		 */
> +		gap_size = 8 - index;
> +		index = 0;
> +		for (j = 0; j < 8; ) {
> +			if (gap_size > 0 && j == gap) {
> +				j += gap_size;
> +			} else {
> +#ifdef __BIG_ENDIAN
> +				v6addr[j++] = tmp_v6addr[index];
> +#else
> +				v6addr[j++] = ((tmp_v6addr[index] & 0xff) << 8) +
> +					      ((tmp_v6addr[index] & 0xff00) >> 8);
> +#endif
> +				index++;
> +			}
> +		}
> +		pred_val = kzalloc(field->size, GFP_KERNEL);
> +		memcpy(pred_val, v6addr, field->size);
> +		pred->val = (u64)pred_val;
> +		pred->fn_num = FILTER_PRED_FN_MEMCMP;
> +		if (pred->op == OP_NE)
> +			pred->not = 1;
> +
>  	} else if (str[i] == '0' && tolower(str[i + 1]) == 'x' &&
>  		   field->size > 8) {
>  		/* For sizes > 8 bytes, we store hex bytes for comparison;