On Thu, Aug 09, 2018 at 10:37:50PM +0200, Sebastian Benoit wrote:
> Claudio Jeker([email protected]) on 2018.08.09 21:59:16 +0200:
> > On Thu, Aug 09, 2018 at 03:10:11PM +0200, Claudio Jeker wrote:
> > > Per rfc6472 AS_SET should no longer be used but some AS still do.
> > > Until now source-as would take the rightmost AS number of an AS_PATH no
> > > matter if it was an AS_SEQUENCE or an AS_SET. Thit is not correct. Also
> > > because AS_SET are used in aggregation source-as should match against the
> > > aggregator AS (which it the rightmost as of the previous AS_PATH segment).
> > > At the same time move the peer-as check out of the loop, there is no
> > > reason to have that inside the for loop.
> > >
> > > This diff implements this and also makes aspath_extract() and the lenght
> > > checks a bit more paranoid.
> >
> > OK, updated diff here that adds three things.
> > a) it passes the filter_as struct to as_compare
> > b) it ensures that if for whatever crazy reason AS 0 is passed to the
> > filter it will not match (no matter what). This is more paranoia.
> > I decided to do that since util.c is shared between bgpd and bgpctl
> > and so the easy way of calling fatalx() is not an option here.
> > c) fix the souce-as 42 >< 4242 case. Currently the >< operator is not
> > correctly implemented it is more a <> as in inside but excluding the
> > edges. I think 42 >< 4242 should match for AS 1 - 41 and 4243 and
> > higher.
> >
> > Is this still OK?
>
> hm. yes. ok.
>
> although this will cause bizarre effects:
>
> deny from any AS 0
>
> wont match even for as == 0.
>
Yes, this rule should never match since there should be never an AS 0 in
any path that makes it to the filter. This is why it is kind of valid to
ingore that rule completly and move to the next one.
> i have to admit that we check for AS 0 in the path somewhere else.
> still...
Yes, aspath with a 0 AS number in them are soft rejected by aspath_verify.
I currently don't want to disallow AS 0 in the parse.y because ROA may
need it in one way or the other. But even there it is a make sure it never
matches thing. Once I have those bits together it may be possible to
restrict this further.
In the end having aspath_extract return 0 is in all current usecases of
bgpd impossible. There is always some additional code that ensures that
there is no overflow but we tripped into that hole at least once and I
would prefer to not do it again. In general I would throw in a fatalx()
there but as mentioned that won't fly in bgpctl which uses this code as
well.
I will commit this without the aspath_extract() chunk and the as == 0
check. Will put that into a new diff for later.
> > --
> > :wq Claudio
> >
> >
> > Index: bgpd.h
> > ===================================================================
> > RCS file: /cvs/src/usr.sbin/bgpd/bgpd.h,v
> > retrieving revision 1.329
> > diff -u -p -r1.329 bgpd.h
> > --- bgpd.h 8 Aug 2018 14:29:05 -0000 1.329
> > +++ bgpd.h 9 Aug 2018 19:43:39 -0000
> > @@ -1156,8 +1156,6 @@ int aspath_snprint(char *, size_t, voi
> > int aspath_asprint(char **, void *, u_int16_t);
> > size_t aspath_strlen(void *, u_int16_t);
> > int aspath_match(void *, u_int16_t, struct filter_as *,
> > u_int32_t);
> > -int as_compare(u_int8_t, u_int32_t, u_int32_t, u_int32_t,
> > - u_int32_t);
> > u_int32_t aspath_extract(const void *, int);
> > int aspath_verify(void *, u_int16_t, int);
> > #define AS_ERR_LEN -1
> > Index: util.c
> > ===================================================================
> > RCS file: /cvs/src/usr.sbin/bgpd/util.c,v
> > retrieving revision 1.28
> > diff -u -p -r1.28 util.c
> > --- util.c 22 Jul 2018 16:52:27 -0000 1.28
> > +++ util.c 9 Aug 2018 19:38:34 -0000
> > @@ -312,6 +312,22 @@ aspath_strlen(void *data, u_int16_t len)
> > return (total_size);
> > }
> >
> > +static int
> > +as_compare(struct filter_as *f, u_int32_t as, u_int32_t match)
> > +{
> > + if (as == 0) /* AS 0 is invalid and therefor never matches */
> > + return (0);
> > + if ((f->op == OP_NONE || f->op == OP_EQ) && as == match)
> > + return (1);
> > + else if (f->op == OP_NE && as != match)
> > + return (1);
> > + else if (f->op == OP_RANGE && as >= f->as_min && as <= f->as_max)
> > + return (1);
> > + else if (f->op == OP_XRANGE && (as < f->as_min || as > f->as_max))
> > + return (1);
> > + return (0);
> > +}
> > +
> > /* we need to be able to search more than one as */
> > int
> > aspath_match(void *data, u_int16_t len, struct filter_as *f, u_int32_t
> > match)
> > @@ -320,7 +336,7 @@ aspath_match(void *data, u_int16_t len,
> > int final;
> > u_int16_t seg_size;
> > u_int8_t i, seg_len;
> > - u_int32_t as;
> > + u_int32_t as = 0, preas;
> >
> > if (f->type == AS_EMPTY) {
> > if (len == 0)
> > @@ -330,27 +346,35 @@ aspath_match(void *data, u_int16_t len,
> > }
> >
> > seg = data;
> > - for (; len > 0; len -= seg_size, seg += seg_size) {
> > +
> > + /* just check the first (leftmost) AS */
> > + if (f->type == AS_PEER && len >= 6) {
> > + as = aspath_extract(seg, 0);
> > + if (as_compare(f, as, match))
> > + return (1);
> > + else
> > + return (0);
> > + }
> > +
> > + for (; len >= 6; len -= seg_size, seg += seg_size) {
> > seg_len = seg[1];
> > seg_size = 2 + sizeof(u_int32_t) * seg_len;
> >
> > final = (len == seg_size);
> >
> > - /* just check the first (leftmost) AS */
> > - if (f->type == AS_PEER) {
> > - as = aspath_extract(seg, 0);
> > - if (as_compare(f->op, as, match, f->as_min, f->as_max))
> > - return (1);
> > - else
> > - return (0);
> > - }
> > /* just check the final (rightmost) AS */
> > if (f->type == AS_SOURCE) {
> > + /* extract the rightmost AS and keep the one before */
> > + preas = as;
> > + as = aspath_extract(seg, seg_len - 1);
> > /* not yet in the final segment */
> > if (!final)
> > continue;
> > - as = aspath_extract(seg, seg_len - 1);
> > - if (as_compare(f->op, as, match, f->as_min, f->as_max))
> > + if (seg[0] == AS_SET)
> > + /* use aggregator AS per rfc6472 */
> > + if (preas)
> > + as = preas;
> > + if (as_compare(f, as, match))
> > return (1);
> > else
> > return (0);
> > @@ -364,32 +388,17 @@ aspath_match(void *data, u_int16_t len,
> > if (final && i == seg_len - 1 && f->type == AS_TRANSIT)
> > return (0);
> > as = aspath_extract(seg, i);
> > - if (as_compare(f->op, as, match, f->as_min, f->as_max))
> > + if (as_compare(f, as, match))
> > return (1);
> > }
> > }
> > return (0);
> > }
> >
> > -int
> > -as_compare(u_int8_t op, u_int32_t as, u_int32_t match, u_int32_t as_min,
> > - u_int32_t as_max)
> > -{
> > - if ((op == OP_NONE || op == OP_EQ) && as == match)
> > - return (1);
> > - else if (op == OP_NE && as != match)
> > - return (1);
> > - else if (op == OP_RANGE && as >= as_min && as <= as_max)
> > - return (1);
> > - else if (op == OP_XRANGE && as > as_min && as < as_max)
> > - return (1);
> > - return (0);
> > -}
> > -
> > /*
> > * Extract the asnum out of the as segment at the specified position.
> > * Direct access is not possible because of non-aligned reads.
> > - * ATTENTION: no bounds checks are done.
> > + * Only works on verified and 4byte ASN paths.
> > */
> > u_int32_t
> > aspath_extract(const void *seg, int pos)
> > @@ -397,6 +406,9 @@ aspath_extract(const void *seg, int pos)
> > const u_char *ptr = seg;
> > u_int32_t as;
> >
> > + /* minimal overflow check, return 0 since that is an invalid ASN */
> > + if (pos >= ptr[1])
> > + return (0);
> > ptr += 2 + sizeof(u_int32_t) * pos;
> > memcpy(&as, ptr, sizeof(u_int32_t));
> > return (ntohl(as));
> >
>
--
:wq Claudio
