Commit Graph

5811 Commits

Author SHA1 Message Date
David Lawrence
d20d22616b A socket result of ISC_R_SUCCESS can still possibly be posted even after
isc_socket_cancel() has been called.  resquery_senddone() has been fixed
to account for this (there was a race condition because of the faulty
assumption that only ISC_R_CANCELED would be returned after
isc_socket_cancel()).

The interaction of resquery_response() with the dispatcher now has to be
closely reviewed to make sure it does not have a similar problem with
a canceled query.  If the faulty assumption hadn't been faulty, it was
fine, but now I'm not so sure.
2000-07-07 19:43:14 +00:00
Brian Wellington
0308460880 Some of the lwres functions weren't calling lwres_conf_parse. 2000-07-07 19:43:10 +00:00
Brian Wellington
674a8f811e Don't die when lwres_conf_parse() returns failure. 2000-07-07 19:19:09 +00:00
Brian Wellington
68aa880dd5 remove unnecessary calls to fcntl(). 2000-07-07 19:10:00 +00:00
Brian Wellington
d8c8722f28 313. [bug] When parsing resolv.conf, don't terminate on an
error.  Instead, parse as much as possible, but
                        still return an error if one was found.
2000-07-07 18:58:46 +00:00
Brian Wellington
03968c0f5e removed dead code 2000-07-07 18:28:58 +00:00
Brian Wellington
1e8c2e72e7 remove extra spaces in I: lines 2000-07-07 18:25:20 +00:00
Brian Wellington
4b7167e96a 312. [bug] Increase the number of allowed elements in the
resolv.conf search path from 6 to 8.  If there
                        are more than this, ignore the remainder rather
                        than returning a failure in lwres_conf_parse.
2000-07-07 18:24:11 +00:00
Brian Wellington
a4b496f2ab 311. [bug] lwres_conf_parse failed when the first line of
resolv.conf was empty or a comment.
2000-07-07 18:16:46 +00:00
James Brister
524c8ea0dc typo in print statement. 2000-07-07 14:30:00 +00:00
James Brister
952e9025b0 310. [func] Changes to named.conf "controls" statement (inet
subtype only)

                          - support "keys" clause

                                controls {
                                   inet * port 1024
                                        allow { any; } keys { "foo"; }
                                }

                          - allow "port xxx" to be left out of statement,
                            in which case it defaults to omapi's default port
                            of 953.
2000-07-07 13:56:14 +00:00
Brian Wellington
6c3bd0545a If the test timed out, R:FAIL was being printed twice, and the error
message was missing the I: prefix.
2000-07-07 07:15:04 +00:00
Brian Wellington
25496cebad If trying to validate a key set that happens to be a security root, the
validation should only consist of checking that each key in the key set
is also in the list of security root keys.

Strangeness occurs when the key set is signed, since the key set is marked
as secure, but the sig set is not, since it wasn't used in the validation
process.  This means that a query for a key set at a security root will
have the AD bit set if the key set is unsigned and not if the key set is signed.
2000-07-07 00:44:01 +00:00
Brian Wellington
20e597d71c remove spaces from I: statements 2000-07-07 00:41:18 +00:00
David Lawrence
3658cf5419 comment cleanup (missing "not", spaces where tabs should have been) 2000-07-07 00:25:47 +00:00
David Lawrence
9fb4b3f08a Do not return an exact match from dns_rbt_findnode() when the exact
match does not have associated data and DNS_RBTFIND_EMPTYDATA is not set.
Return DNS_R_PARTIALMATCH instead if there is a superdomain with
data, or ISC_R_NOTFOUND if no superdomain has data.

Make dns_rbt_findname() honor DNS_RBTFIND_EMPTYDATA, per the rbt.h
documentation that has been saying it will.  It didn't.
2000-07-06 23:54:45 +00:00
Brian Wellington
d52cd4a9fa Properly fill in sa_family and sa_len (if present). Also fix a bug in
test_noop (you'd really think a noop test would be bug free).
2000-07-06 22:28:44 +00:00
Brian Wellington
861df15904 The includes were in the wrong order, causing digital unix to fail. 2000-07-06 22:13:15 +00:00
Brian Wellington
23efeb1a58 Check LWRES_PLATFORM_HAVESALEN, not ISC_PLATFORM_HAVESALEN 2000-07-06 21:42:05 +00:00
Brian Wellington
23ce1c7cca Lots of the comments were wrong. 2000-07-06 20:05:13 +00:00
Michael Sawyer
fec9cc90e6 Corrected cleanup 2000-07-06 19:54:02 +00:00
Michael Sawyer
03967f5ba6 Remove unneeded file 2000-07-06 19:46:20 +00:00
Michael Sawyer
f7b7a2448c use runall.sh and cleanall.sh 2000-07-06 19:40:57 +00:00
Michael Sawyer
2c65255baa Missing quotes 2000-07-06 19:37:47 +00:00
Michael Sawyer
ea1fabfe76 SUBDIRS entry corrected and brought up to date 2000-07-06 19:34:27 +00:00
Andreas Gustafsson
2c9fdfdfbe added script for running all system tests in sequence 2000-07-06 19:32:37 +00:00
Brian Wellington
b67a6d9657 If a zone is specified by the user, use it as the suffix for relative names. 2000-07-06 19:29:40 +00:00
Andreas Gustafsson
c1db6a381f added draft-ietf-dnsext-iana-dns-00.txt (already expired) 2000-07-06 18:32:36 +00:00
Andreas Gustafsson
457a63f3b8 missing colon in glue_test rule 2000-07-06 04:36:55 +00:00
Brian Wellington
ca93ad9b9e If we're asking for a key and need to call dns_view_findzonecut(), set
the DNS_DB_NOEXACT flag.  This means that the resolver will start look for the
key at or above the parent zone, and will return the key set from the parent
zone if key sets are present in both the parent and child.
2000-07-06 02:33:43 +00:00
Andreas Gustafsson
770dd208fa work around NetBSD /bin/sh bug, with the side effect of making the
test script shorter
2000-07-06 02:33:42 +00:00
Brian Wellington
baeb4bcf32 In query_find(), if we're looking up a key and dns_db_find() returns
DNS_R_DELEGATION, don't use the domain & ns set found by dns_db_find().
This means that a recursive server will not immediately ask the child zone
for a key set at the apex; the resolver will determine who to ask.
2000-07-06 02:27:26 +00:00
Michael Sawyer
313b7db2ce Add NSUPDATE to conf.sh.in, not conf.sh... 2000-07-06 02:10:46 +00:00
Andreas Gustafsson
249763bcbf added glue test 2000-07-06 02:06:02 +00:00
Andreas Gustafsson
afb2c39d2d added glue test 2000-07-06 01:28:08 +00:00
Andreas Gustafsson
8993565982 if the file nsN/namedopts exists, use it to set named
command line options
2000-07-06 01:26:05 +00:00
Michael Sawyer
0819ae2ca5 Use limits.h/INT_MAX instead of the deprecated values.h/MAXINT 2000-07-06 01:02:42 +00:00
Andreas Gustafsson
08133a2dbe This is "the glue fix". It changes the additional data lookup
algorithm to more closely follow RFC1035, so that root servers will
provide a more complete set of name server glue addresses in ccTLD
referral responses.

For non-referral responses, the server no longer uses glue as
additional data.

For referral responses, the servers now looks for glue A/AAAA/A6 RRs
in the zone where the NS RRs reside, even in the case where this is
different from the zone where the A/AAAA/A6 RRs would have resided had
they been authoritative data.

A6 chains included as additional info may not yet fully follow these
rules.
2000-07-06 01:00:46 +00:00
Michael Sawyer
c3c6770e53 Add nsupdate/ixfr test suite 2000-07-06 00:54:03 +00:00
Brian Wellington
61e7208704 memory leak 2000-07-05 23:42:08 +00:00
Michael Sawyer
38cf6e52ce Lots of style changes.
Clean up what appears to be a varient of the race condition, this one set
off in nslookup.
2000-07-05 23:28:32 +00:00
Brian Wellington
5ddea03e3e Log a message in addition to setting a failure code when kill -TERM doesn't
kill a server.
2000-07-05 23:28:07 +00:00
Brian Wellington
54b9dc7739 Add support for setting the server and/or zone 2000-07-05 23:24:18 +00:00
Andreas Gustafsson
4defd73fca This is "the glue fix". It changes the additional data lookup
algorithm to more closely follow RFC1035, so that root servers will
provide a more complete set of name server glue addresses in ccTLD
referral responses.

For non-referral responses, the server no longer uses glue as
additional data.

For referral responses, the servers now looks for glue A/AAAA/A6 RRs
in the zone where the NS RRs reside, even in the case where this is
different from the zone where the A/AAAA/A6 RRs would have resided had
they been authoritative data.

A6 chains included as additional info may not yet fully follow these
rules.
2000-07-05 23:10:06 +00:00
Andreas Gustafsson
66465da5d1 work around NetBSD /bin/sh bug, with the side effect of making the
test script shorter
2000-07-05 22:13:56 +00:00
Andreas Gustafsson
c6ab6ca3fd moved linux_keepcaps() call and check for Linux nonroot caps back
to earlier locations; use a separate boolean flag for keeping track of whether a
UID switch was requested rather than overloading the UID value zero
2000-07-05 22:03:42 +00:00
Brian Wellington
58208669bc Should sleep for a while between kill -TERM and kill -KILL 2000-07-05 20:56:11 +00:00
Andreas Gustafsson
b7b5dd0d10 treat all responses to AXFR requests as AXFR responses 2000-07-05 20:28:34 +00:00
Michael Graff
8681403c16 Shut up a gcc warning 2000-07-05 20:26:46 +00:00
Andreas Gustafsson
aade2614cb work around NetBSD /bin/sh bug, with the side effect of making the
test script shorter
2000-07-05 19:42:34 +00:00