-
- Change History (most recent first):
-
-$Log: mDNS.c,v $
-Revision 1.977 2009/07/23 23:30:01 cheshire
-<rdar://problem/7086623> Sleep Proxy: Ten-second maintenance wake not long enough to reliably get network connectivity
-
-Revision 1.976 2009/07/23 09:15:06 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-Fixed silly mistake in checkin 1.974 that broke SendResponses
-
-Revision 1.975 2009/07/21 23:46:19 cheshire
-Improved "DNS Message too short" syslog debugging message
-
-Revision 1.974 2009/07/21 23:41:05 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-Fixed silly mistake in checkin 1.974 that broke SendResponses
-
-svn merge: Revision 1.974 2009/07/21 23:41:05 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-Another refinement: When building a response packet, if we're going to add an OWNER option at the end,
-reserve enough bytes to ensure that we'll be able to do that
-
-svn merge: Revision 1.973 2009/07/16 00:34:18 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-Additional refinement: If we didn't register with a Sleep Proxy when going to sleep,
-we don't need to include our OWNER option in our packets when we re-awaken
-
-svn merge: Revision 1.972 2009/07/16 00:12:23 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-Additional fixes: Only add and send OWNER option if we were already going to send a non-empty packet anyway
-
-svn merge: Revision 1.971 2009/07/15 23:35:40 cheshire
-<rdar://problem/6434656> Sleep Proxy: Put owner OPT records in multicast announcements to avoid conflicts
-
-Revision 1.970.2.1 2009/07/23 23:36:04 cheshire
-<rdar://problem/7086623> Sleep Proxy: Ten-second maintenance wake not long enough to reliably get network connectivity
-
-Revision 1.970 2009/07/11 01:59:27 cheshire
-<rdar://problem/6613674> Sleep Proxy: Add support for using sleep proxy in local network interface hardware
-When going to sleep, try calling ActivateLocalProxy before registering with remote sleep proxy
-
-Revision 1.969 2009/06/30 21:18:19 cheshire
-<rdar://problem/7020041> Plugging and unplugging the power cable shouldn't cause a network change event
-Additional fixes:
-1. Made mDNS_ActivateNetWake_internal and mDNS_DeactivateNetWake_internal more defensive against bad parameters
-2. mDNS_DeactivateNetWake_internal also needs to stop any outstanding Sleep Proxy resolve operations
-
-Revision 1.968 2009/06/29 23:51:09 cheshire
-<rdar://problem/6690034> Can't bind to Active Directory
-
-Revision 1.967 2009/06/27 00:25:27 cheshire
-<rdar://problem/6959273> mDNSResponder taking up 13% CPU with 400 KBps incoming bonjour requests
-Removed overly-complicate and ineffective multi-packet known-answer snooping code
-(Bracketed it with "#if ENABLE_MULTI_PACKET_QUERY_SNOOPING" for now; will delete actual code later)
-
-Revision 1.966 2009/06/26 01:55:55 cheshire
-<rdar://problem/6890712> mDNS: iChat's Buddy photo always appears as the "shadow person" over Bonjour
-Additional refinements -- except for the case of explicit queries for record types we don't have (for names we own),
-add additional NSEC records only when there's space to do that without having to generate an additional packet
-
-Revision 1.965 2009/06/24 22:14:21 cheshire
-<rdar://problem/6911445> Plugging and unplugging the power cable shouldn't cause a network change event
-
-Revision 1.964 2009/06/03 23:07:13 cheshire
-<rdar://problem/6890712> mDNS: iChat's Buddy photo always appears as the "shadow person" over Bonjour
-Large records were not being added in cases where an NSEC record was also required
-
-Revision 1.963 2009/05/28 00:39:19 cheshire
-<rdar://problem/6926465> Sleep is delayed by 10 seconds if BTMM is on
-After receiving confirmation of wide-area record deletion, need to schedule another evaluation of whether we're ready to sleep yet
-
-Revision 1.962 2009/05/19 23:40:37 cheshire
-<rdar://problem/6903507> Sleep Proxy: Retransmission logic not working reliably on quiet networks
-Added m->NextScheduledSPRetry timer for scheduling Sleep Proxy registration retries
-
-Revision 1.961 2009/05/19 23:00:43 cheshire
-Improved comments and debugging messages
-
-Revision 1.960 2009/05/13 17:25:33 mkrochma
-<rdar://problem/6879926> Should not schedule maintenance wake when machine has no advertised services
-Sleep proxy client should only look for services being advertised via Multicast
-
-Revision 1.959 2009/05/12 23:10:31 cheshire
-<rdar://problem/6879926> Should not schedule maintenance wake when machine has no advertised services
-Make new routine mDNSCoreHaveAdvertisedServices so daemon.c can tell whether it needs to schedule a maintenance wake
-
-Revision 1.958 2009/05/12 19:19:20 cheshire
-<rdar://problem/6879925> Sleep Proxy delays sleep by ten seconds when logged in to VPN
-
-Revision 1.957 2009/05/07 23:56:25 cheshire
-<rdar://problem/6601427> Retransmit and retry Sleep Proxy Server requests
-To get negative answers for our AAAA query we need to set the ReturnIntermed flag on the NetWakeResolve question
-
-Revision 1.956 2009/05/07 23:46:27 cheshire
-<rdar://problem/6601427> Retransmit and retry Sleep Proxy Server requests
-
-Revision 1.955 2009/05/07 23:40:54 cheshire
-Minor code rearrangement in preparation for upcoming changes
-
-Revision 1.954 2009/05/01 21:28:34 cheshire
-<rdar://problem/6721680> AppleConnectAgent's reachability checks delay sleep by 30 seconds
-No longer suspend network operations after we've acknowledged that the machine is going to sleep,
-because other software may not have yet acknowledged the sleep event, and may be still trying
-to do unicast DNS queries or other Bonjour operations.
-
-Revision 1.953 2009/05/01 19:17:35 cheshire
-<rdar://problem/6501561> Sleep Proxy: Reduce the frequency of maintenance wakes: ODD, fans, power
-
-Revision 1.952 2009/05/01 19:16:45 mcguire
-<rdar://problem/6846322> Crash: mDNS_vsnprintf + 1844
-
-Revision 1.951 2009/04/28 23:48:19 jessic2
-<rdar://problem/6830541> regservice_callback: instance->request is NULL 0
-
-Revision 1.950 2009/04/25 01:17:10 mcguire
-Fix spurious TCP connect failures uncovered by <rdar://problem/6729406> PPP doesn't automatically reconnect on wake from sleep
-
-Revision 1.949 2009/04/25 01:11:02 mcguire
-Refactor: create separate function: RestartRecordGetZoneData
-
-Revision 1.948 2009/04/24 21:25:16 cheshire
-<rdar://problem/6601002> Special case Net Assistant port so Apple Remote Desktop doesn't wake up every machine on the network
-
-Revision 1.947 2009/04/24 19:41:12 mcguire
-<rdar://problem/6791775> 4 second delay in DNS response
-
-Revision 1.946 2009/04/24 19:28:39 mcguire
-<rdar://problem/6791775> 4 second delay in DNS response
-
-Revision 1.945 2009/04/24 00:30:30 cheshire
-<rdar://problem/3476350> Return negative answers when host knows authoritatively that no answer exists
-Added code to generate and process NSEC records
-
-Revision 1.944 2009/04/23 22:06:29 cheshire
-Added CacheRecord and InterfaceID parameters to MakeNegativeCacheRecord, in preparation for:
-<rdar://problem/3476350> Return negative answers when host knows authoritatively that no answer exists
-
-Revision 1.943 2009/04/22 01:19:56 jessic2
-<rdar://problem/6814585> Daemon: mDNSResponder is logging garbage for error codes because it's using %ld for int 32
-
-Revision 1.942 2009/04/21 02:13:29 cheshire
-<rdar://problem/5270176> Local hostname changed even though there really isn't a name conflict
-Made code less susceptible to being tricked by stale packets echoed back from the network.
-
-Revision 1.941 2009/04/15 22:22:23 mcguire
-<rdar://problem/6768947> uDNS: Treat RCODE 5 (Refused) responses as failures
-Additional fix: protect against deref of NULL
-
-Revision 1.940 2009/04/15 20:42:51 mcguire
-<rdar://problem/6768947> uDNS: Treat RCODE 5 (Refused) responses as failures
-
-Revision 1.939 2009/04/11 00:19:32 jessic2
-<rdar://problem/4426780> Daemon: Should be able to turn on LogOperation dynamically
-
-Revision 1.938 2009/04/06 23:44:57 cheshire
-<rdar://problem/6757838> mDNSResponder thrashing kernel lock in the UDP close path, hurting SPECweb performance
-
-Revision 1.937 2009/04/04 00:14:49 mcguire
-fix logging in BeginSleepProcessing
-
-Revision 1.936 2009/04/04 00:10:59 mcguire
-don't ignore m->SystemWakeOnLANEnabled when going to sleep
-
-Revision 1.935 2009/04/01 17:50:11 mcguire
-cleanup mDNSRandom
-
-Revision 1.934 2009/03/27 17:17:58 cheshire
-Improved "Ignoring suspect uDNS response" debugging message
-
-Revision 1.933 2009/03/21 02:40:21 cheshire
-<rdar://problem/6704514> uDNS: Need to create negative cache entries for "local" SOA
-
-Revision 1.932 2009/03/20 23:53:03 jessic2
-<rdar://problem/6646228> SIGHUP should restart all in-progress queries
-
-Revision 1.931 2009/03/18 19:08:15 cheshire
-Show old/new sleep sequence numbers in logical order
-
-Revision 1.930 2009/03/17 23:40:45 cheshire
-For now only try the highest-ranked Sleep Proxy; fixed come compiler warnings
-
-Revision 1.929 2009/03/17 21:55:56 cheshire
-Fixed mistake in logic for decided when we're ready to go to sleep
-
-Revision 1.928 2009/03/17 19:48:12 cheshire
-<rdar://problem/6688927> Don't cache negative unicast answers for Multicast DNS names
-
-Revision 1.927 2009/03/17 01:22:56 cheshire
-<rdar://problem/6601427> Sleep Proxy: Retransmit and retry Sleep Proxy Server requests
-Initial support for resolving up to three Sleep Proxies in parallel
-
-Revision 1.926 2009/03/17 01:05:07 mcguire
-<rdar://problem/6657640> Reachability fixes on DNS config change
-
-Revision 1.925 2009/03/13 01:35:36 mcguire
-<rdar://problem/6657640> Reachability fixes on DNS config change
-
-Revision 1.924 2009/03/10 23:45:20 cheshire
-Added comments explaining usage of SetSPSProxyListChanged()
-
-Revision 1.923 2009/03/09 21:53:02 cheshire
-<rdar://problem/6650479> Sleep Proxy: Need to stop proxying when it sees an ARP probe from the client
-
-Revision 1.922 2009/03/09 21:30:17 cheshire
-Improved some LogSPS messages; made RestartProbing() subroutine
-
-Revision 1.921 2009/03/06 22:53:31 cheshire
-Don't bother registering with Sleep Proxy if we have no advertised services
-
-Revision 1.920 2009/03/06 20:08:55 cheshire
-<rdar://problem/6601429> Sleep Proxy: Return error responses to clients
-
-Revision 1.919 2009/03/05 21:54:43 cheshire
-Improved "Sleep Proxy Server started / stopped" message
-
-Revision 1.918 2009/03/04 01:37:14 cheshire
-<rdar://problem/6601428> Limit maximum number of records that a Sleep Proxy Server will accept
-
-Revision 1.917 2009/03/03 23:14:25 cheshire
-Got rid of code duplication by making subroutine "SetupOwnerOpt"
-
-Revision 1.916 2009/03/03 23:04:43 cheshire
-For clarity, renamed "MAC" field to "HMAC" (Host MAC, as opposed to Interface MAC)
-
-Revision 1.915 2009/03/03 22:51:53 cheshire
-<rdar://problem/6504236> Sleep Proxy: Waking on same network but different interface will cause conflicts
-
-Revision 1.914 2009/03/03 00:46:09 cheshire
-Additional debugging information in ResolveSimultaneousProbe
-
-Revision 1.913 2009/02/27 03:08:47 cheshire
-<rdar://problem/6547720> Crash while shutting down when "local" is in the user's DNS searchlist
-
-Revision 1.912 2009/02/27 02:31:28 cheshire
-Improved "Record not found in list" debugging message
-
-Revision 1.911 2009/02/21 01:42:11 cheshire
-Updated log messages
-
-Revision 1.910 2009/02/19 01:50:53 cheshire
-Converted some LogInfo messages to LogSPS
-
-Revision 1.909 2009/02/14 00:04:59 cheshire
-Left-justify interface names
-
-Revision 1.908 2009/02/13 19:40:07 cheshire
-Improved alignment of LogSPS messages
-
-Revision 1.907 2009/02/13 18:16:05 cheshire
-Fixed some compile warnings
-
-Revision 1.906 2009/02/13 06:10:17 cheshire
-Convert LogOperation messages to LogInfo
-
-Revision 1.905 2009/02/12 20:57:24 cheshire
-Renamed 'LogAllOperation' switch to 'LogClientOperations'; added new 'LogSleepProxyActions' switch
-
-Revision 1.904 2009/02/11 02:37:29 cheshire
-m->p->SystemWakeForNetworkAccessEnabled renamed to m->SystemWakeOnLANEnabled
-Moved code to send goodbye packets from mDNSCoreMachineSleep into BeginSleepProcessing,
-so that it happens correctly even when we delay re-sleep due to a very short wakeup.
-
-Revision 1.903 2009/02/09 23:34:31 cheshire
-Additional logging for debugging unknown packets
-
-Revision 1.902 2009/02/07 05:57:01 cheshire
-Fixed debugging log message
-
-Revision 1.901 2009/02/07 02:57:31 cheshire
-<rdar://problem/6084043> Sleep Proxy: Need to adopt IOPMConnection
-
-Revision 1.900 2009/02/02 21:29:24 cheshire
-<rdar://problem/4786302> Implement logic to determine when to send dot-local lookups via Unicast
-If Negative response for our special Microsoft Active Directory "local SOA" check has no
-SOA record in the authority section, assume we should cache the negative result for 24 hours
-
-Revision 1.899 2009/01/31 00:37:50 cheshire
-When marking cache records for deletion in response to a uDNS response,
-make sure InterfaceID matches (i.e. it should be NULL for a uDNS cache record)
-
-Revision 1.898 2009/01/30 23:49:20 cheshire
-Exclude mDNSInterface_Unicast from "InterfaceID ... not currently found" test
-
-Revision 1.897 2009/01/30 22:04:49 cheshire
-Workaround to reduce load on root name servers when caching the SOA record for "."
-
-Revision 1.896 2009/01/30 22:00:05 cheshire
-Made mDNS_StartQuery_internal pay attention to mDNSInterface_Unicast
-
-Revision 1.895 2009/01/30 17:46:39 cheshire
-Improved debugging messages for working out why spurious name conflicts are happening
-
-Revision 1.894 2009/01/30 00:22:09 cheshire
-<rdar://problem/6540743> No announcement after probing & no conflict notice
-
-Revision 1.893 2009/01/29 22:27:03 mcguire
-<rdar://problem/6407429> Cleanup: Logs about Unknown DNS packet type 5450
-
-Revision 1.892 2009/01/24 01:38:23 cheshire
-Fixed error in logic for targeted queries
-
-Revision 1.891 2009/01/22 02:14:25 cheshire
-<rdar://problem/6515626> Sleep Proxy: Set correct target MAC address, instead of all zeroes
-
-Revision 1.890 2009/01/22 00:45:02 cheshire
-Improved SPS debugging log messages; we are eligible to start answering ARP requests
-after we send our first announcement, not after we send our last probe
-
-Revision 1.889 2009/01/21 03:43:56 mcguire
-<rdar://problem/6511765> BTMM: Add support for setting kDNSServiceErr_NATPortMappingDisabled in DynamicStore
-
-Revision 1.888 2009/01/20 00:27:43 mcguire
-<rdar://problem/6305725> when removing a uDNS record, if a dup exists, copy information to it
-
-Revision 1.887 2009/01/17 05:14:37 cheshire
-Convert SendQueries Probe messages to LogSPS messages
-
-Revision 1.886 2009/01/17 03:43:09 cheshire
-Added SPSLogging switch to facilitate Sleep Proxy Server debugging
-
-Revision 1.885 2009/01/16 22:44:18 cheshire
-<rdar://problem/6402123> Sleep Proxy: Begin ARP Announcements sooner
-
-Revision 1.884 2009/01/16 21:43:52 cheshire
-Let InitializeLastAPTime compute the correct interval, instead of having it passed in as a parameter
-
-Revision 1.883 2009/01/16 21:11:18 cheshire
-When purging expired Sleep Proxy records, need to check DuplicateRecords list too
-
-Revision 1.882 2009/01/16 19:54:28 cheshire
-Use symbols "SleepProxyServiceType" and "localdomain" instead of literal strings
-
-Revision 1.881 2009/01/14 01:38:38 mcguire
-<rdar://problem/6492710> Write out DynamicStore per-interface SleepProxyServer info
-
-Revision 1.880 2009/01/10 01:51:19 cheshire
-q->CurrentAnswers not being incremented/decremented when answering a question with a local AuthRecord
-
-Revision 1.879 2009/01/10 01:43:52 cheshire
-Changed misleading function name 'AnsweredLOQ' to more informative 'AnsweredLocalQ'
-
-Revision 1.878 2009/01/10 01:38:10 cheshire
-Changed misleading function name 'AnswerLocalOnlyQuestionWithResourceRecord' to more informative 'AnswerLocalQuestionWithLocalAuthRecord'
-
-Revision 1.877 2009/01/10 01:36:08 cheshire
-Changed misleading function name 'AnswerLocalOnlyQuestions' to more informative 'AnswerAllLocalQuestionsWithLocalAuthRecord'
-
-Revision 1.876 2009/01/09 22:56:06 cheshire
-Don't touch rr after calling mDNS_Deregister_internal -- the memory may have been free'd
-
-Revision 1.875 2009/01/09 22:54:46 cheshire
-When tranferring record from DuplicateRecords list to ResourceRecords list,
-need to copy across state of 'Answered Local-Only-Questions' flag
-
-Revision 1.874 2009/01/07 23:07:24 cheshire
-<rdar://problem/6479416> SPS Client not canceling outstanding resolve call before sleeping
-
-Revision 1.873 2008/12/17 00:18:59 mkrochma
-Change some LogMsg to LogOperation before submitting
-
-Revision 1.872 2008/12/12 01:30:40 cheshire
-Update platform-layer BPF filters when we add or remove AddressProxy records
-
-Revision 1.871 2008/12/10 02:25:31 cheshire
-Minor fixes to use of LogClientOperations symbol
-
-Revision 1.870 2008/12/10 02:11:41 cheshire
-ARMv5 compiler doesn't like uncommented stuff after #endif
-
-Revision 1.869 2008/12/05 02:35:24 mcguire
-<rdar://problem/6107390> Write to the DynamicStore when a Sleep Proxy server is available on the network
-
-Revision 1.868 2008/12/04 21:08:51 mcguire
-<rdar://problem/6116863> mDNS: Provide mechanism to disable Multicast advertisements
-
-Revision 1.867 2008/11/26 21:19:36 cheshire
-<rdar://problem/6374334> Sleeping Server should choose the best Sleep Proxy by using advertised metrics
-
-Revision 1.866 2008/11/26 20:32:46 cheshire
-<rdar://problem/6374328> Sleep Proxy: Advertise BSP metrics in service name
-Update advertised name when Sleep Proxy "intent" metric changes
-
-Revision 1.865 2008/11/26 19:49:25 cheshire
-Record originally-requested port in sr->NATinfo.IntPort
-
-Revision 1.864 2008/11/26 19:02:37 cheshire
-Don't answer ARP Probes from owner machine as it wakes up and rejoins the network
-
-Revision 1.863 2008/11/26 03:59:03 cheshire
-Wait 30 seconds before starting ARP Announcements
-
-Revision 1.862 2008/11/25 23:43:07 cheshire
-<rdar://problem/5745355> Crashes at ServiceRegistrationGotZoneData + 397
-Made code more defensive to guard against ServiceRegistrationGotZoneData being called with invalid ServiceRecordSet object
-
-Revision 1.861 2008/11/25 22:46:30 cheshire
-For ease of code searching, renamed ZoneData field of ServiceRecordSet_struct from "nta" to "srs_nta"
-
-Revision 1.860 2008/11/25 05:07:15 cheshire
-<rdar://problem/6374328> Advertise Sleep Proxy metrics in service name
-
-Revision 1.859 2008/11/20 02:07:56 cheshire
-<rdar://problem/6387470> Refresh our NAT mappings on wake from sleep
-
-Revision 1.858 2008/11/20 01:38:36 cheshire
-For consistency with other parts of the code, changed code to only check
-that the first 4 bytes of MAC address are zero, not the whole 6 bytes.
-
-Revision 1.857 2008/11/14 22:55:18 cheshire
-Fixed log messages
-
-Revision 1.856 2008/11/14 21:08:28 cheshire
-Only put owner option in query packet if we have a non-zero MAC address to put
-Only process owner options in received query packets if the MAC address in the option is non-zero
-
-Revision 1.855 2008/11/14 02:29:54 cheshire
-If Sleep Proxy client fails to renew proxy records before they expire, remove them from our m->ResourceRecords list
-
-Revision 1.854 2008/11/14 00:00:53 cheshire
-After client machine wakes up, Sleep Proxy machine need to remove any records
-it was temporarily holding as proxy for that client
-
-Revision 1.853 2008/11/13 19:07:30 cheshire
-Added code to put OPT record, containing owner and lease lifetime, into SPS registration packet
-
-Revision 1.852 2008/11/12 23:23:11 cheshire
-Before waking a host, check to see if it has an SRV record advertising
-a service on the port in question, and if not, don't bother waking it.
-
-Revision 1.851 2008/11/12 01:54:15 cheshire
-<rdar://problem/6338021> Add domain back to end of _services._dns-sd._udp PTR records
-It turns out it is beneficial to have the domain on the end, because it allows better name compression
-
-Revision 1.850 2008/11/11 01:56:57 cheshire
-Improved name conflict log messages
-
-Revision 1.849 2008/11/06 23:50:43 cheshire
-Allow plain (non-SYN) ssh data packets to wake sleeping host
-
-Revision 1.848 2008/11/05 02:40:28 mkrochma
-Change mDNS_SetFQDN syslog mesage to debugf
-
-Revision 1.847 2008/11/04 23:06:50 cheshire
-Split RDataBody union definition into RDataBody and RDataBody2, and removed
-SOA from the normal RDataBody union definition, saving 270 bytes per AuthRecord
-
-Revision 1.846 2008/11/04 22:21:44 cheshire
-Changed zone field of AuthRecord_struct from domainname to pointer, saving 252 bytes per AuthRecord
-
-Revision 1.845 2008/11/03 23:52:05 cheshire
-Improved ARP debugging messages to differentiate ARP Announcements from Requests
-
-Revision 1.844 2008/10/31 23:43:51 cheshire
-Fixed compile error in Posix build
-
-Revision 1.843 2008/10/31 22:55:04 cheshire
-Initial support for structured SPS names
-
-Revision 1.842 2008/10/30 00:12:07 cheshire
-Fixed spin when PutSPSRec fails to put a record because it's too big to fit
-
-Revision 1.841 2008/10/29 23:23:38 cheshire
-Refined cache size reporting to go in steps of 1000 when number is above 1000
-
-Revision 1.840 2008/10/29 21:34:10 cheshire
-Removed some old debugging messages
-
-Revision 1.839 2008/10/29 21:31:32 cheshire
-Five seconds not always enough time for machine to go to sleep -- increased to ten seconds
-
-Revision 1.838 2008/10/28 18:30:37 cheshire
-Added debugging message in mDNSCoreReceiveRawPacket
-
-Revision 1.837 2008/10/24 23:58:05 cheshire
-Wake up for Back to My Mac IPSEC packets, except NAT keepalive packets
-
-Revision 1.836 2008/10/24 23:18:18 cheshire
-If we have a Sleep Proxy Server, don't remove service registrations from the DNS server
-
-Revision 1.835 2008/10/24 23:07:59 cheshire
-Wake SPS client if we receive conflicting mDNS respoonse (record with same name as one of our unique records, but different rdata)
-
-Revision 1.834 2008/10/24 23:03:24 cheshire
-Wake SPS client if we receive a conflicting ARP (some other machine claiming to own that IP address)
-
-Revision 1.833 2008/10/24 23:01:26 cheshire
-To reduce spurious wakeups for now, we'll only wake for incoming TCP SYN packets
-
-Revision 1.832 2008/10/24 22:58:24 cheshire
-For now, since we don't get IPv6 ND or data packets, don't advertise AAAA records for our SPS clients
-
-Revision 1.831 2008/10/24 22:50:41 cheshire
-When waking SPS client, include interface name in syslog message
-
-Revision 1.830 2008/10/24 20:50:34 cheshire
-Use "#if USE_SEPARATE_UDNS_SERVICE_LIST" instead of "#if defined(USE_SEPARATE_UDNS_SERVICE_LIST)"
-
-Revision 1.829 2008/10/23 23:55:57 cheshire
-Fixed some missing "const" declarations
-
-Revision 1.828 2008/10/23 22:25:56 cheshire
-Renamed field "id" to more descriptive "updateid"
-
-Revision 1.827 2008/10/23 03:06:25 cheshire
-Fixed "Waking host" log message
-
-Revision 1.826 2008/10/22 23:21:30 cheshire
-Make sure we have enough bytes before reading into the transport-level header
-
-Revision 1.825 2008/10/22 22:31:53 cheshire
-Log SYN/FIN/RST bits from TCP header, and don't wake for FIN/RST
-
-Revision 1.824 2008/10/22 20:00:31 cheshire
-If we ourselves go to sleep, stop advertising sleep proxy service, then re-advertise after we wake up
-
-Revision 1.823 2008/10/22 19:55:35 cheshire
-Miscellaneous fixes; renamed FindFirstAnswerInCache to FindSPSInCache
-
-Revision 1.822 2008/10/22 01:41:39 cheshire
-Set question->ThisQInterval back to -1 after we cancel our NetWakeResolve
-
-Revision 1.821 2008/10/22 01:12:53 cheshire
-Answer ARP Requests for any IP address we're proxying for
-
-Revision 1.820 2008/10/21 01:11:11 cheshire
-Added mDNSCoreReceiveRawPacket for handling raw packets received by platform layer
-
-Revision 1.819 2008/10/20 22:16:27 cheshire
-Updated comments; increased cache shedding threshold from 3000 to 4000
-
-Revision 1.818 2008/10/16 22:01:54 cheshire
-Fix last checkin: Should be "ar->resrec.rdata->u.data", not "ar->resrec.rdata.u.data"
-
-Revision 1.817 2008/10/16 21:40:49 cheshire
-Need to set ar->resrec.rdlength correctly before calling mDNS_Register_internal()
-
-Revision 1.816 2008/10/15 23:12:36 cheshire
-On receiving SPS registration from client, broadcast ARP Announcements claiming ownership of that IP address
-
-Revision 1.815 2008/10/15 20:46:38 cheshire
-When transferring records to SPS, include Lease Option
-
-Revision 1.814 2008/10/15 19:51:27 cheshire
-Change "NOTE:" to "Note:" so that BBEdit 9 stops putting those lines into the funtion popup menu
-
-Revision 1.813 2008/10/15 00:09:23 cheshire
-When acting as Sleep Proxy Server, handle DNS Updates received from SPS clients on the network
-
-Revision 1.812 2008/10/15 00:01:40 cheshire
-When going to sleep, discover and resolve SPS, and if successful, transfer records to it
-
-Revision 1.811 2008/10/14 23:51:57 cheshire
-Created new routine GetRDLengthMem() to compute the in-memory storage requirements for particular rdata
-
-Revision 1.810 2008/10/14 21:37:55 cheshire
-Removed unnecessary m->BeSleepProxyServer variable
-
-Revision 1.809 2008/10/10 23:45:48 cheshire
-For ForceMCast records, SetTargetToHostName should use the dot-local multicast hostname,
-not a wide-area unicast hostname
-
-Revision 1.808 2008/10/09 18:59:19 cheshire
-Added NetWakeResolve code, removed unused m->SendDeregistrations and m->SendImmediateAnswers
-
-Revision 1.807 2008/10/07 15:56:58 cheshire
-Fixed "unused variable" warnings in non-debug builds
-
-Revision 1.806 2008/10/04 00:53:37 cheshire
-On interfaces that support Wake-On-LAN, browse to discover Sleep Proxy Servers
-
-Revision 1.805 2008/10/03 18:17:28 cheshire
-<rdar://problem/6134215> Sleep Proxy: Mac with Internet Sharing should also offer Sleep Proxy service
-Update advertised Sleep Proxy Server name if user changes computer name
-
-Revision 1.804 2008/10/03 01:26:06 mcguire
-<rdar://problem/6266145> mDNS_FinalExit failed to send goodbye for duplicate uDNS records
-Put back Duplicate Record check
-
-Revision 1.803 2008/10/02 23:38:56 mcguire
-<rdar://problem/6266145> mDNS_FinalExit failed to send goodbye for duplicate uDNS records
-
-Revision 1.802 2008/10/02 23:13:48 cheshire
-<rdar://problem/6134215> Sleep Proxy: Mac with Internet Sharing should also offer Sleep Proxy service
-Need to drop lock before calling "mDNSCoreBeSleepProxyServer(m, mDNSfalse);"
-
-Revision 1.801 2008/10/02 22:51:04 cheshire
-<rdar://problem/6134215> Sleep Proxy: Mac with Internet Sharing should also offer Sleep Proxy service
-Added mDNSCoreBeSleepProxyServer() routine to start and stop Sleep Proxy Service
-
-Revision 1.800 2008/10/02 22:13:15 cheshire
-<rdar://problem/6230680> 100ms delay on shutdown
-Additional refinement: Also need to clear m->SuppressSending
-
-Revision 1.799 2008/09/29 20:12:37 cheshire
-Rename 'AnswerLocalQuestions' to more descriptive 'AnswerLocalOnlyQuestions' and 'AnsweredLocalQ' to 'AnsweredLOQ'
-
-Revision 1.798 2008/09/26 19:53:14 cheshire
-Fixed locking error: should not call mDNS_Deregister_internal within "mDNS_DropLock" section
-
-Revision 1.797 2008/09/25 20:40:59 cheshire
-<rdar://problem/6245044> Stop using separate m->ServiceRegistrations list
-In mDNS_SetFQDN, need to update all AutoTarget SRV records, even if m->MulticastHostname hasn't changed
-
-Revision 1.796 2008/09/25 20:17:10 cheshire
-<rdar://problem/6245044> Stop using separate m->ServiceRegistrations list
-Added defensive code to make sure *all* records of a ServiceRecordSet have
-completed deregistering before we pass on the mStatus_MemFree message
-
-Revision 1.795 2008/09/25 00:30:11 cheshire
-<rdar://problem/6245044> Stop using separate m->ServiceRegistrations list
-
-Revision 1.794 2008/09/24 23:48:05 cheshire
-Don't need to pass whole ServiceRecordSet reference to GetServiceTarget;
-it only needs to access the embedded SRV member of the set
-
-Revision 1.793 2008/09/23 04:11:53 cheshire
-<rdar://problem/6238774> Remove "local" from the end of _services._dns-sd._udp PTR records
-
-Revision 1.792 2008/09/23 02:30:07 cheshire
-Get rid of PutResourceRecordCappedTTL()
-
-Revision 1.791 2008/09/20 00:34:21 mcguire
-<rdar://problem/6129039> BTMM: Add support for WANPPPConnection
-
-Revision 1.790 2008/09/18 22:46:34 cheshire
-<rdar://problem/6230680> 100ms delay on shutdown
-
-Revision 1.789 2008/09/18 06:15:06 mkrochma
-<rdar://problem/6117156> Cleanup: mDNSResponder logging debugging information to console
-
-Revision 1.788 2008/09/16 21:11:41 cheshire
-<rdar://problem/6223969> mDNS: Duplicate TXT record queries being produced by iPhone Remote
-
-Revision 1.787 2008/09/05 22:53:24 cheshire
-Improve "How is rr->resrec.rroriginalttl <= SecsSinceRcvd" debugging message
-
-Revision 1.786 2008/09/05 22:23:28 cheshire
-Moved initialization of "question->LocalSocket" to more logical place
-
-Revision 1.785 2008/08/14 19:20:55 cheshire
-<rdar://problem/6143846> Negative responses over TCP incorrectly rejected
-
-Revision 1.784 2008/08/13 00:47:53 mcguire
-Handle failures when packet logging
-
-Revision 1.783 2008/07/25 07:09:51 mcguire
-<rdar://problem/3988320> Should use randomized source ports and transaction IDs to avoid DNS cache poisoning
-
-Revision 1.782 2008/07/24 20:23:03 cheshire
-<rdar://problem/3988320> Should use randomized source ports and transaction IDs to avoid DNS cache poisoning
-
-Revision 1.781 2008/07/18 21:37:35 mcguire
-<rdar://problem/5736845> BTMM: alternate SSDP queries between multicast & unicast
-
-Revision 1.780 2008/07/18 02:24:36 cheshire
-<rdar://problem/6041178> Only trigger reconfirm on hostname if both A and AAAA query fail to elicit a response
-Additional fix: Don't want to do the ReconfirmAntecedents() stuff if q->RequestUnicast is set (that indicates
-we're still on our first or second query after an interface registration or wake from sleep).
-
-Revision 1.779 2008/07/18 01:05:23 cheshire
-<rdar://problem/6041178> Only trigger reconfirm on hostname if both A and AAAA query fail to elicit a response
-
-Revision 1.778 2008/06/26 17:24:11 mkrochma
-<rdar://problem/5450912> BTMM: Stop listening on UDP 5351 for NAT Status Announcements
-
-Revision 1.777 2008/06/19 01:20:48 mcguire
-<rdar://problem/4206534> Use all configured DNS servers
-
-Revision 1.776 2008/04/17 20:14:14 cheshire
-<rdar://problem/5870023> CurrentAnswers/LargeAnswers/UniqueAnswers counter mismatch
-
-Revision 1.775 2008/03/26 01:53:34 mcguire
-<rdar://problem/5820489> Can't resolve via uDNS when an interface is specified
-
-Revision 1.774 2008/03/17 17:46:08 mcguire
-When activating an LLQ, reset all the important state and destroy any tcp connection,
-so that everything will be restarted as if the question had just been asked.
-Also reset servPort, so that the SOA query will be re-issued.
-
-Revision 1.773 2008/03/14 22:52:36 mcguire
-<rdar://problem/5321824> write status to the DS
-Update status when any unicast LLQ is started
-
-Revision 1.772 2008/03/06 02:48:34 mcguire
-<rdar://problem/5321824> write status to the DS
-
-Revision 1.771 2008/02/26 22:04:44 cheshire
-<rdar://problem/5661661> BTMM: Too many members.mac.com SOA queries
-Additional fixes -- should not be calling uDNS_CheckCurrentQuestion on a
-question while it's still in our 'm->NewQuestions' section of the list
-
-Revision 1.770 2008/02/22 23:09:02 cheshire
-<rdar://problem/5338420> BTMM: Not processing additional records
-Refinements:
-1. Check rdatahash == namehash, to skip expensive SameDomainName check when possible
-2. Once we decide a record is acceptable, we can break out of the loop
-
-Revision 1.769 2008/02/22 00:00:19 cheshire
-<rdar://problem/5338420> BTMM: Not processing additional records
-
-Revision 1.768 2008/02/19 23:26:50 cheshire
-<rdar://problem/5661661> BTMM: Too many members.mac.com SOA queries
-
-Revision 1.767 2007/12/22 02:25:29 cheshire
-<rdar://problem/5661128> Records and Services sometimes not re-registering on wake from sleep
-
-Revision 1.766 2007/12/15 01:12:27 cheshire
-<rdar://problem/5526796> Need to remove active LLQs from server upon question cancellation, on sleep, and on shutdown
-
-Revision 1.765 2007/12/15 00:18:51 cheshire
-Renamed question->origLease to question->ReqLease
-
-Revision 1.764 2007/12/14 00:49:53 cheshire
-Fixed crash in mDNS_StartExit -- the service deregistration loop needs to use
-the CurrentServiceRecordSet mechanism to guard against services being deleted,
-just like the record deregistration loop uses m->CurrentRecord.
-
-Revision 1.763 2007/12/13 20:20:17 cheshire
-Minor efficiency tweaks -- converted IdenticalResourceRecord, IdenticalSameNameRecord, and
-SameRData from functions to macros, which allows the code to be inlined (the compiler can't
-inline a function defined in a different compilation unit) and therefore optimized better.
-
-Revision 1.762 2007/12/13 00:13:03 cheshire
-Simplified RDataHashValue to take a single ResourceRecord pointer, instead of separate rdlength and RDataBody
-
-Revision 1.761 2007/12/13 00:03:31 cheshire
-Improved efficiency in IdenticalResourceRecord() by doing SameRData() check before SameDomainName() check
-
-Revision 1.760 2007/12/08 00:36:19 cheshire
-<rdar://problem/5636422> Updating TXT records is too slow
-Remove unnecessary delays on announcing record updates, and on processing them on reception
-
-Revision 1.759 2007/12/07 22:41:29 cheshire
-<rdar://problem/5526800> BTMM: Need to clean up registrations on shutdown
-Further refinements -- records on the DuplicateRecords list were getting missed on shutdown
-
-Revision 1.758 2007/12/07 00:45:57 cheshire
-<rdar://problem/5526800> BTMM: Need to clean up registrations on shutdown
-
-Revision 1.757 2007/12/06 00:22:27 mcguire
-<rdar://problem/5604567> BTMM: Doesn't work with Linksys WAG300N 1.01.06 (sending from 1026/udp)
-
-Revision 1.756 2007/12/05 01:52:30 cheshire
-<rdar://problem/5624763> BTMM: getaddrinfo_async_start returns EAI_NONAME when resolving BTMM hostname
-Delay returning IPv4 address ("A") results for autotunnel names until after we've set up the tunnel (or tried to)
-
-Revision 1.755 2007/12/03 23:36:45 cheshire
-<rdar://problem/5623140> mDNSResponder unicast DNS improvements
-Need to check GetServerForName() result is non-null before dereferencing pointer
-
-Revision 1.754 2007/12/01 01:21:27 jgraessley
-<rdar://problem/5623140> mDNSResponder unicast DNS improvements
-
-Revision 1.753 2007/12/01 00:44:15 cheshire
-Fixed compile warnings, e.g. declaration of 'rr' shadows a previous local
-
-Revision 1.752 2007/11/14 01:10:51 cheshire
-Fixed LogOperation() message wording
-
-Revision 1.751 2007/10/30 23:49:41 cheshire
-<rdar://problem/5519458> BTMM: Machines don't appear in the sidebar on wake from sleep
-LLQ state was not being transferred properly between duplicate questions
-
-Revision 1.750 2007/10/29 23:58:52 cheshire
-<rdar://problem/5536979> BTMM: Need to create NAT port mapping for receiving LLQ events
-Use standard "if (mDNSIPv4AddressIsOnes(....ExternalAddress))" mechanism to determine whether callback has been invoked yet
-
-Revision 1.749 2007/10/29 21:28:36 cheshire
-Change "Correcting TTL" log message to LogOperation to suppress it in customer build
-
-Revision 1.748 2007/10/29 20:02:50 cheshire
-<rdar://problem/5526813> BTMM: Wide-area records being announced via multicast
-
-Revision 1.747 2007/10/26 22:53:50 cheshire
-Made mDNS_Register_internal and mDNS_Deregister_internal use AuthRecord_uDNS macro
-instead of replicating the logic in both places
-
-Revision 1.746 2007/10/25 22:48:50 cheshire
-Added LogOperation message saying when we restart GetZoneData for record and service registrations
-
-Revision 1.745 2007/10/25 20:48:47 cheshire
-For naming consistency (with AuthRecord's UpdateServer) renamed 'ns' to 'SRSUpdateServer'
-
-Revision 1.744 2007/10/25 20:06:14 cheshire
-Don't try to do SOA queries using private DNS (TLS over TCP) queries
-
-Revision 1.743 2007/10/25 00:12:46 cheshire
-<rdar://problem/5496734> BTMM: Need to retry registrations after failures
-Retrigger service registrations whenever a new network interface is added
-
-Revision 1.742 2007/10/24 22:40:06 cheshire
-Renamed: RecordRegistrationCallback -> RecordRegistrationGotZoneData
-Renamed: ServiceRegistrationZoneDataComplete -> ServiceRegistrationGotZoneData
-
-Revision 1.741 2007/10/24 00:50:29 cheshire
-<rdar://problem/5496734> BTMM: Need to retry registrations after failures
-Retrigger record registrations whenever a new network interface is added
-
-Revision 1.740 2007/10/23 00:38:03 cheshire
-When sending uDNS cache expiration query, need to increment rr->UnansweredQueries
-or code will spin sending the same cache expiration query repeatedly
-
-Revision 1.739 2007/10/22 23:46:41 cheshire
-<rdar://problem/5519458> BTMM: Machines don't appear in the sidebar on wake from sleep
-Need to clear question->nta pointer after calling CancelGetZoneData()
-
-Revision 1.738 2007/10/19 22:08:49 cheshire
-<rdar://problem/5519458> BTMM: Machines don't appear in the sidebar on wake from sleep
-Additional fixes and refinements
-
-Revision 1.737 2007/10/18 23:06:42 cheshire
-<rdar://problem/5519458> BTMM: Machines don't appear in the sidebar on wake from sleep
-Additional fixes and refinements
-
-Revision 1.736 2007/10/18 20:23:17 cheshire
-Moved SuspendLLQs into mDNS.c, since it's only called from one place
-
-Revision 1.735 2007/10/18 00:12:34 cheshire
-Fixed "unused variable" compiler warning
-
-Revision 1.734 2007/10/17 22:49:54 cheshire
-<rdar://problem/5519458> BTMM: Machines don't appear in the sidebar on wake from sleep
-
-Revision 1.733 2007/10/17 22:37:23 cheshire
-<rdar://problem/5536979> BTMM: Need to create NAT port mapping for receiving LLQ events
-
-Revision 1.732 2007/10/17 21:53:51 cheshire
-Improved debugging messages; renamed startLLQHandshakeCallback to LLQGotZoneData
-
-Revision 1.731 2007/10/17 18:37:50 cheshire
-<rdar://problem/5539930> Goodbye packets not being sent for services on shutdown
-Further refinement: pre-increment m->CurrentRecord before calling mDNS_Deregister_internal()
-
-Revision 1.730 2007/10/16 21:16:07 cheshire
-<rdar://problem/5539930> Goodbye packets not being sent for services on shutdown
-
-Revision 1.729 2007/10/05 17:56:10 cheshire
-Move CountLabels and SkipLeadingLabels to DNSCommon.c so they're callable from other files
-
-Revision 1.728 2007/10/04 23:18:14 cheshire
-<rdar://problem/5523706> mDNSResponder flooding DNS servers with unreasonable query level
-
-Revision 1.727 2007/10/04 22:51:57 cheshire
-Added debugging LogOperation message to show when we're sending cache expiration queries
-
-Revision 1.726 2007/10/03 00:14:24 cheshire
-Removed write to null to generate stack trace for SetNextQueryTime locking failure
-
-Revision 1.725 2007/10/02 21:11:08 cheshire
-<rdar://problem/5518270> LLQ refreshes don't work, which breaks BTMM browsing
-
-Revision 1.724 2007/10/02 20:10:23 cheshire
-Additional debugging checks on shutdown -- list all records we didn't send a goodbye for, not just the first one
-
-Revision 1.723 2007/10/02 19:56:54 cheshire
-<rdar://problem/5518310> Double-dispose causes crash changing Dynamic DNS hostname
-
-Revision 1.722 2007/10/01 22:59:46 cheshire
-<rdar://problem/5516303> mDNSResponder did not shut down after 20 seconds
-Need to shut down NATTraversals on exit
-
-Revision 1.721 2007/10/01 18:42:07 cheshire
-To make packet logs appear in a more intuitive order, dump received packets *before* handling them, not after
-
-Revision 1.720 2007/09/29 20:40:19 cheshire
-<rdar://problem/5513378> Crash in ReissueBlockedQuestions
-
-Revision 1.719 2007/09/27 22:23:56 cheshire
-<rdar://problem/4947392> uDNS: Use SOA to determine TTL for negative answers
-Need to clear m->rec.r.resrec.RecordType after we've finished using m->rec
-
-Revision 1.718 2007/09/27 22:02:33 cheshire
-<rdar://problem/5464941> BTMM: Registered records in BTMM don't get removed from server after calling RemoveRecord
-
-Revision 1.717 2007/09/27 21:21:39 cheshire
-Export CompleteDeregistration so it's callable from other files
-
-Revision 1.716 2007/09/27 02:12:21 cheshire
-Updated GrantCacheExtensions degugging message to show new record lifetime
-
-Revision 1.715 2007/09/27 01:20:06 cheshire
-<rdar://problem/5500077> BTMM: Need to refresh LLQs based on lease life and not TTL of response
-
-Revision 1.714 2007/09/27 00:37:01 cheshire
-<rdar://problem/4947392> BTMM: Use SOA to determine TTL for negative answers
-
-Revision 1.713 2007/09/27 00:25:39 cheshire
-Added ttl_seconds parameter to MakeNegativeCacheRecord in preparation for:
-<rdar://problem/4947392> uDNS: Use SOA to determine TTL for negative answers
-
-Revision 1.712 2007/09/26 23:16:58 cheshire
-<rdar://problem/5496399> BTMM: Leopard sending excessive LLQ registration requests to .Mac
-
-Revision 1.711 2007/09/26 22:06:02 cheshire
-<rdar://problem/5507399> BTMM: No immediate failure notifications for BTMM names
-
-Revision 1.710 2007/09/26 00:49:46 cheshire
-Improve packet logging to show sent and received packets,
-transport protocol (UDP/TCP/TLS) and source/destination address:port
-
-Revision 1.709 2007/09/21 21:12:36 cheshire
-<rdar://problem/5498009> BTMM: Need to log updates and query packet contents
-
-Revision 1.708 2007/09/20 23:13:37 cheshire
-<rdar://problem/4038277> BTMM: Not getting LLQ remove events when logging out of VPN or disconnecting from network
-Additional fix: If we have no DNS servers at all, then immediately purge all unicast cache records (including for LLQs)
-
-Revision 1.707 2007/09/20 02:29:37 cheshire
-<rdar://problem/4038277> BTMM: Not getting LLQ remove events when logging out of VPN or disconnecting from network
-
-Revision 1.706 2007/09/20 01:13:19 cheshire
-Export CacheGroupForName so it's callable from other files
-
-Revision 1.705 2007/09/20 01:12:06 cheshire
-Moved HashSlot(X) from mDNS.c to DNSCommon.h so it's usable in other files
-
-Revision 1.704 2007/09/19 22:47:25 cheshire
-<rdar://problem/5490182> Memory corruption freeing a "no such service" service record
-
-Revision 1.703 2007/09/14 01:46:59 cheshire
-Fix Posix build (#ifdef _LEGACY_NAT_TRAVERSAL_ section included a closing curly brace it should not have)
-
-Revision 1.702 2007/09/13 22:06:46 cheshire
-<rdar://problem/5480643> Tully's Free WiFi: DNS fails
-Need to accept DNS responses where the query ID field matches, even if the source address does not
-
-Revision 1.701 2007/09/12 23:22:32 cheshire
-<rdar://problem/5476979> Only accept NAT Port Mapping packets from our default gateway
-
-Revision 1.700 2007/09/12 23:03:08 cheshire
-<rdar://problem/5476978> DNSServiceNATPortMappingCreate callback not giving correct interface index
-
-Revision 1.699 2007/09/12 22:19:28 cheshire
-<rdar://problem/5476977> Need to listen for port 5350 NAT-PMP announcements
-
-Revision 1.698 2007/09/12 22:13:27 cheshire
-Remove DynDNSHostNames cleanly on shutdown
-
-Revision 1.697 2007/09/12 01:44:47 cheshire
-<rdar://problem/5475938> Eliminate "Correcting TTL" syslog messages for unicast DNS records
-
-Revision 1.696 2007/09/12 01:26:08 cheshire
-Initialize LastNATReplyLocalTime to timenow, so that gateway uptime checks work more reliably
-
-Revision 1.695 2007/09/11 19:19:16 cheshire
-Correct capitalization of "uPNP" to "UPnP"
-
-Revision 1.694 2007/09/10 22:06:51 cheshire
-Rename uptime => upseconds and LastNATUptime => LastNATupseconds to make it clear these time values are in seconds
-
-Revision 1.693 2007/09/07 22:24:36 vazquez
-<rdar://problem/5466301> Need to stop spewing mDNSResponderHelper logs
-
-Revision 1.692 2007/09/07 00:12:09 cheshire
-<rdar://problem/5466010> Unicast DNS changes broke efficiency fix 3928456
-
-Revision 1.691 2007/09/05 22:25:01 vazquez
-<rdar://problem/5400521> update_record mDNSResponder leak
-
-Revision 1.690 2007/09/05 21:48:01 cheshire
-<rdar://problem/5385864> BTMM: mDNSResponder flushes wide-area Bonjour records after an hour for a zone.
-Now that we're respecting the TTL of uDNS records in the cache, the LLQ maintenance code needs
-to update the cache lifetimes of all relevant records every time it successfully renews an LLQ,
-otherwise those records will expire and vanish from the cache.
-
-Revision 1.689 2007/09/05 02:29:06 cheshire
-<rdar://problem/5457287> mDNSResponder taking up 100% CPU in ReissueBlockedQuestions
-Additional fixes to code implementing "NoAnswer" logic
-
-Revision 1.688 2007/08/31 22:56:39 cheshire
-<rdar://problem/5407080> BTMM: TTLs incorrect on cached BTMM records
-
-Revision 1.687 2007/08/31 19:53:14 cheshire
-<rdar://problem/5431151> BTMM: IPv6 address lookup should not succeed if autotunnel cannot be setup
-If AutoTunnel setup fails, the code now generates a fake NXDomain error saying that the requested AAAA record does not exist
-
-Revision 1.686 2007/08/30 00:01:56 cheshire
-Added comment about SetTargetToHostName()
-
-Revision 1.685 2007/08/29 01:19:24 cheshire
-<rdar://problem/5400181> BTMM: Tunneled services do not need NAT port mappings
-Set AutoTarget to Target_AutoHostAndNATMAP for non-AutoTunnel wide-area services
-
-Revision 1.684 2007/08/28 23:58:42 cheshire
-Rename HostTarget -> AutoTarget
-
-Revision 1.683 2007/08/28 23:53:21 cheshire
-Rename serviceRegistrationCallback -> ServiceRegistrationZoneDataComplete
-
-Revision 1.682 2007/08/27 20:28:19 cheshire
-Improve "suspect uDNS response" log message
-
-Revision 1.681 2007/08/24 23:37:23 cheshire
-Added debugging message to show when ExtraResourceRecord callback gets invoked
-
-Revision 1.680 2007/08/24 00:15:19 cheshire
-Renamed GetAuthInfoForName() to GetAuthInfoForName_internal() to make it clear that it may only be called with the lock held
-
-Revision 1.679 2007/08/23 21:47:09 vazquez
-<rdar://problem/5427316> BTMM: mDNSResponder sends NAT-PMP packets on public network
-make sure we clean up port mappings on base stations by sending a lease value of 0,
-and only send NAT-PMP packets on private networks; also save some memory by
-not using packet structs in NATTraversals.
-
-Revision 1.678 2007/08/01 16:09:13 cheshire
-Removed unused NATTraversalInfo substructure from AuthRecord; reduced structure sizecheck values accordingly
-
-Revision 1.677 2007/08/01 01:58:24 cheshire
-Added RecordType sanity check in mDNS_Register_internal
-
-Revision 1.676 2007/08/01 00:04:13 cheshire
-<rdar://problem/5261696> Crash in tcpKQSocketCallback
-Half-open TCP connections were not being cancelled properly
-
-Revision 1.675 2007/07/31 02:28:35 vazquez
-<rdar://problem/3734269> NAT-PMP: Detect public IP address changes and base station reboot
-
-Revision 1.674 2007/07/31 01:57:23 cheshire
-Adding code to respect TTL received in uDNS responses turned out to
-expose other problems; backing out change for now.
-
-Revision 1.673 2007/07/30 23:31:26 cheshire
-Code for respecting TTL received in uDNS responses should exclude LLQ-type responses
-
-Revision 1.672 2007/07/28 01:25:56 cheshire
-<rdar://problem/4780038> BTMM: Add explicit UDP event port to LLQ setup request, to fix LLQs not working behind NAT
-
-Revision 1.671 2007/07/27 22:32:54 cheshire
-When processing TTLs in uDNS responses, we'll currently impose a minimum effective TTL
-of 2 seconds, or other stuff breaks (e.g. we end up making a negative cache entry).
-
-Revision 1.670 2007/07/27 20:54:43 cheshire
-Fixed code to respect real record TTL received in uDNS responses
-
-Revision 1.669 2007/07/27 20:09:32 cheshire
-Don't need to dump out all received mDNS packets; they're easily viewed using mDNSNetMonitor
-
-Revision 1.668 2007/07/27 19:58:47 cheshire
-Use symbolic names QC_add and QC_rmv instead of mDNStrue/mDNSfalse
-
-Revision 1.667 2007/07/27 19:52:10 cheshire
-Don't increment m->rrcache_active for no-cache add events
-
-Revision 1.666 2007/07/27 19:30:39 cheshire
-Changed mDNSQuestionCallback parameter from mDNSBool to QC_result,
-to properly reflect tri-state nature of the possible responses
-
-Revision 1.665 2007/07/27 18:44:01 cheshire
-Rename "AnswerQuestionWithResourceRecord" to more informative "AnswerCurrentQuestionWithResourceRecord"
-
-Revision 1.664 2007/07/27 18:38:56 cheshire
-Rename "uDNS_CheckQuery" to more informative "uDNS_CheckCurrentQuestion"
-
-Revision 1.663 2007/07/25 03:05:02 vazquez
-Fixes for:
-<rdar://problem/5338913> LegacyNATTraversal: UPnP heap overflow
-<rdar://problem/5338933> LegacyNATTraversal: UPnP stack buffer overflow
-and a myriad of other security problems
-
-Revision 1.662 2007/07/24 20:22:46 cheshire
-Make sure all fields of main mDNS object are initialized correctly
-
-Revision 1.661 2007/07/21 00:54:45 cheshire
-<rdar://problem/5344576> Delay IPv6 address callback until AutoTunnel route and policy is configured
-
-Revision 1.660 2007/07/20 20:00:45 cheshire
-"Legacy Browse" is better called "Automatic Browse"
-
-Revision 1.659 2007/07/20 00:54:18 cheshire
-<rdar://problem/4641118> Need separate SCPreferences for per-user .Mac settings
-
-Revision 1.658 2007/07/18 02:28:57 cheshire
-Don't set AutoTunnel settings in uDNS_RegisterService; should be done in GetServiceTarget
-
-Revision 1.657 2007/07/18 00:57:10 cheshire
-<rdar://problem/5303834> Automatically configure IPSec policy when resolving services
-Only need to call AddNewClientTunnel() for IPv6 addresses
-
-Revision 1.656 2007/07/16 23:54:48 cheshire
-<rdar://problem/5338850> Crash when removing or changing DNS keys
-
-Revision 1.655 2007/07/16 20:11:37 vazquez
-<rdar://problem/3867231> LegacyNATTraversal: Need complete rewrite
-Init LNT stuff and handle SSDP packets
-
-Revision 1.654 2007/07/12 23:30:23 cheshire
-Changed some 'LogOperation' calls to 'debugf' to reduce verbosity in syslog
-
-Revision 1.653 2007/07/12 02:51:27 cheshire
-<rdar://problem/5303834> Automatically configure IPSec policy when resolving services
-
-Revision 1.652 2007/07/11 23:43:42 cheshire
-Rename PurgeCacheResourceRecord to mDNS_PurgeCacheResourceRecord
-
-Revision 1.651 2007/07/11 22:44:40 cheshire
-<rdar://problem/5328801> SIGHUP should purge the cache
-
-Revision 1.650 2007/07/11 21:34:09 cheshire
-<rdar://problem/5304766> Register IPSec tunnel with IPv4-only hostname and create NAT port mappings
-Need to hold mDNS_Lock when calling mDNS_AddDynDNSHostName/mDNS_RemoveDynDNSHostName
-
-Revision 1.649 2007/07/11 02:52:52 cheshire
-<rdar://problem/5303807> Register IPv6-only hostname and don't create port mappings for AutoTunnel services
-In uDNS_RegisterService, set HostTarget for AutoTunnel services
-
-Revision 1.648 2007/07/09 23:48:12 cheshire
-Add parentheses around bitwise operation for clarity
-
-Revision 1.647 2007/07/06 21:17:55 cheshire
-Initialize m->retryGetAddr to timenow + 0x78000000;
-
-Revision 1.646 2007/07/06 18:55:49 cheshire
-Initialize m->NextScheduledNATOp
-
-Revision 1.645 2007/06/29 22:55:54 cheshire
-Move declaration of DNSServer *s; Fixed incomplete comment.
-
-Revision 1.644 2007/06/29 00:07:29 vazquez
-<rdar://problem/5301908> Clean up NAT state machine (necessary for 6 other fixes)
-
-Revision 1.643 2007/06/20 01:10:12 cheshire
-<rdar://problem/5280520> Sync iPhone changes into main mDNSResponder code
-
-Revision 1.642 2007/06/15 21:54:50 cheshire
-<rdar://problem/4883206> Add packet logging to help debugging private browsing over TLS
-
-Revision 1.641 2007/05/25 00:30:24 cheshire
-When checking for duplicate questions, make sure privacy (or not) status, and long-lived (or not)
-status matches. This is particularly important when doing a private query for an SOA record,
-which will result in a call StartGetZoneData which does a non-private query for the same SOA record.
-If the latter is tagged as a duplicate of the former, then we have deadlock, and neither will complete.
-
-Revision 1.640 2007/05/25 00:25:44 cheshire
-<rdar://problem/5227737> Need to enhance putRData to output all current known types
-
-Revision 1.639 2007/05/23 00:51:33 cheshire
-Increase threshold for shedding cache records from 512 to 3000. The 512 figure was calculated when
-each cache entry took about 700 bytes; now they're only 164 bytes. Also, machines have more RAM these
-days, and there are more services being advertised using DNS-SD, so it makes sense to cache more.
-
-Revision 1.638 2007/05/23 00:43:16 cheshire
-If uDNS UDP response has TC (truncated) bit set, don't interpret it as being the entire RRSet
-
-Revision 1.637 2007/05/14 23:53:00 cheshire
-Export mDNS_StartQuery_internal and mDNS_StopQuery_internal so they can be called from uDNS.c
-
-Revision 1.636 2007/05/10 23:27:15 cheshire
-Update mDNS_Deregister_internal debugging messages
-
-Revision 1.635 2007/05/07 20:43:45 cheshire
-<rdar://problem/4241419> Reduce the number of queries and announcements
-
-Revision 1.634 2007/05/04 22:09:08 cheshire
-Only do "restarting exponential backoff sequence" for mDNS questions
-In mDNS_RegisterInterface, only retrigger mDNS questions
-In uDNS_SetupDNSConfig, use ActivateUnicastQuery() instead of just setting q->ThisQInterval directly
-
-Revision 1.633 2007/05/04 21:45:12 cheshire
-Get rid of unused q->RestartTime; Get rid of uDNS_Close (synonym for uDNS_Sleep)
-
-Revision 1.632 2007/05/04 20:20:50 cheshire
-<rdar://problem/5167331> RegisterRecord and RegisterService need to cancel StartGetZoneData
-Need to set srs->nta = mDNSNULL; when regState_NoTarget
-
-Revision 1.631 2007/05/04 00:39:42 cheshire
-<rdar://problem/4410011> Eliminate looping SOA lookups
-When creating a cascade of negative SOA cache entries, CacheGroup pointer cg needs to be updated
-each time round the loop to reference the right CacheGroup for each newly fabricated SOA name
-
-Revision 1.630 2007/05/03 22:40:38 cheshire
-<rdar://problem/4669229> mDNSResponder ignores bogus null target in SRV record
-
-Revision 1.629 2007/05/03 00:15:51 cheshire
-<rdar://problem/4410011> Eliminate looping SOA lookups
-
-Revision 1.628 2007/05/02 22:21:33 cheshire
-<rdar://problem/5167331> RegisterRecord and RegisterService need to cancel StartGetZoneData
-
-Revision 1.627 2007/04/30 19:29:13 cheshire
-Fix display of port number in "Updating DNS Server" message
-
-Revision 1.626 2007/04/30 04:21:13 cheshire
-Can't safely call AnswerLocalQuestions() from within mDNS_Deregister() -- need to defer it until mDNS_Execute time
-
-Revision 1.625 2007/04/28 01:34:21 cheshire
-Fixed crashing bug: We need to update rr->CRActiveQuestion pointers for *all* questions
-(Code was explicitly ignoring wide-area unicast questions, leading to stale pointers and crashes)
-
-Revision 1.624 2007/04/27 21:04:30 cheshire
-On network configuration change, need to call uDNS_RegisterSearchDomains
-
-Revision 1.623 2007/04/27 19:28:01 cheshire
-Any code that calls StartGetZoneData needs to keep a handle to the structure, so
-it can cancel it if necessary. (First noticed as a crash in Apple Remote Desktop
--- it would start a query and then quickly cancel it, and then when
-StartGetZoneData completed, it had a dangling pointer and crashed.)
-
-Revision 1.622 2007/04/26 16:09:22 cheshire
-mDNS_StopQueryWithRemoves should ignore kDNSRecordTypePacketNegative records
-
-Revision 1.621 2007/04/26 15:43:22 cheshire
-Make sure DNSServer *s is non-null before using value in LogOperation
-
-Revision 1.620 2007/04/26 13:11:05 cheshire
-Fixed crash when logging out of VPN
-
-Revision 1.619 2007/04/26 00:35:15 cheshire
-<rdar://problem/5140339> uDNS: Domain discovery not working over VPN
-Fixes to make sure results update correctly when connectivity changes (e.g. a DNS server
-inside the firewall may give answers where a public one gives none, and vice versa.)
-
-Revision 1.618 2007/04/25 19:26:01 cheshire
-m->NextScheduledQuery was getting set too early in SendQueries()
-Improved "SendQueries didn't send all its queries" debugging message
-
-Revision 1.617 2007/04/25 17:48:22 cheshire
-Update debugging message
-
-Revision 1.616 2007/04/25 16:38:32 cheshire
-If negative cache entry already exists, reactivate it instead of creating a new one
-
-Revision 1.615 2007/04/25 02:14:38 cheshire
-<rdar://problem/4246187> uDNS: Identical client queries should reference a single shared core query
-Additional fixes to make LLQs work properly
-
-Revision 1.614 2007/04/23 21:52:45 cheshire
-<rdar://problem/5094009> IPv6 filtering in AirPort base station breaks Wide-Area Bonjour
-
-Revision 1.613 2007/04/23 04:58:20 cheshire
-<rdar://problem/5072548> Crash when setting extremely large TXT records
-
-Revision 1.612 2007/04/22 20:39:38 cheshire
-<rdar://problem/4633194> Add 20 to 120ms random delay to browses
-
-Revision 1.611 2007/04/22 18:16:29 cheshire
-Removed incorrect ActiveQuestion(q) check that was preventing suspended questions from getting reactivated
-
-Revision 1.610 2007/04/22 06:02:02 cheshire
-<rdar://problem/4615977> Query should immediately return failure when no server
-
-Revision 1.609 2007/04/20 21:17:24 cheshire
-For naming consistency, kDNSRecordTypeNegative should be kDNSRecordTypePacketNegative
-
-Revision 1.608 2007/04/20 19:45:31 cheshire
-In LogClientOperations mode, dump out unknown DNS packets in their entirety
-
-Revision 1.607 2007/04/19 23:56:25 cheshire
-Don't do cache-flush processing for LLQ answers
-
-Revision 1.606 2007/04/19 22:50:53 cheshire
-<rdar://problem/4246187> Identical client queries should reference a single shared core query
-
-Revision 1.605 2007/04/19 20:06:41 cheshire
-Rename field 'Private' (sounds like a boolean) to more informative 'AuthInfo' (it's a DomainAuthInfo pointer)
-
-Revision 1.604 2007/04/19 18:03:04 cheshire
-Add "const" declaration
-
-Revision 1.603 2007/04/06 21:00:25 cheshire
-Fix log message typo
-
-Revision 1.602 2007/04/05 22:55:35 cheshire
-<rdar://problem/5077076> Records are ending up in Lighthouse without expiry information
-
-Revision 1.601 2007/04/04 21:48:52 cheshire
-<rdar://problem/4720694> Combine unicast authoritative answer list with multicast list
-
-Revision 1.600 2007/04/04 01:31:33 cheshire
-Improve debugging message
-
-Revision 1.599 2007/04/04 00:03:26 cheshire
-<rdar://problem/5089862> DNSServiceQueryRecord is returning kDNSServiceErr_NoSuchRecord for empty rdata
-
-Revision 1.598 2007/04/03 19:43:16 cheshire
-Use mDNSSameIPPort (and similar) instead of accessing internal fields directly
-
-Revision 1.597 2007/03/31 00:32:32 cheshire
-After skipping OPT and TSIG, clear m->rec.r.resrec.RecordType
-
-Revision 1.596 2007/03/28 20:59:26 cheshire
-<rdar://problem/4743285> Remove inappropriate use of IsPrivateV4Addr()
-
-Revision 1.595 2007/03/26 23:48:16 cheshire
-<rdar://problem/4848295> Advertise model information via Bonjour
-Refinements to reduce unnecessary transmissions of the DeviceInfo TXT record
-
-Revision 1.594 2007/03/26 23:05:05 cheshire
-<rdar://problem/5089257> Don't cache TSIG records
-
-Revision 1.593 2007/03/23 17:40:08 cheshire
-<rdar://problem/4060169> Bug when auto-renaming Computer Name after name collision
-
-Revision 1.592 2007/03/22 18:31:48 cheshire
-Put dst parameter first in mDNSPlatformStrCopy/mDNSPlatformMemCopy, like conventional Posix strcpy/memcpy
-
-Revision 1.591 2007/03/22 00:49:19 cheshire
-<rdar://problem/4848295> Advertise model information via Bonjour
-
-Revision 1.590 2007/03/21 23:05:59 cheshire
-Rename uDNS_HostnameInfo to HostnameInfo; deleted some unused fields
-
-Revision 1.589 2007/03/20 15:37:19 cheshire
-Delete unnecessary log message
-
-Revision 1.588 2007/03/20 00:24:44 cheshire
-<rdar://problem/4175213> Should deliver "name registered" callback slightly *before* announcing PTR record
-
-Revision 1.587 2007/03/16 22:10:56 cheshire
-<rdar://problem/4471307> mDNS: Query for *either* type A or AAAA should return both types
-
-Revision 1.586 2007/03/10 03:26:44 cheshire
-<rdar://problem/4961667> uDNS: LLQ refresh response packet causes cached records to be removed from cache
-
-Revision 1.585 2007/03/10 02:02:58 cheshire
-<rdar://problem/4961667> uDNS: LLQ refresh response packet causes cached records to be removed from cache
-Eliminate unnecessary "InternalResponseHndlr responseCallback" function pointer
-
-Revision 1.584 2007/02/28 01:51:27 cheshire
-Added comment about reverse-order IP address
-
-Revision 1.583 2007/01/27 03:19:33 cheshire
-Need to initialize question->sock
-
-Revision 1.582 2007/01/25 00:40:16 cheshire
-Unified CNAME-following functionality into cache management code (which means CNAME-following
-should now also work for mDNS queries too); deleted defunct pktResponseHndlr() routine.
-
-Revision 1.581 2007/01/23 02:56:11 cheshire
-Store negative results in the cache, instead of generating them out of pktResponseHndlr()
-
-Revision 1.580 2007/01/19 21:17:33 cheshire
-StartLLQPolling needs to call SetNextQueryTime() to cause query to be done in a timely fashion
-
-Revision 1.579 2007/01/19 18:39:10 cheshire
-Fix a bunch of parameters that should have been declared "const"
-
-Revision 1.578 2007/01/10 22:51:57 cheshire
-<rdar://problem/4917539> Add support for one-shot private queries as well as long-lived private queries
-
-Revision 1.577 2007/01/10 02:05:21 cheshire
-Delay uDNS_SetupDNSConfig() until *after* the platform layer
-has set up the interface list and security credentials
-
-Revision 1.576 2007/01/09 02:40:57 cheshire
-uDNS_SetupDNSConfig() shouldn't be called from mDNSMacOSX.c (platform support layer);
-moved it to mDNS_Init() in mDNS.c (core code)
-
-Revision 1.575 2007/01/09 00:17:25 cheshire
-Improve "ERROR m->CurrentRecord already set" debugging messages
-
-Revision 1.574 2007/01/05 08:30:41 cheshire
-Trim excessive "$Log" checkin history from before 2006
-(checkin history still available via "cvs log ..." of course)
-
-Revision 1.573 2007/01/05 06:34:03 cheshire
-Improve "ERROR m->CurrentQuestion already set" debugging messages
-
-Revision 1.572 2007/01/04 23:11:11 cheshire
-<rdar://problem/4720673> uDNS: Need to start caching unicast records
-When an automatic browsing domain is removed, generate appropriate "remove" events for legacy queries
-
-Revision 1.571 2007/01/04 21:45:20 cheshire
-Added mDNS_DropLockBeforeCallback/mDNS_ReclaimLockAfterCallback macros,
-to do additional lock sanity checking around callback invocations
-
-Revision 1.570 2007/01/04 20:57:47 cheshire
-Rename ReturnCNAME to ReturnIntermed (for ReturnIntermediates)
-
-Revision 1.569 2007/01/04 20:27:27 cheshire
-Change a LogMsg() to debugf()
-
-Revision 1.568 2007/01/04 02:39:53 cheshire
-<rdar://problem/4030599> Hostname passed into DNSServiceRegister ignored for Wide-Area service registrations
-
-Revision 1.567 2006/12/21 00:01:37 cheshire
-Tidy up code alignment
-
-Revision 1.566 2006/12/20 04:07:34 cheshire
-Remove uDNS_info substructure from AuthRecord_struct
-
-Revision 1.565 2006/12/19 22:49:23 cheshire
-Remove uDNS_info substructure from ServiceRecordSet_struct
-
-Revision 1.564 2006/12/19 02:38:20 cheshire
-Get rid of unnecessary duplicate query ID field from DNSQuestion_struct
-
-Revision 1.563 2006/12/19 02:18:48 cheshire
-Get rid of unnecessary duplicate "void *context" field from DNSQuestion_struct
-
-Revision 1.562 2006/12/16 01:58:31 cheshire
-<rdar://problem/4720673> uDNS: Need to start caching unicast records
-
-Revision 1.561 2006/12/01 07:38:53 herscher
-Only perform cache workaround fix if query is wide-area
-
-Revision 1.560 2006/11/30 23:07:56 herscher
-<rdar://problem/4765644> uDNS: Sync up with Lighthouse changes for Private DNS
-
-Revision 1.559 2006/11/27 08:20:57 cheshire
-Preliminary support for unifying the uDNS and mDNS code, including caching of uDNS answers
-
-Revision 1.558 2006/11/10 07:44:03 herscher
-<rdar://problem/4825493> Fix Daemon locking failures while toggling BTMM
-
-Revision 1.557 2006/11/10 01:12:51 cheshire
-<rdar://problem/4829718> Incorrect TTL corrections
-
-Revision 1.556 2006/11/10 00:54:14 cheshire
-<rdar://problem/4816598> Changing case of Computer Name doesn't work
-
-Revision 1.555 2006/10/30 20:03:37 cheshire
-<rdar://problem/4456945> After service restarts on different port, for a few seconds DNS-SD may return stale port number
-
-Revision 1.554 2006/10/20 05:35:04 herscher
-<rdar://problem/4720713> uDNS: Merge unicast active question list with multicast list.
-
-Revision 1.553 2006/10/05 03:42:43 herscher
-Remove embedded uDNS_info struct in DNSQuestion_struct
-
-Revision 1.552 2006/09/15 21:20:15 cheshire
-Remove uDNS_info substructure from mDNS_struct
-
-Revision 1.551 2006/08/14 23:24:22 cheshire
-Re-licensed mDNSResponder daemon source code under Apache License, Version 2.0
-
-Revision 1.550 2006/07/27 17:58:34 cheshire
-Improved text of "SendQueries didn't send all its queries; will try again" debugging message
-
-Revision 1.549 2006/07/20 22:07:31 mkrochma
-<rdar://problem/4633196> Wide-area browsing is currently broken in TOT
-More fixes for uninitialized variables
-
-Revision 1.548 2006/07/20 19:30:19 mkrochma
-<rdar://problem/4633196> Wide-area browsing sometimes doesn't work in TOT
-
-Revision 1.547 2006/07/15 02:31:30 cheshire
-<rdar://problem/4630812> Suppress log messages for certain old devices with inconsistent TXT RRSet TTLs
-
-Revision 1.546 2006/07/07 01:09:09 cheshire
-<rdar://problem/4472013> Add Private DNS server functionality to dnsextd
-Only use mallocL/freeL debugging routines when building mDNSResponder, not dnsextd
-
-Revision 1.545 2006/07/05 23:10:30 cheshire
-<rdar://problem/4472014> Add Private DNS client functionality to mDNSResponder
-Update mDNSSendDNSMessage() to use uDNS_TCPSocket type instead of "int"
-
-Revision 1.544 2006/06/29 07:42:14 cheshire
-<rdar://problem/3922989> Performance: Remove unnecessary SameDomainName() checks
-
-Revision 1.543 2006/06/29 01:38:43 cheshire
-<rdar://problem/4605285> Only request unicast responses on wake from sleep and network connection
-
-Revision 1.542 2006/06/27 23:40:29 cheshire
-Fix typo in comment: mis-spelled "compile"
-
-Revision 1.541 2006/06/27 19:46:24 cheshire
-Updated comments and debugging messages
-
-Revision 1.540 2006/06/15 21:35:16 cheshire
-Move definitions of mDNS_vsnprintf, mDNS_SetupResourceRecord, and some constants
-from mDNS.c to DNSCommon.c, so they can be accessed from dnsextd code
-
-Revision 1.539 2006/06/08 23:45:46 cheshire
-Change SimultaneousProbe messages from debugf() to LogOperation()
-
-Revision 1.538 2006/03/19 17:13:06 cheshire
-<rdar://problem/4483117> Need faster purging of stale records
-Shorten kDefaultReconfirmTimeForNoAnswer to five seconds
-and reconfirm whole chain of antecedents ot once
-
-Revision 1.537 2006/03/19 02:00:07 cheshire
-<rdar://problem/4073825> Improve logic for delaying packets after repeated interface transitions
-
-Revision 1.536 2006/03/08 23:29:53 cheshire
-<rdar://problem/4468716> Improve "Service Renamed" log message
-
-Revision 1.535 2006/03/02 20:41:17 cheshire
-<rdar://problem/4111464> After record update, old record sometimes remains in cache
-Minor code tidying and comments to reduce the risk of similar programming errors in future
-
-Revision 1.534 2006/03/02 03:25:46 cheshire
-<rdar://problem/4111464> After record update, old record sometimes remains in cache
-Code to harmonize RRSet TTLs was inadvertently rescuing expiring records
-
-Revision 1.533 2006/02/26 00:54:41 cheshire
-Fixes to avoid code generation warning/error on FreeBSD 7
-
-*/