summaryrefslogtreecommitdiffstats
path: root/contrib/ntp/TODO
blob: e72ee7bef65a813347d0b2b8a9f3d46cbe26632c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122

      *** IF YOU CAN HELP FIX ANY OF THESE THINGS, PLEASE DO! ***

010402: Look harder at -lm and -lelf - they are needed less and less...

970711: Look Real Hard at changing the key stuff from u_long to u_int32.

970711: Make sure it's safe to convert proto_config's 2nd argument from
	u_long to u_int32.  Watch "set" in ntp_request.c:setclr_flags().

970318: in hourly_stats(?), squawk if the magnitude of the drift is,
	say, >400.

970301: Implement the utmp/wtmp timestamping on time steps.

970210: Find a way to dump the current configuration to either syslog or
	a file.

Problems that need to be fixed:

- Get rid of the old SYS_* macros:
  (It's worth noting that any code that would have been "enabled" by any
  of these macros has not been used since 5.83, and there have been very
  few complaints...)

  SYS_44BSD:
   authstuff/md5driver.c

  SYS_BSDI:
   authstuff/md5driver.c

  SYS_DECOSF1:
   util/ntptime.c

  SYS_DOMAINOS:
   parseutil/dcfd.c
   xntpd/ntpd.c

  SYS_HPUX:
   kernel/sys/ppsclock.h
   ntpdate/ntpdate.c
   ntptrace/ntptrace.c

  SYS_PTX:
   libntp/machines.c

  SYS_SOLARIS:
   libparse/parse.c
   libparse/clk_trimtsip.c
   xntpd/ntp_io.c
   xntpd/ntp_refclock.c
   xntpd/ntpd.c

  SYS_SUNOS4:
   libparse/parse.c
   libparse/clk_trimsip.c

  SYS_WINNT:
   include/ntp.h
   include/ntp_fp.h
   include/ntp_if.h
   include/ntp_machine.h
   include/ntp_select.h
   include/ntp_syslog.h
   include/ntp_unixtime.h
   include/ntpd.h
   libntp/libntp.mak
   libntp/machines.c
   libntp/mexit.c
   libntp/msyslog.c
   libntp/systime.c
   ntpdate/ntpdate.c
   ntpdate/ntpdate.mak
   ntpq/ntpq.c
   ntpq/ntpq.mak
   ntpq/ntpq_ops.c
   ntptrace/ntptrace.c
   ntptrace/ntptrace.mak
   xntpd/ntp_config.c
   xntpd/ntp_filegen.c
   xntpd/ntp_intres.c
   xntpd/ntp_io.c
   xntpd/ntp_loopfilter.c
   xntpd/ntp_peer.c
   xntpd/ntp_proto.c
   xntpd/ntp_refclock.c
   xntpd/ntp_timer.c
   xntpd/ntp_unixclock.c
   xntpd/ntp_util.c
   xntpd/ntpd.c
   xntpd/xntpd.mak
   xntpdc/ntpdc.c
   xntpdc/xntpdc.mak

- config.guess might need help to identify:

  Fujitsu's UXP				--enable-adjtime-is-accurate
					--enable-step-slew

  Unixware				--enable-adjtime-is-accurate
					--enable-tick=10000
					--enable-tickadj=80
					--enable-udp-wildcard
					--disable-step-slew

  DomainOS				--enable-adjtime-is-accurate
					--disable-kmem
					--enable-tick=1000000

  OpenVMS				--enable-slew-always
					--enable-hourly-todr-sync

  Is adjtime accurate on ALL sysv4* machines?
  Can we identify DomainOS with *-apollo-* ?
  Do we catch all Unixware machines with *-univel-sysv* ?

- Combine enable-step-slew and enable-ntpdate-step

- Make sure enable-hourly-todr-sync is always disabled
  What about NextStep and OpenVMS, where hourly TODR sync used to be enabled?

- Check dcfd.c for variables that need to be volatile.
OpenPOWER on IntegriCloud