summaryrefslogtreecommitdiff
path: root/ntpshm.c
Commit message (Collapse)AuthorAgeFilesLines
* uBlox -> u-blox; the company does not biicapitalize its name.Eric S. Raymond2013-11-091-1/+1
|
* Code is fully cppcheck, splint, and Coverity clean.Eric S. Raymond2013-11-061-0/+1
| | | | All regression tests pass. PPS is live.
* Partial cppcheck and Coverity cleanup.Eric S. Raymond2013-11-061-1/+3
| | | | | | Coverity found a potential overrun bug. All regression tests pass. PPS is live.
* Unify time-difference computations. All regression tests pass, PPS is live.Eric S. Raymond2013-11-061-2/+1
|
* splint cleanup.Eric S. Raymond2013-11-061-1/+3
|
* Clarifying name change. All regression test pass. PPS is live.Eric S. Raymond2013-11-061-3/+3
|
* Update for the existence of the barrier() function.Eric S. Raymond2013-11-061-2/+5
|
* Sets precision inside ntpshm_put().Michael Tatarinov2013-11-061-3/+10
| | | | Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
* Simplify the PPS hook.Michael Tatarinov2013-11-061-6/+5
| | | | Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
* ntpshm_pps() is gone.Michael Tatarinov2013-11-061-107/+23
| | | | Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
* C variables are case insensitive, but still nice to have matching case.Gary E. Miller2013-11-051-13/+13
| | | | | Should be no functional change. Thanks to: Michael Tatarinov <kukabu@gmail.com>
* Get some ntpshm coments to align with new realityGary E. Miller2013-11-051-3/+1
|
* Use the new timedrift_t to simplify PPS.Gary E. Miller2013-11-051-26/+18
|
* Fix a PPS debug report nit.Gary E. Miller2013-11-051-1/+1
|
* Partial splint cleanup. All regression tests pass. PPS is live.Eric S. Raymond2013-11-051-1/+7
|
* Be a little more paranoid about shmTime.Gary E. Miller2013-11-051-2/+2
| | | | Fix per Harlan Stenn <stenn@ntp.org>
* Fix typo. Thanks to Mick Durkin <mick.durkin@napierdeltic.demon.co.uk>Gary E. Miller2013-11-051-1/+1
|
* Insert memory barriers where requested, since we actually have them.Eric S. Raymond2013-11-051-4/+8
|
* Fix a rounding problem in last patch.Gary E. Miller2013-11-051-9/+7
| | | | ntpshm uses the same sec for usec and nsec, so do not round.
* Add new ntpd nSec fields to SHM interfaceGary E. Miller2013-11-051-1/+7
|
* Use timedrift_t for all hook arguments.Eric S. Raymond2013-11-031-26/+28
| | | | | | | | | No logic changes in this commit. We push naosecond precision closer to where time hints are shipped. Changes to ppsthread.c are kept minimal; it needs a cleanup but that will be done in another change that can be separately reviewed. All regression tests pass. PPS is live in gpsmon and reported by gpsd.
* Eliminate an enable flag that always gets set true.Eric S. Raymond2013-11-021-4/+2
| | | | All regression tests pass. PPS is live.
* Namespace cleanup. All regression tests pass, PPS is live.Eric S. Raymond2013-11-021-2/+2
|
* Finish decoupling the PPS watcher code from NTP shared-segment management.Eric S. Raymond2013-11-021-2/+13
| | | | Also, further contain the latter.
* Eliminate the race in KPPS thread startup.Eric S. Raymond2013-10-311-7/+1
| | | | | | | | | The previous change (ntpd_link_activate() moving to after device-open time) made it possible to lift the attempts at KPPS initialization out of the PPS-watcher thread, because it no longer has to await the device open. This also greatly simplifies the generic thread-watcher interface. All regression tests pass, PPS is live.
* PPS on a pty doesn't work.Eric S. Raymond2013-10-311-1/+1
| | | | | | | | | This had been breaking the regression tests for a while (since "Unconditionally enable PPS at startup." to be exact) but I didn't realize it because an unrelated bug in the build system was usually turning PPS off for me. Live tests were unaffected. All regression tests pass with pps either on or off. Live PPS works.
* Revert "Separate ntpd support from chrony support, step 1."Eric S. Raymond2013-10-301-0/+505
| | | | We'll have to sneak up on this change more slowly.
* Separate ntpd support from chrony support, step 1.Eric S. Raymond2013-10-301-505/+0
| | | | | | | | | | | | | ntp_offset becomes time_offset. There is a new config symbol CHRONY_ENABLE; most time service code is npw controlled by TIMESERVICE_ENABLE. The file ntpshm.c becomes timeexport.c This change is not complete. More disentanglement has to be done inside timeexport.c itself; at the moment enabling one but not both of ntpshm or chrony will probably break its compile. The point of getting this commit out is so Gary will see the new baseline code ASAP. All regression tests pass.
* Re-enable compilation with ntpshm=no.Eric S. Raymond2013-10-301-0/+3
|
* splint cleanup. Regression tests pass and PPS live-tests correctly.Eric S. Raymond2013-10-291-5/+11
|
* Separate policy from mechanism.Eric S. Raymond2013-10-291-10/+24
|
* ntpshm.c splits up; thread management now lives in ppsthread.cEric S. Raymond2013-10-291-677/+6
|
* Now we can handle a PPS thread per session. Tested with live PPS and works.Eric S. Raymond2013-10-291-6/+4
|
* A big step - ntpd/chrony-specific code becomes method ponters.Eric S. Raymond2013-10-291-9/+24
|
* Eliminate shmTimePPS, no longer used.Eric S. Raymond2013-10-291-2/+2
|
* Unconditionally enable PPS at startup.Eric S. Raymond2013-10-291-8/+1
| | | | | | | | Before this change, ntpshm_init() had an argument depebding on the state of the -n flag. I think this was some sort of remnant from before we had the daemon initialization sequence quite right. The new code works (and ships PPS) whether -n is on or not.
* Another factoring step in ntpshm.c.Eric S. Raymond2013-10-291-16/+27
|
* Make chrony_send() and ntpshm_pps() take the same arguments in theGary E. Miller2013-10-281-58/+53
| | | | same order.
* Remove broken and unused ONCORE PPS codeGary E. Miller2013-10-281-11/+0
|
* Another step at using nSec instead of uSec in ntpshm.cGary E. Miller2013-10-281-51/+54
|
* Comment polishing.Eric S. Raymond2013-10-281-1/+3
|
* More code rearrangement. No logic changes.Eric S. Raymond2013-10-281-11/+11
|
* Move some code and remove unneeded deoendency.Eric S. Raymond2013-10-281-75/+73
| | | | All regression tesrs pass.
* More factoring out of chrony code.Eric S. Raymond2013-10-271-15/+16
|
* Attempting to disentangle the thread monitor from ntpd and chrony reporting.Eric S. Raymond2013-10-271-16/+31
| | | | Gary, you should review and test carefully.
* Begin the painful process of factoring the thread processing.Eric S. Raymond2013-10-271-41/+54
|
* Build system symplifications fixing minor bugs exposed by 13.10.Eric S. Raymond2013-10-271-0/+2
|
* Add a clarifying comment.Eric S. Raymond2013-10-251-0/+8
|
* Remove a redundant test, and explain the redundancy.Eric S. Raymond2013-10-251-6/+12
|
* Anoter chrony socket error reporting tweak.Gary E. Miller2013-10-251-2/+2
|
View Lorry Controller Status