Michael Davidsaver
041423092f
osdnetIntf.c cleanup indentation
2015-08-18 15:51:38 -04:00
Michael Davidsaver
869410d6f6
libCom/osi: osiLocalAddr() avoid race on initialization
2015-08-18 15:46:22 -04:00
Ambroz Bizjak
6862ef6580
timerQueueActive.patch from lp:1479316
2015-08-18 14:39:27 -04:00
Ambroz Bizjak
b6aea68304
epicsSingletonMutex.patch from lp:1479316
2015-08-18 14:38:18 -04:00
Michael Davidsaver
9d00978176
src/ca/client: no need to remove const
2015-08-18 14:33:44 -04:00
Ambroz Bizjak
b6fbea9610
repeaterSubscribeTimer.patch from lp:1479316
2015-08-18 14:24:10 -04:00
Michael Davidsaver
180f40c1f7
dbLock: fix unlock w/o lock during iocInit
2015-08-18 11:09:35 -04:00
Michael Davidsaver
d5832354e8
iocInit: Don't free LSET until scan tasks have stopped
2015-08-18 11:09:35 -04:00
Michael Davidsaver
44980a1dac
dbContextReadNotifyCache: variable length for CAC gets
2015-08-18 10:46:20 -04:00
Michael Davidsaver
af1daea3e7
dbContext: local CA variable length
2015-08-18 10:46:20 -04:00
Michael Davidsaver
b2716f0a19
dbCa: subscribe to variable length arrays
2015-08-18 10:46:20 -04:00
Michael Davidsaver
050b9f9a48
ioc/db/test: dbChannelTest use dbUnitTest.h
2015-08-18 09:07:18 -04:00
Michael Davidsaver
4ab6aa79e3
ioc/db/test: fixup arrShorthandTest to use dbUnitTest
2015-08-18 09:07:18 -04:00
Michael Davidsaver
1e39224836
dbCa: simplify shutdown
...
Don't need partial shutdown case anymore.
Also avoid race when destorying startStopEvent
2015-08-18 09:07:18 -04:00
Michael Davidsaver
f99fbe1964
move dbCaShutdown earlier
...
shutdown dbCa thread before free'ing records
2015-08-18 09:07:18 -04:00
Michael Davidsaver
aa04941119
ioc/db/test: wrong variable for -I
2015-08-18 09:07:18 -04:00
Andrew Johnson
46853d6862
Clean up compiler warnings
2015-08-16 17:43:10 -05:00
Andrew Johnson
fe4f607e63
libCom/errlog: Don't store stderr in pvtData.console
...
Some VxWorks BSPs close the stderr stream that the shell running
the startup script created and open a new one for the interactive
shell. This change makes pvtData.console==NULL mean use stderr
instead of storing the stderr value in pvtData.console at init.
2015-08-13 10:44:52 -05:00
Michael Davidsaver
02ac91aa2a
epicsReadline: ensure readline context is zero'd
...
Otherwise ->osd != NULL for the default RTEMS readline impl
2015-08-12 10:17:31 -05:00
Andrew Johnson
9e82a96700
dbTest: Remove special ENUM handling from dbpf()
...
This was preventing the value check in cvt_st_menu() from working.
2015-08-10 15:04:11 -05:00
Andrew Johnson
ef7399159c
dbAccess: Make dbPut() always do special processing after put
...
If the value gets rejected by a put to SCAN, we must still call
dbPutSpecial() to add the record back onto its old scan thread.
2015-08-10 15:03:48 -05:00
Andrew Johnson
d7b3293ba3
libCom: Use MSVC's struct timespec when available
2015-08-07 09:41:48 -05:00
Andrew Johnson
9014ca899d
dbStatic: Fix invalid memory read
2015-07-30 11:57:50 -05:00
Andrew Johnson
f1b403b59a
Merged calinktest branch
2015-07-28 14:22:06 -05:00
Andrew Johnson
877a409de1
Fix podToHtml build issue on Windows
2015-07-24 15:59:38 -05:00
Andrew Johnson
7395d1d88d
Make dbCa routines macros to dbLink wrappers
2015-07-24 14:06:45 -05:00
Andrew Johnson
4d3831ce58
Merged changes from 3.15 branch to revno 12699
2015-07-24 12:01:53 -05:00
Andrew Johnson
c5130468cd
Let dbdToHtml work with wrong newlines
2015-07-24 10:42:09 -05:00
Andrew Johnson
7dd067e887
Merged dbCa bug-fix from 3.14 branch, revno 12595
2015-07-23 18:02:54 -05:00
Andrew Johnson
c56091978c
db: Fixed obscure dbCa bug
...
This could be triggered if dbCaRemoveLink() is called on a link for
which there is an outstanding dbCaPutCallback().
Normally a dbCaPutCallback() callback routine is passed the associated
userPvt pointer as an argument, but in the event that dbCaRemoveLink()
gets used on the same link between the put and its completion callback,
the callback routine was being called with the link pointer as the
argument instead.
For all the existing Asyn Soft Channel device supports this is not a
problem as they all pass the link pointer as their userPvt argument, but
that won't necessarily always be the case.
Also updated the comments describing the process of removing links.
2015-07-23 17:57:30 -05:00
Andrew Johnson
2f8e6bf17e
Fix smoothing for Nan/Inf values in devAiSoftCallback
2015-07-23 13:05:38 -05:00
Andrew Johnson
144281e0a3
Removed extraneous .db file
2015-07-23 12:34:05 -05:00
Andrew Johnson
83c2414ad0
Merged changes from 3.14 branch to revno 12594
2015-07-23 10:13:07 -05:00
Andrew Johnson
1d99ad6cfc
Move closer to having pluggable link types.
...
All link types now have a plink->precord pointer.
Add link support entry tables for const, DB and CA link types.
Route all link I/O options through the new plink->lset.
2015-07-22 14:56:48 -05:00
Andrew Johnson
bd4d8ec26b
Merged lockopt branch
2015-08-21 22:44:11 -05:00
Michael Davidsaver
504665bf09
iocInit: move dbCaShutdown earlier
...
Switch dbCaShutdown from an exit hook to
be called from iocShutdown. Place it
after callback shutdown.
2015-07-21 19:08:07 -04:00
Michael Davidsaver
1411522a11
dbCa: only add dbCaExit once
2015-07-21 15:53:37 -04:00
Andrew Johnson
ae3d3904d9
Fix minor locking issue in dbEvent.c
2015-07-17 18:09:14 -05:00
Andrew Johnson
002bafdf07
Add testHarnessDone
2015-07-17 18:06:58 -05:00
Michael Davidsaver
8bfa40d858
dbScan: avoid race on shutdown
2015-07-16 17:05:41 -04:00
Michael Davidsaver
138e2f1ad5
callback: fix race on shutdown
...
epicsEventDestory() in the main thread could be called
before epicsEventTrigger() completes in the worker.
2015-07-16 17:05:41 -04:00
Andrew Johnson
a5e1f367d6
Merged cbshutdown changes from 3.15 branch, revno 12690
2015-07-16 11:54:40 -05:00
Michael Davidsaver
fefe6fd1fc
Fix shutdown issues with scan and callback.
...
The main reason for this merge proposal is the change to "public" API functions.
Use atomic counter to resolve data race on threadsRunning in callback.
Split up callbackShutdown() and scanShutdown() into two phases *Stop() and
*Cleanup(). The *Stop() functions signal worker threads, and wait for them to
exit. The *Cleanup() functions actually reclaim global resources.
These two mechanisms have couplings which are quite complex. I/O Intr scans
involve both scan lists and callbacks.
2015-07-16 11:48:29 -05:00
Michael Davidsaver
eb31fcedeb
iocInit: dbCa shutdown order
2015-07-13 17:51:46 -04:00
Michael Davidsaver
dd6edb103c
scanIoTest: set # of tests
2015-07-13 17:51:46 -04:00
Andrew Johnson
50e6a305b1
Merged changes from 3.15 branch to revno 12689
2015-07-13 13:42:10 -05:00
Andrew Johnson
5eb49ebaf0
Merged changes from 3.14 branch to revno 12590
2015-07-13 13:05:33 -05:00
Andrew Johnson
58602cfa80
Clean up in src/ioc/db/test/scanIoTest.c
2015-07-10 16:22:31 -05:00
Andrew Johnson
1e68f3899d
Merged gen-vcs-version branch
2015-07-10 14:02:01 -05:00
Michael Davidsaver
8b3d37d392
genVersionHeader: work with git repo w/o tags
...
--always is necessary for (new) repositories w/o any tags.
--abbrev=20 to get more than the default 7 characters of the hash
2015-07-10 14:19:44 -04:00
