46ec38756b
[PATCH 06/12] use epics extern macro
2010-05-25 13:58:46 +01:00
e21cde013d
[PATCH 05/12] reorganize devLib.h
...
Group functions by bus type with general functions at the top
add macro to exclude deprecated api
2010-05-25 13:58:45 +01:00
83adda7e6c
[PATCH 04/12] move "virtual os" definition to a seperate file
...
The virtual os table is not something ordinary devLib users
should know or care about.
2010-05-25 13:58:44 +01:00
ad8cd28f7d
[PATCH 03/12] mark all implimentation functions as static
2010-05-25 13:58:43 +01:00
acb905fcfc
[PATCH 02/12] add devInterruptInUseVME to function table
...
Add test for interrupt handler function to the virtual os table.
Adds a stub implementation to devLib.c
Renames OS implementations functions for RTEMS and vxWorks to avoid symbol name conflicts
2010-05-25 13:58:43 +01:00
62d035c310
[PATCH 01/12] add jumping VME interrupt routines to independent implementation
...
Adds implementations which use the virtual os table to invoke the
OS specific call.
rename conflicting functions
2010-05-25 13:58:42 +01:00
498ef7075e
Fix missing parallel build dependency.
...
Ron Sluiter pointed out that gdd/dbMapper.h includes db_access.h so
gdd has to be built after ca, simplifying the cas dependency too.
2010-05-19 14:46:09 -05:00
6790acf78f
Fix for hang from Ron Sluiter.
2010-05-18 16:12:36 -05:00
6eb36e1224
Fix build on Windows.
2010-05-18 16:10:07 -05:00
6b64304570
Optimize epicsThreadOnce() routines.
...
Don't call epicsThreadGetIdSelf() unnecessarily, check for the value
EPICS_THREAD_ONCE_DONE first, since that's the most common state.
2010-05-14 17:26:54 -05:00
6cd22d4efe
Don't call errlogPrintf() in epicsThreadGetOsiPriorityValue()
...
This is a partial fix of a WIN32-only problem Mark Rivers reported.
A WIN32 thread created by a manufacturer's library is used to call one
of his callback routines, which tries to use an epics facility that has
an epicsThreadOnce(). The new version of epicsThreadOnce() uses
epicsThreadGetIdSelf() to detect recursive initialization; the thread
doesn't have an epicsThreadId yet, so epicsThreadImplicitCreate() is
asked to make one. It calls epicsThreadGetOsiPriorityValue(), but
osdPriority is 15 which is not known by that code, so it reports an
error by calling errlogPrintf(). That checks epicsThreadIsOkToBlock()
which calls epicsThreadOnce() but that needs our epicsThreadId...
This doesn't solve the issue of the unknown osdPriority value 15
(priorityClass = 32), but it allows the IOC to continue working with
just a warning message at init time.
It also points out that calling epicsThreadGetIdSelf() every time we
check an epicsThreadOnce() is probably not good for performance.
2010-05-14 16:49:32 -05:00
e9593e4b6c
Run epicsMessageQueueTest last, since it takes 5 minutes.
2010-05-05 15:43:58 -05:00
831f48e719
Fix the macLib buffer overflow issue, Bug 551909.
...
This is mostly Eric's patches, without the valend name change.
2010-05-04 15:43:20 -05:00
5868aae543
Define EPICS_THREAD_ONCE_DONE properly on RTEMS.
2010-05-04 14:39:37 -05:00
12dc840eda
Don't print anything from epicsThreadShow() when tid=0.
...
Also don't pass level > 1 to vxWorks taskShow() routine.
2010-05-03 17:28:04 -05:00
7f534cbc74
Rename 'thread' iocsh command to 'epicsThreadShow'.
2010-05-03 17:26:41 -05:00
50d755a79f
Fix compiler warning about missing sentinel in execlp() call.
2010-05-03 12:29:13 -05:00
d192f51858
Don't use back-tick chars as open quotes, they aren't any more.
2010-04-28 17:29:04 -05:00
618e189898
'make cvsclean' should delete old editor backup files too.
...
Not that we're using CVS for Base any more...
2010-04-28 17:19:36 -05:00
b4be728539
cvtFastPerform measures performance, it doesn't do functional testing.
...
* Cleaned up some warning messages in it on darwin
* Removed it from the epicsRunLibComTests function.
2010-04-26 17:19:10 -05:00
6ece3235c9
Added libCom/test code for epicsThreadOnce() implementations.
...
Also fixed subsystems that were not using epicsThreadOnce correctly.
2010-04-26 15:48:42 -05:00
c5a27fa32e
Rework epicsThreadOnce() using ideas from Michael Davidsaver.
...
An epicsThreadOnceId is now an epicsThreadId.
During initialization, it is set to the thread running the init routine
which can now detect a recursive initialization attempt and suspend.
EPICS_THREAD_ONCE_INIT is still zero, the implementations now define a
new private value for EPICS_THREAD_ONCE_DONE. This is deliberately not
made public.
2010-04-26 15:38:11 -05:00
4c70951869
Reduce warnings when built with Perl 5.10.
2010-04-21 17:35:26 -05:00
32e34f536c
Move STATIC_ASSERT statements.
...
This macro can only be used where a variable declaration is allowed,
the vxWorks compiler objects otherwise.
2010-04-20 17:13:54 -05:00
20483795e4
Remove "dbRecordHead: tempList not empty" errors.
...
Loading a .db file that uses an unknown record type leaves items
on the parser's tempList. This patch empties that list when the
error is reported so future dbLoadRecords() commands can run.
2010-04-12 11:31:24 -05:00
e8f56d08b8
Ensure the string VAL and OVAL have the same storage length
...
There won't be any problems with overflow as long as dbPut
behaves correctly when setting VAL.
2010-04-06 14:48:48 -04:00
b63e381fbe
rec: update last/prev./old value fields after dset init_record
...
This must be done if init_record is allowed to change values
or monitors may not be posted when values change.
2010-04-05 14:49:18 -04:00
6518adf901
Merged Ralph Lange's 'compress-fix' branch.
2010-04-02 18:03:24 -04:00
c3b19955f5
Merged Michael Davidsaver's 'minor-fixes' branch.
2010-04-02 16:51:04 -05:00
c6cf442abd
Fixed crash when ALG (algorithm) was changed to Average at runtime. ( Fixes #552803 )
2010-04-02 17:26:17 -04:00
e56d4c2337
The monitor last value field needs to be initialized.
...
Device support init_record is allowed to set VAL to 1.
If uninitialized then mlst is 0.
During the first process() this will miss a monitor if VAL is changed to 0,
and cause a spurious monitor if VAL remains 1.
- mdavidsaver
2010-03-24 13:21:38 -05:00
356ee58927
use osi epicsMath.h
2010-03-24 13:17:41 -04:00
cadbcb43a2
use C++ name make.h
2010-03-24 11:34:00 -04:00
77d09de2c4
quiet compiler warning
2010-03-16 21:09:53 -04:00
d9a9a1f6ad
remove unused iocRegister stub
2010-03-16 21:09:53 -04:00
75a15809fa
epicsTimer: check that delay time is finite
2010-03-16 21:09:52 -04:00
4fb5c60493
softIoc: extra dep for generated header
...
This is needed to make the initial build work.
2010-03-16 21:09:52 -04:00
244796bbaa
remove STATIC
2010-03-16 21:09:51 -04:00
36980c5a1a
doc: note for dbScanLock
2010-03-16 21:09:51 -04:00
3b0036b389
Improve code readablity, add more numeric tests.
2010-03-09 10:49:21 -06:00
526cfdc9e7
Warn if the IOC's TOP changed since building.
...
If the IOC's registerRecordDeviceDriver routine sees a different path
to TOP than the one used at build-time, we tell the user.
2010-03-08 17:31:19 -06:00
a782ed68b9
Fix UDF processing so we respect IVOA.
2010-03-01 13:19:43 -06:00
4bad7073e1
J field should not be PP, to match others.
2010-03-01 13:17:11 -06:00
c29551f247
Rename the time provider name routines, don't rely on the #define.
2010-02-09 10:49:08 -06:00
215bbed49c
Include standard headers before defining epicsExportSharedSymbols
2010-01-11 17:56:17 -06:00
f8e95a001b
Ensure EPICS_HOST_ARCH is defined.
...
Added support for := and ?= assignments.
Warn when include files not found.
2009-12-22 23:55:02 +00:00
023cf2c548
Added redirection support for puts() and putchar().
2009-12-22 23:47:19 +00:00
b55cad9c16
Disable "redo last put on reconnection" for STRING writing to ENUM.
2009-12-22 23:43:52 +00:00
fd5773a3fc
ESHUTDOWN is not in FreeBSD and hence not in RTEMS.
2009-12-22 17:43:25 +00:00
68b62dd4b3
Add rule to create depends file in O.<T_A> dir. Simplify Makefile.
2009-12-21 19:45:12 +00:00
Michael Davidsaver