epics/pvAccess
0
0
Fork 0
Code Issues Pull Requests Projects Wiki Activity
1,839 Commits 1 Branch 41 Tags
b06aa1831f3ccfa79ea7abee683aaafa6ee1e047
Commit Graph

159 Commits

Author SHA1 Message Date
Michael Davidsaver
b06aa1831f minor 
disambiguate error messages
 2019-07-15 14:03:50 -07:00
Michael Davidsaver
a6e7d7f5fc Remove dead code 
aliveNotification() and unresponsiveTransport() never called.
Remove these, and code only referenced through these methods.
 2019-05-21 08:54:14 -07:00
Michael Davidsaver
64453e1f1a split client/server protocol version 2019-05-21 08:51:01 -07:00
Michael Davidsaver
dfeb9fd5a5 more Troubleshoot "no more data in UDP packet" 2019-05-21 08:51:00 -07:00
Michael Davidsaver
ecee708d3a redo hexDump 
ludicrously inefficient (cf. toHex() ), and inflexible.
 2019-05-21 08:51:00 -07:00
Michael Davidsaver
0084b6e60a de-virt Transport::getRevision() 
Makes clear that this is/was only for TCP,
and eliminate redundant storage.  Also apply
locking to this storage.

Use header version directly in search handler.
Correct as only UDP search is actually used atm.
 2019-05-20 17:48:10 -07:00
Michael Davidsaver
be1bbe4aa8 const-ness 2019-05-20 17:48:10 -07:00
Michael Davidsaver
32ba2f32d0 avoid deprecated ByteBuffer::getArray() 
use identical, but not overloaded, getBuffer()
 2019-05-19 14:09:38 -07:00
Michael Davidsaver
736e0c169a Merge branch 'auth' 
* auth:
  fix osdGetRoles
  oops
  Add showauth executable
  getgrouplist() on OSX has different sig.
  auth fix mingw static
  fix search PeerInfo
  PeerInfo in pva/server.h and pva/sharedstate.h
  make PeerInfo available during search phase
  auth status void accidental success.
  minor doc
  always need to link with netapi32 now
  authorize with local group lookup
  Add osdGetRoles()
  authorization framework
  pvasr show PeerInfo
  redo security (aka. access control)
  remove asCheck

# Conflicts:
#	src/server/pv/responseHandlers.h
#	src/server/responseHandlers.cpp
 2019-05-12 14:40:29 -07:00
Michael Davidsaver
3fa1e83b55 avoid redundant sockaddr -> string 2019-04-17 09:35:43 -07:00
Michael Davidsaver
fa484a883a redo security (aka. access control) 
Break up monolithic SecurityPlugin API into three parts.

1. Authentication.
2. Authorization.
3. Access Control decision.  Left to ChannelProviders
 2019-01-12 15:14:19 -08:00
Michael Davidsaver
3b4a52a360 remove usage of NoDefaultMethods 2019-01-12 12:02:25 -08:00
Michael Davidsaver
31261cee24 client context less noise 2018-09-25 18:55:21 -07:00
Michael Davidsaver
d424202323 client: exception propagation from callbacks 
Swallowing exceptions here hides some of our errors.
Particularly unrecoverable introspectionRegistry misses.
Instead, propagate up to Thread and close connection.
 2018-06-22 13:28:49 -07:00
Michael Davidsaver
e08de8c2df detect Structure caching error 
The assert in the client side copy is hit during
some kind of failure to keep the Field cache
up to date.
 2018-06-22 13:28:49 -07:00
Michael Davidsaver
6ccca1ce58 const-ify string constants 
yes really...
 2018-05-18 10:12:34 -07:00
Michael Davidsaver
a75720120c client: call ChannelSearchManager::cancel() 2018-04-24 12:33:54 -07:00
Michael Davidsaver
1e0523b7b9 client context. Prevent transport dtor under channel mutex. 
m_channelMutex is locked recursively, so places like
disconnect(bool,bool) which purport to unlock for the
Transport detory actually don't as they are called
with m_channelMutex already locked.

Make a rather ugly hack of saving a ref to m_channel
at the various entry point.
 2018-04-04 19:25:04 -07:00
Michael Davidsaver
c9c8fb369a client context explicitly close() timer queue 2018-04-04 18:53:22 -07:00
Michael Davidsaver
d4822f141c simplify InternalClientContextImpl::destroy() 
Collapse internalDestroy() into destroy()
 2018-04-04 18:50:58 -07:00
Michael Davidsaver
128d4b67f9 simplify InternalChannelImpl::destroy() 
Collapse channel destroy(bool) and destroyChannel() into destroy()
 2018-04-04 18:48:48 -07:00
Michael Davidsaver
e97354db61 client context collapse channel createChannel() into searchResponse() 2018-04-04 18:44:11 -07:00
Michael Davidsaver
192955e7b5 client context: don't call cancel() 
it doesn't do anything!
 2018-04-04 14:55:41 -07:00
Michael Davidsaver
18491f6eb3 const-ify AbstractClientResponseHandler::_context 2018-04-04 10:12:44 -07:00
Michael Davidsaver
9e090fa191 client context: avoid lock order violations 
avoid ordering violation of calling checkAndGetTransport(),
which locks channel mutex, while request mutex is locked.
 2018-04-04 10:12:44 -07:00
Michael Davidsaver
8068ab3454 eliminate unnecessary m_nullMonitorElement 2018-04-04 10:12:44 -07:00
Michael Davidsaver
40ecc93e9d client context: use mutex guards 
whether this locking is necessary is another question...
 2018-04-04 10:12:44 -07:00
Michael Davidsaver
fe5780521b Context::getSecurityPlugins() no return of non-const ref. 
Return of map as non-const ref. allows callers to modify.
Probably unintended.
 2018-04-04 09:09:43 -07:00
Michael Davidsaver
6c2d20353a minor, avoid extra string copys 2018-03-28 13:59:33 -07:00
Michael Davidsaver
735dc22e2d collapse SimpleChannelSearchManagerImpl into ChannelSearchManager 
Virtual base with only one sub-class.  So avoid some unnecessary
indirection.
 2018-03-28 13:37:42 -07:00
Michael Davidsaver
1a100a0955 client fix init order leading to *NULL 
Avoid a race between the main thread alloc of SimpleChannelSearchManagerImpl
the the UDP socket worker trying to use the same.

Split allocation and initialization of search manager
around socket creation and worker start.

Also in-line initializeUDPTransport() to only call site.
 2018-03-28 13:21:46 -07:00
Michael Davidsaver
aec1c7c9d7 client search: cleanup context pointer handling 2018-03-28 11:25:54 -07:00
Michael Davidsaver
79ada524fa getRequester() may return NULL 2018-02-20 11:28:32 -08:00
Michael Davidsaver
5f9508a3b1 demote getTransport() fialure message to debug 
the connect() will be retried, so this isn't fatal.
 2018-02-01 20:04:03 -08:00
Michael Davidsaver
9b828852c0 Transport::getRemoteAddress() return reference 
make it clear that result is always !NULL
 2018-02-01 19:09:44 -08:00
Michael Davidsaver
f1defe4e9f mark NamedLockPattern as deprecated 
just too hard to use (exception) safely
 2018-02-01 19:00:38 -08:00
Michael Davidsaver
58f996fd98 update transportRegistry 
avoid leaks of SOCKET and leaving mutex locked
when exceptions are thrown.
 2018-02-01 18:54:25 -08:00
Michael Davidsaver
d33356ced6 drop deprecated 2018-01-31 12:48:59 -08:00
Michael Davidsaver
9daf7a0cad remove unnecessary includes 2018-01-31 11:51:13 -08:00
Michael Davidsaver
903d117485 pvAccessLog() add EPICS_PRINTF_STYLE() 
and fixup incorrect specifiers
 2017-12-28 10:49:54 -06:00
Michael Davidsaver
0161799c93 must zero osiSockAddr prior to use! 
RTEMS depends on this or eg. bind()
will fail because comparison with
list of interface addresses is done
with memcmp()!
 2017-11-28 17:54:25 -06:00
Michael Davidsaver
d4eb31e729 silence warning 
the 'done' flag avoids actual use of uninitialized 'size'.
 2017-11-28 17:54:25 -06:00
Michael Davidsaver
3f51c74ba1 fix mingw build error 2017-11-16 18:47:39 -06:00
Michael Davidsaver
cf7827ceab quiet mingw include order warning 
#warning Please include winsock2.h before windows.h
 2017-11-16 18:45:44 -06:00
Michael Davidsaver
e626ca0aaf client const-ify where possible 2017-11-16 17:38:11 -06:00
Michael Davidsaver
f45de1de68 client minor locking 2017-11-16 17:38:11 -06:00
Michael Davidsaver
8e846b02bc tighten up validation 2017-11-16 17:38:11 -06:00
Michael Davidsaver
cc9c666731 Improve client pvRequest option parsing 
Give some notice of parse errors
 2017-11-16 17:38:11 -06:00
Michael Davidsaver
00e01d5211 tighten up client operation send state handling 2017-11-16 17:03:10 -06:00
Michael Davidsaver
1d56ee7283 pva client ignore CMD_DESTROY_REQUEST reply. 
Presently CMD_DESTROY_REQUEST and CMD_CANCEL_REQUEST aren't acknowledged.
This allows ambiguity in when the request ioid goes out of scope.

Currently, the client forgets about an ioid as soon as the destroy request is sent.
So any in-flight response will be erroneously flagged as "Undecipherable message".

Similarly, a cancel request can cross paths with a reply.
This creates confusion in the event the client starts a new
request immediately after canceling the first.

Let's build some forward compatibility by ignoring replies to
cancel and destroy requests.
 2017-11-16 17:03:10 -06:00