summaryrefslogtreecommitdiff
path: root/vxi11_server.c
AgeCommit message (Collapse)Author
1999-12-31clear MAV bit after sending VXI messageroot
2013-01-31Don't dereference null pointers, don't go passing them to the parser (stops ↵daniel
lots of glib asserts).
2013-01-30Comment out a call to vxi11_fireinterrupt I put in for debuggingdaniel
2013-01-30Data lengths sent to the server shouldn't include the terminator.daniel
Fix up how the handle gets copied out into perl
2013-01-30Fire an interrupt when a new interrupt channel is requested and one exists,daniel
this is to check if the existing connection is dead and get the RPC stuff to set the error
2013-01-30Make interrupt timeout configurabledaniel
2013-01-29debugging printf on vxi11_fireinterruptMichael J. Chudobiak
1999-12-31cancel vxi panel lockout if all vxi connections are gone?root
1999-12-31Merge branch 'vxi' of ulmo:Instrument into vxiroot
2013-01-28added vxi remote/local functionalityMichael J. Chudobiak
2013-01-29Actually lock the device if requested in the link creation.daniel
2013-01-29Merge branch 'vxi' of grenfell.avtechpulse.com:Instrument into vxidaniel
2013-01-29Endian fixes.daniel
2013-01-28Add a little check to kill off an existing interrupt channel if it's died.daniel
2013-01-28Implement the abort behaviour as noted in the specdaniel
2013-01-26move locking bits around and free the lock if the link destroyeddaniel
2013-01-26Add a "hack" to report back to the vxi11 service code when a link dies.daniel
It's a hack because linux/glibc's RPC implementation doesn't seem to have the ability to set a recv error callback. The hack is to add an extra event to the events poll'd.. that event tells us if the socket died.
2013-01-25docmd and better termination fixdaniel
2013-01-25String termination fix try 2daniel
2013-01-25force termination of the strings coming from the remotedaniel
1999-12-31device_trigger now returns ERR_OPERATIONNOTSUPPORTEDroot
1999-12-31add hex dump of received data for debuggingroot
2013-01-25Add some code to track when a link was last active and some code in create ↵daniel
link and the lock check to see if a link hasn't been active for a long time and if so remove it. This should make deadlocks less of a problem.. before if a link held the lock and then disappeared no one else would ever be able to lock again. If 8 clients connected and didn't shutdown their link no more clients would be able to connect.
2013-01-24make read cleanerdaniel
2013-01-24Add a max recv value for completeness.daniel
2013-01-24make the wait loop simplerdaniel
2013-01-24fix logic for io waitdaniel
2013-01-24memset some of the static structures so that info doesn't get leaked to the ↵daniel
next caller
2013-01-24Remove the unneeded string copy, use g_malloc, g_free, bounds check on link ↵daniel
array
2013-01-24Fix the last commit that didn't actually compile because of the time being ↵daniel
messed up on my BB
2013-01-24Fix a leakdaniel
2013-01-24Copy the buffer into a fresh allocation and free the copy in globals..daniel
1999-12-31added a debugging printf in response to vxi readsroot
2013-01-17If a client disconnects while holding a lock, clear the lockdaniel
2013-01-13clean up compiler warnings in the serverdaniel
2013-01-13Would help to actually pass the data read back outdaniel
2013-01-13fire the interrupt from the right placedaniel
2013-01-12fix double freedaniel
2013-01-11most of it seems to work in perl nowdaniel
2013-01-11testing interrupt channel, worksdaniel
2013-01-11add the missing srq enable, disable checks in gpib for nowdaniel
2013-01-11more wiring updaniel
2013-01-11start feeding the stuff from write into the parserdaniel
2013-01-11Merge branch 'master' into vxidaniel
Conflicts: instr-daemon.c
2013-01-10add all of the locking calls to the calls that can be locked outdaniel
2013-01-10more filling out on the server sidedaniel
2013-01-10the interrupt channel works now.daniel
2013-01-10add some more code to the server sidedaniel
2013-01-10Interrupt channel workdaniel
2013-01-10working abort channeldaniel