Los Alamos National Laboratory
SNS Division
BESSY
Copyright © 2002 The University of Chicago, as Operator of
Argonne National Laboratory.
Copyright © 2002 The Regents of the University of California, as Operator of
Los Alamos National Laboratory.
Copyright © 2002 Berliner Elektronenspeicherringgesellschaft für
Synchrotronstrahlung.
EPICS BASE Versions 3.13.7 and higher are distributed subject to a Software License Agreement found in the file LICENSE that is included with this distribution.
Modified on 2004/01/27 20:26:20
Typically reasons to reconfigure EPICS Channel Access:
All Channel Access (CA) configuration occurs through EPICS environment variables. When searching for an EPICS environment variable EPICS first looks in the environment using the ANSI C getenv() call. If no matching variable exists then the default specified in the EPICS build system configuration files is used.
Name | Range | Default |
---|---|---|
EPICS_CA_ADDR_LIST | {N.N.N.N N.N.N.N:P ...} | <none> |
EPICS_CA_AUTO_ADDR_LIST | {YES, NO} | YES |
EPICS_CA_CONN_TMO | r > 0.1 seconds | 30.0 |
EPICS_CA_BEACON_PERIOD | r > 0.1 seconds | 15.0 |
EPICS_CA_REPEATER_PORT | i > 5000 | 5065 |
EPICS_CA_SERVER_PORT | i > 5000 | 5064 |
EPICS_CA_MAX_ARRAY_BYTES | i >= 16384 | 16384 |
EPICS_TS_MIN_WEST | -720 < i <720 minutes | 360 |
Environment variables are set differently depending on the command line shell that is in use.
C shell | setenv EPICS_CA_ADDR_LIST 1.2.3.4 |
bash | export EPICS_CA_ADDR_LIST=1.2.3.4 |
vxWorks shell | putenv ( "EPICS_CA_ADDR_LIST =1.2.3.4" ) |
DOS command line | set EPICS_CA_ADDR_LIST=1.2.3.4 |
Windows NT / 2000 / XP | control panel / system / environment tab |
Normally in a local area network (LAN) environment CA discovers the address of the host for an EPICS process variable by broadcasting frames containing a list of channel names ( CA search messages ) and waiting for responses from the servers that host the channels identified. Likewise CA clients efficiently discover that CA servers have recently joined the LAN or disconnected from the LAN by monitoring periodically broadcasted beacons sent out by the servers. Since hardware broadcasting requires special hardware capabilities, we are required to provide additional configuration information when EPICS is extended to operate over a wide area network (WAN).
Channel Access is implemented using internet protocols (IP). IP addresses are divided into host and network portions. The boundary between each portion is determined by the IP netmask. Portions of the IP address corresponding to zeros in the netmask specify the hosts address within an IP subnet. Portions of the IP address corresponding to binary ones in the netmask specify the address of a host's IP subnet. Normally the scope of a broadcasted frame will be limited to one IP subnet. Addresses with the host address portion set to all zeros or all ones are special. Modern IP kernel implementations reserve destination addresses with the host portion set to all ones for the purpose of addressing broadcasts to a particular subnet. In theory we can issue a broadcast frame on any broadcast capable LAN within the interconnected internet by specifying the proper subnet address combined with a host portion set to all ones. In practice these "directed broadcasts" are frequently limited by the default router configuration. The proper directed broadcast address required to reach a particular host can be obtained by logging into that host and typing the command required by your local operating environment. Ignore the loop back interface and use the broadcast address associated with an interface connected to a path through the network to your client. Typically there will be only one Ethernet interface.
UNIX | ifconfig -a |
vxWorks | ifShow |
Windows | ipconfig |
IP ports are positive integers. The IP address, port number, and protocol type uniquely identify the source and destination of a particular frame transmitted between computers. Servers are typically addressed by a well known port number. Clients are assigned a unique ephemeral port number during initialization. IP ports below 1024 are reserved for servers that provide standardized facilities such as mail or file transfer. Port number between 1024 and 5000 are typically reserved for ephemeral port number assignments.
The two default IP port numbers used by Channel Access may be reconfigured. This might occur when a site decides to set up two or more completely independent control systems that will share the same network. For instance, a site might set up an operational control system and a test control system on the same network. In this situation it is desirable for the test system and the operational system to use identical PV names without fear of collision. A site might also configure the CA port numbers because some other facility is already using the default port numbers. The default Channel Access port numbers have been registered with IANA.
Purpose | Default | Environment Variable |
---|---|---|
CA Server | 5064 | EPICS_CA_SERVER_PORT |
CA Beacons (sent to CA repeater daemon) | 5065 | EPICS_CA_REPEATER_PORT |
If a client needs to communicate with two servers that are residing at different port numbers then an extended syntax may be used with the EPICS_CA_ADDRESS_LIST environment variable. See WAN Environment below.
When the CA client library connects a channel it must first determine the IP address of the server the channels Process Variable resides on. To accomplish this the client sends name resolution (search) requests to a list of server destination addresses. These server destination addresses can be IP unicast addresses (individual host addresses) or IP broadcast addresses. Each name resolution (search) request contains a list of Process Variable names.If one of the servers reachable by this address list knows the IP address of a CA server that can service one or more of the specified Process Variables, then it sends back a response containing the server's IP address and port number.
During initialization CA builds the list of server destination addresses used when sending CA client name resolution (search) requests. This table is initialized by introspecting the network interfaces attached to the host. For each interface found that is attached to a broadcast capable IP subnet, the broadcast address of that subnet is added to the list. For each point to point interface found, the destination address of that link is added to the list. This automatic server address list initialization can be disabled if the EPICS environment variable "EPICS_CA_AUTO_ADDR_LIST" exists and its value is either of "no" or "NO". The typical default is to enable network interface introspection driven initialization with "EPICS_CA_AUTO_ADDR_LIST" set to "YES" or "yes".
Following network interface introspection, any IP addresses specified in the EPICS environment variable EPICS_CA_ADDR_LIST are added to the list of destination addresses for CA client name resolution requests. In an EPICS system crossing multiple subnets the EPICS_CA_ADDR_LIST must be set so that CA name resolution ( search requests ) frames pass from CA clients to the targeted CA servers unless a CA proxy (gateway) is installed. The addresses in EPICS_CA_ADDR_LIST may be dotted IP addresses or host names if the local OS has support for host name to IP address translation. When multiple names are added to EPICS_CA_ADDR_LIST they must be separated by white space. There is no requirement that the addresses specified in the EPICS_CA_ADDR_LIST be a broadcast addresses, but this will often be the most convenient choice.
C shell | setenv EPICS_CA_ADDR_LIST "1.2.3.255 8.9.10.255" |
bash | export EPICS_CA_ADDR_LIST="1.2.3.255 8.9.10.255" |
vxWorks | putenv ( "EPICS_CA_ADDR_LIST=1.2.3.255 8.9.10.255" ) |
If a client needs to communicate with two servers that are residing at different port numbers then an extended syntax may be used with the EPICS_CA_ADDRESS_LIST environment variable. Each host name or IP address in the EPICS_CA_ADDR_LIST may be immediately followed by a colon and an IP port number without intervening whitespace. Entries that do not specify a port number will default to EPICS_CA_SERVER_PORT.
C shell | setenv EPICS_CA_ADDR_LIST "1.2.3.255 8.9.10.255:10000" |
Frequently vxWorks systems boot by default with routes limiting access only to the local subnet. If a EPICS system is operating in a WAN environment it may be necessary to configure routes into the vxWorks system which enable a vxWorks based CA server to respond to requests originating outside it's subnet. These routing restrictions can also apply to vxWorks base CA clients communicating with off subnet servers. An EPICS system manager can implement an rudimentary, but robust, form of access control for a particular host by not providing routes in that host that reach outside of a limited set of subnets. See "routeLib" in the vxWorks reference manual.
If the CA client library does not see a beacon from a server that it is connected to for EPICS_CA_CONN_TMO seconds then an state-of-health message is sent to the server over TCP/IP. If this state-of-health message isn't promptly replied to then the client will assume that the server is no longer present on the network and disconnect. Disconnecting implies notification of client side application programs via function callbacks. The parameter EPICS_CA_CONN_TMO is specified in floating point seconds. The default is typically 30 seconds. For efficient operation it is recommended that EPICS_CA_CONN_TMO be set to no less than twice the value specified for EPICS_CA_BEACON_PERIOD.
The CA client library will continuously attempt to connect any CA channels that an application has created until it is successful. The library periodically queries the server destination address list described above with name resolution requests for any unresolved channels. Since this address list frequently contains broadcast addresses, and because nonexistent process variable names are frequently configured, or servers may be temporarily unavailable, then it is necessary for the CA client library internals to carefully schedule these requests in time to avoid introducing excessive load on the network and the servers.
When the CA client library has many channels to connect, and most of its name resolution requests are responded to, then it sends name resolution requests at an interval that is twice the estimated round trip interval for the set of servers responding, or at the minimum delay quantum for the operating system - whichever is greater. The number of UDP frames per interval is also dynamically adjusted based on the past success rate.
If name resolution requests are not responded to, then the client library doubles the delay between name resolution attempts and reduces the number of requests per interval. The delay between attempts is initially limited by a maximum however, after some long interval, if the client library does not receive any responses it stops sending name resolution attempts altogether until it sees a beacon anomaly.
The CA client library continually estimates the beacon period of all server beacons received. If a particular server's beacon period becomes significantly shorter or longer then the client is said to detect a beacon anomaly. When a client sees a beacon anomaly then it resumes search requests but with a longer initial interval between requests than is used when the application creates a channel. An initial delay based on the client's ephemeral port number is also imposed before the first name resolution request to avoid all clients responding to a beacon anomaly at the same instant. The program "casw" prints a message on standard out each time that a CA client will detect a beacon anomaly.
Two conclusions deserve special emphasis. First, if a client does not see the server's beacons, then it will use additional network and server resources sending periodic state-of-health messages. Second, if a client does not see the server's beacons, then it may not connect to a newly introduced server that was initially inaccessible if the client timed out attempting to find it. The typical situation where a client would not see the server's beacon might be when the client isnt on the same IP subnet as the server, and the EPICS_CA_ADDR_LIST was modified to include a destination address for the server, but the server's beacon address list was not modified so that it's beacons are received by the client.
When several client processes run on the same host it is not possible for all of them to directly receive a copy of the server beacon messages when the beacon messages are sent to unicast addresses, or when legacy IP kernels are still in use. To avoid confusion over these restrictions a special UDP server, the CA Repeater, is automatically spawned by the CA client library when it is not found to be running. This program listens for server beacons sent to the UDP port specified in the EPICS_CA_REPEATER_PORT parameter and fans any beacons received out to any CA client program running on the same host that have registered themselves with the CA Repeater. If the CA Repeater is not already running on a workstation, then the "caRepeater" program must be in your path before using the CA client library for the first time. If a host based IOC is run on the same workstation with standalone CA client processes, then it is probably best to start the caRepeater process when the workstation is booted. Otherwise it is possible for the standalone CA client processes to become dependent on a CA repeater started within the confines of the host based IOC. As long as the host based IOC continues to run there is nothing wrong with this situation, but problems could arise if this host based IOC process exits before the standalone client processes which are relying on its CA repeater for services exit.
Note: Starting with EPICS R3.14 all of the libraries in the EPICS base distribution rely on facilities built into the operating system to determine the correct time zone. Nevertheless, several programs commonly used with EPICS still use the original "tssubr" library and therefore they still rely on proper configuration of EPICS_TS_MIN_WEST.
While the CA client library does not translate in between the local time and the time zone independent internal storage of EPICS time stamps, many EPICS client side applications call core EPICS libraries which provide these services. To set the correct time zone users must compute the number of positive minutes west of GMT (maximum 720 inclusive) or the negative number of minutes east of GMT (minimum -720 inclusive). This integer value is then placed in the variable EPICS_TS_MIN_WEST.
Time Zone | EPICS_TS_MIN_WEST |
---|---|
USA Eastern | 300 |
USA Central | 360 |
USA Mountain | 420 |
USA Pacific | 480 |
Alaska | 540 |
Hawaii | 600 |
Japan | -540 |
Germany | -120 |
United Kingdom | 0 |
The environment variable EPICS_CA_MAX_ARRAY_BYTES determines the size of the largest array that may pass through CA. This parameter must be set appropriately for both the CA client and the CA server. In EPICS R3.14 CA maintains a free list of 16384 byte network buffers that are used for ordinary communication. If EPICS_CA_MAX_ARRAY_BYTES is larger than 16384 then a second free list of larger data buffers is established when clients request transportation of large arrays.
Name | Range | Default |
---|---|---|
EPICS_CAS_SERVER_PORT | i > 5000 | EPICS_CA_SERVER_PORT |
EPICS_CAS_AUTO_BEACON_ADDR_LIST | {YES, NO} | EPICS_CA_AUTO_ADDR_LIST |
EPICS_CAS_BEACON_ADDR_LIST | {N.N.N.NN.N.N.N:P...} | EPICS_CA_ADDR_LIST1 |
EPICS_CAS_BEACON_PERIOD | r > 0.1 seconds | 15.0 |
EPICS_CAS_BEACON_PORT | i > 5000 | EPICS_CA_REPEATER_PORT |
EPICS_CAS_INTF_ADDR_LIST | {N.N.N.NN.N.N.N:P...} | <none> |
EPICS_CAS_IGNORE_ADDR_LIST | {N.N.N.NN.N.N.N:P...} | <none> |
The server configures its port number from the EPICS_CAS_SERVER_PORT environment variable if it is specified. Otherwise the EPICS_CA_SERVER_PORT environment variable determines the server's port number. Two servers can share the same UDP port number on the same machine, but there are restrictions - see a discussion of unicast addresses and two servers sharing the same UDP port on the same host.
The EPICS_CAS_BEACON_PERIOD parameter determines the server's beacon period and is specified in floating point seconds. The default is typically 15 seconds. See also EPICS_CA_CONN_TMO and Dynamic Changes in the CA Client Library Search Interval.
CA servers build a list of addresses to send beacons to during initialization. If EPICS_CAS_AUTO_BEACON_ADDR_LIST has the value "YES" then the beacon address list will contain at least the broadcast address of all LAN interfaces found in the host and the destination address of all point-to-point interfaces found in the host.
If EPICS_CAS_BEACON_ADDR_LIST is defined then its contents will be used to augment this list. Individual entries in EPICS_CAS_BEACON_ADDR_LIST may override the destination port number if ":nnn" follows the host name or IP address there. Alternatively, if EPICS_CAS_BEACON_ADDR_LIST is not defined, EPICS_CA_ADDR_LIST is defined, and EPICS_CAS_INTF_ADDR_LIST is not defined, then the contents of EPICS_CA_ADDR_LIST will be used to augment the list. Otherwise, the list is not augmented.
The EPICS_CAS_BEACON_PORT parameter specifies the destination port for server beacons. The only exception to this occurs when ports are specified in EPICS_CAS_BEACON_ADDR_LIST or possibly in EPICS_CA_ADDR_LIST. If EPICS_CAS_BEACON_PORT is not specified then beacons are sent to the port specified in EPICS_CA_REPEATER_PORT.
The parameter EPICS_CAS_INTF_ADDR_LIST allows a ca server to bind itself to, and therefore accept messages only from, a limited set of network interfaces (each specified by it's IP address). Specifically, UDP search messages addressed to both the IP addresses in EPICS_CAS_INTF_ADDR_LIST and also to the broadcast addresses of the corresponding LAN interfaces will be accepted by the server. By default, the CA server is accessible from all network interfaces configured into its host. In R3.14 and previous releases the CA server employed by iocCore does not implemet this feature.
Name resolution requests originating from any of the IP addresses specified in the EPICS_CAS_IGNORE_ADDR_LIST parameter are not replied to. In R3.14 and previous releases the CA server employed by iocCore does not implemet this feature.
See also Configuring the Maximum Array Size.
See also Routing Restrictions on vxWorks Systems.
acctst <PV name> [progress logging level] [channel duplication count] [test repetition count] [enable preemptive callback]
Channel Access Client Library regression test.
The PV used with the test must be native type DBR_DOUBLE or DBR_FLOAT, and modified only by acctst while the test is running. Therefore, periodically scanned hardware attached analog input records do not work well. Test failure is indicated if the program stops prior to printing "test complete". If unspecified the progress logging level is zero, and no messages are printed while the test is progressing. If unspecified, the channel duplication count is 20000. If unspecified, the test repetition count is once only. If unspecified, preemptive callback is disabled.
catime <PV name> [channel count] [append number to pv name if true]
Channel Access Client Library performance test.
If unspecified, the channel count is 10000. If the "append number to pv name if true" argument is specified and it is greater than zero then the channel names in the test are numbered as follows.
<PV name>000000, <PV name>000001, ... <PV name>nnnnnn
casw [-i <interest level>]
CA server "beacon anomaly" logging.
CA server beacon anomalies occur when a new server joins the network, a server is rebooted, network connectivity to a server is reestablished, or if a server's CPU exits a CPU load saturated state.
CA clients with unresolved channels reset their search request scheduling timers whenever they see a beacon anomaly.
This program can be used to detect situations where there are too many beacon anomalies. IP routing configuration problems may result in false beacon anomalies that might cause CA clients to use unnecessary additional network bandwidth and server CPU load when searching for unresolved channels.
If there are no new CA servers appearing on the network, and network connectivity remains constant, then casw should print no messages at all.
caEventRate <PV name> [subscription count]
Connect to the specified PV, subscribe for monitor updates the specified number of times (default once), and periodically log the current sampled event rate, average event rate, and the standard deviation of the event rate in Hertz to standard out.
ca_test <PV name> [value to be written]
If a value is specified it is written to the PV. Next, the current value of the PV is converted to each of the many external data type that can be specified at the CA client library interface, and each of these is formated and then output to the console.
Note: The CA Command Line Tools are currently under development. Thus only the first of the tools is included in EPICS Base R3.14.5, while the others are still being worked on. These tools have not yet proven to be stable and reliable. The user interface might still change, so please be careful relying on these early versions in scripts and other crucial places. Please report any bugs or unexpected behaviour to the author: [email protected]
caget [options] <PV name> ...
Get and print value for PV(s).
The values for one or multiple PVs are read and printed to stdout. The DBR_... format in which the data is read, the output format, and a number of details of how integer and float values are represented can be controlled using command line options.
Option | Description | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
-h | Print usage information | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CA options: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-w <sec> | Wait time, specifies longer CA timeout, default is 1.0 second | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-c | Asynchronous get (use ca_get_callback instead of ca_get) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Format and data type options: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Default output format is "name value" | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-t | Terse mode - print only value, without name | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-a | Wide mode "name timestamp value stat sevr" (read PVs as DBR_TIME_xxx) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-n | Print DBF_ENUM values as number (default are enum string values) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-d <type> | Request specific dbr type; use string (DBR_ prefix may be omitted)
or number of one of the following types:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Arrays: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Value format: Print number of requested values, then list of values | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Default: | Print all values | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-# <count> | Print first <count> elements of an array | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Floating point type format: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Default: | Use %g format | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-f <nr> | Use %f format, with <nr> digits after the decimal point | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-e <nr> | Use e format, with <nr> digits after the decimal point | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Integer number format: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Default: | Print as decimal number | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-0x | Print as hex number | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-0o | Print as octal number | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-0b | Print as binary number |
camonitor [options] <PV name> ...
Subscribe to and print value updates for PV(s).
Not implemented yet. Under development.
caput [options] <PV name> <value>
Put value to a PV.
Not implemented yet. Under development.
cainfo [options] <PV name> ...
Get and print channel and connection information for PV(s).
Not implemented yet. Under development.
Verify that the broadcast addresses are identical on the server's host and on the client's host. This can be checked on UNIX with "netstat -i" or "ifconfig -a"; on vxWorks with ifShow; and on windows with ipconfig. It is normal for the broadcast addresses to not be identical if the client and server are not directly attached to the same IP subnet, and in this situation the EPICS_CA_ADDR_LIST must be set. Otherwise, if the client and server are intended to be on the same IP subnet, then the problem may be that the IP netmask is incorrectly set in the network interface configuration. On most operating systems, when the host's IP address is configured, the host's IP subnet mask is also configured.
Verify that the client and server are using the same UDP port. Check the server's port by running "netstat -a | grep nnn" where nnn is the port number configured in the client. If you do not set EPICS_CA_SERVER_PORT or EPICS_CAS_SERVER_PORT then the default port will be 5064.
Two servers can run on the same host with the same server port number, but there are restrictions. If the host has a modern IP kernel it is possible to have two or more servers share the same UDP port. It is not possible for these servers to run on the same host using the same TCP port. If the CA server library detects that a server is attempting to start on the same port as an existing CA server then both servers will use the same UDP port, and the 2nd server will be allocated an ephemeral TCP port. Clients can be configured to use the same port number for both servers. They will locate the 2nd server via the shared UDP port, and transparently connect to the 2nd server's ephemeral TCP port. Be aware however that If there are two server's running on the same host sharing the same UDP port then they will both receive UDP search requests sent as broadcasts, but unfortunately (due to a weakness of most IP kernel implementations) only one of the servers will typically receive UDP search requests sent to unicast addresses (i.e. a single specific host's ip address).
See Dynamic Changes in the CA Client Library Search Interval.
When communication over a virtual circuit times out, then each channel attached to the circuit enters a disconnected state and the disconnect callback handler specified for the channel is called. However, the circuit is not disconnected until TCP/IP's internal, typically long duration, keep alive timer expires. The disconnected channels remain attached to the beleaguered circuit and no attempt is made to search for, or to reestablish, a new circuit. If, at some time in the future, the circuit becomes responsive again, then the attached channels enter a connected state again and reconnect call back handlers are called. Any monitor subscriptions that received an update message while the channel was disconnected are also refreshed. If at any time the library receives an indication from the operating system that a beleaguered circuit has shutdown or was disconnected then the library will immediately reattempt to find servers for each channel and connect circuits to them.
A well known negative side effect of the above behavior is that CA clients will wait the full (typically long) duration of TCP/IP's internal keep alive timer prior to reconnecting under the following scenario (all of the following occur):
It is unlikely that any rational organization will advocate the above scenario in a production system. Nevertheless, there are opportunities for users to become confused during control system development, but it is felt that the robustness improvements justify isolated confusion during the system integration and checkout activities where the above scenarios are most likely to occur.
Contrast the above behavior with the CA client library behavior of releases prior to R3.14.5 where the beleaguered circuit was immediately closed when communication over it timed out. Any attached channels were immediately searched for, and after successful search responses arrived then attempts were made to build a new circuit. This behavior could result in undesirable resource consumption resulting from periodic circuit setup and teardown overhead (thrashing) during periods of CPU / network / IP kernel buffer congestion.
Many Berkley UNIX derived Internet Protocol (IP) kernels use a memory management scheme with a fixed sized low level memory allocation quantum called an "mbuf". Messages about "ENOBUFS" are an indication that your IP kernel is running low on mbuf buffers. An IP kernel mbuf starvation situation may lead to temporary IP communications stalls or reduced throughput. This issue has to date been primarily associated with vxWorks systems where mbuf starvation on earlier vxWorks versions is rumored to lead to permanent IP communications stalls which are resolved only by a system reboot. IP kernels that use mbufs frequently allow the initial and maximum number of mbufs to be configured. Consult your OS's documentation for configuration procedures which vary between OS and even between different versions of the same OS.
Significant performance gains can be realized when the CA client library doesn't wait for a response to return from the server after each request. All requests which require interaction with a CA server are accumulated (buffered) and not forwarded to the IOC until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called allowing several operations to be efficiently sent over the network together. Any process variable values written into your program's variables by ca_get() should not be referenced by your program until ECA_NORMAL has been received from ca_pend_io().
If successful, the routines described here return the status code ECA_NORMAL. Unsuccessful status codes returned from the client library are listed with each routine in this manual. Operations that appear to be valid to the client can still fail in the server. Writing the string "off" to a floating point field is an example of this type of error. If the server for a channel is located in a different address space than the client then the ca_xxx() operations that communicate with the server return status indicating the validity of the request and whether it was successfully enqueued to the server, but communication of completion status is deferred until a user callback is called, or lacking that an exception handler is called. An error number and the error's severity are embedded in CA status (error) constants. Applications shouldn't test the success of a CA function call by checking to see if the returned value is zero as is the UNIX convention. Below are several methods to test CA function returns. See ca_signal() and SEVCHK for more information on this topic.
status = ca_XXXX();
SEVCHK( status, "ca_XXXX() returned failure status");
if ( status & CA_M_SUCCESS ) {
printf ( "The requested ca_XXXX() operation didn't complete successfully");
}
if ( status != ECA_NORMAL ) {
printf("The requested ca_XXXX() operation didn't complete successfully because \"%s\"\n",
ca_message ( status ) );
}
Arguments of type chtype specifying the data type you wish to transfer. They expect one of the set of DBR_XXXX data type codes defined in db_access.h. There are data types for all of the C primitive types, and there are also compound (C structure) types that include various process variable properties such as units, limits, time stamp, or alarm status. The primitive C types follow a naming convention where the C typedef dbr_xxxx_t corresponds to the DBR_XXXX data type code. The compound (C structure) types follow a naming convention where the C structure tag dbr_xxxx corresponds to the DBR_XXXX data type code. The following table provides more details on the structure of the CA data type space. Since data addresses are passed to the CA client library as typeless "void *" pointers then care should be taken to ensure that you have passed the correct C data type corresponding to the DBR_XXXX type that you have specified. Architecture independent types are provided in db_access.h to assist programmers in writing portable code. For example "dbr_short_t" should be used to send or receive type DBR_SHORT.
CA Type Code | Read / Write | Primitive C Data Type | Process Variable Properties |
---|---|---|---|
DBR_<PRIMITIVE TYPE> | RW | dbr_<primitive type>_t | value |
DBR_STS_<PRIMITIVE TYPE> | R | struct dbr_sts_<primitive type> | value, alarm status, and alarm severity |
DBR_TIME_<PRIMITIVE TYPE> | R | struct dbr_time_<primitive type> | value, alarm status, alarm severity, and time stamp |
DBR_GR_<PRIMITIVE TYPE> | R | struct dbr_gr_<primitive type> | value, alarm status, alarm severity, units, display precision, and graphic limits |
DBR_CTRL_<PRIMITIVE TYPE> | R | struct dbr_ctrl_<primitive type> | value, alarm status, alarm severity, units, display precision, graphic limits, and control limits |
DBR_PUT_ACKT | W | dbr_put_ackt_t | Used for global alarm acknowledgement. Do transient alarms have to be acknowledged? (0,1) means (no, yes). |
DBR_PUT_ACKS | W | dbr_put_acks_t | Used for global alarm acknowledgement. The highest alarm severity to acknowledge. If the current alarm severity is less then or equal to this value the alarm is acknowledged. |
DBR_STSACK_STRING | R | struct dbr_stsack_string | value, alarm status, alarm severity, ackt, ackv |
DBR_CLASS_NAME | R | dbr_class_name_t | name of enclosing interface (name of the record if channel is attached to EPICS run time database) |
Channel value arrays can also be included within the structured CA data types. If more than one element is requested, then the individual elements can be accessed in an application program by indexing a pointer to the value field in the DBR_XXX structure. For example, the following code computes the sum of the elements in a array process variable and prints its time stamp. The dbr_size_n function can be used to determine the correct number of bytes to reserve when there are more than one value field elements in a structured CA data type.
#include <stdio.h>
#include <stdlib.h>
#include "cadef.h"
int main ( int argc, char ** argv )
{
struct dbr_time_double * pTD;
const dbr_double_t * pValue;
unsigned nBytes;
unsigned elementCount;
char timeString[32];
unsigned i;
chid chan;
double sum;
int status;
if ( argc != 2 ) {
fprintf ( stderr, "usage: %s <channel name>", argv[0] );
return -1;
}
status = ca_create_channel ( argv[1], 0, 0, 0, & chan );
SEVCHK ( status, "ca_create_channel()" );
status = ca_pend_io ( 15.0 );
if ( status != ECA_NORMAL ) {
fprintf ( stderr, "\"%s\" not found.\n", argv[1] );
return -1;
}
elementCount = ca_element_count ( chan );
nBytes = dbr_size_n ( DBR_TIME_DOUBLE, elementCount );
pTD = ( struct dbr_time_double * ) malloc ( nBytes );
if ( ! pTD ) {
fprintf ( stderr, "insufficient memory to complete request\n" );
return -1;
}
status = ca_array_get ( DBR_TIME_DOUBLE, elementCount, chan, pTD );
SEVCHK ( status, "ca_array_get()" );
status = ca_pend_io ( 15.0 );
if ( status != ECA_NORMAL ) {
fprintf ( stderr, "\"%s\" didnt return a value.\n", argv[1] );
return -1;
}
pValue = & pTD->value;
sum = 0.0;
for ( i = 0; i < elementCount; i++ ) {
sum += pValue[i];
}
epicsTimeToStrftime ( timeString, sizeof ( timeString ),
"%a %b %d %Y %H:%M:%S.%f", & pTD->stamp );
printf ( "The sum of elements in %s at %s was %f\n",
argv[1], timeString, sum );
ca_clear_channel ( chan );
ca_task_exit ();
free ( pTD );
return 0;
}
Certain CA client initiated requests asynchronously execute an application
supplied call back in the client process when a response arrives. The
functions ca_put_callback, ca_get_callback, and ca_add_event all request
notification of asynchronous completion via this mechanism. The
event_handler_args
structure is passed by value to the
application supplied callback. In this structure the dbr
field
is a void pointer to any data that might be returned. The
s
tatus
field will be set to one of the CA error
codes in caerr.h and will indicate the status of the operation performed in
the IOC. If the status field isn't set to ECA_NORMAL or data isn't normally
returned from the operation (i.e. put call back) then you should expect that
the dbr
field will be set to a nill pointer (zero). The fields
usr
, chid
, and type
are set to the
values specified when the request was made by the application.
typedef struct event_handler_args {
void *usr; /* user argument supplied with request */
chanId chid; /* channel id */
long type; /* the type of the item returned */
long count; /* the element count of the item returned */
const void *dbr; /* a pointer to the item returned */
int status; /* ECA_XXX status of the requested op from the server */
} evargs;
void myCallback ( struct event_handler_args args )
{
if ( args.status != ECA_NORMAL ) {
}
if ( args.type == DBR_TIME_DOUBLE ) {
const struct dbr_time_double * pTD =
( const struct dbr_time_double * ) args.dbr;
}
}
When the server detects a failure, and there is no client call back function attached to the request, then an exception handler is executed in the client. The default exception handler prints a message on the console and exits if the exception condition is severe. Certain internal exceptions within the CA client library, and failures detected by the SEVCHK macro may also cause the exception handler to be invoked. To modify this behavior see ca_add_exception_event().
If the Process Variable's server and it's client are colocated within the same memory address space and the same host then the ca_xxx() operations bypass the server and directly interact with the server tool component (commonly the IOC's function block database). In this situation the ca_xxx() routines frequently return the completion status of the requested operation directly to the caller with no opportunity for asynchronous notification of failure via an exception handler. Likewise, callbacks may be directly invoked by the CA library functions that request them.
For routines that require an argument specifying the number of array elements, no more than the process variable's maximum native element count may be requested. The process variable's maximum native element count is available from ca_element_count() when the channel is connected. If less elements than the process variable's native element count are requested the requested values will be fetched beginning at element zero. By default CA limits the number of elements in an array to be no more than approximately 16k divided by the size of one element in the array. Starting with EPICS R3.14 the maximum array size may be configured in the client and in the server.
Application programs should assume that CA server may be restarted, and that network connectivity is transient. When you create a CA channel it's initial connection state will most commonly be disconnected. If the Process Variable's server is available the library will immediately initiate the necessary actions to make a connection with it. Otherwise, the client library will monitor the state of servers on the network and immediately connect or reconnect with the process variable's server when it becomes available.
Two methods may be used to determine if a channel has connected: the
application program can block in ca_pend_io
, or the application program can
install a connection callback handler when it calls ca_create_channel
. The ca_pend_io
approach is best suited to simple
command line programs with a short runtime duration, and the connection
callback method is best suited to toolkit components with a long runtime
duration. If a connection state change call back function is not
installed when ca_create_channel
is called (if a nil function
pointer is supplied) then the application program must wait for
successful status from ca_pend_io
prior to using the channel for the first time. Otherwise, if a connection
state change call back function is supplied, then one of the
arguments to this function distinguishes between connect and disconnect
events, and ca_pend_io
will not
block for the channel to connect. The user's connection state change
function will be run immediately from within ca_create_channel
if the CA client and
the server are both hosted within the same address space (within the same
process).
Once the channel connects the application program can freely perform IO operations through the channel, but it should expect that the channel might disconnect at any time due to network connectivity disruptions or server restarts.
Starting with EPICS R3.14 the CA client libraries are fully thread safe on all OS (in past releases the library was thread safe only on vxWorks). When the client library is initialized the programmer may specify if preemptive call back is enabled. Preemptive call back is disabled by default. If preemptive call back is enabled then the user's call back functions might be called by CA's auxiliary threads when the main initiating channel access thread is not inside of a function in the channel access client library. Otherwise, the user's call back functions will be called only when the main initiating channel access thread is executing inside of the CA client library. When the CA client library invokes a user's call back function it will always wait for the current callback to complete prior to executing another call back function.
To set up a traditional single threaded client you will need code like this (see ca_context_create and CA Client Contexts and Application Specific Auxiliary Threads) .
SEVCHK ( ca_context_create(ca_disable_preemptive_callback ),
"application pdq calling ca_context_create" );
To set up a preemptive callback enabled CA client context you will need code like this (see ca_context_createand CA Client Contexts and Application Specific Auxiliary Threads) .
SEVCHK ( ca_context_create(ca_enable_preemptive_callback ),
"application pdq calling ca_context_create" );
It may be necessary for several CA client side tools running in the same address space (process) to be independent of each other. For example, the database CA links and the sequencer are designed to not use the same CA client library threads, network circuits, and data structures. Each thread that calls ca_context_create() for the first time either directly, or implicitly when calling a CA routine for the first time, creates a CA client library context. A CA client library context contains all of the threads, network circuits, and data structures required to connect and communicate with the channels that a CA client application has created. The priority of auxiliary threads spawned by the CA client library are at fixed offsets from the priority of the thread that called ca_context_create(). An application specific auxiliary thread can join a CA context by calling ca_attach_context() using the CA context identifier that was returned from ca_current_context() when it was called by the thread that called ca_context_create(). A CA client library context can be shut down and cleaned up, after destroying any channels or application specific threads that are attached to it, by calling ca_context_destroy().
If preemptive call back is not enabled, then for proper operation CA must periodically be polled to take care of background activity. This requires that your application must either wait in one of ca_pend_event(), ca_pend_io(), or ca_sg_block() or alternatively it must call ca_poll() at least every 100 milli-seconds. In single threaded applications a file descriptor manager like Xt or the interface described in fdManager.h can be used to monitor both mouse clicks and also CA's file descriptors so that ca_poll() can be called immediately when CA server messages arrives over the network.
With the embryonic releases of EPICS it was a common practice to examine a
channel's connection state, its native type, and its native element count by
directly accessing fields in a structure using a pointer stored in type
chid
. Likewise, a user private pointer in the per channel
structure was also commonly set by directly accessing fields in the channel
structure. A number of difficulties arise from this practice, which has long
since been deprecated. For example, prior to release 3.13 it was recognized
that transient changes in certain private fields in the per channel structure
would make it difficult to reliably test the channels connection state using
these private fields directly. Therefore, in release 3.13 the names of
certain fields were changed to discourage this practice. Starting with
release 3.14 codes written this way will not compile. Codes intending to
maintain the highest degree of portability over a wide range of EPICS
versions should be especially careful. For example you should replace all
instances off channel_id->count
with
ca_element_count(channel_id)
. This approach should be reliable
on all versions of EPICS in use today. The construct ca_puser(chid) =
xxxx
is particularly problematic. The best mechanisms for setting the
per channel private pointer will be to pass the user private pointer in when
creating the channel. This approach is implemented on all versions.
Otherwise, you can also use ca_set_puser(CHID,PUSER)
, but this
function is available only after the first official (post beta) release of
EPICS 3.13.
ca_context_create()
#include <cadef.h>
enum ca_preemptive_callback_select
{ ca_disable_preemptive_callback, ca_enable_preemptive_callback };
int ca_context_create ( enum ca_preemptive_callback_select SELECT );
This function should be called once prior to making any of the other channel access calls.
If ca_disable_preemptive_callback
is specified then
additional threads are not allowed to join the CA context using
ca_context_attach() because allowing other threads to join implies that CA
callbacks will be called preemptively from more than one thread.
SELECT
ca_disable_preemptive_callback
. If
ca_enable_preemptive_callback is specified then CA client background
activities, such as connection management, will proceed even if the
thread that calls this routine is not executing in the CA client
library.ECA_NORMAL - Normal successful completion
ECA_ALLOCMEM - Failed, unable to allocate space in pool
ca_context_destroy()
ca_context_destroy()
#include <cadef.h> void ca_context_destroy();
Shut down a channel access client context and free any resources allocated. On most operating systems this is performed automatically at process exit.
ECA_NORMAL - Normal successful completion
ca_create_channel()
#include <cadef.h>
typedef void ( *pCallBack ) (
struct connection_handler_args );
int ca_create_channel
(
const char *PROCESS_VARIABLE_NAME,
caCh *USERFUNC,
void *PUSER,
capri priority,
chid *PCHID
);
This function creates a CA channel. The CA client library will attempt to establish and maintain a virtual circuit between the caller's application and a named process variable in a CA server. Each call to ca_create_channel allocates resources in the CA client library and potentially also a CA server. The function ca_clear_channel() is used to release these resources. If successful, the routine writes a channel identifier into the user's variable of type "chid". This identifier can be used with any channel access call that operates on a channel.
The circuit may be initially connected or disconnected depending on the state of the network and the location of the channel. A channel will only enter a connected state after server's address is determined, and only if channel access successfully establishes a virtual circuit through the network to the server. Channel access routines that send a request to a server will return ECA_DISCONNCHID if the channel is currently disconnected.
There are two ways to obtain asynchronous notification when a channel enters a connected state.
The function ca_state(CHID) can be used to test the connection state of a channel. Valid connections may be isolated from invalid ones with this function if ca_pend_io() times out.
Due to the inherently transient nature of network connections the order of connection call backs relative to the order that ca_create_channel() calls are made by the application can't be guaranteed, and application programs may need to be prepared for a connected channel to enter a disconnected state at any time.
See caExample.c in the example application created by makeBaseApp.pl.
PROCESS_VARIABLE_NAME
USERFUNC
The following structure is passed by value to the user's
connection connection callback function. The op
field will
be set by the CA client library to CA_OP_CONN_UP
when the
channel connects, and to CA_OP_CONN_DOWN
when the channel
disconnects. See ca_puser
if the
PUSER
argument is required in your callback
handler.
struct ca_connection_handler_args {
chanId chid; /* channel id */
long op; /* one of CA_OP_CONN_UP or CA_OP_CONN_DOWN */
};
PUSER
PRIORITY
PCHID
ECA_NORMAL - Normal successful completion
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_STRTOBIG - Unusually large string
ECA_ALLOCMEM - Unable to allocate memory
ca_clear_channel()
#include <cadef.h>
int ca_clear_channel (evid CHID);
Shutdown and reclaim resources associated with a channel created by ca_create_channel().
All remote operation requests such as the above are accumulated (buffered) and not forwarded to the IOC until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently sent over the network in one message.
Clearing a channel does not cause its disconnect handler to be called, but clearing a channel does shutdown and reclaim any channel state change event subscriptions (monitors) registered with the channel.
CHID
ECA_NORMAL - Normal successful completion
ECA_BADCHID - Corrupted CHID
ca_put()
#include <cadef.h>
int ca_put ( chtype TYPE,
chid CHID, void *PVALUE );
int ca_array_put ( chtype TYPE,
unsigned long COUNT,
chid CHID, const void *PVALUE);
typedef void ( *pCallBack ) (struct event_handler_args );
int ca_put_callback ( chtype TYPE,
chid CHID, const void *PVALUE,
pCallBack PFUNC, void *USERARG );
int ca_array_put_callback ( chtype TYPE,
unsigned long COUNT,
chid CHID, const void *PVALUE,
pCallBack PFUNC, void *USERARG );
Write a scalar or array value to a process variable.
When ca_array_put or ca_put are invoked the client will receive no response unless the request can not be fulfilled in the server. If unsuccessful an exception handler is run on the client side. If a connection is lost and then resumed outstanding ca_array_put or ca_put requests are not automatically reissued following reconnect, and no additional notification are provided to the user for each put request.
When ca_array_put_callback are invoked the user supplied asynchronous call back is called only after the initiated write operation and all actions resulting from the initiating write operation complete. If unsuccessful the call back function is invoked indicating bad status. If the channel disconnects before a put callback request can be completed, then the client's call back function is called with bad status, but this does not guarantee that the server did not receive and process the request before the disconnect.
All of these functions return ECA_DISCONN if the channel is currently disconnected.
All put requests are accumulated (buffered) and not forwarded to the IOC until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently combined into one message.
TYPE
COUNT
CHID
PVALUE
PFUNC
USERARG
ECA_NORMAL - Normal successful completion
ECA_BADCHID - Corrupted CHID
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_BADCOUNT - Requested count larger than native element count
ECA_STRTOBIG - Unusually large string supplied
ECA_NOWTACCESS - Write access denied
ECA_ALLOCMEM - Unable to allocate memory
ECA_DISCONN - Channel is disconnected
ca_pend_event()
ca_get()
#include <cadef.h>
int ca_get ( chtype TYPE,
chid CHID, void *PVALUE );
int ca_array_get ( chtype TYPE, unsigned long COUNT,
chid CHID, void *PVALUE );
typedef void ( *pCallBack ) (struct event_handler_args );
int ca_get_callback ( chtype TYPE,
chid CHID, pCallBack USERFUNC, void *USERARG);
int ca_array_get_callback ( chtype TYPE, unsigned long COUNT,
chid CHID,
pCallBack USERFUNC, void *USERARG );
Read a scalar or array value from a process variable.
When ca_get or ca_array_get are invoked the returned channel value cant be assumed to be stable in the application supplied buffer until after ECA_NORMAL is returned from ca_pend_io. If a connection is lost outstanding get requests are not automatically reissued following reconnect.
When ca_get_callback or ca_array_get_callback are invoked a value is read from the channel and then the user's callback is invoked with a pointer to the retrieved value. Note that ca_pend_io will not block for the delivery of values requested by ca_get_callback. If the channel disconnects before a get callback request can be completed, then the clients call back function is called with bad status.All of these functions return ECA_DISCONN if the channel is currently disconnected.
All get requests are accumulated (buffered) and not forwarded to the IOC until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently sent over the network in one message.
See caExample.c in the example application created by makeBaseApp.pl.
TYPE
COUNT
CHID
PVALUE
USERFUNC
USERARG
ECA_NORMAL - Normal successful completion
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_BADCHID - Corrupted CHID
ECA_BADCOUNT - Requested count larger than native element count
ECA_GETFAIL - A local database get failed
ECA_NORDACCESS - Read access denied
ECA_ALLOCMEM - Unable to allocate memory
ECA_DISCONN - Channel is disconnected
ca_pend_event()
ca_create_subscription()
#include <cadef.h>
typedef void ( *pCallBack ) (
struct event_handler_args );
int ca_create_subscription ( chtype TYPE,
unsigned long COUNT, chid CHID,
unsigned long MASK, pCallBack USERFUNC, void *USERARG,
evid *PEVID );
Register a state change subscription and specify a call back function to be invoked whenever the process variable undergoes significant state changes. A significant change can be a change in the process variable's value, alarm status, or alarm severity. In the process control function block database the deadband field determines the magnitude of a significant change for for the process variable's value. Each call to this function consumes resources in the client library and potentially a CA server until one of ca_clear_channel or ca_clear_event is called.
Subscriptions may be installed or canceled against both connected and disconnected channels. The specified USERFUNC is called once immediately after the subscription is installed with the process variable's current state if the process variable is connected. Otherwise, the specified USERFUNC is called immediately after establishing a connection (or reconnection) with the process variable. The specified USERFUNC is called immediately with the process variable's current state from within ca_add_event() if the client and the process variable share the same address space.
If a subscription is installed on a channel in a disconnected state then the requested count will be set to the native maximum element count of the channel if the requested count is larger.
All subscription requests such as the above are accumulated (buffered) and not forwarded to the IOC until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently sent over the network in one message.
If at any time after subscribing, read access to the specified process variable is lost, then the call back will be invoked immediately indicating that read access was lost via the status argument. When read access is restored normal event processing will resume starting always with at least one update indicating the current state of the channel.
A better name for this function might have been ca_subscribe.
See caMonitor.c in the example application created by makeBaseApp.pl.
TYPE
COUNT
CHID
USRERFUNC
USERARG
RESERVED
PEVID
MASK
For functions above that do not include a trigger specification, events will be triggered when there are significant changes in the channel's value or when there are changes in the channel's alarm state. This is the same as "DBE_VALUE | DBE_ALARM."
ECA_NORMAL - Normal successful completion
ECA_BADCHID - Corrupted CHID
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_ALLOCMEM - Unable to allocate memory
ECA_ADDFAIL - A local database event add failed
ca_flush_io()
ca_clear_subscription()
#include <cadef.h>
int ca_clear_subscription ( evid EVID );
Cancel a subscription.
All ca_clear_event() requests such as the above are accumulated (buffered) and not forwarded to the server until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently sent together in one message.
ECA_NORMAL - Normal successful completion
ECA_BADCHID - Corrupted CHID SEE ALSO ca_add_event()
ca_pend_io()
#include <cadef.h>
int ca_pend_io ( double TIMEOUT );
This function flushes the send buffer and then blocks until outstanding ca_get requests complete, and until channels created specifying nill connection handler function pointers connect for the first time.
If ECA_TIMEOUT is returned then get requests may be reissued followed by a subsequent call to ca_pend_io(). Specifically, the function will block only for outstanding ca_get requests issued, and also any channels created specifying a nill connection handler function pointer, after the last call to ca_pend_io() or ca client context creation whichever is later. Note that ca_create_channel requests generally should not be reissued for the same process variable unless ca_clear_channel is called first.
If no ca_get or connection state change events are outstanding then ca_pend_io() will flush the send buffer and return immediately without processing any outstanding channel access background activities.
The delay specified to ca_pend_io() should take into account worst case network delays such as Ethernet collision exponential back off until retransmission delays which can be quite long on overloaded networks.
Unlike ca_pend_event
, this routine will
not process CA's background activities if none of the selected IO requests
are pending.
TIMEOUT
interval of
zero specifies forever.ECA_NORMAL - Normal successful completion
ECA_TIMEOUT - Selected IO requests didnt complete before specified timeout
ECA_EVDISALLOW - Function inappropriate for use within an event handler
ca_test_io()
#include <cadef.h>
int ca_test_io();
This function tests to see if all ca_get requests are complete and channels created specifying a nill connection callback function pointer are connected. It will report the status of outstanding ca_get requests issued, and channels created specifying a nill connection callback function pointer, after the last call to ca_pend_io() or CA context initialization whichever is later.
ECA_IODONE - All IO operations completed
ECA_IOINPROGRESS - IO operations still in progress
ca_pend_event()
#include <cadef.h> int ca_pend_event ( double TIMEOUT ); int ca_poll ();
When ca_pend_event is invoked the send buffer is flushed and CA background activity is processed for TIMEOUT seconds.
When ca_poll is invoked the send buffer is flushed and any outstanding CA background activity is processed.
This routine will not return before the specified time-out
expires and all unfinished channel access labor has been processed, and
unlike ca_pend_io
it does not
indicate anything about the status of pending IO requests when it
returns ECA_NORMAL.
See also Thread Safety and Preemptive Callback to User Code.
TIMEOUT
ECA_NORMAL - Normal successful completion
ECA_TIMEOUT - The operation timed out
ECA_EVDISALLOW - Function inappropriate for use within a call back handler
ca_flush_io()
#include <cadef.h> int ca_flush_io();
Flush outstanding IO requests to the server. This routine might be useful to users who need to flush requests prior to performing client side labor in parallel with labor performed in the server.
Outstanding requests are also sent whenever the buffer which holds them becomes full.
ECA_NORMAL - Normal successful completion
ca_signal()
#include <cadef.h>
int ca_signal ( long CA_STATUS, const char * CONTEXT_STRING );
void SEVCHK( CA_STATUS, CONTEXT_STRING );
Provide the error message character string associated with the supplied channel access error code and the supplied error context to diagnostics. If the error code indicates an unsuccessful operation a stack dump is printed, if this capability is available on the local operating system, and execution is terminated.
SEVCHK is a macro envelope around ca_signal which only calls ca_signal() if the supplied error code indicates an unsuccessful operation. SEVCHK is the recommended error handler for simple applications which do not wish to write code testing the status returned from each channel access call.
status = ca_context_create (...);
SEVCHK ( status, "Unable to create a CA client context" );
If the application only wishes to print the message associated with an error code or test the severity of an error there are also functions provided for this purpose.
CA_STATUS
CONTEXT_STRING
ECA_NORMAL - Normal successful completion
ca_add_exception_event()
#include <cadef.h> typedef void (*pCallback) ( struct exception_handler_args HANDLERARGS ); int ca_add_exception_event ( pCallback USERFUNC, void *USERARG );
Replace the currently installed CA context global exception handler call back.
When an error occurs in the server asynchronous to the clients thread then information about this type of error is passed from the server to the client in an exception message. When the client receives this exception message an exception handler callback is called.The default exception handler prints a diagnostic message on the client's standard out and terminates execution if the error condition is severe.
Note that certain fields in "struct exception_handler_args" are not applicable in the context of some error messages. For instance, a failed get will supply the address in the client task where the returned value was requested to be written. For other failed operations the value of the addr field should not be used.
USERFUNC
op
field can be one of
CA_OP_GET, CA_OP_PUT, CA_OP_CREATE_CHANNEL, CA_OP_ADD_EVENT,
CA_OP_CLEAR_EVENT, or CA_OP_OTHER.
struct exception_handler_args {
void *usr; /* user argument supplied when installed */
chanId chid; /* channel id (may be nill) */
long type; /* type requested */
long count; /* count requested */
void *addr; /* user's address to write results of CA_OP_GET */
long stat; /* channel access ECA_XXXX status code */
long op; /* CA_OP_GET, CA_OP_PUT, ..., CA_OP_OTHER */
const char *ctx; /* a character string containing context info */
sonst char *pFile; /* source file name (may be NULL) */
unsigned lineNo; /* source file line number (may be zero) */
};
USERARG
void ca_exception_handler (
struct exception_handler_args args)
{
char buf[512];
char *pName;
if ( args.chid ) {
pName = ca_name ( args.chid );
}
else{
pName = "?";
}
sprintf ( buf,
"%s - with request chan=%s op=%d data type=%s count=%d",
args.ctx, pName, args.op, dbr_type_to_text ( args.type ), args.count );
ca_signal ( args.stat, buf );
}
ca_add_exception_event ( ca_exception_handler , 0 );
ECA_NORMAL - Normal successful completion
ca_replace_printf_handler
()
#include <cadef.h>
typedef int caPrintfFunc ( const char *pFromat, va_list args );
int ca_replace_printf_handler ( caPrintfFunc *PFUNC );
Replace the default handler for formatted diagnostic message output. The default handler uses fprintf to send messages to 'stderr'.
PFUNC
int my_printf ( char *pformat, va_list args ) {
int status;
status = vfprintf( stderr, pformat, args);
return status;
}
status = ca_replace_printf_handler ( my_printf );
SEVCHK ( status, "failed to install my printf handler" );
ECA_NORMAL - Normal successful completion
ca_replace_access_rights_event()
#include <cadef.h>
typedef void ( *pCallBack )( struct access_rights_handler_args );
int ca_replace ( chid CHAN, pCallBack PFUNC );
Install or replace the access rights state change callback handler for the specified channel.
The callback handler is called in the following situations.
When a channel is created no access rights handler is installed.
CHAN
PFUNC
typedef struct ca_access_rights {
unsigned read_access:1;
unsigned write_access:1;
} caar;
/* arguments passed to user access rights handlers */
struct access_rights_handler_args {
chanId chid; /* channel id */
caar ar; /* new access rights state */
};
ECA_NORMAL - Normal successful completion
ca_modify_user_name()
ca_modify_host_name()
ca_field_type()
#include <cadef.h>
chtype ca_field_type ( CHID );
Return the native type in the server of the process variable.
CHID
TYPE
ca_element_count()
#include <cadef.h>
unsigned ca_element_count ( CHID );
Return the maximum array element count in the server for the specified IO channel.
CHID
COUNT
ca_name()
#include <cadef.h>
char * ca_name ( CHID );
Return the name provided when the supplied channel id was created.
CHID
PNAME
ca_set_puser()
#include <cadef.h>
void ca_set_puser ( chid CHID, void *PUSER );
Set a user private void pointer variable retained with each channel for use at the users discretion.
ca_puser()
#include <cadef.h>
void * ca_puser ( CHID );
Return a user private void pointer variable retained with each channel for use at the users discretion.
CHID
PUSER
ca_state()
#include <cadef.h>
enum channel_state {
cs_never_conn, /* valid chid, server not found or unavailable */
cs_prev_conn, /* valid chid, previously connected to server */
cs_conn, /* valid chid, connected to server */
cs_closed }; /* channel deleted by user */
enum channel_state ca_state ( CHID );
Returns an enumerated type indicating the current state of the specified IO channel.
CHID
STATE
ca_message()
#include <cadef.h>
const char * ca_message ( STATUS );
return a message character string corresponding to a user specified CA status code.
STATUS
INGca_host_name()
#include <cadef.h>
char * ca_host_name ( CHID );
Return a character string which contains the name of the host to which a channel is currently connected.
CHID
STRING
ca_read_access()
#include <cadef.h>
int ca_read_access ( CHID );
Returns boolean true if the client currently has read access to the specified channel and boolean false otherwise.
CHID
STRING
ca_write_access()
#include <cadef.h>
int ca_write_access ( CHID );
Returns boolean true if the client currently has write access to the specified channel and boolean false otherwise.
CHID
STRING
dbr_size[]
#include <db_access.h>
extern unsigned dbr_size[/*TYPE*/];
An array that returns the size in bytes for a DBR_XXXX type.
TYPE
SIZE
dbr_size_n()
#include <db_access.h>
unsigned dbr_size_n ( TYPE, COUNT );
Returns the size in bytes for a DBR_XXXX type with COUNT elements. If the DBR type is a structure then the value field is the last field in the structure. If COUNT is greater than one then COUNT-1 elements are appended to the end of the structure so that they can be addressed as an array through a pointer to the value field.
TYPE
COUNT
SIZE
dbr_value_size[]
#include <db_access.h>
extern unsigned dbr_value_size[/* TYPE */];
The array dbr_value_size[TYPE] returns the size in bytes for the value stored in a DBR_XXXX type. If the type is a structure the size of the value field is returned otherwise the size of the type is returned.
TYPE
SIZE
dbr_type_to_text()
#include <db_access.h>
const char * dbr_type_text ( chtype TYPE );
Returns a constant null terminated string corresponding to the specified dbr type.
TYPE
STRING
ca_test_event()
#include <cadef.h>
void ca_test_event ( struct event_handler_args );
A built-in subscription update call back handler for debugging purposes that prints diagnostics to standard out.
void ca_test_event ();
status = ca_add_event ( type, chid, ca_test_event, NULL, NULL );
SEVCHK ( status, .... );
ca_sg_create()
#include <cadef.h>
int ca_sg_create ( CA_SYNC_GID *PGID );
Create a synchronous group and return an identifier for it.
A synchronous group can be used to guarantee that a set of channel access requests have completed. Once a synchronous group has been created then channel access get and put requests may be issued within it using ca_sg_get() and ca_sg_put() respectively. The routines ca_sg_block() and ca_sg_test() can be used to block for and test for completion respectively. The routine ca_sg_reset() is used to discard knowledge of old requests which have timed out and in all likelihood will never be satisfied.
Any number of asynchronous groups can have application requested operations outstanding within them at any given time.
PGID
CA_SYNC_GID gid;
status = ca_sg_create ( &gid );
SEVCHK ( status, Sync group create failed );
ECA_NORMAL - Normal successful completion
ECA_ALLOCMEM - Failed, unable to allocate memory
ca_sg_block()
ca_sg_test()
ca_sg_reset()
ca_sg_put()
ca_sg_get()
ca_sg_delete()
#include <cadef.h>
int ca_sg_delete ( CA_SYNC_GID GID );
Deletes a synchronous group.
CA_SYNC_GID gid;
status = ca_sg_delete ( gid );
SEVCHK ( status, Sync group delete failed );
ECA_NORMAL - Normal successful completion
ECA_BADSYNCGRP - Invalid synchronous group
ca_sg_block
()
#include <cadef.h>
int ca_sg_block ( CA_SYNC_GID GID, double timeout );
Flushes the send buffer and then waits until outstanding requests complete or the specified time out expires. At this time outstanding requests include calls to ca_sg_array_get() and calls to ca_sg_array_put(). If ECA_TIMEOUT is returned then failure must be assumed for all outstanding queries. Operations can be reissued followed by another ca_sg_block(). This routine will only block on outstanding queries issued after the last call to ca_sg_block(), ca_sg_reset(), or ca_sg_create() whichever occurs later in time. If no queries are outstanding then ca_sg_block() will return immediately without processing any pending channel access activities.
Values written into your program's variables by a channel access synchronous group request should not be referenced by your program until ECA_NORMAL has been received from ca_sg_block(). This routine will process pending channel access background activity while it is waiting.
CA_SYNC_GID gid;
status = ca_sg_block(gid);
SEVCHK(status, Sync group block failed);
ECA_NORMAL - Normal successful completion
ECA_TIMEOUT - The operation timed out
ECA_EVDISALLOW - Function inappropriate for use within an event handler
ECA_BADSYNCGRP - Invalid synchronous group
ca_sg_test()
ca_sg_reset()
ca_sg_test()
#include <cadef.h>
int ca_sg_test ( CA_SYNC_GID GID )
Test to see if all requests made within a synchronous group have completed.
GID
Test to see if all requests made within a synchronous group have completed.
CA_SYNC_GID gid;
status = ca_sg_test ( gid );
ECA_IODONE - IO operations completed
ECA_IOINPROGRESS - Some IO operations still in progress
ca_sg_reset()
#include <cadef.h>
int ca_sg_reset ( CA_SYNC_GID GID )
Reset the number of outstanding requests within the specified synchronous group to zero so that ca_sg_test() will return ECA_IODONE and ca_sg_block() will not block unless additional subsequent requests are made.
GID
CA_SYNC_GID gid;
status = ca_sg_reset(gid);
ECA_NORMAL - Normal successful completion
ECA_BADSYNCGRP - Invalid synchronous group
ca_sg_put()
#include <cadef.h>
int ca_sg_array_put ( CA_SYNC_GID GID, chtype TYPE,
unsigned long COUNT, chid CHID, void *PVALUE );
Write a value, or array of values, to a channel and increment the outstanding request count of a synchronous group.
All remote operation requests such as the above are accumulated (buffered) and not forwarded to the server until one of ca_flush_io(), ca_pend_io(), ca_pend_event(), or ca_sg_pend() are called. This allows several requests to be efficiently sent in one message.
If a connection is lost and then resumed outstanding puts are not reissued.
GID
TYPE
COUNT
CHID
PVALUE
ECA_NORMAL - Normal successful completion
ECA_BADSYNCGRP - Invalid synchronous group
ECA_BADCHID - Corrupted CHID
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_BADCOUNT - Requested count larger than native element count
ECA_STRTOBIG - Unusually large string supplied
ECA_PUTFAIL - A local database put failed
ca_sg_get()
#include <cadef.h>
int ca_sg_array_get ( CA_SYNC_GID GID,
chtype TYPE, unsigned long COUNT,
chid CHID, void *PVALUE );
Read a value from a channel and increment the outstanding request count of a synchronous group.
The values written into your program's variables by ca_sg_get should not be referenced by your program until ECA_NORMAL has been received from ca_sg_block , or until ca_sg_test returns ECA_IODONE.
All remote operation requests such as the above are accumulated (buffered) and not forwarded to the server until one of ca_flush_io, ca_pend_io, ca_pend_event, or ca_sg_pend are called. This allows several requests to be efficiently sent in one message.
If a connection is lost and then resumed outstanding gets are not reissued.
GID
TYPE
COUNT
CHID
PVALUE
ECA_NORMAL - Normal successful completion
ECA_BADSYNCGRP - Invalid synchronous group
ECA_BADCHID - Corrupted CHID
ECA_BADCOUNT - Requested count larger than native element count
ECA_BADTYPE - Invalid DBR_XXXX type
ECA_GETFAIL - A local database get failed
ca_client_status()
int ca_client_status ( unsigned level ); int ca_context_status ( struct ca_client_context *, unsigned level );
Prints information about the client context including, at higher interest levels, status for each channel. Lacking a CA context pointer, ca_client_status() prints information about the calling threads CA context.
CONTEXT
LEVEL
struct ca_client_context * ca_current_context ();
Returns a pointer to the current thread's CA context. If none then nil is returned.
ca_attach_context()
int ca_attach_context (struct ca_client_context *CONTEXT);
Become a member of the specified CA context. If
ca_disable_preemptive_callback
is specified when
ca_context_create() is called (or if ca_task_initialize() is called) then
additional threads are not allowed to join the CA context because
allowing other threads to join implies that CA callbacks will be called
preemptively from more than one thread.
CONTEXT
ECA_ISATTACHED - already attached to a CA context
ca_current_context()
void ca_dump_dbr (
chtype TYPE, unsigned COUNT, const
void * PDBR );
Dumps the specified dbr data type to standard out.
TYPE
COUNT
PDBR
$Id: CAref.html,v 1.58.2.1 2003/09/03 22:31:48 jhill Exp $