You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@plc4x.apache.org by cd...@apache.org on 2018/03/14 15:52:04 UTC

[incubator-plc4x] branch master updated: - Added some more documentation on PDU fragmenting.

This is an automated email from the ASF dual-hosted git repository.

cdutz pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-plc4x.git


The following commit(s) were added to refs/heads/master by this push:
     new 3a00e8d  - Added some more documentation on PDU fragmenting.
3a00e8d is described below

commit 3a00e8d4893c35c76208f71b8fef75f8bdb955ee
Author: Christofer Dutz <ch...@c-ware.de>
AuthorDate: Wed Mar 14 16:52:01 2018 +0100

    - Added some more documentation on PDU fragmenting.
---
 src/site/asciidoc/protocols/s7/s7comm.adoc | 35 +++++++++++++++++++++---------
 1 file changed, 25 insertions(+), 10 deletions(-)

diff --git a/src/site/asciidoc/protocols/s7/s7comm.adoc b/src/site/asciidoc/protocols/s7/s7comm.adoc
index 7b7bcde..78f23f0 100644
--- a/src/site/asciidoc/protocols/s7/s7comm.adoc
+++ b/src/site/asciidoc/protocols/s7/s7comm.adoc
@@ -97,27 +97,42 @@ During the connection to a S7 PLC the client and PLC agree on 3 important parame
 - Max AMQ Callee
 
 The PDU Size is the size of a data packet the PLC is willing to accept.
-Here note, that in reality there are two PDU sizes involved: On `ISO TP/COTP` protocol level and a second time on the `S7` protocol level.
+Here note, that in reality there are two PDU sizes involved: On `ISO TP/COTP` protocol level (TPDU Size) and a second time on the `S7` protocol level (PDU Size).
 Most implementations treat them as somewhat equal, but this doesn't have to be the case.
 A PLC could accept a higher PDU size on `ISO TP` level than on `S7` level.
 
-The `Max AMQ`  parameters define how many unacknowledged requests a PLC (Callee) is able to accept from a client (Caller).
+The `Max AMQ` parameters define how many unacknowledged requests a PLC (Callee) is able to accept from a client (Caller).
 
-If the `ISO TP` PDU size is bigger than the `S7` one, then theoretically multiple `S7` PDUs can be contained in one `ISO TP` packet.
+If the `ISO TP` TPDU size is bigger than the `S7` PDU size, then theoretically multiple `S7` PDUs can be contained in one `ISO TP` packet.
 But at max `Max AMQ` packets.
-Most drivers don't utilize this option however.
+Most drivers don't utilize this option however (We won't either).
 
-When issuing a read request there are other things that have to be taken into consideration:
+When issuing a read request there are other things that have to be taken into account:
 
-The data retrieved by a read request has to fit into a PDU.
-So if a block of data requested in a read request, that would read a big block of data that would exceed the agreed upon PDU size, the PLC will respond with an `Access Violation` and not return any data.
+If on `ISO TP` level a max PDU size has been agreed on, the max PDU Size of the `S7` packet will be smaller.
+So if a TPDU size of 256 bytes has been agreed upon, then the ISO TP header takes 3 bytes and the header of the S7 packet takes 10 bytes.
+The header of a `Read Var` parameter takes 2 bytes.
+So if the TPDU size is 256 bytes, this leaves 256 - (3 + 10 + 2) bytes = 241 bytes
+
+An `S7ANY` type variable specification takes 12 bytes, so the maximum number of memory areas that can be accessed in one S7 request packet is: 241 / 12 = 20
+That's also the reason why most available drivers limit the number of addresses to 20 (Some go down to 18 because some devices seem to calculate the boundaries differently).
+
+To make things even more complicated, we have to ensure the data retrieved by a read request fits into a PDU.
+So if a block of data requested in a read request would would exceed the agreed upon PDU size, the PLC will respond with an `Access Violation` and not return any data.
+So the size limit for reading data with a 256 byte PDU size would be 241 - 12 = 229 bytes.
 
 Another example would be if you read different memory blocks, each smaller than the max PDU size, the read PDU could be quite small, but if these blocks are quite large, the read response could exceed the max PDU size.
 In this case the item responses that exceed the size will simply be responded with an `Access violation` by the PLC.
 
-When communicating with a `S7 1200` for example the `ISO TP` PDU size is set to 256 bytes, then the maximum size of the `S7` is , this limits the maximum amount of separate blocks accessible in the read request to 18.
-The typical `Max AMQ` of `8` further complicates things as we can't simply split up the one message into multiple ones.
-If the number of messages exceeds this number, we have to queue these PDUs and wait till the PLC confirmed these and can then continue sending further fragments.
+The typical `Max AMQ` of `8` further complicates things as we can't simply split up the one message into multiple ones and blindly fire them at the PLC.
+If the number of messages exceeds this number, we have to queue the excess PDUs and wait till the PLC confirmed some and can then continue sending further fragments.
+
+Most commercial drivers reduce the stress by rearranging and changing the requests.
+If for example a request would read 8 consecutive bits, it would automatically change the 8 items into reading one byte.
+Also if the gap between addresses is smaller than the overhead for a new definition, it would extend the first to include the second memory area.
+Same should sometimes be possible in the other direction.
+So if we wanted to read 300 bytes with a 256 byte TPDU size, the driver would split this up into two smaller byte arrays and internally merge them. However when reading Real (Number) values there could be issues with this.
+However in this case the driver has to completely take care of this optimization.
 
 === Links
 

-- 
To stop receiving notification emails like this one, please contact
cdutz@apache.org.