generic modbus driver by dtsecon · Pull Request #1052 · networkupstools/nut

dtsecon · 2021-06-24T15:53:04Z

This is a generic modbus driver expected to work with contact (direct) signal UPS devices, connected via modbus RIO (remote I/O) either serial or TCP/IP. The driver has been tested against PULS UPS (model UB40.241) via MOXA ioLogikR1212 (RS485) and ioLogikE1212 (TCP/IP).

The PULS UPS UB40.241 supports the following signals:

Ready contact (DO) <--> HB
Buffering contact (DO) <--> OL | OB
Battery-low (DO) <--> LB
Replace Battery (DO) <--> RB
Inhibit (DI) <--> FSD

DO and DI assumes device perspective

The driver's concept is to map the UPS states (as defined in nut) onto UPS contacts' states

The driver has an extended configuration interface implemented using variables defined in ups.conf.

[generic_modbus]
# info parameters
driver = generic_modbus
port = /dev/ttyUSB0
# port = 192.168.2.104
# port = 192.168.2.104:1502
desc = "generic ups driver"
# device info
device_mfr = "PULS"
device_model = "UB40.241"
# serial settings
ser_baud_rate = 9600
ser_parity = N
ser_data_bit = 8
ser_stop_bit = 1
modbus slave id
rio_slave_id = 5
# UPS signal state attributes
OB_addr = 0x0
OB_regtype = 1
OB_noro = 0
LB_addr = 0x1
LB_regtype = 1
HB_addr = 0x2
HB_regtype = 1
RB_addr = 0x3
RB_regtype = 1
# CHRG_addr = 0x0
# CHRG_regtype = 1
# CHRG_noro = 1
DISCHRG_addr = 0x0
DISCHRG_regtype = 1
DISCHRG_noro = 0
FSD_addr = 0x140
FSD_regtype = 0
FSD_pulse_duration = 180

XXX_addr denote the modbus address of RIO where the UPS signal is connected
XXX_regtype is the type of modbus register (0:COIL, 1:INPUT_B, 2:INPUT_R, 3:HOLDING)
XXX_norois the configuration for NO/NC (normally open, normally closed contact)
FSD_pulse_duration defines the duration in ms of the inhibit pulse. if it's not defined signal has only one transition depending on noro configuration.

Regards
Dimitris.

donnib · 2021-07-05T15:24:02Z

@dtsecon does this mean that i can setup my own Modbus TCP server and set NUT server to point to it and then i fill out data into the registries which NUT server would read , is that correct understood ?

dtsecon · 2021-07-05T22:34:01Z

@dtsecon does this mean that i can setup my own Modbus TCP server and set NUT server to point to it and then i fill out data into the registries which NUT server would read , is that correct understood ?

Yes, your assumption is correct although never tested. However, OL (online), OB (on battery), LB (low battery), HB (high battery), RB (replace battery) and FSD (forced shutdown) statuses are currently supported.

dtsecon · 2021-07-06T16:57:34Z

@jimklimov hello,
I noticed that travis-ci is stuck and waiting for a status report although it has been setup and run a successful build for the PR repo. Is there something else needed to be done in order to proceed ?

Regards,
Dimitris.

jimklimov · 2021-08-26T23:44:51Z

Sorry for the lag on my side. Travis CI has gone away from free FOSS builds, so a new dedicated NUT CI is being built. There are still some rough edges, but it should be usable by now and I'm bumping the existing PRs to build there. Thanks for waiting :\

jimklimov · 2021-09-13T15:01:33Z

Hello, it seems that I may not update the PR source (your "master" branch) because it has github branch protection rules, probably: Couldn't update "master": Required status check "Travis CI - Branch" is expected.

Can you please disable that config point, seems Travis is not coming back...

dtsecon · 2021-09-14T08:37:53Z

Hello, it seems that I may not update the PR source (your "master" branch) because it has github branch protection rules, probably: Couldn't update "master": Required status check "Travis CI - Branch" is expected.

Can you please disable that config point, seems Travis is not coming back...

done!

jimklimov · 2021-09-15T09:39:43Z

With CI finally somewhat rectified, it finds code issues in this PR - mostly due to platforms/standards mixup I think. A few that caught my eye in orange balls at https://ci.networkupstools.org/blue/organizations/jenkins/nut%2Fnut/detail/PR-1052/1 are:

[2021-09-15T05:45:20.637Z] In file included from generic_modbus.c:24:
[2021-09-15T05:45:20.637Z] generic_modbus.h:46:14: error: ISO C forbids forward references to 'enum' types [-Werror=pedantic]
[2021-09-15T05:45:20.637Z]    46 | typedef enum regtype regtype_t;
[2021-09-15T05:45:20.637Z]       |              ^~~~~~~

[2021-09-15T05:45:20.637Z] generic_modbus.h:55:14: error: ISO C forbids forward references to 'enum' types [-Werror=pedantic]
[2021-09-15T05:45:20.637Z]    55 | typedef enum devstate devstate_t;
[2021-09-15T05:45:20.637Z]       |              ^~~~~~~~
[2021-09-15T05:45:20.637Z] cc1: all warnings being treated as errors

Some uses of ulong - these should become an explicit-bitness type like uint32_t or uint64_t, or a size_t if some (time-related) library accepts that in the method signatures:

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:14: error: expected ';' after expression
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]              ^
[2021-09-15T03:16:06.347Z]              ;

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:9: error: use of undeclared identifier 'ulong'
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]         ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:15: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]               ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:438:34: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_usec -= 1000000 * nsec;
[2021-09-15T03:16:06.347Z]                                  ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:439:23: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_sec += nsec;
[2021-09-15T03:16:06.347Z]                       ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:14: error: expected ';' after expression
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]              ^
[2021-09-15T03:16:06.347Z]              ;

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:9: error: use of undeclared identifier 'ulong'
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]         ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:15: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]               ^
[2021-09-15T03:16:06.347Z] generic_modbus.c:443:34: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_usec += 1000000 * nsec;
[2021-09-15T03:16:06.347Z]                                  ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:444:23: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_sec -= nsec;
[2021-09-15T03:16:06.347Z]                       ^
[2021-09-15T03:16:06.347Z] 10 errors generated.
[2021-09-15T03:16:06.347Z] *** Error code 1

dtsecon · 2021-09-16T10:07:36Z

With CI finally somewhat rectified, it finds code issues in this PR - mostly due to platforms/standards mixup I think. A few that caught my eye in orange balls at https://ci.networkupstools.org/blue/organizations/jenkins/nut%2Fnut/detail/PR-1052/1 are:

[2021-09-15T05:45:20.637Z] In file included from generic_modbus.c:24:
[2021-09-15T05:45:20.637Z] generic_modbus.h:46:14: error: ISO C forbids forward references to 'enum' types [-Werror=pedantic]
[2021-09-15T05:45:20.637Z]    46 | typedef enum regtype regtype_t;
[2021-09-15T05:45:20.637Z]       |              ^~~~~~~

[2021-09-15T05:45:20.637Z] generic_modbus.h:55:14: error: ISO C forbids forward references to 'enum' types [-Werror=pedantic]
[2021-09-15T05:45:20.637Z]    55 | typedef enum devstate devstate_t;
[2021-09-15T05:45:20.637Z]       |              ^~~~~~~~
[2021-09-15T05:45:20.637Z] cc1: all warnings being treated as errors

Some uses of ulong - these should become an explicit-bitness type like uint32_t or uint64_t, or a size_t if some (time-related) library accepts that in the method signatures:

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:14: error: expected ';' after expression
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]              ^
[2021-09-15T03:16:06.347Z]              ;

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:9: error: use of undeclared identifier 'ulong'
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]         ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:437:15: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         ulong nsec = (end.tv_usec - start->tv_usec) / 1000000 + 1;
[2021-09-15T03:16:06.347Z]               ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:438:34: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_usec -= 1000000 * nsec;
[2021-09-15T03:16:06.347Z]                                  ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:439:23: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_sec += nsec;
[2021-09-15T03:16:06.347Z]                       ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:14: error: expected ';' after expression
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]              ^
[2021-09-15T03:16:06.347Z]              ;

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:9: error: use of undeclared identifier 'ulong'
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]         ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:442:15: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         ulong nsec = (start->tv_usec - end.tv_usec) / 1000000;
[2021-09-15T03:16:06.347Z]               ^
[2021-09-15T03:16:06.347Z] generic_modbus.c:443:34: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_usec += 1000000 * nsec;
[2021-09-15T03:16:06.347Z]                                  ^

[2021-09-15T03:16:06.347Z] generic_modbus.c:444:23: error: use of undeclared identifier 'nsec'
[2021-09-15T03:16:06.347Z]         end.tv_sec -= nsec;
[2021-09-15T03:16:06.347Z]                       ^
[2021-09-15T03:16:06.347Z] 10 errors generated.
[2021-09-15T03:16:06.347Z] *** Error code 1

done!

jimklimov · 2021-09-16T21:33:59Z

Looks good to me and CI, hopefully tested with HW on your side at least? :)

One question about the side fixes for asciidoc builds -- did your tools complain about the lack of equals sign? Seems those versions on CI farm are okay with it either way, at least they do not fail due to CLI :)
Did you verify that the results of those docs builds are equivalent?

dtsecon · 2021-09-17T08:33:58Z

Looks good to me and CI, hopefully tested with HW on your side at least? :)

One question about the side fixes for asciidoc builds -- did your tools complain about the lack of equals sign? Seems those versions on CI farm are okay with it either way, at least they do not fail due to CLI :)
Did you verify that the results of those docs builds are equivalent?

the driver supports any UPS with relay contact signals via a remote I/O device (with RS485 or TCP/IP interface). It has been thoroughly tested against PULS UB40.241 but is expected to work with any other similar UPS.

there is a missing '=' in a2x --xsltproc-opts switch argument, specifically in conficure.ac and docs/man/Makefile.am. It seems that "=" is mandatory for a2x to run successfully otherwise building procedure (with --with-doc(s)) couldn't compile man pages at all.

jimklimov · 2021-09-19T12:47:29Z

Thanks for clarifications. Just to know: what version of a2x and on what OS did not work for you with the older Makefiles? Because that part did not change for a few years AFAIK and worked on many different platforms. Wondering if yours is too new or too old? Or some fork? Either way, new CLI syntax seems to work for CI builders too, so thanks!

…

On Fri, Sep 17, 2021, 10:34 dtsecon ***@***.***> wrote: Looks good to me and CI, hopefully tested with HW on your side at least? :) One question about the side fixes for asciidoc builds -- did your tools complain about the lack of equals sign? Seems those versions on CI farm are okay with it either way, at least they do not fail due to CLI :) Did you verify that the results of those docs builds are equivalent? the driver supports any UPS with relay contact signals via a remote I/O device (with RS485 or TCP/IP interface). It has been thoroughly tested against PULS UB40.241 but is expected to work with any other similar UPS. there is a missing '=' in a2x --xsltproc-opts switch argument, specifically in conficure.ac and docs/man/Makefile.am. It seems that "=" is mandatory for a2x to run successfully otherwise building procedure (with --with-doc(s)) couldn't compile man pages at all. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1052 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAMPTFF2S733G24U6IXLJFTUCL4QBANCNFSM47IFNFEA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

dtsecon · 2021-09-20T12:00:10Z

Thanks for clarifications. Just to know: what version of a2x and on what OS did not work for you with the older Makefiles? Because that part did not change for a few years AFAIK and worked on many different platforms. Wondering if yours is too new or too old? Or some fork? Either way, new CLI syntax seems to work for CI builders too, so thanks!
…
On Fri, Sep 17, 2021, 10:34 dtsecon @.***> wrote: Looks good to me and CI, hopefully tested with HW on your side at least? :) One question about the side fixes for asciidoc builds -- did your tools complain about the lack of equals sign? Seems those versions on CI farm are okay with it either way, at least they do not fail due to CLI :) Did you verify that the results of those docs builds are equivalent? the driver supports any UPS with relay contact signals via a remote I/O device (with RS485 or TCP/IP interface). It has been thoroughly tested against PULS UB40.241 but is expected to work with any other similar UPS. there is a missing '=' in a2x --xsltproc-opts switch argument, specifically in conficure.ac and docs/man/Makefile.am. It seems that "=" is mandatory for a2x to run successfully otherwise building procedure (with --with-doc(s)) couldn't compile man pages at all. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1052 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMPTFF2S733G24U6IXLJFTUCL4QBANCNFSM47IFNFEA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

Here is the platform used for building:

OS: Ubuntu 20.04, a2x 9.0.0rc1 (provided by asciidoc-base_9.0.0~rc1-1), GNU Make 4.2.1.

jimklimov · 2021-09-20T15:55:33Z

Still looks good, merging! Thanks for the driver and for a2x fixes to move forward :)

jimklimov · 2021-09-23T18:08:28Z

Cheers, one more question as I'm going over the code to address some style and stricter compiler warnings: in many places you use arch-dependent "int" and "uint", shouldn't there be more definitively-sized types (uint16_t and so on) to correspond to some wire protocol and library structures?

dtsecon added 4 commits June 24, 2021 17:17

generic_modbus sources add in buidling procedure

0de9f47

initial commit of generic_modbus beta version source files

239ea53

remove unnecessary stuff

97744a6

minor changes and tidying up

64e112f

This comment has been minimized.

Sign in to view

time_ellapsed fixed

81536c7

This comment has been minimized.

Sign in to view

removing unnecessary code

ca163a7

This comment has been minimized.

Sign in to view

dtsecon added 2 commits June 29, 2021 01:54

variable ctx changed to mbctx

b552eb3

minor changes to comply with nut development guidelines

f4e9f17

jimklimov mentioned this pull request Jul 5, 2021

Is there an API to add a custom UPS ? #1056

Closed

Merge branch 'master' into master

2a085ee

dtsecon added 5 commits July 6, 2021 13:20

just one space less

2194116

generic_modbus.h added to Makefile.am

7ea36a5

Merge branch 'master' of https://github.com/dtsecon/nut

a0d18f2

fix 'for' loop initial declarations for compatibility out of C99 mode

2e1a387

fix serial options in modbus_open

070d7f3

dtsecon closed this Jul 6, 2021

dtsecon reopened this Jul 6, 2021

dtsecon and others added 5 commits July 16, 2021 17:48

man page text created, added in build procedure

a26a0e8

improve man page

8d95c4c

fix man pages broken build process

f4e8c45

shutdown configurable repetition with increasing time interval added

cb6ea97

Merge branch 'master' into master

0036eda

Merge branch 'master' into master

b853f9e

Merge branch 'master' into master

f9090eb

dtsecon and others added 3 commits September 14, 2021 17:20

update nut.dic

c096564

update nut.dict

df37702

Merge branch 'master' into master

e129613

dtsecon added 2 commits September 16, 2021 13:04

fix forward declarations of enum occurances

544a4b8

change ulon to uint_32

e6d0733

jimklimov added the ready / gonna merge The PR is in final cycles leading to merge unless someone logs an objection before we hit the button label Sep 16, 2021

Merge branch 'master' into master

987f3d7

jimklimov merged commit 8d0cc97 into networkupstools:master Sep 20, 2021

jimklimov mentioned this pull request Sep 23, 2021

drivers/generic_modbus.{c,h}: fix compilation warnings #1107

Merged

bremenpl mentioned this pull request Apr 18, 2023

Server protocol implementation #1873

Open

jimklimov added the modbus label Apr 18, 2023

Uh oh!

Conversation

dtsecon commented Jun 24, 2021

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

donnib commented Jul 5, 2021

Uh oh!

dtsecon commented Jul 5, 2021

Uh oh!

dtsecon commented Jul 6, 2021

Uh oh!

jimklimov commented Aug 26, 2021

Uh oh!

jimklimov commented Sep 13, 2021

Uh oh!

dtsecon commented Sep 14, 2021

Uh oh!

jimklimov commented Sep 15, 2021

Uh oh!

dtsecon commented Sep 16, 2021

Uh oh!

jimklimov commented Sep 16, 2021

Uh oh!

dtsecon commented Sep 17, 2021

Uh oh!

jimklimov commented Sep 19, 2021 via email • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dtsecon commented Sep 20, 2021 • edited by jimklimov Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jimklimov commented Sep 20, 2021

Uh oh!

jimklimov commented Sep 23, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

jimklimov commented Sep 19, 2021 via email •

edited

Loading

dtsecon commented Sep 20, 2021 •

edited by jimklimov

Loading