Skip to content
Snippets Groups Projects
Forked from cms_tk_ph2 / Ph2_ACF
2775 commits behind the upstream repository.
Fabio Ravera's avatar
Fabio Ravera authored
Removed access to container with index and added exception handler

See merge request cms_tk_ph2/Ph2_ACF!399
9e302c41
History

CMS Ph2 ACF (Acquisition & Control Framework)

Contains:

  • A middleware API layer, implemented in C++, which wraps the firmware calls and handshakes into abstracted functions
  • A C++ object-based library describing the system components (CBCs, RD53, Hybrids, Boards) and their properties (values, status)

A short guide to write the GoldenImage to the SD card

  1. Connect the SD card
  2. Download the golden firmware from the cms-tracker-daq webpage
  3. sudo fdisk -l - find the name of the SD card (for example, /dev/mmcblk0)
  4. sudo chmod 744 /dev/sd_card_name - to be able to play with it
  5. Go to the folder were you saved the sdgoldenimage.img file
  6. dd if=sdgoldenimage.img of=/dev/sd_card_name bs=512 - to write the image to the SD card. If the SD card is partitioned (formatted), pay attention to write on the block device (e.g. /dev/mmcblk0) and not inside the partition (e.g. /dev/mmcblk0p1)
  7. Once the previous command is done, you can list the SD card: ./imgtool /dev/sd_card_name list - there should be a GoldenImage.bin, with 20MB block size
  8. Insert the SD card into the FC7

Alternatively, instead of the dd command above, to only copy the needed bytes you can do:

imageName=sdgoldenimage.img
dd if=$imageName bs=512 iflag=count_bytes of=somefile_or_device count=$(ls -s --block-size=1 $imageName | awk '{print $1}')

If you installed the command pv (sudo yum install -y pv), then the best way is the following (replacing /dev/mmcblk0 with your target device):

pv sdgoldenimage.img | sudo dd of=/dev/mmcblk0

=x= Middleware for the Inner-Tracker (IT) system =x=

+ Last change made to this section: 20/03/2023

Suggested software and firmware versions:

  • Software git branch / tag : Dev / v4-13
  • Firmware tag: v4-06

Important webpages and information:

FC7 setup:

  1. Install wireshark in order to figure out which is the MAC address of your FC7 board (sudo yum install wireshark, then run sudo tshark -i ethernet_card, where ethernet_card is the name of the ethernet card of your PC to which the FC7 is connected to)
  2. In /etc/ethers put mac_address fc7.board.1 and in /etc/hosts put 192.168.1.80 fc7.board.1
  3. Restart the network: sudo /etc/init.d/network restart
  4. Install the rarpd daemon (version for CENTOS6 should work just fine even for CENTOS7): sudo yum install rarp_file_name.rpm from here
  5. Start the rarpd daemon: sudo systemctl start rarpd or sudo rarp -e -A (to start rarpd automatically after bootstrap: sudo systemctl enable rarpd)

More details on the hardware needed to setup the system can be found here

Firmware setup:

  1. Check whether the DIP switches on FC7 board are setup for the use of a microSD card (out-in-in-in-out-in-in-in)
  2. Insert a microSD card in the PC and run /sbin/fdisk -l to understand to which dev it's attached to (/dev/sd_card_name)
  3. Upload a golden firmware on the microSD card (read FC7 manual or run dd if=sdgoldenimage.img of=/dev/sd_card_name bs=512)
  4. Download the proper IT firmware version from here
  5. Plug the microSD card in the FC7
  6. From Ph2_ACF use the command fpgaconfig to upload the proper IT firmware (see instructions: IT-DAQ setup and run before running this command)

N.B.: a golden firmware is any stable firmware either from IT or OT, and it's needed just to initialize the IPbus communication at bootstrap (in order to create and image of the microSD card you can use the command: dd if=/dev/sd_card_name conv=sync,noerror bs=128K | gzip -c > sdgoldenimage.img.gz)
A golden firmware can be downloaded from the cms-tracker-daq webpage
A detailed manual about the firmware can be found here

IT-DAQ setup and run:

  1. Folow instructions below to install all needed software packages (like pugixml, boost, python. etc ...)
  2. mkdir choose_a_name
  3. cp settings/RD53Files/CMSIT_RD53A/B.txt choose_a_name
  4. cp settings/CMSIT_RD53A/B.xml choose_a_name
  5. cd choose_a_name
  6. Edit the file CMSIT_RD53A/B.xml in case you want to change some parameters needed for the calibrations or for configuring the chip
  7. Run the command: CMSITminiDAQ -f CMSIT_RD53A/B.xml -r to reset the FC7 (just once)
  8. Run the command: CMSITminiDAQ -f CMSIT_RD53A/B.xml -c name_of_the_calibration (or CMSITminiDAQ --help for help)

N.B.: to speed up the IPbus communication you can implement this trick

Basic list of commands for the fpgaconfig program (run from the choose_a_name directory):

  • Run the command: fpgaconfig -c CMSIT_RD53A/B.xml -l to check which firmware is on the microSD card
  • Run the command: fpgaconfig -c CMSIT_RD53A/B.xml -f firmware_file_name_on_the_PC -i firmware_file_name_on_the_microSD to upload a new firmware to the microSD card
  • Run the command: fpgaconfig -c CMSIT_RD53A/B.xml -i firmware_file_name_on_the_microSD to load a new firmware from the microSD card to the FPGA
  • Run the command: fpgaconfig --help for help

The program CMSITminiDAQ is the portal for all calibrations and for data taking. Through CMSITminiDAQ, and with the right command line option, you can run the following scans/ calibrations/ operation mode:

1. Latency scan
2. PixelAlive
3. Noise scan
4. SCurve and cross-talk scan
5. Gain scan
6. Threshold equalization
7. Gain optimization
8. Threshold minimization
9. Threshold adjustment
10. Injection delay scan
11. Clock delay scan
12. Bit Error Rate test
13. Data read back optimisation
14. Chip internal voltage tuning
15. Generic DAC-DAC scan
16. Physics

It might be useful to create one CMSIT.xml file for each "set" of calibrations, for instance noise, gain, and "the rest".

=x= End of Inner-Tracker section =x=

Setup

Firmware for the FC7 can be found in /firmware. Since the "old" FMC flavour is deprecated, only new FMCs (both connectors on the same side) are supported. You'll need Xilinx Vivado and a Xilinx Platform Cable USB II (http://uk.farnell.com/xilinx/hw-usb-ii-g/platform-cable-configuration-prog/dp/1649384). For more information on the firmware, please check the doc directory of https://gitlab.cern.ch/cms_tk_ph2/d19c-firmware

Gitlab CI setup for Developers (required to submit merge requests!!!)

  1. Enable shared Runners (if not enabled) i. from settings > CI/CD expand the Runners section ii. click the Allow shared Runners button

Setup on CentOs7

  1. Install devtoolset 10
sudo yum install -y centos-release-scl-rh
sudo yum install -y devtoolset-10
  1. On CC7 you also need to install boost v1.53 headers (default on this system) and pugixml as they don't ship with uHAL any more:
sudo yum install -y boost-devel pugixml-devel json-devel
  1. Install uHAL. SW tested with uHAL version up to 2.7.1

    Follow instructions from https://ipbus.web.cern.ch/ipbus/doc/user/html/software/install/yum.html

  2. Install CERN ROOT

sudo yum install -y root
sudo yum install -y root-net-http root-net-httpsniff  root-graf3d-gl root-physics root-montecarlo-eg root-graf3d-eve root-geom libusb-devel xorg-x11-xauth.x86_64
  1. Install CMAKE3 > 3.0:
sudo yum install -y cmake3
  1. Install python3
sudo yum install -y python3 python3-devel
  1. Install protobuf:

    Follow instructions from https://gitlab.cern.ch/cms_tk_ph2/MessageUtils/-/blob/master/README.md

  2. Install pybind11 (if installed in the same directoory when you plan to install the Ph2_ACF, the setup.sh will point to the correct location)

wget https://github.com/pybind/pybind11/archive/refs/tags/v2.9.2.tar.gz
tar zxvf v2.9.2.tar.gz

Run in docker container

Docker container are provided to facilitate users and developers in setting up the framework.

All docker containers can be found here: https://gitlab.cern.ch/cms_tk_ph2/docker_exploration/container_registry

Do run using one of the container, use the command:

docker run --rm -ti -v $PWD:$PWD -w $PWD <image>

Suggested images are:

  • For users (comes with Ph2_ACF of Dev branch installed): gitlab-registry.cern.ch/cms_tk_ph2/docker_exploration/cmstkph2_user_c7:latest
  • For developers (no Ph2_ACF, just environment and libraries): gitlab-registry.cern.ch/cms_tk_ph2/docker_exploration/cmstkph2_udaq_c7:latest

Specific tags can be pulled substituting latest with ph2_acf_<Ph2_ACF tag> (i.e. ph2_acf_v4-05)

clang-format (required to submit merge requests!!!)

  1. install 7.0 llvm toolset:
yum install centos-release-scl
yum install llvm-toolset-7.0
  1. if you already sourced the environment, you should be able to run the command to format the Ph2_ACF (to be done before each merge request!!!):
formatAll

The Ph2_ACF software

Follow these instructions to install and compile the libraries (provided you installed the latest version of gcc, µHal, mentioned above):

  1. Clone the GitHub repo and run cmake
git clone --recurse-submodules https://gitlab.cern.ch/cms_tk_ph2/Ph2_ACF.git # N.B. to syncrhonize only the submodule: `git submodule sync; git submodule update --init --recursive --remote`
cd Ph2_ACF
source setup.sh
mkdir build
cd build
cmake .. # add -D CMAKE_BUILD_TYPE=Debug if you plan to use gdb for debugging, if you yum-instanlled `cmake3`, you might need to tall it `cmake3 ..`
  1. Do a make -jN in the build/ directory or alternatively do make -C build/ -jN in the Ph2_ACF root directory.

  2. Don't forget to source setup.sh to set all the environment variables correctly.

  3. Launch

systemtest --help

to test the parsing of the HWDescription.xml file.

  1. Launch
datatest --help

to test if you can correctly read data

  1. Launch
calibrate --help

to calibrate a hybrid,

hybridtest --help

to test a hybird's I2C registers and input channel connectivity

cmtest --help

to run the CM noise study

pulseshape --help

to measure the analog pulseshape of the cbc

configure --help

to apply a configuration to the CBCs

  1. Launch
commission --help

to do latency & threshold scans

  1. Launch
fpgaconfig --help

to upload a new FW image to the GLIB

  1. Launch
miniDAQ --help

to save binary data from the GLIB to file

  1. Launch
miniDQM --help

to run the DQM code from the June '15 beamtest

Setup on CentOs8 (deprecated)

The following procedure will install (in order):

  1. the boost and pugixml libraries
  2. the cactus libraries for ipBus (using these instructions)
  3. root with all its needed libraries
  4. cmake, tools for clang, including clang-format and git-extras

Libraries needed by Ph2_ACF

sudo yum install -y boost-devel pugixml-devel json-devel

uHAL libraries (cactus)

sudo curl https://ipbus.web.cern.ch/doc/user/html/_downloads/ipbus-sw.centos8.x86_64.repo \
  -o /etc/yum.repos.d/ipbus-sw.repo
sudo yum-config-manager --enable powertools
sudo yum clean all
sudo yum groupinstall uhal

ROOT

sudo yum install -y root root-net-http root-net-httpsniff root-graf3d-gl root-physics \
  root-montecarlo-eg root-graf3d-eve root-geom libusb-devel xorg-x11-xauth.x86_64

Build tools and some nice git extras

sudo yum install -y cmake3
sudo yum install -y clang-tools-extra
sudo yum install -y git-extras

Install devtoolset 10

sudo yum makecache --refresh
sudo yum -y install gcc-toolset-10

Install python3

sudo yum install -y python3 python3-devel

Install protobuf:

Follow instructions to install protobuf from (Just install section is needed) https://gitlab.cern.ch/cms_tk_ph2/MessageUtils/-/blob/master/README.md

Install pybind11 (if installed in the same directoory when you plan to install the Ph2_ACF, the setup.sh will point to the correct location)

wget https://github.com/pybind/pybind11/archive/refs/tags/v2.9.2.tar.gz
tar zxvf v2.9.2.tar.gz

Setup on RHEL 9.1 or AlmaLinux 9.1

The following procedure will install (in order):

  1. complete the cern installation
  2. when installing rarpd use the version for Fedora
  3. the boost and pugixml libraries
  4. erlang (using these instructions)
  5. the cactus libraries for ipBus (using these instructions)
  6. root with all its needed libraries
  7. cmake, tools for clang, including clang-format and git-extras

Complete the CERN installation

Make sure that the CERN installation is complete by running

sudo dnf --repofrompath=cern9el,http://linuxsoft.cern.ch/internal/repos/cern9el-stable/x86_64/os --repo=cern9el install cern-release

Libraries needed by Ph2_ACF

sudo yum install -y boost-devel pugixml-devel json-devel

Erlang (needed by uHAL)

This installs erlang from a specific rpm. It would be nice if in the future, the correct version of erlang could be made available via the CERN repository

wget https://github.com/rabbitmq/erlang-rpm/releases/download/v25.1.2/erlang-25.1.2-1.el9.x86_64.rpm
sudo yum -y install erlang-25.1.2-1.el9.x86_64.rpm

uHAL libraries (cactus)

sudo curl https://ipbus.web.cern.ch/doc/user/html/_downloads/ipbus-sw.el9.repo -o /etc/yum.repos.d/ipbus-sw.repo
sudo yum clean all
sudo yum groupinstall -y uhal controlhub

ROOT

sudo yum install -y root root-net-http root-net-httpsniff root-graf3d-gl root-physics \
  root-montecarlo-eg root-graf3d-eve root-geom libusb-devel xorg-x11-xauth.x86_64

Build tools and some nice git extras

sudo yum install -y cmake3 clang-tools-extra git-extras

devtoolset 12

sudo yum makecache --refresh
sudo yum -y install gcc-toolset-12

python3

sudo yum install -y python3 python3-devel

protobuf Follow instructions to install protobuf from (Just install section is needed) https://gitlab.cern.ch/cms_tk_ph2/MessageUtils/-/blob/master/README.md

pybind11 (if installed in the same directoory when you plan to install the Ph2_ACF, the setup.sh will point to the correct location)

wget https://github.com/pybind/pybind11/archive/refs/tags/v2.9.2.tar.gz
tar zxvf v2.9.2.tar.gz

Nota Bene

When you write a register in the Glib or the Cbc, the corresponding map of the HWDescription object in memory is also updated, so that you always have an exact replica of the HW Status in the memory.

Register values are:

  • 8-bit unsigend integers for the CBCs that should be edited in hex notation, i.e. '0xFF'
  • 32-bit unsigned integers for the GLIB: decimal values

For debugging purpose, you can activate DEV_FLAG in the sources or in the Makefile and also activate the uHal log in RegManager.cc.

External clock and trigger

Please see the D19C FW documentation for instructions on how to use external clock and trigger with the various FMCs (DIO5 and CBC3 FMC)

Known issues

uHAL exceptions and UDP timeouts when reading larger packet sizes from the GLIB board: this can happen for some users (cause not yet identified) but can be circumvented by changing the line

ipbusudp-2.0://192.168.000.175:50001

in the connections.xml file to

chtcp-2.0://localhost:10203?target=192.168.000.175:50001

and then launching the CACTUS control hub by the command:

/opt/cactus/bin/controlhub_start

This uses TCP protocol instead of UDP which accounts for packet loss but decreases the performance.

Support, suggestions?

For any support/suggestions, mail to fabio.raveraSPAMNOT@cern.ch, mauro.dinardoSPAMNOT@cern.ch

Firmware repository for OT tracker

https://udtc-ot-firmware.web.cern.ch/