Build-setup issue
alsrbok opened this issue · 4 comments
Background Work
- Yes, I searched the mailing list
- Yes, I searched prior issues
- Yes, I searched the documentation
Chipyard Version and Hash
Release: 1.5.0
Hash: a6a6a6
OS Setup
Ex: Output of uname -a
+ lsb_release -a
+ printenv
+ conda list
x86_64 GNU/Linux
LSB Version: :core-4.1-amd64:core-4.1-noarch:cxx-4.1-amd64:cxx-4.1-noarch:desktop-4.1-amd64:desktop-4.1-noarch:languages-4.1-amd64:languages-4.1-noarch:printing-4.1-amd64:printing-4.1-noarch
Distributor ID: RedHatEnterprise
Description: Red Hat Enterprise Linux release 8.10 (Ootpa)
Release: 8.10
Codename: Ootpa
Other Setup
Ex: Prior steps taken / Documentation Followed / etc...
Current Behavior
Build-setup issue
When we clone new version and try to run build-setup.sh with conda env, it failed.
- build-setup.sh try to use .conda-lock-env, not .conda-env. This is mismatch with generated env.sh file/
- There are changes on init-submodules-no-riscv-tools-nolog.sh file.
The current version does not seem to download some required executables from .conda-env/bin.
init-submodules-no-riscv-tools-nolog.csv
I upload an older version that I have utilized successfully in the past.
I resolved this problem by manually fixing lots of codes and files, but I hope we can correctly use it without any modification.
log.docx
This log file only contains the first error I encountered while using the current version.
After fixing this environment activation issue, I get many of the other errors mentioned above.
I think you can test it by cloning the project on new folder.
Thank you.
Expected Behavior
.
Other Information
No response
Can you upload the logs as a raw text file instead of .docx?
$ git clone https://github.com/ucb-bar/chipyard.git
Cloning into 'chipyard'...
remote: Enumerating objects: 40993, done.
remote: Counting objects: 100% (1046/1046), done.
remote: Compressing objects: 100% (557/557), done.
remote: Total 40993 (delta 534), reused 825 (delta 422), pack-reused 39947 (from 1)
Receiving objects: 100% (40993/40993), 51.82 MiB | 36.85 MiB/s, done.
Resolving deltas: 100% (23062/23062), done.
$ conda create -n base_vpu python=3.10
(base_vpu) $ conda install conda-libmamba-solver
(base_vpu) $ conda config --set solver libmamba
(base_vpu) $ ./build-setup.sh riscv-tools
========== BEGINNING STEP 1: Conda environment setup ==========
Channels:
- conda-forge
- defaults
Platform: linux-64
Collecting package metadata (repodata.json): ...working... done
Solving environment: ...working... done
Package Plan
environment location: /home/ mgpark/chipyard-test/chipyard/.conda-lock-env
added / updated specs:
- conda-lock=2.5.7
Downloading and Extracting Packages: ...working... done
Preparing transaction: ...working... done
Verifying transaction: ...working... done
Executing transaction: ...working... done
To activate this environment, use
$ conda activate /home/ mgpark/chipyard-test/chipyard/.conda-lock-env
To deactivate an active environment, use
$ conda deactivate
Using lockfile for conda: /home/ mgpark/chipyard-test/chipyard/conda-reqs/conda-lock-reqs/conda-requirements-riscv-tools-linux-64.conda-lock.yml
Storing main conda environment in /home/ mgpark/chipyard-test/chipyard/.conda-env
Usage: conda-lock install [OPTIONS] [LOCK_FILE]
Try 'conda-lock install --help' for help.
Error: Got unexpected extra arguments ({ \local cmd="${1-missing}"; case "$cmd" in activate | deactivate) __conda_activate "$@" ;; install | update | upgrade | remove | uninstall) __conda_exe "$@" || \return; __conda_reactivate ;; *) __conda_exe "$@" ;; esac } /home/ mgpark/chipyard-test/chipyard/conda-reqs/conda-lock-reqs/conda-requirements-riscv-tools-linux-64.conda-lock.yml)
build-setup.sh: Build script failed with exit code 2 at step 1: Conda environment setup
here it is.
The error is unclear still. Can you run build-setup.sh
with the -v
flag for more output. (Also, you shouldn't have to create a new environment (base_vpu
) to run build-setup.sh
). My 1st guess is that there might be some version mismatch between conda-lock
and conda
but I don't know how that would happen. You could try running the conda-lock
command yourself to debug as well:
conda-lock install --conda $(which conda) -p $CYDIR/.conda-env $CYDIR/conda-reqs/conda-lock-reqs/conda-requirements-$TOOLCHAIN_TYPE-linux-64.conda-lock.yml
https://github.com/ucb-bar/chipyard/blob/51b66dddb99146eacab3e5c7864e0fbc86859125/scripts/build-setup.sh#L190C5-L190C71 is where we create the conda
environment with conda-lock
.
Thanks for your help. I think I resolve this issue.
I'll explain the workaround I used for other users
In my case few things interrupted me to succesfully run ./build-setup.sh riscv-tools
- Change to GLOBAL_ENV_NAME=".conda-env" from GLOBAL_ENV_NAME=""
- Utilize CONDA_ENV_ARG="-p $GLOBAL_ENV_NAME" instaed of CONDA_ENV_ARG="-n $GLOBAL_ENV_NAME"
(make sure to use -p option) - Remove --conda $(which conda) from conda-lock install
(In my case 'which conda' made unexpected argement error) - Utilize conda activate .conda-env/ instead of conda activate $CONDA_ENV_NAME
(In my case it gave wrong command, leading to error)
Also, we sometimes need to modify env.sh to ensure correct activate.
In that case skip some stages or disable replace_content command in build-setup.sh.