(Chipyard 1.13) GLIBC `riscv-isa-sim` build error on non-RHEL9 OS

[zzulb]

Background Work

• Yes, I searched the mailing list
• Yes, I searched prior issues
• Yes, I searched the documentation

Chipyard Version and Hash

Release: 1.13.0
latest

OS Setup

ubuntu 20.0

Other Setup

Current Behavior

Execute “module_all riscv-isa-sim --prefix="${RISCV}" --with-boost=no --with-boost-asio=no --with-boost-regex=no” and an error is reported. It is prompted to give the --host parameter.

Expected Behavior

"module_all riscv-isa-sim --prefix="${RISCV}" --with-boost=no --with-boost-asio=no --with-boost-regex=no --host=x86_64-linux-gnu" This can be executed. But I don't know if it is correct.

"module_all riscv-isa-sim --prefix="${RISCV}" --with-boost=no --with-boost-asio=no --with-boost-regex=no --host=riscv${XLEN}-unknown-elf" This still reports an error, missing libpthread, I don't know how to deal with it.

Other Information

No response

[zzulb]

This is a screenshot of the error.

[MuzamilFarid]

I am facing similar issue in Step 3, Building Toolchain Collateral , this is the exact snipped from the log file

========== BEGINNING STEP 3: Building toolchain collateral ==========
Installing extra toolchain utilities/tests to /home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/.conda-env/riscv-tools
==>  Installing Spike
=>  Starting riscv-isa-sim build
==>  Initializing riscv-isa-sim submodule
+ git submodule update --init --recursive /home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/toolchains/riscv-tools/riscv-isa-sim
Cloning into '/home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/toolchains/riscv-tools/riscv-isa-sim'...
Submodule path 'toolchains/riscv-tools/riscv-isa-sim': checked out '76858cc3aa178446584eae52f82e0a933e71f8db'
==>  cd /home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/toolchains/riscv-tools/riscv-isa-sim
+ mkdir -p build
==>  Configuring riscv-isa-sim
+ cd build
+ ../configure --prefix=/home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/.conda-env/riscv-tools --with-boost=no --with-boost-asio=no --with-boost-regex=no
checking build system type... x86_64-pc-linux-gnu
checking host system type... x86_64-pc-linux-gnu
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables... 
checking whether we are cross compiling... configure: error: in '/home/idsproj01/trusted/projects/TrEB/Design/TReB_ASIC_01/_cad/Digital/user/farid/chipyard/chipyard/toolchains/riscv-tools/riscv-isa-sim/build':
configure: error: cannot run C compiled programs.
If you meant to cross compile, use '--host'.
See 'config.log' for more details
build-setup.sh: Build script failed with exit code 77 at step 3: Building toolchain collateral
``

[u-su]

This is due to the fact that there is a hard requirement of GLIBC = 2.34 within the script.
This is contradictory; considering EDA tools such as VCS (as of today) have no official support for RHEL9; which contains GLIBC 2.34.

A temporary workaround would be to use 'patchelf' with desired GLIBC libraries built separately. This is a somewhat tedious process.

In the long run, scripts should be updated to accommodate GLIBC 2.28 as well (which is, the version that ships with RHEL8).

[Fi50]

[changed title to be clearer(?), feel free to modify >_<]
Vague thoughts: hm, seems this is coming up for ubuntu users in particular..? We recently upgraded to rhel9 so it's possible some changes got upstreamed that are breaking for others..? A quick check might be for someone who's running into this to revert to some older working commits and compare (e.g. test if the only change is GLIBC =, etc). If there's a minimal changeset that works maybe it can be implemented as a fallback..

[suhwanhong999]

Seems like the merge found here is the splitting point. #2174

The problem seems to appear from this commit and onwards, while the commit before this (#2166) seems to not have this issue with the GLIBC version.

[suhwanhong999]

Seems that changing the .conda-reqs/chipyard-base.yaml file line from "sysroot_linux-64=2.34" to "sysroot_linux-64=2.28" doesn't fix anything as far as I can tell. Still failing on step 3.

[marie-anne-xu]

Try running generate-conda-lockfiles.sh in the scripts directory to update the lockfiles to use 2.28, then try building again?

[suhwanhong999]

doesn't seem to fix the issue as far as I can tell. I know the OS being old is the issue, since all of this works perfectly well on WSL (running on my laptop)