Hi Siva,
An SE simulation implicitly runs at EL0, and reading MPIDR_EL1 is undefined at
EL0 (unless FEAT_IDST is implemented, which is not currently the case in gem5).
It looks like gem5 is segfaulting instead of failing more gracefully in this
case, but in general you won’t be able to read
Hi Yifan,
I believe Linux boot with FEAT_VHE enabled is not supported in gem5 yet.
The VIRTUALIZATION feature is supported, but there were some problem edge cases
so it was not enabled by default in gem5-21.2 and earlier. We expect it will be
enabled by default in the next stable version of
Hi Lucas,
As Ayaz mentioned, it is possible to run many instances of gem5. Aside from
system resources like available memory, the main caveat is that you will need
to keep any writable resources separate per simulation. In practice this is
most likely to be the gem5 output directory.
You can
Hello,
You can use the special "m5ops" instructions to communicate between your
workload and the simulator. The documentation is here:
https://www.gem5.org/documentation/general_docs/m5ops/
The available instructions are listed here, and you can build these wrapper
functions into a library to
Hello,
The default python-config probably picks up your default Python installation
(3.11) even if you choose a different Python version to invoke scons.
You could check this by running python-config and checking which link options
it produces.
As Gabriel mentioned, Python 3.11 isn’t supported
Hi Derek,
We have fixed this problem on the `develop` branch, but have not backported it
to the stable branch yet.
You should be able to cherry-pick this commit to fix the problem:
https://gem5-review.googlesource.com/c/public/gem5/+/70237
For a venv, the problem is that the python3-config