Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix macos #1980

Merged
merged 6 commits into from
Jun 26, 2024
Merged

Fix macos #1980

merged 6 commits into from
Jun 26, 2024

Conversation

d3zd3z
Copy link
Member

@d3zd3z d3zd3z commented Jun 12, 2024

Several fixes to allow the simulator to run on MacOS (aarch64). Several of these are actual problems with taking addresses of stack variables, or values that are later moved.

There is also a problem with the LLVM linker and the way the FIH_LABEL labels are added, which results in large blocks of code being discarded, and resulting in nonsensical code flow. For now, disable these labels on Apple targets, but as Rust is moving to the LLVM linker on other targets, this is likely to be more important.

Fixes #1979.

Copy link
Member

@utzig utzig left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There's a typo in the last commit message: conflicing -> conflicting.

sim/mcuboot-sys/src/c.rs Outdated Show resolved Hide resolved
This struct was having addresses taken of fields within it, and then being
returned.  It is platform-specific whether this causes a move.  It seems to be
working on x86_64, but causes a segfault on aarch64.  Box the struct so that it
isn't moved after being initialized.

Signed-off-by: David Brown <[email protected]>
Since there are references to this struct passed to C code, put it into an Rc so
that it won't move around when the data is moved.

Signed-off-by: David Brown <[email protected]>
The symbols injected here cause some kind of poor interaction with the linker on
MacOS, which results in most of the code becoming hopelessly corrupt.  For now,
just disable these symbols on this target.

Signed-off-by: David Brown <[email protected]>
Many of these extern functions are missing the "C". It doesn't seem to matter on
any of our targets, but this does make the code more correct, and might be a
problem in the future.

Signed-off-by: David Brown <[email protected]>
Cargo prints a warning about conflicting resolvers being used.  Fix this
by explicitly setting the resolver in the project file.

Signed-off-by: David Brown <[email protected]>
Increase the size of the jmpbuf to accomodate other architectures.
Unfortunately, the size of this is not available in the libc crate.
Increase this so encompass any platforms we wish to support, including
aarch64 on both Linux and MacOS.

Increasing an array beyond 32 means there is no default offered, so
implement this manually.

Signed-off-by: David Brown <[email protected]>
Copy link
Member

@utzig utzig left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@d3zd3z d3zd3z merged commit a706317 into mcu-tools:main Jun 26, 2024
58 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

sim: AreaDesc::get_c violates stack
3 participants