Conversation
|
I'll edit the CHANGELOG and update the PR. |
fneddy
force-pushed
the
big_endian
branch
3 times, most recently
from
7bd088e to
8abe282
Compare
|
Ok. I think all the tests are actually not my fault. this is ready for review. |
theotherjimmy
left a comment
theotherjimmy
left a comment
There was a problem hiding this comment.
Looks good. I appreciate explaining the clearing of errno in one of the commit messages.
|
Thanks for the contribution! We've never added support for a target platform that we cannot run in the rust-vmm CI. AFAIK there's no such CI runner yet in the rust-vmm CI, is there? Otherwise, we'd be merging code without any quality guarantee that it works (I, for one, don't have such a machine to test even) so I'm afraid we can't proceed unless support is added for it in the CI (like we did for riscv also: rust-vmm/rust-vmm-ci#159) |
|
thx I will look into it! |
|
rust-vmm doesn't support big endian architectures. |
|
there are other projects that depend on seccompiler that are interesting for big-endian. I would request an action runner at https://github.com/IBM/actionspz/ if that's ok. |
|
Under supported platform: https://github.com/rust-vmm/seccompiler/blob/main/README.md#supported-platforms, should add |
big-endian: - added endian switch to SeccompCondition::get_data_offsets everything else is endianness independent. - fixed tests to have endianness specific data offsets Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
s390: - add s390x to generate_syscall_tables and the generates code - add flags and switches for s390x to the backend test: - there is no guarantee that errno is 0 at main entry or thread entry in fact on s390x its 2 ENOENT as rust startup code will try to getauxval(AT_MINSIGSTKSZ) that is not defined. Also getpid WILL NOT reset errno to 0 if it suceeded. Therefore add errno crate and set errno at start of the test to 0. Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
added s390x to supported platforms section in readme. Marked it as Experimental Signed-off-by: Eddy (Eduard) Stefes <eddy@linux.ibm.com>
We shouldn't do this without CI validating that it works. Otherwise, what stops us from adding all targets we can think of, not testing them, and just releasing possibly broken code (even if documented as experimental) |
5 participants
Summary of the PR
added big endian support and as first big-endian platform s390x.
Requirements
Before submitting your PR, please make sure you addressed the following
requirements:
git commit -s), and the commit message has max 60 characters for thesummary and max 75 characters for each description line.
test.
Release" section of CHANGELOG.md (if no such section exists, please create one).
unsafecode is properly documented.TODO
-- s390x: Introduce s390x CI container rust-vmm-container#145