[RuntimeDyldChecker][AArch32] Add a PC offset to next_PC for ARM targets #91746
Conversation
There was a problem hiding this comment.
Hey! Thanks for your proposal. I think it's valuable. Your second patch seemed very much to the point. Maybe with a more verbose comment like:
We could then clean up these comments in the tests. What do you think?| "expected '+' for offset or ')' if no offset"), | ||
| ""); | ||
| } | ||
|
|
There was a problem hiding this comment.
This is only used in a single test, right? If we really want to use it there, we should rather fix the test by adding another label. Otherwise, if we really want to go down this road, it would probably be best to support full expressions here again.
There was a problem hiding this comment.
I was a little doubtful with adding the feature for a single case and I think you are right, this solves a problem we might not need to have. I will update the test instead.
Thanks for the review! Sounds good, I was conservative with line estate as the checker deals with all targets but more explanation would not hurt definitely. |
In ARM mode, the Program Counter (PC) points to the current instruction's address + 8 instead of + 4. This offset is added to RuntimeDyldChecker to use `next_pc` expression in JITLink tests.
|
Flang failed to build on Windows with "compiler is out of heap space" -- once again an issue with the infra. Since all our tests passed, I think this is ready to land. Thanks again for working on this! |
Thanks for the green light! |
…ets (llvm#91746) In ARM mode, the Program Counter (PC) points to the current instruction's address + 8 instead of + 4. An offset is added to RuntimeDyldChecker to use `next_pc` expression in JITLink tests with both Thumb and Arm.
In ARM mode, the Program Counter (PC) points to the current instruction's
address + 8 instead of + 4. This offset is added to RuntimeDyldChecker to
use
next_pcexpression in JITLink tests.