Port the D64906 technique to EM_X86_64.
Details
Details
- Reviewers
grimar peter.smith ruiu • espindola - Commits
- rGd4306e90cb18: [ELF][X86] Allow PT_LOAD to have overlapping p_offset ranges on EM_X86_64
rL371958: [ELF][X86] Allow PT_LOAD to have overlapping p_offset ranges on EM_X86_64
rLLD371958: [ELF][X86] Allow PT_LOAD to have overlapping p_offset ranges on EM_X86_64
Diff Detail
Diff Detail
- Repository
- rLLD LLVM Linker
- Build Status
Buildable 38041 Build 38040: arc lint + arc unit
Event Timeline
Comment Actions
This patch is the (awful) reason I posted D67325 before this one. 🤔
The dependency order D67324 (omagic & nmagic) -> D67325 (image base) -> D67482 (EM_X86_64) -> D67481 (-z separate-segments) will cause the least number of redundantly updated tests. Say, I do D67482 before D67325, the following tests will be updated redundantly.
Failing Tests (6): lld :: ELF/avoid-empty-program-headers.s lld :: ELF/common-page.s lld :: ELF/gnu-ifunc-empty.s lld :: ELF/image-base.s lld :: ELF/linkerscript/page-size.s lld :: ELF/linkerscript/symbol-reserved.s ``` 🤔
Comment Actions
Just a quick, no objections from me. I'll leave the decision to go ahead to people most concerned and knowledgeable about x86_64.
Comment Actions
Thanks for the math suggestion. Will push shortly. We should soon get rid of the bool enabled variable in Writer<ELFT>::fixSectionAlignments().
test/ELF/x86-64-tls-ie.s | ||
---|---|---|
33 | It should be 0x2023C8. It is another use of tls0. |