[BOLT] Testcase to repro dyn reloc bug
authorRafael Auler <rafaelauler@fb.com>
Fri, 7 Oct 2022 18:26:51 +0000 (11:26 -0700)
committerRafael Auler <rafaelauler@fb.com>
Fri, 7 Oct 2022 18:27:23 +0000 (11:27 -0700)
commit696b8ea05fb18309bf3a68ae1d7e9d1659561ed8
treee27643543136aa4015701fe99c3fcbbb92a49c51
parent07c0a41b53f532902ed977589363d2ba51748e75
[BOLT] Testcase to repro dyn reloc bug

Add a new testcase that shows a bug in BOLT when writing out
dynamic relocations. This is currently marked as XFAIL as we work on
solving it. This bug happens when the current strategy fails to
recognize that the original dynamic relocation in the input should
reference the original .bolt.org.rodata section instead of the new one
.rodata created by BOLT after moving jump tables. This bug started
happening after 729d29e167a5.

Reviewed By: Amir

Differential Revision: https://reviews.llvm.org/D125941
bolt/test/X86/Inputs/define_bar.s [new file with mode: 0644]
bolt/test/X86/dynrelocs.s [new file with mode: 0644]