[BOLT][NFC] Fix test X86/dynamic-relocs-on-entry.s

[yozhu]

No description provided.

[llvmbot]

@llvm/pr-subscribers-bolt

Author: YongKang Zhu (yozhu)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/125264.diff

1 Files Affected:

• (modified) bolt/test/X86/dynamic-relocs-on-entry.s (+2-2)

diff --git a/bolt/test/X86/dynamic-relocs-on-entry.s b/bolt/test/X86/dynamic-relocs-on-entry.s
index 2a29a43c4939a2..a5bf89c8824434 100644
--- a/bolt/test/X86/dynamic-relocs-on-entry.s
+++ b/bolt/test/X86/dynamic-relocs-on-entry.s
@@ -9,7 +9,7 @@
 # RUN: FileCheck %s --input-file=%t.out.txt
 
 ## Check if the new address in `chain` is correctly updated by BOLT
-# CHECK: Relocation section '.rela.dyn' at offset 0x{{.*}} contains 1 entry:
+# CHECK: Relocation section '.rela.dyn' at offset 0x{{.*}} contains 1 entries:
 # CHECK: {{.*}} R_X86_64_RELATIVE [[#%x,ADDR:]]
 # CHECK: [[#ADDR]]: c3 retq
 	.text
@@ -29,4 +29,4 @@ _start:
 
 	.data
 .Lfoo:
-	.quad Label
\ No newline at end of file
+	.quad Label

[dcci]

Thanks @yozhu -- I wonder where this was failing?

[yozhu]

Thanks @yozhu -- I wonder where this was failing?

It fails in a "ninja check-bolt" run.

[dcci]

I see what's happening now. I think a better way of fixing this test would be that of using llvm-readelf rather than the system readelf. This is used to dump parts of the executable, so the format isn't particularly relevant.

[dcci]

LG if it passes CI. Thanks for the fix @yozhu