[COMMITTED] gas: testsuite: fix regression in cfi-common-10.d
Alan Modra
amodra@gmail.com
Tue Aug 19 23:54:44 GMT 2025
On Tue, Aug 19, 2025 at 02:13:02PM -0700, Indu Bhagat via Binutils wrote:
> A previous commit 09292f4ae2c introduced the new test cfi-common-10.d.
> The testcase needs some adjustments for it to be useful on a variety of
> targets.
Thanks for looking at it, but unfortunately that didn't cure all of
the regressions. You still have these:
arc-elf +FAIL: CFI common 10
arc-linux-uclibc +FAIL: CFI common 10
mips64el-openbsd +FAIL: CFI common 10
mips64-linux-gnuabi64 +FAIL: CFI common 10
mips64-openbsd +FAIL: CFI common 10
sh4-linux-gnu +FAIL: CFI common 10
shle-unknown-netbsdelf +FAIL: CFI common 10
sh-nto +FAIL: CFI common 10
sh-rtems +FAIL: CFI common 10
The arc and sh failures are caused by "DW_CFA_advance_loc: 4 to 0+0004"
not matching. They get different addresses due to peculiarities in
their reloc howto descriptions, resulting in .eh_frame not being
relocated properly for objdump. The mips64 fails are due to FDE size
not matching.
I'm going to commit the following. The .long could also be removed
from the .s file but I decided it didn't matter and left that as-is.
Re: gas: testsuite: fix regression in cfi-common-10.d
Just match what the test is actually trying to verify, to fix more
regressions.
diff --git a/gas/testsuite/gas/cfi/cfi-common-10.d b/gas/testsuite/gas/cfi/cfi-common-10.d
index 2261da47dfd..80900d96e64 100644
--- a/gas/testsuite/gas/cfi/cfi-common-10.d
+++ b/gas/testsuite/gas/cfi/cfi-common-10.d
@@ -1,19 +1,8 @@
#objdump: -Wf
#name: CFI common 10
-#...
-Contents of the .eh_frame section:
-0+0000 0+00(10|14) 0+0000 CIE
- Version: 1
- Augmentation: "zR"
- Code alignment factor: .*
- Data alignment factor: .*
- Return address column: .*
- Augmentation data: [01][abc]
#...
-
-0+00(14|18) 0+00(18|1c) 0+00(18|1c) FDE cie=0+0000 pc=.*
+Contents of the \.eh_frame section:
#...
- DW_CFA_advance_loc: 4 to 0+0004
DW_CFA_def_cfa_offset: 1099511627808
-#...
+#pass
--
Alan Modra
More information about the Binutils
mailing list