Re: [PATCH v3 55/58] objtool/klp: Add test for absolute and __ADDRESSABLE symbols

On Mon, Sep 14, 2026 at 5:53 PM [off-list ref] wrote:

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] objtool: test suite build failure on 32-bit architectures due to .quad
--

commit 8413836c45f39cb1efbf4a2f69693cc56d13f861
Author: Song Liu [off-list ref]

objtool/klp: Add test for absolute and __ADDRESSABLE symbols

This commit adds a test fixture and script to ensure that objtool's
livepatching features correctly handle absolute symbols and __ADDRESSABLE
pointers without failing or crashing. It uses inline assembly to emit
these structures for the test framework.

quoted

diff --git a/tools/objtool/tests/generic/fixtures/abs_and_addressable.c b/tools/objtool/tests/generic/fixtures/abs_and_addressable.c
new file mode 100644
index 0000000000000..6392ff99af421
--- /dev/null
+++ b/tools/objtool/tests/generic/fixtures/abs_and_addressable.c

[ ... ]

quoted

+int helper(int x);
+int helper(int x) { return x + 1; }
+
+/* The shape of __ADDRESSABLE(helper). */
+__asm__(".pushsection .discard.addressable, \"aw\"\n"
+     ".balign 8\n"
+     ".quad helper\n"

[Severity: Medium]
Does using ".quad" here cause a build failure on 32-bit architectures?

When running the generic klp test suite on a 32-bit host (like 32-bit x86),
the assembler cannot emit a 64-bit absolute relocation for a symbol and fails
with an error like "cannot represent relocation type BFD_RELOC_64".

Would using a pointer-sized assembly directive instead of hardcoding a 64-bit
directive prevent this from causing a hard failure in build_one() when
running `make tests` on 32-bit targets?

klp-build is for 64 bit only, so let's not worry about 32 bit targets.

Thanks,
Song