Re: [PATCH v3 16/19] perf test cs-etm: Make disassembly test use kcore

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The arm-cs-trace-disasm.py script fails to use the explicitly provided kcore file for dynamically loaded kernel modules, silently falling back to the unpatched build-id ELF.
--

commit fb1d5bc3ba8d47081f18c5b9c0c07af6646b6389
Author: James Clark <james.clark@linaro.org>

perf test cs-etm: Make disassembly test use kcore

Hits in modules return empty disassembly with vmlinux as an input to objdump.
Make the disassembly test more reliable by always using kcore. And update the
comments to say that this is supported by the script.

> diff --git a/tools/perf/scripts/python/arm-cs-trace-disasm.py b/tools/perf/scripts/python/arm-cs-trace-disasm.py
> index ba208c90d6311..8f6fa4a007b42 100755
> --- a/tools/perf/scripts/python/arm-cs-trace-disasm.py
> +++ b/tools/perf/scripts/python/arm-cs-trace-disasm.py
> @@ -18,29 +18,29 @@ from perf_trace_context import perf_sample_srccode, perf_config_get
>  
>  # Below are some example commands for using this script.
>  # Note a --kcore recording is required for accurate decode
> -# due to the alternatives patching mechanism. However this
> -# script only supports reading vmlinux for disassembly dump,
> -# meaning that any patched instructions will appear
> -# as unpatched, but the instruction ranges themselves will
> -# be correct. In addition to this, source line info comes
> -# from Perf, and when using kcore there is no debug info. The
> -# following lists the supported features in each mode:
> +# due to the alternatives patching mechanism. In addition to this,
> +# source line info comes from Perf, and when using kcore there is
> +# no debug info. The following lists the supported features in each mode:
>  #
>  # +-----------+-----------------+------------------+------------------+
>  # | Recording | Accurate decode | Source line dump | Disassembly dump |
>  # +-----------+-----------------+------------------+------------------+
>  # | --kcore   | yes             | no               | yes              |
> -# | normal    | no              | yes              | yes              |
> +# | normal    | no              | yes (inaccurate) | yes (inaccurate) |
>  # +-----------+-----------------+------------------+------------------+

[Severity: Medium]
While this commit advertises kcore support, does arm-cs-trace-disasm.py
actually use the kcore file for dynamically loaded kernel modules?

Looking at get_dso_file_path() in
tools/perf/scripts/python/arm-cs-trace-disasm.py, if a module (e.g., [ext4])
is encountered, it bypasses options.vmlinux completely and attempts to use
the unpatched ELF from the build-id directory:

	if (dso_name == "[kernel.kallsyms]" or dso_name == "vmlinux"):
		if (options.vmlinux):
			return options.vmlinux;
		else:
			return find_vmlinux() if find_vmlinux() else dso_name

	if (dso_name == "[vdso]") :
		append = "/vdso"
	else:
		append = "/elf"

	dso_path = os.environ['PERF_BUILDID_DIR'] + "/" + dso_name + "/" + dso_build_id + append;

Additionally, if get_dso_file_path() were changed to return the kcore file
for modules, would process_event() still fail?

process_event() sets dso_vm_start to the module's map start address:

		if (dso == "[kernel.kallsyms]" or dso_start == 0x400000):
			dso_vm_start = 0
			map_pgoff = 0
		else:
			dso_vm_start = int(dso_start)

The subsequent read_disam() call then subtracts this from the sample address
to pass relative file offsets to objdump. This works for module ELFs but
breaks kcore, which requires absolute virtual addresses.

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260603-james-cs-context-tracking-fix-v3-0-c392945d9ed5@linaro.org?part=16