Re: [PATCH v12 06/19] perf symbol: Avoid use of machine__is

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Cross-platform analysis regression due to host architecture fallback for guest kernels
--

commit 5ffd66d0549571c3b1c8d4a0baf650be921efbdd
Author: Ian Rogers <irogers@google.com>

perf symbol: Avoid use of machine__is

This commit switches from using machine__is() to dynamically checking the
ELF machine type via the DSO or running machine environment. This avoids
imprecise string comparisons and reliance on the perf_env arch string.

> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index 714b6e6048fa2..2ce512f08a1d6 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -851,6 +851,23 @@ static int maps__split_kallsyms_for_kcore(struct maps *kmaps, struct dso *dso)
>  	return count;
>  }
>  
> +static uint16_t machine_or_dso_e_machine(struct machine *machine, struct dso *dso)
> +{
> +	uint16_t e_machine = EM_NONE;
> +	/* DSO should be most accurate */
> +	if (dso)
> +		e_machine = dso__e_machine(dso, machine, /*e_flags=*/NULL);
> +
> +	if (e_machine != EM_NONE)
> +		return e_machine;
> +
> +	/* Check the global environment next. */
> +	if (machine && machine->env && machine->env->e_machine != EM_NONE)
> +		return machine->env->e_machine;
> +
> +	return perf_env__e_machine(machine ? machine->env : NULL, /*e_flags=*/NULL);

[Severity: High]
Does this fallback logic inadvertently break cross-platform analysis for
guest kernels?

When analyzing guest kernels, machine->env is NULL. If dso__e_machine()
returns EM_NONE (e.g. for kernel pseudo-files, where open() fails), this
falls through to call perf_env__e_machine() with a NULL environment.

A NULL environment passed to perf_env__e_machine() delegates to
perf_env__e_machine_nocache(), which invokes uname() to unconditionally
substitute the host's architecture. The previous code using machine__is()
safely returned "unknown" for guests without falling back to the host.

> +}
> +

[ ... ]

> @@ -1716,7 +1732,7 @@ int dso__load(struct dso *dso, struct map *map)
>  			ret = dso__load_guest_kernel_sym(dso, map);
>  
>  		machine = maps__machine(map__kmaps(map));
> -		if (machine__is(machine, "x86_64"))
> +		if (machine && machine_or_dso_e_machine(machine, dso) == EM_X86_64)

[Severity: High]
If an x86_64 host analyzes a cross-platform non-x86 guest kernel, will the
host fallback cause EM_X86_64 to match here?

Due to the host fallback described above, this appears like it would
mistakenly trigger machine__map_x86_64_entry_trampolines() for the non-x86
guest, potentially injecting bogus memory maps and corrupting the guest's
layout analysis.

>  			machine__map_x86_64_entry_trampolines(machine, dso);
>  		goto out;
>  	}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260602152516.2831152-1-irogers@google.com?part=6