Re: [PATCH 2/6] perf dwarf-aux: Add die_same_file() and die_get_type_die()

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] perf dwarf-aux: truncation of u64 die_off in pr_debug on 32-bit platforms
--

commit c7e2660460f55978ed41b85fca70ec2e99b2a867
Author: Arnaldo Carvalho de Melo <acme@redhat.com>

perf dwarf-aux: Add die_same_file() and die_get_type_die()

This commit introduces die_same_file() to compare the debug files of two DIEs
and die_get_type_die() to resolve a type DIE offset in the correct debug file.
This ensures proper offset resolution for DWARF files and DWZ common files.

> diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
> index 2c135ebb1551f..54f8b5ec74a2f 100644
> --- a/tools/perf/util/dwarf-aux.c
> +++ b/tools/perf/util/dwarf-aux.c

[ ... ]

> +Dwarf_Die *die_get_type_die(Dwarf *dbg, u64 die_off, int die_tag, bool from_alt,
> +			    Dwarf_Die *die_mem)
> +{
> +	Dwarf *target = dbg;
> +	Dwarf_Die die;
> +
> +	if (from_alt) {
> +		/* No fallback: an alt offset here parses whatever is at it. */
> +		target = dwarf_getalt(dbg);
> +		if (target == NULL) {
> +			pr_debug("DWARF: no alt (dwz) debug file to resolve the type DIE at offset 0x%lx in\n",
> +				 (unsigned long)die_off);

[Severity: Low]
Could this truncate the DWARF file offset in tools/perf/util/dwarf-aux.c
within die_get_type_die() on 32-bit platforms if the offset exceeds 4GB?
The 64-bit integer die_off is cast to unsigned long when logging a failed
alt offset resolution, which is only 32 bits on 32-bit architectures,
causing debug messages to display an incorrect offset. Should this use
PRIx64 instead?

> +			return NULL;
> +		}
> +	}
> +
> +	if (dwarf_offdie(target, die_off, &die) && dwarf_tag(&die) == die_tag) {
> +		*die_mem = die;
> +		return die_mem;
> +	}
> +
> +	pr_debug("DWARF: no DIE with tag %d at offset 0x%lx in the %s debug file\n",
> +		 die_tag, (unsigned long)die_off, from_alt ? "alt" : "main");

[Severity: Low]
Similarly, when logging a missing DIE at the end of die_get_type_die(),
does casting the u64 die_off to unsigned long truncate the upper 32 bits
on 32-bit architectures? Consider formatting with PRIx64 to ensure the full
64-bit value is printed correctly.

> +	return NULL;
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260925150657.1826942-1-acme@kernel.org?part=2