From: Namhyung Kim <namhyung@kernel.org>
To: Zecheng Li <zecheng@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Mark Rutland <mark.rutland@arm.com>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Jiri Olsa <jolsa@kernel.org>, Ian Rogers <irogers@google.com>,
Adrian Hunter <adrian.hunter@intel.com>,
"Liang, Kan" <kan.liang@linux.intel.com>,
Masami Hiramatsu <mhiramat@kernel.org>,
Xu Liu <xliuprof@google.com>,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 01/10] perf dwarf-aux: Use signed variable types in match_var_offset
Date: Wed, 27 Aug 2025 23:52:02 -0700 [thread overview]
Message-ID: <aK_8kt5Yf9MDoPdu@google.com> (raw)
In-Reply-To: <20250825195412.223077-2-zecheng@google.com>
On Mon, Aug 25, 2025 at 07:54:03PM +0000, Zecheng Li wrote:
> match_var_offset compares address offsets to determine if an access
> falls within a variable's bounds. The offsets involved for those
> relative to base registers from DW_OP_breg can be negative.
>
> The current implementation uses unsigned types (u64) for these offsets,
> which rejects almost all negative values.
Right, I thought it cannot get negative offsets except for stack access
(e.g. fbreg). But it turns out that container_of() trick can generate
them with optimizing compilers.
>
> Change the signature of match_var_offset to use signed types (s64). This
> ensures correct behavior when addr_offset or addr_type are negative.
>
> Signed-off-by: Zecheng Li <zecheng@google.com>
I've confirmed it produced slightly better results on my test sets.
Reviewed-by: Namhyung Kim <namhyung@kernel.org>
Thanks,
Namhyung
> ---
> tools/perf/util/dwarf-aux.c | 11 ++++++-----
> 1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
> index 559c953ca172..920054425578 100644
> --- a/tools/perf/util/dwarf-aux.c
> +++ b/tools/perf/util/dwarf-aux.c
> @@ -1388,18 +1388,19 @@ struct find_var_data {
> #define DWARF_OP_DIRECT_REGS 32
>
> static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
> - u64 addr_offset, u64 addr_type, bool is_pointer)
> + s64 addr_offset, s64 addr_type, bool is_pointer)
> {
> Dwarf_Die type_die;
> Dwarf_Word size;
> + s64 offset = addr_offset - addr_type;
>
> - if (addr_offset == addr_type) {
> + if (offset == 0) {
> /* Update offset relative to the start of the variable */
> data->offset = 0;
> return true;
> }
>
> - if (addr_offset < addr_type)
> + if (offset < 0)
> return false;
>
> if (die_get_real_type(die_mem, &type_die) == NULL)
> @@ -1414,11 +1415,11 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
> if (dwarf_aggregate_size(&type_die, &size) < 0)
> return false;
>
> - if (addr_offset >= addr_type + size)
> + if ((u64)offset >= size)
> return false;
>
> /* Update offset relative to the start of the variable */
> - data->offset = addr_offset - addr_type;
> + data->offset = offset;
> return true;
> }
>
> --
> 2.51.0.261.g7ce5a0a67e-goog
>
next prev parent reply other threads:[~2025-08-28 6:52 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-25 19:54 [PATCH v2 00/10] perf tools: Some improvements on data type profiler Zecheng Li
2025-08-25 19:54 ` [PATCH v2 01/10] perf dwarf-aux: Use signed variable types in match_var_offset Zecheng Li
2025-08-28 6:52 ` Namhyung Kim [this message]
2025-09-03 15:49 ` Arnaldo Carvalho de Melo
2025-09-03 22:05 ` Arnaldo Carvalho de Melo
2025-09-05 19:50 ` Namhyung Kim
2025-08-25 19:54 ` [PATCH v2 02/10] perf dwarf-aux: More accurate variable type match for breg Zecheng Li
2025-08-28 7:18 ` Namhyung Kim
2025-08-28 18:36 ` Zecheng Li
2025-08-30 0:53 ` Namhyung Kim
2025-08-25 19:54 ` [PATCH v2 03/10] perf dwarf-aux: Better variable collection for insn tracking Zecheng Li
2025-08-30 1:22 ` Namhyung Kim
2025-08-25 19:54 ` [PATCH v2 04/10] perf annotate: Skip annotating data types to lea instructions Zecheng Li
2025-08-30 6:41 ` Namhyung Kim
2025-08-25 19:54 ` [PATCH v2 05/10] perf dwarf-aux: Find pointer type to a type Zecheng Li
2025-08-30 6:48 ` Namhyung Kim
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aK_8kt5Yf9MDoPdu@google.com \
--to=namhyung@kernel.org \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=kan.liang@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mhiramat@kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=xliuprof@google.com \
--cc=zecheng@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).