linux-kernel - Re: [PATCH v4 1/9] perf annotate: Skip annotating data types to lea instructions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <aPW6sUywwZzDeOkW@google.com>
Date: Mon, 20 Oct 2025 13:29:37 +0900
From: Namhyung Kim <namhyung@...nel.org>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Zecheng Li <zecheng@...gle.com>, Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...hat.com>, Mark Rutland <mark.rutland@....com>,
	Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
	Jiri Olsa <jolsa@...nel.org>, Ian Rogers <irogers@...gle.com>,
	Adrian Hunter <adrian.hunter@...el.com>,
	Masami Hiramatsu <mhiramat@...nel.org>,
	Xu Liu <xliuprof@...gle.com>, linux-perf-users@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 1/9] perf annotate: Skip annotating data types to lea
 instructions

On Tue, Oct 14, 2025 at 03:40:33PM -0300, Arnaldo Carvalho de Melo wrote:
> On Mon, Oct 13, 2025 at 06:15:58PM +0000, Zecheng Li wrote:
> > Introduce a helper function is_address_gen_insn() to check
> > arch-dependent address generation instructions like lea in x86. Remove
> > type annotation on these instructions since they are not accessing
> > memory. It should be counted as `no_mem_ops`.
> > 
> > Signed-off-by: Zecheng Li <zecheng@...gle.com>
> > ---
> >  tools/perf/util/annotate.c | 20 ++++++++++++++++++++
> >  1 file changed, 20 insertions(+)
> > 
> > diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c
> > index a2e34f149a07..fb60467fa877 100644
> > --- a/tools/perf/util/annotate.c
> > +++ b/tools/perf/util/annotate.c
> > @@ -2698,6 +2698,20 @@ static bool is_stack_canary(struct arch *arch, struct annotated_op_loc *loc)
> >  	return false;
> >  }
> >  
> > +/**
> > + * Returns true if the instruction has a memory operand without
> > + * performing a load/store
> > + */
> > +static bool is_address_gen_insn(struct arch *arch, struct disasm_line *dl)
> > +{
> > +	if (arch__is(arch, "x86")) {
> > +		if (!strncmp(dl->ins.name, "lea", 3))
> > +			return true;
> > +	}
> 
> Can't we turn this into:
> 
> tatic bool disasm_line__is_address_gen_insn(const struct disasm_line *dl)
> {
> 		return dl->ins.address_gen;
> }

Probably better to add ins->ops->addr_gen() instead of adding a new
field to the struct ins.  Then it could be:

static bool ins__is_addr_gen(const struct ins *ins)
{
	return ins->ops == &addr_gen_ops;
}

But this requires more changes and it can be done later.

Thanks,
Namhyung

> 
> I.e. at some initial step when setting dl->ins, cache this series of
> string operations and then use it s result?
> 
> - Arnaldo
> 
> > +
> >  static struct disasm_line *
> >  annotation__prev_asm_line(struct annotation *notes, struct disasm_line *curr)
> >  {
> > @@ -2806,6 +2820,12 @@ __hist_entry__get_data_type(struct hist_entry *he, struct arch *arch,
> >  		return &stackop_type;
> >  	}
> >  
> > +	if (is_address_gen_insn(arch, dl)) {
> > +		istat->bad++;
> > +		ann_data_stat.no_mem_ops++;
> > +		return NO_TYPE;
> > +	}
> > +
> >  	for_each_insn_op_loc(&loc, i, op_loc) {
> >  		struct data_loc_info dloc = {
> >  			.arch = arch,
> > -- 
> > 2.51.0.788.g6d19910ace-goog