linux-kernel - Re: [PATCH v2 26/26] perf hist: Fix srcline memory leak

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: <ZIeOY4OZM2BeaUhC@kernel.org>
Date:   Mon, 12 Jun 2023 18:30:11 -0300
From:   Arnaldo Carvalho de Melo <acme@...nel.org>
To:     Andi Kleen <ak@...ux.intel.com>
Cc:     Ian Rogers <irogers@...gle.com>,
        John Garry <john.g.garry@...cle.com>,
        Will Deacon <will@...nel.org>,
        James Clark <james.clark@....com>,
        Mike Leach <mike.leach@...aro.org>,
        Leo Yan <leo.yan@...aro.org>,
        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>,
        Namhyung Kim <namhyung@...nel.org>,
        Adrian Hunter <adrian.hunter@...el.com>,
        Suzuki K Poulose <suzuki.poulose@....com>,
        "Naveen N. Rao" <naveen.n.rao@...ux.vnet.ibm.com>,
        Kan Liang <kan.liang@...ux.intel.com>,
        German Gomez <german.gomez@....com>,
        Ali Saidi <alisaidi@...zon.com>,
        Jing Zhang <renyu.zj@...ux.alibaba.com>,
        Athira Rajeev <atrajeev@...ux.vnet.ibm.com>,
        Miguel Ojeda <ojeda@...nel.org>,
        ye xingchen <ye.xingchen@....com.cn>,
        Liam Howlett <liam.howlett@...cle.com>,
        Dmitrii Dolgov <9erthalion6@...il.com>,
        Yang Jihong <yangjihong1@...wei.com>,
        K Prateek Nayak <kprateek.nayak@....com>,
        Changbin Du <changbin.du@...wei.com>,
        Ravi Bangoria <ravi.bangoria@....com>,
        Sean Christopherson <seanjc@...gle.com>,
        "Steinar H. Gunderson" <sesse@...gle.com>,
        Yuan Can <yuancan@...wei.com>,
        Brian Robbins <brianrob@...ux.microsoft.com>,
        liuwenyu <liuwenyu7@...wei.com>,
        Ivan Babrou <ivan@...udflare.com>,
        Fangrui Song <maskray@...gle.com>,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-perf-users@...r.kernel.org, coresight@...ts.linaro.org
Subject: Re: [PATCH v2 26/26] perf hist: Fix srcline memory leak

Em Mon, Jun 12, 2023 at 02:16:34PM -0700, Andi Kleen escreveu:
> On Mon, Jun 12, 2023 at 02:23:41PM -0300, Arnaldo Carvalho de Melo wrote:
> > Em Mon, Jun 12, 2023 at 07:46:14AM -0700, Ian Rogers escreveu:
> > > On Mon, Jun 12, 2023 at 7:16 AM Arnaldo Carvalho de Melo
> > > <acme@...nel.org> wrote:
> > > >
> > > > Em Mon, Jun 12, 2023 at 11:13:59AM -0300, Arnaldo Carvalho de Melo escreveu:
> > > > > Em Thu, Jun 08, 2023 at 04:28:23PM -0700, Ian Rogers escreveu:
> > > > > > srcline isn't freed if it is SRCLINE_UNKNOWN. Avoid strduping in this
> > > > > > case as such strdups are redundant and leak memory.
> > > > >
> > > > > The patch is ok as its what the rest of the code is doing, i.e. strcmp()
> > > > > to check if a srcline is the unknown one, but how about the following
> > > > > patch on top of yours?
> > > >
> > > > [acme@...co perf-tools-next]$ strings ~/bin/perf | grep '??:0'
> > > > ??:0
> > > > SRCLINE_UNKNOWN ((char *) "??:0")
> > > > [acme@...co perf-tools-next]$
> > > 
> > > Agreed, the strcmps make me nervous as they won't distinguish heap
> > > from a global meaning we could end up with things like pointers to
> > > freed memory. The comparison with the global is always going to be
> > > same imo.
> > > 
> > > Acked-by: Ian Rogers <irogers@...gle.com>
> > 
> > Thanks, applied and added your Acked-by.
> 
> Actually was there another patch that turned it into a explicit global? 
> 
> At least in my tree it isn't:
> 
> util/srcline.h
> 28:#define SRCLINE_UNKNOWN  ((char *) "??:0")
> 
> Without any explicit global it's a bit dangerous because you rely on the
> linker doing string deduplication.  While it normally does that there might be
> odd corner cases where it doesn't.

Yeah, several gcc and clang versions complained with:

2    65.38 almalinux:9                   : FAIL clang version 15.0.7 (Red Hat 15.0.7-2.el9)
    builtin-diff.c:1381:17: error: result of comparison against a string literal is unspecified (use an explicit string comparison function instead) [-Werror,-Wstring-compare]
            if (start_line != SRCLINE_UNKNOWN &&

So I added the following on top:

diff --git a/tools/perf/util/srcline.c b/tools/perf/util/srcline.c
index 48a04f42b308b080..aec596a0b0bbec0f 100644
--- a/tools/perf/util/srcline.c
+++ b/tools/perf/util/srcline.c
@@ -23,6 +23,8 @@
 
 bool srcline_full_filename;
 
+char *srcline__unknown = (char *)"??:0";
+
 static const char *dso__name(struct dso *dso)
 {
 	const char *dso_name;
diff --git a/tools/perf/util/srcline.h b/tools/perf/util/srcline.h
index a15c7db9058ece96..167645bcff0755e2 100644
--- a/tools/perf/util/srcline.h
+++ b/tools/perf/util/srcline.h
@@ -25,7 +25,8 @@ char *srcline__tree_find(struct rb_root_cached *tree, u64 addr);
 /* delete all srclines within the tree */
 void srcline__tree_delete(struct rb_root_cached *tree);
 
-#define SRCLINE_UNKNOWN  ((char *) "??:0")
+extern char *srcline__unknown;
+#define SRCLINE_UNKNOWN srcline__unknown
 
 struct inline_list {
 	struct symbol		*symbol;