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: <20200226161132.2aeea72b71d48e9988aa27d9@kernel.org>
Date:   Wed, 26 Feb 2020 16:11:32 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     He Zhe <zhe.he@...driver.com>
Cc:     <peterz@...radead.org>, <mingo@...hat.com>, <acme@...nel.org>,
        <mark.rutland@....com>, <alexander.shishkin@...ux.intel.com>,
        <jolsa@...hat.com>, <namhyung@...nel.org>,
        <kstewart@...uxfoundation.org>, <tglx@...utronix.de>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] perf: probe-file: Check return value of
 strlist__add

On Wed, 26 Feb 2020 10:49:19 +0800
He Zhe <zhe.he@...driver.com> wrote:

> 
> 
> On 2/26/20 6:49 AM, Masami Hiramatsu wrote:
> > On Tue, 25 Feb 2020 22:41:43 +0800
> > <zhe.he@...driver.com> wrote:
> >
> >> From: He Zhe <zhe.he@...driver.com>
> >>
> >> strlist__add may fail with -ENOMEM or -EEXIST. Check it and give debugging
> >> hint when necessary.
> >>
> >> Signed-off-by: He Zhe <zhe.he@...driver.com>
> >> ---
> >>  tools/perf/builtin-probe.c   | 30 ++++++++++++++++--------------
> >>  tools/perf/util/probe-file.c | 26 +++++++++++++++++++++-----
> >>  2 files changed, 37 insertions(+), 19 deletions(-)
> >>
> >> diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c
> >> index 26bc5923e6b5..8b4511c70fed 100644
> >> --- a/tools/perf/builtin-probe.c
> >> +++ b/tools/perf/builtin-probe.c
> >> @@ -442,24 +442,26 @@ static int perf_del_probe_events(struct strfilter *filter)
> >>  	}
> >>  
> >>  	ret = probe_file__get_events(kfd, filter, klist);
> >> -	if (ret == 0) {
> >> -		strlist__for_each_entry(ent, klist)
> >> -			pr_info("Removed event: %s\n", ent->s);
> >> +	if (ret < 0)
> >> +		goto out;
> > No, this is ignored by design.
> > Since probe_file__get_events() returns -ENOENT when no event is matched,
> > this should be just ignored, and goto uprobe event matching.
> 
> Thanks for pointing it out. However when strlist__add in probe_file__get_events
> returns a -ENOMEM and we ignore that, though it happens not very likely, we
> would miss some entries. So I add checks here and in probe_file__get_events to
> give a heads-up in advance.

If you are aware of -ENOMEM ( I guess in such case you'll see OOM killer
sooner or later ), please just catch it. I mean 

if (ret == -ENOMEM)
	goto out;

will be good.

Thank you,

> 
> And the same reason is for the checks below for probe_cache__load,
> probe_cache__add_entry and probe_cache__scan_sdt.
> 
> 
> Regards,
> Zhe
> 
> >
> >>  
> >> -		ret = probe_file__del_strlist(kfd, klist);
> >> -		if (ret < 0)
> >> -			goto error;
> >> -	}
> >> +	strlist__for_each_entry(ent, klist)
> >> +		pr_info("Removed event: %s\n", ent->s);
> >> +
> >> +	ret = probe_file__del_strlist(kfd, klist);
> >> +	if (ret < 0)
> >> +		goto error;
> >>  
> >>  	ret2 = probe_file__get_events(ufd, filter, ulist);
> >> -	if (ret2 == 0) {
> >> -		strlist__for_each_entry(ent, ulist)
> >> -			pr_info("Removed event: %s\n", ent->s);
> >> +	if (ret2 < 0)
> >> +		goto out;
> > Ditto.
> >
> > Thank you,
> >
> >>  
> >> -		ret2 = probe_file__del_strlist(ufd, ulist);
> >> -		if (ret2 < 0)
> >> -			goto error;
> >> -	}
> >> +	strlist__for_each_entry(ent, ulist)
> >> +		pr_info("Removed event: %s\n", ent->s);
> >> +
> >> +	ret2 = probe_file__del_strlist(ufd, ulist);
> >> +	if (ret2 < 0)
> >> +		goto error;
> >>  
> >>  	if (ret == -ENOENT && ret2 == -ENOENT)
> >>  		pr_warning("\"%s\" does not hit any event.\n", str);
> >> diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
> >> index cf44c05f89c1..00f086cba88f 100644
> >> --- a/tools/perf/util/probe-file.c
> >> +++ b/tools/perf/util/probe-file.c
> >> @@ -307,10 +307,14 @@ int probe_file__get_events(int fd, struct strfilter *filter,
> >>  		p = strchr(ent->s, ':');
> >>  		if ((p && strfilter__compare(filter, p + 1)) ||
> >>  		    strfilter__compare(filter, ent->s)) {
> >> -			strlist__add(plist, ent->s);
> >> -			ret = 0;
> >> +			ret = strlist__add(plist, ent->s);
> >> +			if (ret < 0) {
> >> +				pr_debug("strlist__add failed (%d)\n", ret);
> >> +				goto out;
> >> +			}
> >>  		}
> >>  	}
> >> +out:
> >>  	strlist__delete(namelist);
> >>  
> >>  	return ret;
> >> @@ -517,7 +521,11 @@ static int probe_cache__load(struct probe_cache *pcache)
> >>  				ret = -EINVAL;
> >>  				goto out;
> >>  			}
> >> -			strlist__add(entry->tevlist, buf);
> >> +			ret = strlist__add(entry->tevlist, buf);
> >> +			if (ret < 0) {
> >> +				pr_debug("strlist__add failed (%d)\n", ret);
> >> +				goto out;
> >> +			}
> >>  		}
> >>  	}
> >>  out:
> >> @@ -678,7 +686,12 @@ int probe_cache__add_entry(struct probe_cache *pcache,
> >>  		command = synthesize_probe_trace_command(&tevs[i]);
> >>  		if (!command)
> >>  			goto out_err;
> >> -		strlist__add(entry->tevlist, command);
> >> +		ret = strlist__add(entry->tevlist, command);
> >> +		if (ret < 0) {
> >> +			pr_debug("strlist__add failed (%d)\n", ret);
> >> +			goto out_err;
> >> +		}
> >> +
> >>  		free(command);
> >>  	}
> >>  	list_add_tail(&entry->node, &pcache->entries);
> >> @@ -859,7 +872,10 @@ int probe_cache__scan_sdt(struct probe_cache *pcache, const char *pathname)
> >>  			break;
> >>  		}
> >>  
> >> -		strlist__add(entry->tevlist, buf);
> >> +		ret = strlist__add(entry->tevlist, buf);
> >> +		if (ret < 0)
> >> +			pr_debug("strlist__add failed (%d)\n", ret);
> >> +
> >>  		free(buf);
> >>  		entry = NULL;
> >>  	}
> >> -- 
> >> 2.24.1
> >>
> >
> 


-- 
Masami Hiramatsu <mhiramat@...nel.org>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ