linux-kernel - Re: [PATCH] perf jevents: Fix resource leak in process

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

Date:   Wed, 16 Oct 2019 18:58:36 +0800
From:   Yunfeng Ye <yeyunfeng@...wei.com>
To:     John Garry <john.garry@...wei.com>, <peterz@...radead.org>,
        <mingo@...hat.com>, <acme@...nel.org>, <mark.rutland@....com>,
        <alexander.shishkin@...ux.intel.com>, <jolsa@...hat.com>,
        <namhyung@...nel.org>, <ak@...ux.intel.com>,
        <lukemujica@...gle.com>, <kan.liang@...ux.intel.com>,
        <yuzenghui@...wei.com>
CC:     <linux-kernel@...r.kernel.org>, <hushiyuan@...wei.com>,
        <linfeilong@...wei.com>
Subject: Re: [PATCH] perf jevents: Fix resource leak in process_mapfile()



On 2019/10/16 18:40, John Garry wrote:
> On 16/10/2019 08:47, Yunfeng Ye wrote:
>> There are memory leaks and file descriptor resource leaks in
>> process_mapfile().
>>
>> Fix this by adding free() and fclose() on the error paths.
>>
>> Fixes: 80eeb67fe577 ("perf jevents: Program to convert JSON file")
>> Signed-off-by: Yunfeng Ye <yeyunfeng@...wei.com>
>> ---
>>  tools/perf/pmu-events/jevents.c | 9 +++++++--
>>  1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/tools/perf/pmu-events/jevents.c b/tools/perf/pmu-events/jevents.c
>> index e2837260ca4d..6e60d4cff592 100644
>> --- a/tools/perf/pmu-events/jevents.c
>> +++ b/tools/perf/pmu-events/jevents.c
>> @@ -758,6 +758,7 @@ static int process_mapfile(FILE *outfp, char *fpath)
>>      char *line, *p;
>>      int line_num;
>>      char *tblname;
>> +    int ret = 0;
>>
>>      pr_info("%s: Processing mapfile %s\n", prog, fpath);
>>
>> @@ -769,6 +770,7 @@ static int process_mapfile(FILE *outfp, char *fpath)
>>      if (!mapfp) {
>>          pr_info("%s: Error %s opening %s\n", prog, strerror(errno),
>>                  fpath);
>> +        free(line);
>>          return -1;
>>      }
>>
>> @@ -795,7 +797,8 @@ static int process_mapfile(FILE *outfp, char *fpath)
>>              /* TODO Deal with lines longer than 16K */
>>              pr_info("%s: Mapfile %s: line %d too long, aborting\n",
>>                      prog, fpath, line_num);
>> -            return -1;
>> +            ret = -1;
>> +            goto out;
> 
> There's a subtle change of behaviour here, i.e. now calling print_mapping_table_suffix(), but I don't think that it makes any difference.
> 
yes, I know that "goto out" will run print_mapping_table_suffix(outfp), because the error path before is done like this.
so I think it should be use "goto out" to run run print_mapping_table_suffix(outfp).

> However, does outfp remain open also in this case:
> 
Because it has a comment that "Make build fail", so I am not handle the outfp, only modify the process_mapfile() function.

> main(int argc, char *argv[])
> {
> ...
> 
> if (process_mapfile(eventsfp, mapfile)) {
>     pr_info("%s: Error processing mapfile %s\n", prog, mapfile);
>     /* Make build fail */
>     return 1;
> }
> 
> return 0;
> 
> empty_map:
>     fclose(eventsfp);
>     ...
> }
> 
> I think that this code works on the basis that the program exits on any sort of error and releases resources automatically. Having said that, it is a good practice to tidy up.
> 
I agree with you, when program exits, it will releases resources automatically. It's just to make the program clearer and more correct.

> John
> 
>>          }
>>          line[strlen(line)-1] = '\0';
>>
>> @@ -825,7 +828,9 @@ static int process_mapfile(FILE *outfp, char *fpath)
>>
>>  out:
>>      print_mapping_table_suffix(outfp);
>> -    return 0;
>> +    fclose(mapfp);
>> +    free(line);
>> +    return ret;
>>  }
>>
>>  /*
>>
> 
> 
> 
> .
>