linux-kernel - Re: [PATCH -next] tools build: Check return value of fwrite_unlocked in jvmti

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]

Message-ID: <F549EBCC-2BC0-43D4-B470-A53E9DF37A29@gmail.com>
Date:   Wed, 29 Jul 2020 22:12:46 -0300
From:   Arnaldo Carvalho de Melo <arnaldo.melo@...il.com>
To:     Ian Rogers <irogers@...gle.com>,
        Wang ShaoBo <bobo.shaobowang@...wei.com>
CC:     cj.chengjian@...wei.com, Li Bin <huawei.libin@...wei.com>,
        Jiri Olsa <jolsa@...hat.com>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        linux-perf-users <linux-perf-users@...r.kernel.org>
Subject: Re: [PATCH -next] tools build: Check return value of fwrite_unlocked in jvmti_agent.c



On July 29, 2020 8:47:36 PM GMT-03:00, Ian Rogers <irogers@...gle.com> wrote:
>On Fri, Jul 24, 2020 at 3:07 AM Wang ShaoBo
><bobo.shaobowang@...wei.com> wrote:
>>
>> Function jvmti_write_code called by compiled_method_load_cb may
>return
>> error in using fwrite_unlocked, this failure should be captured and
>> warned.
>>
>> Signed-off-by: Wang ShaoBo <bobo.shaobowang@...wei.com>
>> ---
>>  tools/perf/jvmti/jvmti_agent.c | 23 +++++++++++++++--------
>>  1 file changed, 15 insertions(+), 8 deletions(-)
>>
>> diff --git a/tools/perf/jvmti/jvmti_agent.c
>b/tools/perf/jvmti/jvmti_agent.c
>> index 88108598d6e9..a1fe6aa16b6d 100644
>> --- a/tools/perf/jvmti/jvmti_agent.c
>> +++ b/tools/perf/jvmti/jvmti_agent.c
>> @@ -363,7 +363,7 @@ jvmti_write_code(void *agent, char const *sym,
>>         struct jr_code_load rec;
>>         size_t sym_len;
>>         FILE *fp = agent;
>> -       int ret = -1;
>> +       int ret;
>>
>>         /* don't care about 0 length function, no samples */
>>         if (size == 0)
>> @@ -401,16 +401,23 @@ jvmti_write_code(void *agent, char const *sym,
>>         rec.code_index = code_generation++;
>>
>>         ret = fwrite_unlocked(&rec, sizeof(rec), 1, fp);
>> -       fwrite_unlocked(sym, sym_len, 1, fp);
>> +       if (ret)
>> +               goto error;
>
>Sorry, it seems I lost a reply to this. Won't ret here be the number
>of items written and not an error code? Consequently all writes will
>immediately goto error?

Good thing this is in tmp.perf/core, you're right, the test has to be (ret < 0)... I'll fix, thanks!

Also we need some 'perf test' for this :-/

- Arnaldo
>
>Thanks,
>Ian
>
>> +       ret = fwrite_unlocked(sym, sym_len, 1, fp);
>> +       if (ret)
>> +               goto error;
>>
>> -       if (code)
>> -               fwrite_unlocked(code, size, 1, fp);
>> +       if (code) {
>> +               ret = fwrite_unlocked(code, size, 1, fp);
>> +               if (ret)
>> +                       goto error;
>> +       }
>>
>>         funlockfile(fp);
>> -
>> -       ret = 0;
>> -
>> -       return ret;
>> +       return 0;
>> +error:
>> +       funlockfile(fp);
>> +       return -1;
>>  }
>>
>>  int
>> --
>> 2.17.1
>>

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.