public inbox for ltp@lists.linux.it
 help / color / mirror / Atom feed
From: Xiao Yang <yangx.jy@cn.fujitsu.com>
To: ltp@lists.linux.it
Subject: [LTP] [PATCH] lib/tst_timer_test.c: fix unsigned int overflow on RHEL5.11GA
Date: Thu, 20 Jul 2017 10:21:53 +0800	[thread overview]
Message-ID: <597013C1.8030405@cn.fujitsu.com> (raw)
In-Reply-To: <20170719121557.GG1015@rei.lan>

On 2017/07/19 20:15, Cyril Hrubis wrote:
> Hi!
>> On RHEL5.11GA, pselect01 gets signal SIGSEGV due to unsigned int
>> overflow when all members of the samples array are less than usec
>> in do_timer_test().  all tests which apply timer measurement library
>> could trigger this issue on RHEL5.11GA regularly.
> Ah righ, I've forgotten to add check for the array boundary there as
> well.
>
>> Signed-off-by: Xiao Yang<yangx.jy@cn.fujitsu.com>
>> ---
>>   lib/tst_timer_test.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/lib/tst_timer_test.c b/lib/tst_timer_test.c
>> index f30ad73..178e232 100644
>> --- a/lib/tst_timer_test.c
>> +++ b/lib/tst_timer_test.c
>> @@ -295,9 +295,9 @@ void do_timer_test(long long usec, unsigned int nsamples)
>>   			i, samples[0], samples[i-1]);
>>   	}
>>
>> -	for (i = nsamples - 1; samples[i]<  usec; i--);
>> +	for (i = nsamples - 1; i<  nsamples&&  samples[i]<  usec; i--);
> But this does not seem right, we are decrementing i, hence we should
> check that i>  0, otherwise we overflow anyway. With i<  nsamples we
> will break the cycle after the overflow, which is wrong and works only
> by accident (since i is unsigned we overflow to UINT_MAX if we decrement
> 0).
>
>
> This should be the correct patch to fix the problem:
>
> diff --git a/lib/tst_timer_test.c b/lib/tst_timer_test.c
> index f30ad73dc..0b675f78d 100644
> --- a/lib/tst_timer_test.c
> +++ b/lib/tst_timer_test.c
> @@ -295,7 +295,7 @@ void do_timer_test(long long usec, unsigned int nsamples)
>                          i, samples[0], samples[i-1]);
>          }
>
> -       for (i = nsamples - 1; samples[i]<  usec; i--);
> +       for (i = nsamples - 1; samples[i]<  usec&&  i>  0; i--);
Hi Cyril,

It sounds better, but this change leads that samples[0] could not be 
checked if the whole samples array
is less than usec.

Could we fix this issue as below:
------------------------------------------------------------------------------------------------------------------------
-    for (i = nsamples - 1; samples[i] < usec; i--);
+    for (i = nsamples - 1; samples[i] < usec && i > 0; i--);

      if (i < nsamples - 1) {
-        tst_res(TFAIL, "%s woken up early %u times range: [%lli,%lli]",
-            scall, nsamples - 1 - i,
-            samples[i+1], samples[nsamples-1]);
+        if (i == 0 && samples[i] < usec) {
+            tst_res(TFAIL, "%s woken up early %u times range: [%lli,%lli]",
+                scall, nsamples - i, samples[i],
+                samples[nsamples-1]);
+        } else {
+            tst_res(TFAIL, "%s woken up early %u times range: [%lli,%lli]",
+                scall, nsamples - 1 - i, samples[i+1],
+                samples[nsamples-1]);
+        }
------------------------------------------------------------------------------------------------------------------------
Thanks,
Xiao Yang
>
>          if (i<  nsamples - 1) {
>                  tst_res(TFAIL, "%s woken up early %u times range: [%lli,%lli]",
>




  reply	other threads:[~2017-07-20  2:21 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-07-17  6:53 [LTP] [PATCH] lib/tst_timer_test.c: fix unsigned int overflow on RHEL5.11GA Xiao Yang
2017-07-19 12:15 ` Cyril Hrubis
2017-07-20  2:21   ` Xiao Yang [this message]
2017-07-20  9:13     ` Cyril Hrubis
2017-07-20  9:25       ` Xiao Yang
2017-07-20  9:38       ` [LTP] [PATCH v2] " Xiao Yang
2017-07-20 11:29         ` Cyril Hrubis

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=597013C1.8030405@cn.fujitsu.com \
    --to=yangx.jy@cn.fujitsu.com \
    --cc=ltp@lists.linux.it \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox