posix/tst-spawn: Fix racy tests in spawned processes.

Message ID cebdb748-a39f-7c67-8d56-8ffbaab64874@linux.ibm.com
State Committed
Headers

Commit Message

Stefan Liebler Jan. 28, 2019, 2:15 p.m. UTC
  On 01/28/2019 02:41 PM, Stefan Liebler wrote:
> Hi,
> 
> from time to time I get fails in tst-spawn like:
> tst-spawn.c:111: numeric comparison failure
>     left: 0 (0x0); from: xlseek (fd2, 0, SEEK_CUR)
>    right: 28 (0x1c); from: strlen (fd2string)
> error: 1 test failures
> tst-spawn.c:252: numeric comparison failure
>     left: 1 (0x1); from: WEXITSTATUS (status)
>    right: 0 (0x0); from: 0
> error: 1 test failures
> 
> It turned out, that a child process is testing it's open file 
> descriptors with e.g. a sequence of testing the current position, 
> setting the position to zero and reading a specific amount of bytes.
> 
> Unfortunately starting with commit 
> 2a69f853c03034c2e383e0f9c35b5402ce8b5473 the test is spawning a second 
> child process which is sharing some of the file descriptors.  If the 
> test sequence as mentioned above is running in parallel it leads to test 
> failures.
> 
> As the second call of posix_spawn shall test a NULL pid argument, this 
> patch is just moving the waitpid of the first child before the 
> posix_spawn of the second child.
> 
> Okay for commit?
> (I assume 2.30)
> 
> Bye,
> Stefan
> 
> ChangeLog:
> 
>      * posix/tst-spawn do_test(): Move waitpid before posix_spawn.
> 

Sorry. Forgot to attach the patch.
  

Comments

Adhemerval Zanella Netto Jan. 28, 2019, 4:48 p.m. UTC | #1
On 28/01/2019 12:15, Stefan Liebler wrote:
> On 01/28/2019 02:41 PM, Stefan Liebler wrote:
>> Hi,
>>
>> from time to time I get fails in tst-spawn like:
>> tst-spawn.c:111: numeric comparison failure
>>     left: 0 (0x0); from: xlseek (fd2, 0, SEEK_CUR)
>>    right: 28 (0x1c); from: strlen (fd2string)
>> error: 1 test failures
>> tst-spawn.c:252: numeric comparison failure
>>     left: 1 (0x1); from: WEXITSTATUS (status)
>>    right: 0 (0x0); from: 0
>> error: 1 test failures
>>
>> It turned out, that a child process is testing it's open file descriptors with e.g. a sequence of testing the current position, setting the position to zero and reading a specific amount of bytes.
>>
>> Unfortunately starting with commit 2a69f853c03034c2e383e0f9c35b5402ce8b5473 the test is spawning a second child process which is sharing some of the file descriptors.  If the test sequence as mentioned above is running in parallel it leads to test failures.
>>
>> As the second call of posix_spawn shall test a NULL pid argument, this patch is just moving the waitpid of the first child before the posix_spawn of the second child.
>>
>> Okay for commit?
>> (I assume 2.30)
>>
>> Bye,
>> Stefan
>>
>> ChangeLog:
>>
>>      * posix/tst-spawn do_test(): Move waitpid before posix_spawn.
>>
> 
> Sorry. Forgot to attach the patch.

LGTM to 2.30, thanks for catching it.
  

Patch

commit b6b802488e1c04b0de9235580538f295b1264541
Author: Stefan Liebler <stli@linux.ibm.com>
Date:   Mon Jan 28 14:16:55 2019 +0100

    posix/tst-spawn: Fix racy tests in spawned processes.
    
    From time to time I get fails in tst-spawn like:
    tst-spawn.c:111: numeric comparison failure
       left: 0 (0x0); from: xlseek (fd2, 0, SEEK_CUR)
      right: 28 (0x1c); from: strlen (fd2string)
    error: 1 test failures
    tst-spawn.c:252: numeric comparison failure
       left: 1 (0x1); from: WEXITSTATUS (status)
      right: 0 (0x0); from: 0
    error: 1 test failures
    
    It turned out, that a child process is testing it's open file descriptors
    with e.g. a sequence of testing the current position, setting the position
    to zero and reading a specific amount of bytes.
    
    Unfortunately starting with commit 2a69f853c03034c2e383e0f9c35b5402ce8b5473
    the test is spawning a second child process which is sharing some of the
    file descriptors.  If the test sequence as mentioned above is running in parallel
    it leads to test failures.
    
    As the second call of posix_spawn shall test a NULL pid argument,
    this patch is just moving the waitpid of the first child
    before the posix_spawn of the second child.
    
    ChangeLog:
    
            * posix/tst-spawn do_test(): Move waitpid before posix_spawn.

diff --git a/posix/tst-spawn.c b/posix/tst-spawn.c
index eea5addbf3..9aa7e621e6 100644
--- a/posix/tst-spawn.c
+++ b/posix/tst-spawn.c
@@ -237,6 +237,12 @@  do_test (int argc, char *argv[])
   TEST_COMPARE (posix_spawn (&pid, argv[1], &actions, NULL, spargv, environ),
 		0);
 
+  /* Wait for the children.  */
+  TEST_COMPARE (xwaitpid (pid, &status, 0), pid);
+  TEST_VERIFY (WIFEXITED (status));
+  TEST_VERIFY (!WIFSIGNALED (status));
+  TEST_COMPARE (WEXITSTATUS (status), 0);
+
   /* Same test but with a NULL pid argument.  */
   TEST_COMPARE (posix_spawn (NULL, argv[1], &actions, NULL, spargv, environ),
 		0);
@@ -246,11 +252,6 @@  do_test (int argc, char *argv[])
   free (name3_copy);
 
   /* Wait for the children.  */
-  TEST_COMPARE (xwaitpid (pid, &status, 0), pid);
-  TEST_VERIFY (WIFEXITED (status));
-  TEST_VERIFY (!WIFSIGNALED (status));
-  TEST_COMPARE (WEXITSTATUS (status), 0);
-
   xwaitpid (-1, &status, 0);
   TEST_VERIFY (WIFEXITED (status));
   TEST_VERIFY (!WIFSIGNALED (status));