Re: [PATCH] stdlib: Fixing test-*atexit*-race tests on ia64

[Adhemerval Zanella <adhemerval dot zanella at linaro dot org>]

On 01/02/2018 11:04, Dmitry V. Levin wrote:
> On Thu, Feb 01, 2018 at 10:41:05AM -0200, Adhemerval Zanella wrote:
>> The tests requires a new thread stack size to a value (0x20000) lower than
>> architecture minimum (0x3000).  Set PTHREAD_STACK_MIN in this case.
> 
> These tests require a new thread stack size set to a value (0x20000)
> lower than the architecture minimum (0x30000).  Set the stack size
> to PTHREAD_STACK_MIN in this case.
> 
>> Checked on ia64-linux-gnu.
>>
>> 	* stdlib/test-atexit-race-common.c (do_test): Check stack size
>> 	against PTHREAD_STACK_MIN.
>> ---
>>  ChangeLog                        | 5 +++++
>>  stdlib/test-atexit-race-common.c | 5 ++++-
>>  2 files changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/stdlib/test-atexit-race-common.c b/stdlib/test-atexit-race-common.c
>> index 4d7f911..02d459c 100644
>> --- a/stdlib/test-atexit-race-common.c
>> +++ b/stdlib/test-atexit-race-common.c
>> @@ -34,9 +34,11 @@
>>  #include <stdio.h>
>>  #include <stdlib.h>
>>  #include <support/xthread.h>
>> +#include <sys/param.h>
> 
> #include <limits.h>
> 
>>  const size_t kNumThreads = 1024;
>>  const size_t kNumHandlers = 1024;
>> +const size_t kStacksize = 0x20000;
> 
> const size_t kStacksize =
> 	0x20000 < PTHREAD_STACK_MIN ? PTHREAD_STACK_MIN : 0x20000;

Alright, I just set the logic to be near the comment, I do not have a strong
preference here.

> 
>>  static void *
>>  threadfunc (void *unused)
>> @@ -60,7 +62,8 @@ do_test (void)
>>    /* With default 8MiB Linux stack size, creating 1024 threads can cause
>>       VM exhausiton on 32-bit machines.  Reduce stack size of each thread to
>>       128KiB for a maximum required VM size of 128MiB.  */
>> -  xpthread_attr_setstacksize (&attr, 128 * 1024);
>> +  size_t ss = kStacksize < PTHREAD_STACK_MIN ? PTHREAD_STACK_MIN : kStacksize;
>> +  xpthread_attr_setstacksize (&attr, ss);
>>  
>>    for (i = 0; i < kNumThreads; ++i) {
>>      xpthread_create (&attr, threadfunc, NULL);
> 
> ... and this hunk won't be needed at all.
> 
> 

I will commit with your suggestions.

Attachment: signature.asc
Description: OpenPGP digital signature