selftests broke upstream

selftests broke upstream

[Ingo Molnar]

Hi,

The x86 selftests build broke upstream:

 triton:~/tip/tools/testing/selftests/x86> make
 Makefile:44: warning: overriding recipe for target 'clean'
 ../lib.mk:51: warning: ignoring old recipe for target 'clean'
 gcc -m32 -o /single_step_syscall_32 -O2 -g -std=gnu99 -pthread -Wall  
 single_step_syscall.c -lrt -ldl -lm
 /usr/bin/ld: cannot open output file /single_step_syscall_32: Permission denied
 collect2: error: ld returned 1 exit status
 Makefile:47: recipe for target '/single_step_syscall_32' failed
 make: *** [/single_step_syscall_32] Error 1

It used to be possible to build only the x86 testcases in that directory.

Another problem is that the pkeys testcases are still very noisy:

gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/protection_keys_64 -O2 -g -std=gnu99 -pthread -Wall  protection_keys.c -lrt -ldl
In file included from protection_keys.c:45:0:
pkey-helpers.h: In function ‘sigsafe_printf’:
pkey-helpers.h:41:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
   write(1, dprint_in_signal_buffer, len);
   ^
protection_keys.c: In function ‘dumpit’:
protection_keys.c:407:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
   write(1, buf, nr_read);
   ^
gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/test_vdso_64 -O2 -g -std=gnu99 -pthread -Wall  test_vdso.c -lrt -ldl
test_vdso.c: In function ‘main’:
test_vdso.c:98:37: warning: ‘node’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   if (!ret_vsys && (cpu_vsys != cpu || node_vsys != node))
                                     ^
test_vdso.c:78:12: note: ‘node’ was declared here
   unsigned node;
            ^

I thought all this got already fixed?

Thanks,

	Ingo

Re: selftests broke upstream

[Shuah Khan]

On 03/01/2017 02:24 AM, Ingo Molnar wrote:
> 
> Hi,
> 
> The x86 selftests build broke upstream:
> 
>  triton:~/tip/tools/testing/selftests/x86> make
>  Makefile:44: warning: overriding recipe for target 'clean'

This is odd. It should work. Are there any changes that came in through
x86 tree?

>  ../lib.mk:51: warning: ignoring old recipe for target 'clean'
>  gcc -m32 -o /single_step_syscall_32 -O2 -g -std=gnu99 -pthread -Wall  
>  single_step_syscall.c -lrt -ldl -lm
>  /usr/bin/ld: cannot open output file /single_step_syscall_32: Permission denied
>  collect2: error: ld returned 1 exit status
>  Makefile:47: recipe for target '/single_step_syscall_32' failed
>  make: *** [/single_step_syscall_32] Error 1
> 
> It used to be possible to build only the x86 testcases in that directory.
> 
> Another problem is that the pkeys testcases are still very noisy:
> 
> gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/protection_keys_64 -O2 -g -std=gnu99 -pthread -Wall  protection_keys.c -lrt -ldl
> In file included from protection_keys.c:45:0:
> pkey-helpers.h: In function ‘sigsafe_printf’:
> pkey-helpers.h:41:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
>    write(1, dprint_in_signal_buffer, len);
>    ^
> protection_keys.c: In function ‘dumpit’:
> protection_keys.c:407:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
>    write(1, buf, nr_read);
>    ^
> gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/test_vdso_64 -O2 -g -std=gnu99 -pthread -Wall  test_vdso.c -lrt -ldl
> test_vdso.c: In function ‘main’:
> test_vdso.c:98:37: warning: ‘node’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>    if (!ret_vsys && (cpu_vsys != cpu || node_vsys != node))
>                                      ^
> test_vdso.c:78:12: note: ‘node’ was declared here
>    unsigned node;
>             ^
> 
> I thought all this got already fixed?

Hmm. I included all the patches Dave Hansen sent me in my pull request.
I also added couple of cleanup patches I did for pkeys.

I will take a look and see what's going on and let you know. I usually
wait until rc1 comes out for the dust to settle before I go in run sanity
checks on selftests. But I will do some test runs today and let you know
what I find.

Thanks for letting me know.

-- Shuah

Re: selftests broke upstream

[Shuah Khan]

On 03/01/2017 07:29 AM, Shuah Khan wrote:
> On 03/01/2017 02:24 AM, Ingo Molnar wrote:
>>
>> Hi,
>>
>> The x86 selftests build broke upstream:
>>
>>  triton:~/tip/tools/testing/selftests/x86> make
>>  Makefile:44: warning: overriding recipe for target 'clean'
> 
> This is odd. It should work. Are there any changes that came in through
> x86 tree?
> 
>>  ../lib.mk:51: warning: ignoring old recipe for target 'clean'
>>  gcc -m32 -o /single_step_syscall_32 -O2 -g -std=gnu99 -pthread -Wall  
>>  single_step_syscall.c -lrt -ldl -lm
>>  /usr/bin/ld: cannot open output file /single_step_syscall_32: Permission denied
>>  collect2: error: ld returned 1 exit status
>>  Makefile:47: recipe for target '/single_step_syscall_32' failed
>>  make: *** [/single_step_syscall_32] Error 1
>>
>> It used to be possible to build only the x86 testcases in that directory.
>>
>> Another problem is that the pkeys testcases are still very noisy:
>>
>> gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/protection_keys_64 -O2 -g -std=gnu99 -pthread -Wall  protection_keys.c -lrt -ldl
>> In file included from protection_keys.c:45:0:
>> pkey-helpers.h: In function ‘sigsafe_printf’:
>> pkey-helpers.h:41:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
>>    write(1, dprint_in_signal_buffer, len);
>>    ^
>> protection_keys.c: In function ‘dumpit’:
>> protection_keys.c:407:3: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
>>    write(1, buf, nr_read);
>>    ^
>> gcc -m64 -o /home/mingo/tip/tools/testing/selftests/x86/test_vdso_64 -O2 -g -std=gnu99 -pthread -Wall  test_vdso.c -lrt -ldl
>> test_vdso.c: In function ‘main’:
>> test_vdso.c:98:37: warning: ‘node’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>>    if (!ret_vsys && (cpu_vsys != cpu || node_vsys != node))
>>                                      ^
>> test_vdso.c:78:12: note: ‘node’ was declared here
>>    unsigned node;
>>             ^
>>
>> I thought all this got already fixed?
> 
> Hmm. I included all the patches Dave Hansen sent me in my pull request.
> I also added couple of cleanup patches I did for pkeys.
> 
> I will take a look and see what's going on and let you know. I usually
> wait until rc1 comes out for the dust to settle before I go in run sanity
> checks on selftests. But I will do some test runs today and let you know
> what I find.
> 
> Thanks for letting me know.
> 
> -- Shuah
> 

Sent in the fix. You are cc'ed on it. Thanks again for reporting it. It ended
to be an easy fix.

thanks,
-- Shuah