svn commit: r314075 - head/tests/sys/kern

[Eric Badger]

On 02/22/2017 10:05 PM, Ian Lepore wrote:
> On Wed, 2017-02-22 at 04:35 +0000, Eric Badger wrote:
>> Author: badger
>> Date: Wed Feb 22 04:35:07 2017
>> New Revision: 314075
>> URL: https://svnweb.freebsd.org/changeset/base/314075
>>
>> Log:
>>   Fix world build for archs where __builtin_debugtrap() does not
>> work.
>>
>>   The offending code was introduced in r313992.
>>
>>   Reported by:	rpokala
>>   Approved by:	kib (mentor)
>>
>> Modified:
>>   head/tests/sys/kern/ptrace_test.c
>>
>> Modified: head/tests/sys/kern/ptrace_test.c
>> =====================================================================
>> =========
>> --- head/tests/sys/kern/ptrace_test.c	Wed Feb 22 04:28:10 2017
>> 	(r314074)
>> +++ head/tests/sys/kern/ptrace_test.c	Wed Feb 22 04:35:07 2017
>> 	(r314075)
>> @@ -37,6 +37,7 @@ __FBSDID("$FreeBSD$");
>>  #include <sys/user.h>
>>  #include <sys/wait.h>
>>  #include <errno.h>
>> +#include <machine/cpufunc.h>
>>  #include <pthread.h>
>>  #include <semaphore.h>
>>  #include <signal.h>
>> @@ -1690,7 +1691,7 @@ ATF_TC_BODY(ptrace__PT_KILL_breakpoint,
>>  	ATF_REQUIRE((fpid = fork()) != -1);
>>  	if (fpid == 0) {
>>  		trace_me();
>> -		__builtin_debugtrap();
>> +		breakpoint();
>>  		exit(1);
>>  	}
>>
>>
>
> This fixes only x86 and sparc64.  All other arches have breakpoint()
> under the #ifdef KERNEL wrapper (I have no idea why).  If fixing this
> is going to take any longer, can we disconnect this test from the build
> until it gets worked out?
>
> -- Ian
>

Yes, that was my error. In my haste to fix things, I misread the headers 
and thought I had breakpoint() everywhere (and only tested 
sparc64/amd64).  Sorry about that. It should be fixed in r314118.

Eric