Re: [PATCH 6/7] selftests: only compile userfaultfd for x86 and powperpc

From: Bamvor Zhang Jian
Date: Tue Sep 08 2015 - 08:28:59 EST


Hi, Michael

On 09/08/2015 05:54 PM, Michael Ellerman wrote:
> On Tue, 2015-09-08 at 17:15 +0800, Bamvor Zhang Jian wrote:
>> Hi, Michael
>>
>> I thought I reply to you, but ...
>>
>> On 08/31/2015 11:26 AM, Michael Ellerman wrote:
>>> On Fri, 2015-08-14 at 21:43 +0800, Bamvor Jian Zhang wrote:
>>>> Signed-off-by: Bamvor Jian Zhang <bamvor.zhangjian@xxxxxxxxxx>
>>>> ---
>>>> tools/testing/selftests/vm/Makefile | 12 ++++++++++++
>>>> 1 file changed, 12 insertions(+)
>>>>
>>>> diff --git a/tools/testing/selftests/vm/Makefile b/tools/testing/selftests/vm/Makefile
>>>> index bb888c6..4dd6e4f 100644
>>>> --- a/tools/testing/selftests/vm/Makefile
>>>> +++ b/tools/testing/selftests/vm/Makefile
>>>> @@ -1,5 +1,15 @@
>>>> # Makefile for vm selftests
>>>>
>>>> +uname_M := $(shell uname -m 2>/dev/null || echo not)
>>>> +ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/ -e s/ppc.*/powerpc/)
>>>> +
>>>> +ifeq ($(ARCH),powerpc)
>>>> +support_userfaultfd = yes
>>>> +endif
>>>> +ifeq ($(ARCH),x86)
>>>> +support_userfaultfd = yes
>>>> +endif
>>>> +
>>>> CFLAGS = -Wall
>>>> BINARIES = compaction_test
>>>> BINARIES += hugepage-mmap
>>>> @@ -9,7 +19,9 @@ BINARIES += mlock2-tests
>>>> BINARIES += on-fault-limit
>>>> BINARIES += thuge-gen
>>>> BINARIES += transhuge-stress
>>>> +ifdef support_userfaultfd
>>>> BINARIES += userfaultfd
>>>> +endif
>>>>
>>>> all: $(BINARIES)
>>>> %: %.c
>>>
>>>
>>> This is nasty. It means when userfaultfd gets implemented for other arches
>>> someone has to remember to update the logic here, which they won't.
>>>
>>> Instead the C program should just do nothing when __NR_userfaultfd is not defined, eg:
>>>
>>> #ifdef __NR_userfaultfd
>>>
>>> int main(int argc, char **argv)
>>> {
>>> ...
>>> }
>>>
>>> #else
>>>
>>> int main(void)
>>> {
>>> printf("skip: Skipping userfaultfd test\n");
>>> return 0;
>>> }
>>> #endif
>>>
>>>
>>> This way when the syscall is implemented for other arches the test will just
>>> start working.
>>>
>>> cheers
>>>
>>>
>> When read the following code, It seems that sometimes __NR_userfaultfd is not
>> defined but the syscall is exist. I am not sure why these piece is needed.
>> cc'd c
>>
>> #ifndef __NR_userfaultfd
>> #ifdef __x86_64__
>> #define __NR_userfaultfd 323
>> #elif defined(__i386__)
>> #define __NR_userfaultfd 374
>> #elif defined(__powewrpc__)
>> #define __NR_userfaultfd 364
>> #else
>> #error "missing __NR_userfaultfd definition"
>> #endif
>> #endif
>>
>> Do you mean that we should remove the above code?
>
> Well yes, it would need to be removed to make the logic I suggested work.
>
> I'm not sure those #defines actually help in practice, because if the syscall
> number is not defined then linux/userfaultfd.h will not exist and the whole
> test will not compile anyway.
>
> I was suggesting something like this, which has the properties of:
> - not breaking the build on arches that don't have the syscall
> - still printing a notice on arches that don't have the syscall, both at build
> time and runtime.
> - building correctly on an arch as soon as that arch implements the syscall,
> with no extra changes required.
Ok, I agree with you. I will send the updated patch later.
> cheers
>
>
> diff --git a/tools/testing/selftests/vm/userfaultfd.c b/tools/testing/selftests/vm/userfaultfd.c
> index 2bf1fc3f562b..652c9d805006 100644
> --- a/tools/testing/selftests/vm/userfaultfd.c
> +++ b/tools/testing/selftests/vm/userfaultfd.c
> @@ -64,19 +64,10 @@
> #include <sys/syscall.h>
> #include <sys/ioctl.h>
> #include <pthread.h>
> -#include <linux/userfaultfd.h>
>
> -#ifndef __NR_userfaultfd
> -#ifdef __x86_64__
> -#define __NR_userfaultfd 323
> -#elif defined(__i386__)
> -#define __NR_userfaultfd 374
> -#elif defined(__powewrpc__)
> -#define __NR_userfaultfd 364
> -#else
> -#error "missing __NR_userfaultfd definition"
> -#endif
> -#endif
> +#ifdef __NR_userfaultfd
> +
> +#include <linux/userfaultfd.h>
>
> static unsigned long nr_cpus, nr_pages, nr_pages_per_cpu, page_size;
>
> @@ -636,3 +627,15 @@ int main(int argc, char **argv)
> nr_pages, nr_pages_per_cpu);
> return userfaultfd_stress();
> }
> +
> +#else /* ! __NR_userfaultfd */
> +
> +#warning "missing __NR_userfaultfd definition"
> +
> +int main(void)
> +{
> + printf("skip: Skipping userfaultfd test (missing __NR_userfaultfd)\n");
> + return 0;
> +}
> +
> +#endif
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/