[LTP] [PATCH v3] syscalls: Add set_mempolicy numa tests.

[Jan Stancek <jstancek@redhat.com>]

----- Original Message -----
> Hi!
> > > +
> > > +
> > > +void tst_numa_alloc_parse(struct tst_nodemap *nodes, const char *path,
> > > +                          unsigned int pages)
> > 
> > I'd change this to a function that only counts allocated pages,
> > and increases counters.
> 
> Fair enough.
> 
> > > +{
> > > +	size_t page_size = getpagesize();
> > > +	char *ptr;
> > > +	int fd = -1;
> > > +	int flags = MAP_PRIVATE|MAP_ANONYMOUS;
> > > +	int node;
> > > +	unsigned int i;
> > > +
> > > +	if (path) {
> > > +		fd = SAFE_OPEN(path, O_CREAT | O_EXCL | O_RDWR, 0666);
> > > +		SAFE_FTRUNCATE(fd, pages * page_size);
> > > +		flags = MAP_SHARED;
> > > +	}
> > > +
> > > +	ptr = SAFE_MMAP(NULL, pages  * page_size,
> > > +	                PROT_READ|PROT_WRITE, flags, fd, 0);
> > > +
> > > +	if (path) {
> > > +		SAFE_CLOSE(fd);
> > > +		SAFE_UNLINK(path);
> > > +	}
> > > +
> > > +	memset(ptr, 'a', pages * page_size);
> > 
> > I think all the alloc stuff shouldn't be in library. It's quite
> > possible tests will need different ways to allocate memory.
> > 
> > Maybe the test will want to mmap, but not touch any pages,
> > or pass different flags to mmap. Allocate with other functions
> > than mmap, etc.
> 
> It should be put into some kind of common place though since it's used
> from several tests even at this point.

set_mempolicy sub-dir?

> 
> What about splitting the tst_numa_alloc_parse() to tst_numa_alloc() and
> tst_numa_parse() and keeping the functionality in the library?

That looks more flexible.

Though I'm not 100% sold on single alloc func in library. My concern is
it will be difficult for it to cover all possible scenarios and API
will keep changing as we find more.

Regards,
Jan

> 
> --
> Cyril Hrubis
> chrubis@suse.cz
>