[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH v3 06/16] Add a test suite for cgroups functionality



On Wed, Apr 10, 2013 at 09:04:33PM -0600, Eric Blake wrote:
> On 04/10/2013 04:08 AM, Daniel P. Berrange wrote:
> > From: "Daniel P. Berrange" <berrange redhat com>
> > 
> > Some aspects of the cgroups setup / detection code are quite subtle
> > and easy to break. It would greatly benefit from unit testing, but
> > this is difficult because the test suite won't have privileges to
> > play around with cgroups. The solution is to use monkey patching
> > via LD_PRELOAD to override the fopen, open, mkdir, access functions
> > to redirect access of cgroups files to some magic stubs in the
> > test suite.
> > 
> > Using this we provide custom content for the /proc/cgroup and
> > /proc/self/mounts files which report a fixed cgroup setup. We
> > then override open/mkdir/access so that access to the cgroups
> > filesystem gets redirected into files in a temporary directory
> > tree in the test suite build dir.
> 
> Do you also need to override openat/mkdirat/faccessat, in case we (or
> even libc on our behalf) ever uses the newer *at syscalls?

Well, if glibc did try todo such magic, there'd probably be more
that needed overriding besides hte *at() functions. So we should
just wait until that day, rather than second guessing it I think.

> Wow, this looks complicated, so I'll have to defer my review to sometime
> earlier in my day when I'm thinking straight.  But the premise is
> useful, and a passing 'make check' even on a system with no cgroups
> mounted is a pretty good indication of whether you got it right.
> 
> > +/*
> > + * The plan:
> > + *
> > + * We fake out /proc/mounts, so make it look as is cgroups
> > + * are mounted on /not/really/sys/fs/cgroup. We don't
> > + * use /sys/fs/cgroup, because we want to make it easy to
> > + * detect places where we've not mocked enough syscalls.
> 
> and so that the testsuite will run and pass even on systems without
> cgroups mounted.

Correct.

> > +mymain(void)
> > +{
> > +    int ret = 0;
> > +    char *fakesysfsdir;
> > +
> > +    if (!(fakesysfsdir = strdup(FAKESYSFSDIRTEMPLATE))) {
> > +        fprintf(stderr, "Out of memory\n");
> > +        abort();
> > +    }
> > +
> > +    if (!mkdtemp(fakesysfsdir)) {
> 
> Does this compile on mingw, or do you need to modify bootstrap.conf to
> pull in the mkdtemp gnulib module?  [Then again, it won't compile on
> mingw in the first place, since the Makefile.am limits it to platforms
> with LD_PRELOAD support]

Hmm, I've not tested with win32 build. I think we actually ought to
make sure this is run on Linux only, since it doesn't make sense on
BSD either.


Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]