Re: [PATCH 03/14] add mach_host tests

[Luca Dariz]

Il 29/12/23 14:44, Samuel Thibault ha scritto:
> Luca Dariz, le jeu. 28 déc. 2023 20:42:50 +0100, a ecrit:
> > ---
> >   tests/test-mach_host.c | 54 ++++++++++++++++++++++++++++++++++++++++++
> >   1 file changed, 54 insertions(+)
> >   create mode 100644 tests/test-mach_host.c
> >
> > diff --git a/tests/test-mach_host.c b/tests/test-mach_host.c
> > new file mode 100644
> > index 00000000..99fc4aca
> > --- /dev/null
> > +++ b/tests/test-mach_host.c
>
> Such files are also non-trivial, and thus deserve a copyright header too.
>
> > @@ -0,0 +1,54 @@
> > +
> > +#include <testlib.h>
> > +
> > +#include <mach_host.user.h>
> > +
> > +void test_kernel_version()
> > +{
> > +  int err;
> > +  kernel_version_t kver;
> > +  err = host_get_kernel_version(mach_host_self(), kver);
> > +  ASSERT_RET(err, "host_kernel_info");
> > +  printf("kernel version: %s\n", kver);
> > +}
> > +
> > +void test_host_info()
> > +{
> > +  int err;
> > +  mach_msg_type_number_t count;
> > +  mach_port_t thishost = mach_host_self();
> > +
> > +  host_basic_info_data_t binfo;
> > +  count = HOST_BASIC_INFO_COUNT;
> > +  err = host_info(thishost, HOST_BASIC_INFO, (host_info_t)&binfo, &count);
> > +  ASSERT_RET(err, "host_basic_info");
> > +  ASSERT(count == HOST_BASIC_INFO_COUNT, "");
>
> Perhaps also make some basic checks, such as binfo.avail_cpus <=
> binfo.max_cpus.

ok

> > +  const int maxcpus = 255;
> > +  int proc_slots[maxcpus];
> > +  count = maxcpus;
> > +  err = host_info(thishost, HOST_PROCESSOR_SLOTS, (host_info_t)&proc_slots, 
> > &count);
> > +  ASSERT_RET(err, "host_processor_slots");
> > +  ASSERT((1 <= count) && (count <= maxcpus), "");
> > +
> > +  host_sched_info_data_t sinfo;
> > +  count = HOST_SCHED_INFO_COUNT;
> > +  err = host_info(thishost, HOST_SCHED_INFO, (host_info_t)&sinfo, &count);
> > +  ASSERT_RET(err, "host_sched_info");
> > +  ASSERT(count == HOST_SCHED_INFO_COUNT, "");
>
> Here you can probably check that the min_* values is smaller than e.g.
> 1000, to catch randomly-filled values.

ok

> Also, it would be useful to compile the tests with
> -ftrivial-auto-var-init=pattern so as to fill the structures with random
> values before making the gnumach calls.

good idea, I didn't know this feature.


Luca