On Thu, Dec 09, 2021 at 04:17:08PM +0000, Bruce Richardson wrote:
> On Thu, Dec 09, 2021 at 08:15:01AM -0500, Aaron Conole wrote:
> > Jerin Jacob <jerinjac...@gmail.com> writes:
> >
> > > On Thu, Dec 9, 2021 at 12:30 AM Jie Zhou <j...@linux.microsoft.com> wrote:
> > >>
> > >> Even though test_interrupts.c can compile on Windows, skip interrupt
> > >> tests for now since majority of eal_interrupt on Windows are stubs.
> > >> Will remove the skip after interrupt being fully enabled on Windows.
> > >>
> > >> Signed-off-by: Jie Zhou <j...@linux.microsoft.com>
> > >> Acked-by: Dmitry Kozlyuk <dmitry.kozl...@gmail.com>
> > >>
> > >> ---
> > >> app/test/test_interrupts.c | 10 ++++++++++
> > >> 1 file changed, 10 insertions(+)
> > >>
> > >> diff --git a/app/test/test_interrupts.c b/app/test/test_interrupts.c
> > >> index 2a05399f96..eec9b2805b 100644
> > >> --- a/app/test/test_interrupts.c
> > >> +++ b/app/test/test_interrupts.c
> > >> @@ -12,6 +12,15 @@
> > >>
> > >> #include "test.h"
> > >>
> > >> +#ifdef RTE_EXEC_ENV_WINDOWS
> > >
> > > Across the series,
> > > Instead of adding conditional compilation everywhere, Why not disable
> > > specific file
> > > for compilation for windows?
> > > Purpose of EAL to abstract the differences in execution environment
> > > and application
> > > should not know that.
> >
> > I think this was done because there would be two test lists in the meson
> > unit test file. But this is the second comment about these ifdef's, and
> > maybe we should revisit that discussion. Is there a different way to
> > accomplish not running the tests which are not appropriate for windows
> > builds, while not having two overlapping lists of unit tests in the
> > meson build file?
> >
> I'm wondering if a reasonable compromise solution might be to have the
> build system expose a usable RTE_EXEC_ENV symbol that can be used in C-code
> if statements rather than just in ifdefs. That would allow us to easily add
> e.g.
>
> if (RTE_EXEC_ENV == rte_env_linux)
> return TEST_SKIPPED;
>
> into each test function needing it. Two lines of C-code is a lot easier to
> add and manage than #ifdefs covering the whole file, or alternative lists
> in meson.
>
Quick patch to allow C-code comparisons:
diff --git a/lib/eal/meson.build b/lib/eal/meson.build
index 1722924f67..b5b9fa14b4 100644
--- a/lib/eal/meson.build
+++ b/lib/eal/meson.build
@@ -10,6 +10,12 @@ if not is_windows
subdir('unix')
endif
+exec_envs = {'freebsd': 0, 'linux': 1, 'windows': 2}
+foreach env, id:exec_envs
+ dpdk_conf.set('RTE_ENV_' + env.to_upper(), id)
+endforeach
+dpdk_conf.set('RTE_EXEC_ENV', exec_envs[exec_env])
+
dpdk_conf.set('RTE_EXEC_ENV_' + exec_env.to_upper(), 1)
subdir(exec_env)
A slightly simpler patch would just expose the environment as a string as
e.g. "linux", but I think numeric ids just make the code better rather than
having string comparisons. Alternatively, this could also be done via
C-code with ifdefs in EAL, but as it stands this meson change allows:
if (RTE_EXEC_ENV == RTE_ENV_WINDOWS)
...
or:
switch (RTE_EXEC_ENV) {
case RTE_ENV_LINUX: ... ; break;
case RTE_ENV_FREEBSD: ... ; break;
case RTE_ENV_WINDOWS: ... ; break;
}
Thoughts?
/Bruce
