Re: [PATCH nft 3/4] tests: add meta time test cases

[Phil Sutter <phil@nwl.cc>]

On Tue, Nov 12, 2019 at 10:19:57PM +0100, Phil Sutter wrote:
> On Tue, Nov 12, 2019 at 08:35:57PM +0100, Florian Westphal wrote:
> > Phil Sutter <phil@nwl.cc> wrote:
> > > Hi,
> > > 
> > > On Thu, Aug 29, 2019 at 04:09:03PM +0200, Florian Westphal wrote:
> > > [...]
> > > > diff --git a/tests/py/any/meta.t.payload b/tests/py/any/meta.t.payload
> > > > index 1d8426de9632..402caae5cad8 100644
> > > > --- a/tests/py/any/meta.t.payload
> > > > +++ b/tests/py/any/meta.t.payload
> > > [...]
> > > > +# meta hour "17:00" drop
> > > > +ip test-ip4 input
> > > > +  [ meta load hour => reg 1 ]
> > > > +  [ cmp eq reg 1 0x0000d2f0 ]
> > > > +  [ immediate reg 0 drop ]
> > > 
> > > Does this pass for you? I'm getting such warnings:
> > > 
> > > | 7: WARNING: line 3: 'add rule ip test-ip4 input meta hour "17:00" drop':
> > > | '[ cmp eq reg 1 0x0000d2f0 ]' mismatches '[ cmp eq reg 1 0x0000e100 ]'
> > > 
> > > On my system, "17:00" consistently translates into 0xe100.
> > 
> > Argh, DST :-(
> > 
> > We will need to add change the test so nft-test.py runs with a fixed
> > time zone.
> 
> Isn't this thing translating to UTC before submitting to kernel? I would
> assume netlink debug output to be consistent between different
> timezones.

Timezones and DST obviously exceed my abilities. It just took a while
for me to realize that "17:00" in DST means a different point in time
than the same value in non-DST. Anyway, this seems to work:

| diff --git a/tests/py/nft-test.py b/tests/py/nft-test.py
| index ce42b5ddb1cca..1b1db7aa596f3 100755
| --- a/tests/py/nft-test.py
| +++ b/tests/py/nft-test.py
| @@ -24,6 +24,7 @@ import tempfile
|  
|  TESTS_PATH = os.path.dirname(os.path.abspath(__file__))
|  sys.path.insert(0, os.path.join(TESTS_PATH, '../../py/'))
| +os.environ['TZ'] = 'UTC-2'
|  
|  from nftables import Nftables
|  

Choosing UTC-2 fixes most of the warnings. Not all of them, who knows
why. (Choosing the right offset was already challenging enough.)

Cheers, Phil