From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH 5/5] make test: Also run zbd tests References: <20190813020136.15798-1-bvanassche@acm.org> <20190813020136.15798-6-bvanassche@acm.org> From: Jens Axboe Message-ID: <80f13ecd-34b1-c2c2-5802-da2a8058ab42@kernel.dk> Date: Tue, 13 Aug 2019 14:22:55 -0600 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit To: Sitsofe Wheeler , Bart Van Assche Cc: fio , Damien Le Moal List-ID: On 8/13/19 2:10 PM, Sitsofe Wheeler wrote: > On Tue, 13 Aug 2019 at 16:06, Bart Van Assche wrote: >> >> On 8/12/19 9:39 PM, Sitsofe Wheeler wrote: >>> On Tue, 13 Aug 2019 at 03:02, Bart Van Assche wrote: >>>> test: fio >>>> ./fio --minimal --thread --exitall_on_error --runtime=1s --name=nulltest --ioengine=null --rw=randrw --iodepth=2 --norandommap --random_generator=tausworthe64 --size=16T --name=verifyfstest --filename=fiotestfile.tmp --unlink=1 --rw=write --verify=crc32c --verify_state_save=0 --size=16K >>>> + if modinfo null_blk >/dev/null 2>&1 && \ >>>> + sudo modprobe null_blk && \ >>>> + [ -e /sys/kernel/config/nullb ]; then \ >>>> + git clone https://github.com/hgst/libzbc && \ >>>> + (cd libzbc && \ >>>> + ./autogen.sh && \ >>>> + ./configure --prefix=/usr && \ >>>> + make -j && \ >>>> + sudo make install) && \ >>>> + sudo t/zbd/run-tests-against-regular-nullb && \ >>>> + if modinfo null_blk | grep -q '^parm:[[:blank:]]*zoned:'; then \ >>>> + sudo t/zbd/run-tests-against-zoned-nullb; \ >>>> + fi; \ >>>> + fi >>>> >>>> install: $(PROGS) $(SCRIPTS) tools/plot/fio2gnuplot.1 FORCE >>>> $(INSTALL) -m 755 -d $(DESTDIR)$(bindir) >>>> >>> >>> I don't feel comfortable trying to clone a repo, run sudo, installing >>> stuff as root and loading a kernel module just because someone ran >>> "make test". Maybe we should introduce another target for this? >> >> Hi Sitsofe, >> >> How about executing the new code only on Travis, e.g. by checking >> whether the TRAVIS environment variable has been set? > > That's a bit better but I still think we could introduce a test_travis > target, have travis call that and run any extra commands we want... I agree with that, but let's call it something like fulltest instead. That gives travis/CI the option to run test and fulltest, where a "normal" user would probably just do make test. If they do want fulltest, then it's available to run. -- Jens Axboe