On Wed, Nov 02, 2022 at 05:17:18AM +0000, ruansy.fnst@xxxxxxxxxxx wrote: > > 在 2022/11/2 8:45, Darrick J. Wong 写道: > > On Sun, Oct 30, 2022 at 05:31:43PM +0800, Shiyang Ruan wrote: > > FWIW things have calmed down as of 6.1-rc3 -- if I disable reflink, > > fstests runs without complaint. Now it only seems to be affecting > > reflink=1 filesystems. > > >> And how is your recent test? Still failed with those dmesg warnings? If so, > >> could you zip the test result and send it to me? > > > > https://djwong.org/docs/kernel/daxbad.zip > > Thanks for your info! > > (To Dave) I need your recent test result too. If cases won't fail when > reflink disabled, I'll focusing on solving the warning when reflink enabled. My first run on 6.1-rc3 with reflink disabled was clean. Then I ran a few tests with reflink enabled, and they all failed as expected. Then I ran the no-reflink tests again, and then they all failed, too. Nothing changed between test configs, I didn't even reboot the test machine: $ history |tail -5 500 sudo ./run_check.sh --mkfs-opts "-m rmapbt=1,reflink=0" --run-opts "-s xfs_dax xfs/55[12]" 501 sudo ./run_check.sh --mkfs-opts "-m rmapbt=1,reflink=0" --run-opts "-s xfs_dax_noreflink generic/051 generic/068 generic/074 generic/075 generic/083 generic/112 generic/231 generic/232 generic/269 generic/270 generic/340 generic/388 generic/461 generic/471 generic/476 generic/519 generic/560 generic/561 generic/617 generic/650 generic/656 xfs/011 xfs/013 xfs/017 xfs/073 xfs/297 xfs/305 xfs/517 xfs/538" 502 sudo ./run_check.sh --mkfs-opts "-m rmapbt=1" --run-opts "-s xfs_dax generic/051 generic/068 generic/074 generic/075 generic/083 generic/112 generic/231 generic/232 generic/269 generic/270 generic/340 generic/388 generic/461 generic/471 generic/476 generic/519 generic/560 generic/561 generic/617 generic/650 generic/656 xfs/011 xfs/013 xfs/017 xfs/073 xfs/297 xfs/305 xfs/517 xfs/538" 503 sudo ./run_check.sh --mkfs-opts "-m rmapbt=1,reflink=0" --run-opts "-s xfs_dax_noreflink generic/051 generic/068 generic/074 generic/075 generic/083 generic/112 generic/231 generic/232 generic/269 generic/270 generic/340 generic/388 generic/461 generic/471 generic/476 generic/519 generic/560 generic/561 generic/617 generic/650 generic/656 xfs/011 xfs/013 xfs/017 xfs/073 xfs/297 xfs/305 xfs/517 xfs/538" 504 history |tail -5 $ The first noreflink run: SECTION -- xfs_dax_noreflink ========================= Failures: generic/471 generic/519 xfs/073 Failed 3 of 29 tests Which are typical failures for this config. The first reflink enabled run I killed almost immediately as it threw multiple warnings in the first couple of tests: Running: MOUNT_OPTIONS= ./check -R xunit -b -s xfs_dax generic/051 generic/068 generic/074 generic/075 generic/083 generic/112 generic/231 generic/232 generic/269 generic/270 generic/340 generic/388 generic/461 generic/471 generic/476 generic/519 generic/560 generic/561 generic/617 generic/650 generic/656 xfs/011 xfs/013 xfs/017 xfs/073 xfs/297 xfs/305 xfs/517 xfs/538 SECTION -- xfs_dax FSTYP -- xfs (debug) PLATFORM -- Linux/x86_64 test3 6.1.0-rc3-dgc+ #1649 SMP PREEMPT_DYNAMIC Wed Nov 2 07:58:17 AEDT 2022 MKFS_OPTIONS -- -f -m rmapbt=1 /dev/pmem1 MOUNT_OPTIONS -- -o dax=always -o context=system_u:object_r:root_t:s0 /dev/pmem1 /mnt/scratch generic/051 79s ... _check_dmesg: something found in dmesg (see /home/dave/src/xfstests-dev/results//xfs_dax/generic/051.dmesg) generic/068 43s ... ^C real 1m46.278s user 0m17.465s sys 2m9.981s $ And then I ran the noreflink tests again to make sure the first run wasn't a fluke: SECTION -- xfs_dax_noreflink ========================= Failures: generic/051 generic/068 generic/231 generic/269 generic/270 generic/340 generic/388 generic/461 generic/471 generic/476 generic/519 generic/560 generic/561 xfs/011 xfs/013 xfs/073 xfs/297 xfs/305 xfs/517 xfs/538 Failed 20 of 29 tests It was a fluke - most of the tests failed this time with dax mapping warnings. dmesg from the entire set of test runs is attached. -Dave. -- Dave Chinner david@xxxxxxxxxxxxx
Attachment:
dmesg.log.gz
Description: application/gzip
-- dm-devel mailing list dm-devel@xxxxxxxxxx https://listman.redhat.com/mailman/listinfo/dm-devel