On Wed, Apr 01, 2020 at 12:15:32AM -0400, Qian Cai wrote: > > > > On Apr 1, 2020, at 12:14 AM, Chandan Rajendra <chandan@xxxxxxxxxxxxx> wrote: > > > > On Wednesday, April 1, 2020 3:27 AM Qian Cai wrote: > >> Ever since two days ago, linux-next starts to trigger xfs metadata corruption > >> during compilation workloads on both powerpc and arm64, > > > > Can you please provide the filesystem geometry information? > > You can get that by executing "xfs_info <mount-point>" command. > > Hmm. Do the arm/ppc systems have 64k pages? kconfigs might be a good starting place. Also, does the xfs for-next branch exhibit this problem, or is it just the big -next branch that Stephen Rothwell puts out? --D > == arm64 == > # xfs_info /home/ > meta-data=/dev/mapper/rhel_hpe--apollo--cn99xx--11-home isize=512 agcount=4, agsize=113568256 blks > = sectsz=4096 attr=2, projid32bit=1 > = crc=1 finobt=1, sparse=1, rmapbt=0 > = reflink=1 > data = bsize=4096 blocks=454273024, imaxpct=5 > = sunit=0 swidth=0 blks > naming =version 2 bsize=4096 ascii-ci=0, ftype=1 > log =internal log bsize=4096 blocks=221813, version=2 > = sectsz=4096 sunit=1 blks, lazy-count=1 > realtime =none extsz=4096 blocks=0, rtextents=0 > > > == powerpc == > # xfs_info /home/ > meta-data=/dev/mapper/rhel_ibm--p9wr--01-home isize=512 agcount=4, agsize=118489856 blks > = sectsz=4096 attr=2, projid32bit=1 > = crc=1 finobt=1, sparse=1, rmapbt=0 > = reflink=1 > data = bsize=4096 blocks=473959424, imaxpct=5 > = sunit=0 swidth=0 blks > naming =version 2 bsize=4096 ascii-ci=0, ftype=1 > log =internal log bsize=4096 blocks=231425, version=2 > = sectsz=4096 sunit=1 blks, lazy-count=1 > realtime =none extsz=4096 blocks=0, rtextents=0 > > == x86 (not yet reproduced) == > meta-data=/dev/mapper/rhel_hpe--dl380gen9--01-home isize=512 agcount=16, agsize=3283776 blks > = sectsz=512 attr=2, projid32bit=1 > = crc=1 finobt=1, sparse=1, rmapbt=0 > = reflink=1 > data = bsize=4096 blocks=52540416, imaxpct=25 > = sunit=64 swidth=64 blks > naming =version 2 bsize=4096 ascii-ci=0, ftype=1 > log =internal log bsize=4096 blocks=25664, version=2 > = sectsz=512 sunit=0 blks, lazy-count=1 > realtime =none extsz=4096 blocks=0, rtextents=0
