| Message ID | 20170902170626.GC4073@magnolia (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
On Sat, Sep 02, 2017 at 10:06:26AM -0700, Darrick J. Wong wrote: > In commit fd26a88093ba we added a worst case estimate for rmapbt blocks > needed to satisfy the block mapping request. Since then, we added the > ability to reserve enough space in each AG such that we should never run > out of blocks to grow the rmapbt, which makes this calculation > unnecessary. Revert the commit because it makes the extra delalloc > indlen accounting unnecessary and incorrect. Do you remember any details of why we added it and what is supposed to fix it? I have memories of various issues in this area, but I can't remember the details. -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, Sep 03, 2017 at 12:43:58AM -0700, Christoph Hellwig wrote: > On Sat, Sep 02, 2017 at 10:06:26AM -0700, Darrick J. Wong wrote: > > In commit fd26a88093ba we added a worst case estimate for rmapbt blocks > > needed to satisfy the block mapping request. Since then, we added the > > ability to reserve enough space in each AG such that we should never run > > out of blocks to grow the rmapbt, which makes this calculation > > unnecessary. Revert the commit because it makes the extra delalloc > > indlen accounting unnecessary and incorrect. > > Do you remember any details of why we added it and what is supposed > to fix it? I have memories of various issues in this area, but I > can't remember the details. We'd fill the fs up with delalloc reservations until there wasn't any space, and once the fs fragmented badly then we suddenly needed more than just the indlen to satisfy bmbt + rmapbt expansion. This indlen patch was a hack to try to ENOSPC out of write_begin/page_mkwrite before we ran the fs totally out of blocks back when we were still trying to cram the bmbt and rmap updates into a single huge transaction. Deferred ops broke that, and perag reservations made it unnecesary, so now we can rip it out. --D > -- > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, Sep 03, 2017 at 08:40:50AM -0700, Darrick J. Wong wrote: > On Sun, Sep 03, 2017 at 12:43:58AM -0700, Christoph Hellwig wrote: > > On Sat, Sep 02, 2017 at 10:06:26AM -0700, Darrick J. Wong wrote: > > > In commit fd26a88093ba we added a worst case estimate for rmapbt blocks > > > needed to satisfy the block mapping request. Since then, we added the > > > ability to reserve enough space in each AG such that we should never run > > > out of blocks to grow the rmapbt, which makes this calculation > > > unnecessary. Revert the commit because it makes the extra delalloc > > > indlen accounting unnecessary and incorrect. > > > > Do you remember any details of why we added it and what is supposed > > to fix it? I have memories of various issues in this area, but I > > can't remember the details. > > We'd fill the fs up with delalloc reservations until there wasn't any > space, and once the fs fragmented badly then we suddenly needed more > than just the indlen to satisfy bmbt + rmapbt expansion. This indlen > patch was a hack to try to ENOSPC out of write_begin/page_mkwrite before > we ran the fs totally out of blocks back when we were still trying to > cram the bmbt and rmap updates into a single huge transaction. Deferred > ops broke that, and perag reservations made it unnecesary, so now we can > rip it out. > It's not totally clear to me what the original patch would have done, since we technically could return all of the indlen on delay -> real extent conversion before the rmap operation would have even run. Perhaps it could hide a problem by reserving blocks only to release them just before we'd actually need to allocate rmapbt blocks..? Anyways, the patch seems fine to me: Reviewed-by: Brian Foster <bfoster@redhat.com> > --D > > > -- > > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in > > the body of a message to majordomo@vger.kernel.org > > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 459f4b4f0..044a363 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -49,7 +49,6 @@ #include "xfs_rmap.h" #include "xfs_ag_resv.h" #include "xfs_refcount.h" -#include "xfs_rmap_btree.h" #include "xfs_icache.h" @@ -192,12 +191,8 @@ xfs_bmap_worst_indlen( int maxrecs; /* maximum record count at this level */ xfs_mount_t *mp; /* mount structure */ xfs_filblks_t rval; /* return value */ - xfs_filblks_t orig_len; mp = ip->i_mount; - - /* Calculate the worst-case size of the bmbt. */ - orig_len = len; maxrecs = mp->m_bmap_dmxr[0]; for (level = 0, rval = 0; level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK); @@ -205,20 +200,12 @@ xfs_bmap_worst_indlen( len += maxrecs - 1; do_div(len, maxrecs); rval += len; - if (len == 1) { - rval += XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - + if (len == 1) + return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - level - 1; - break; - } if (level == 0) maxrecs = mp->m_bmap_dmxr[1]; } - - /* Calculate the worst-case size of the rmapbt. */ - if (xfs_sb_version_hasrmapbt(&mp->m_sb)) - rval += 1 + xfs_rmapbt_calc_size(mp, orig_len) + - mp->m_rmap_maxlevels; - return rval; }
In commit fd26a88093ba we added a worst case estimate for rmapbt blocks needed to satisfy the block mapping request. Since then, we added the ability to reserve enough space in each AG such that we should never run out of blocks to grow the rmapbt, which makes this calculation unnecessary. Revert the commit because it makes the extra delalloc indlen accounting unnecessary and incorrect. Reported-by: Eryu Guan <eguan@redhat.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- fs/xfs/libxfs/xfs_bmap.c | 17 ++--------------- 1 file changed, 2 insertions(+), 15 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html