c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-12-25 02:49:32 +03:00
Code Issues Packages Projects Releases Wiki Activity

DMU: Do not pre-read holes during write

Browse Source 
dmu_tx_check_ioerr() pre-reads blocks that are going to be dirtied
as part of transaction to both prefetch them and check for errors.
But it makes no sense to do it for holes, since there are no disk
reads to prefetch and there can be no errors.  On the other side
those blocks are anonymous, and they are freed immediately by the
dbuf_rele() without even being put into dbuf cache, so we just
burn CPU time on decompression and overheads and get absolutely
no result at the end.

Use of dbuf_hold_impl() with fail_sparse parameter allows to skip
the extra work, and on my tests with sequential 8KB writes to empty
ZVOL with 32KB blocks shows throughput increase from 1.7 to 2GB/s.

Reviewed-by: Brian Atkinson <batkinson@lanl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by:	Alexander Motin <mav@FreeBSD.org>
Sponsored by:	iXsystems, Inc.
Closes #15371
This commit is contained in:
Alexander Motin 2023-10-11 19:37:21 -04:00 committed by Tony Hutter
parent 3f67e012e4
commit e82e68400a
1 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
module/zfs/dmu_tx.c
View File

@ -210,10 +210,12 @@ dmu_tx_check_ioerr(zio_t *zio, dnode_t *dn, int level, uint64_t blkid)
dmu_buf_impl_t *db; dmu_buf_impl_t *db;
rw_enter(&dn->dn_struct_rwlock, RW_READER); rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold_level(dn, level, blkid, FTAG); err = dbuf_hold_impl(dn, level, blkid, TRUE, FALSE, FTAG, &db);
rw_exit(&dn->dn_struct_rwlock); rw_exit(&dn->dn_struct_rwlock);
if (db == NULL) if (err == ENOENT)
return (SET_ERROR(EIO)); return (0);
if (err != 0)
return (err);
/* /*
* PARTIAL_FIRST allows caching for uncacheable blocks. It will * PARTIAL_FIRST allows caching for uncacheable blocks. It will
* be cleared after dmu_buf_will_dirty() call dbuf_read() again. * be cleared after dmu_buf_will_dirty() call dbuf_read() again.