Revision c4f675cd40d955d539180506c09515c90169b15b authored by Sergei Trofimovich on 20 May 2011, 20:20:30 UTC, committed by Chris Mason on 23 May 2011, 17:24:14 UTC
Observed as a large delay when --mixed filesystem is filled up.
Test example:
1. create tiny --mixed FS:
   $ dd if=/dev/zero of=2G.img seek=$((2048 * 1024 * 1024 - 1)) count=1 bs=1
   $ mkfs.btrfs --mixed 2G.img
   $ mount -oloop 2G.img /mnt/ut/
2. Try to fill it up:
   $ dd if=/dev/urandom of=10M.file bs=10240 count=1024
   $ seq 1 256 | while read file_no; do echo $file_no; time cp 10M.file ${file_no}.copy; done

Up to '200.copy' it goes fast, but when disk fills-up each -ENOSPC
message takes 3 seconds to pop-up _every_ ENOSPC (and in usermode linux
it's even more: 30-60 seconds!). (Maybe, time depends on kernel's timer resolution).

No IO, no CPU load, just rescheduling. Some debugging revealed busy spinning
in shrink_delalloc.

Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
Reviewed-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
1 parent 0f3b708
History
File Mode Size
Kconfig -rw-r--r-- 3.3 KB
Makefile -rw-r--r-- 1.1 KB
README -rw-r--r-- 7.9 KB
bitmap.c -rw-r--r-- 37.0 KB
dir.c -rw-r--r-- 8.8 KB
do_balan.c -rw-r--r-- 57.0 KB
file.c -rw-r--r-- 10.3 KB
fix_node.c -rw-r--r-- 77.2 KB
hashes.c -rw-r--r-- 3.6 KB
ibalance.c -rw-r--r-- 35.2 KB
inode.c -rw-r--r-- 92.0 KB
ioctl.c -rw-r--r-- 5.4 KB
item_ops.c -rw-r--r-- 18.7 KB
journal.c -rw-r--r-- 122.3 KB
lbalance.c -rw-r--r-- 41.4 KB
lock.c -rw-r--r-- 2.6 KB
namei.c -rw-r--r-- 44.1 KB
objectid.c -rw-r--r-- 6.9 KB
prints.c -rw-r--r-- 20.7 KB
procfs.c -rw-r--r-- 14.7 KB
resize.c -rw-r--r-- 6.2 KB
stree.c -rw-r--r-- 64.3 KB
super.c -rw-r--r-- 62.3 KB
tail_conversion.c -rw-r--r-- 9.1 KB
xattr.c -rw-r--r-- 26.6 KB
xattr_acl.c -rw-r--r-- 12.7 KB
xattr_security.c -rw-r--r-- 3.0 KB
xattr_trusted.c -rw-r--r-- 1.4 KB
xattr_user.c -rw-r--r-- 1.3 KB

README

back to top