Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Wait iput_async before evict_inodes to prevent race #4854

Closed
wants to merge 2 commits into from
Closed

Wait iput_async before evict_inodes to prevent race #4854

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6fb105e
Wait iput_async before evict_inodes to prevent race
Jul 15, 2016
4df8fe0
Fix NULL pointer in zfs_preumount from 1d9b3bd
Jul 19, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 28 additions & 2 deletions module/zfs/zfs_vfsops.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1369,7 +1369,8 @@ zfs_domount(struct super_block *sb, zfs_mntopts_t *zmo, int silent)
dmu_objset_set_user(zsb->z_os, zsb);
mutex_exit(&zsb->z_os->os_user_ptr_lock);
} else {
error = zfs_sb_setup(zsb, B_TRUE);
if ((error = zfs_sb_setup(zsb, B_TRUE)))
goto out;
}

/* Allocate a root inode for the filesystem. */
Expand All @@ -1395,6 +1396,11 @@ zfs_domount(struct super_block *sb, zfs_mntopts_t *zmo, int silent)
if (error) {
dmu_objset_disown(zsb->z_os, zsb);
zfs_sb_free(zsb);
/*
* make sure we don't have dangling sb->s_fs_info which
* zfs_preumount will use.
*/
sb->s_fs_info = NULL;
}

return (error);
Expand All @@ -1413,8 +1419,28 @@ zfs_preumount(struct super_block *sb)
{
zfs_sb_t *zsb = sb->s_fs_info;

if (zsb)
/* zsb is NULL when zfs_domount fails during mount */
if (zsb) {
zfsctl_destroy(sb->s_fs_info);
/*
* wait for iput_async before entering evict_inodes in
* generic_shutdown_super. The reason we must finish before
* evict_inodes is iput when lazytime is on or when zfs_purgedir
* calls zfs_zget would bump i_count from 0 to 1. This would
* race with the i_count check in evict_inodes, so it could
* destroy the inode while we are still using it.
*
* We wait for two pass, xattr dirs in the first pass may add
* xattr entries in zfs_purgedir, so we wait for second pass
* for them. Also, we don't use taskq_wait here is because it's
* a pool wise taskq, so other live filesystem can constantly do
* iput_async, there's no guarantee it will finish wait.
*/
taskq_wait_outstanding(dsl_pool_iput_taskq(
dmu_objset_pool(zsb->z_os)), 0);
taskq_wait_outstanding(dsl_pool_iput_taskq(
dmu_objset_pool(zsb->z_os)), 0);
}
}
EXPORT_SYMBOL(zfs_preumount);

Expand Down