2022-11-15 16:46:02

by ChenXiaoSong

Subject: [PATCH v4 3/3] btrfs: qgroup: fix sleep from invalid context bug in update_qgroup_limit_item()

Syzkaller reported BUG as follows:

BUG: sleeping function called from invalid context at
include/linux/sched/mm.h:274
Call Trace:
<TASK>
dump_stack_lvl+0xcd/0x134
__might_resched.cold+0x222/0x26b
kmem_cache_alloc+0x2e7/0x3c0
update_qgroup_limit_item+0xe1/0x390
btrfs_qgroup_inherit+0x147b/0x1ee0
create_subvol+0x4eb/0x1710
btrfs_mksubvol+0xfe5/0x13f0
__btrfs_ioctl_snap_create+0x2b0/0x430
btrfs_ioctl_snap_create_v2+0x25a/0x520
btrfs_ioctl+0x2a1c/0x5ce0
__x64_sys_ioctl+0x193/0x200
do_syscall_64+0x35/0x80

Fix this by delaying the limit item updates until unlock the spin lock.

Signed-off-by: ChenXiaoSong <[email protected]>
---
fs/btrfs/qgroup.c | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index ca609a70d067..f84507ca3b99 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -2867,6 +2867,8 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
bool need_rescan = false;
u32 level_size = 0;
u64 nums;
+ bool update_limit = false;
+ int err;

/*
* There are only two callers of this function.
@@ -2957,10 +2959,7 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
dstgroup->max_excl = inherit->lim.max_excl;
dstgroup->rsv_rfer = inherit->lim.rsv_rfer;
dstgroup->rsv_excl = inherit->lim.rsv_excl;
-
- ret = btrfs_update_quoto_limit(trans, dstgroup, fs_info);
- if (ret)
- goto unlock;
+ update_limit = true;
}

if (srcid) {
@@ -2987,6 +2986,7 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
dstgroup->max_excl = srcgroup->max_excl;
dstgroup->rsv_rfer = srcgroup->rsv_rfer;
dstgroup->rsv_excl = srcgroup->rsv_excl;
+ update_limit = false;

qgroup_dirty(fs_info, dstgroup);
qgroup_dirty(fs_info, srcgroup);
@@ -3055,6 +3055,11 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,

unlock:
spin_unlock(&fs_info->qgroup_lock);
+ if (update_limit) {
+ err = btrfs_update_quoto_limit(trans, dstgroup, fs_info);
+ if (err)
+ ret = err;
+ }
if (!ret)
ret = btrfs_sysfs_add_one_qgroup(fs_info, dstgroup);
out:
--
2.31.1

2022-11-15 16:46:20

by ChenXiaoSong

Subject: Re: [PATCH v4 3/3] btrfs: qgroup: fix sleep from invalid context bug in update_qgroup_limit_item()

在 2022/11/16 1:17, ChenXiaoSong 写道:
> @@ -2987,6 +2986,7 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
> dstgroup->max_excl = srcgroup->max_excl;
> dstgroup->rsv_rfer = srcgroup->rsv_rfer;
> dstgroup->rsv_excl = srcgroup->rsv_excl;
> + update_limit = false;
>
> qgroup_dirty(fs_info, dstgroup);
> qgroup_dirty(fs_info, srcgroup);

No need to call update_qgroup_limit_item() when condition "if (srcid)"
is true according to Qu Wenruo's suggestions, btrfs_run_qgroups() will
update the quota tree to reflect the result after calling qgroup_dirty().

2022-11-15 23:13:01

by Qu Wenruo

Subject: Re: [PATCH v4 3/3] btrfs: qgroup: fix sleep from invalid context bug in update_qgroup_limit_item()

On 2022/11/16 01:17, ChenXiaoSong wrote:
> Syzkaller reported BUG as follows:
>
> BUG: sleeping function called from invalid context at
> include/linux/sched/mm.h:274
> Call Trace:
> <TASK>
> dump_stack_lvl+0xcd/0x134
> __might_resched.cold+0x222/0x26b
> kmem_cache_alloc+0x2e7/0x3c0
> update_qgroup_limit_item+0xe1/0x390
> btrfs_qgroup_inherit+0x147b/0x1ee0
> create_subvol+0x4eb/0x1710
> btrfs_mksubvol+0xfe5/0x13f0
> __btrfs_ioctl_snap_create+0x2b0/0x430
> btrfs_ioctl_snap_create_v2+0x25a/0x520
> btrfs_ioctl+0x2a1c/0x5ce0
> __x64_sys_ioctl+0x193/0x200
> do_syscall_64+0x35/0x80
>
> Fix this by delaying the limit item updates until unlock the spin lock.
>
> Signed-off-by: ChenXiaoSong <[email protected]>
> ---
> fs/btrfs/qgroup.c | 13 +++++++++----
> 1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
> index ca609a70d067..f84507ca3b99 100644
> --- a/fs/btrfs/qgroup.c
> +++ b/fs/btrfs/qgroup.c
> @@ -2867,6 +2867,8 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
> bool need_rescan = false;
> u32 level_size = 0;
> u64 nums;
> + bool update_limit = false;
> + int err;
>
> /*
> * There are only two callers of this function.
> @@ -2957,10 +2959,7 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
> dstgroup->max_excl = inherit->lim.max_excl;
> dstgroup->rsv_rfer = inherit->lim.rsv_rfer;
> dstgroup->rsv_excl = inherit->lim.rsv_excl;
> -
> - ret = btrfs_update_quoto_limit(trans, dstgroup, fs_info);
> - if (ret)
> - goto unlock;
> + update_limit = true;

Nope, just call qgroup_dirty() for @dstgroup.

Thanks,
Qu
> }
>
> if (srcid) {
> @@ -2987,6 +2986,7 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
> dstgroup->max_excl = srcgroup->max_excl;
> dstgroup->rsv_rfer = srcgroup->rsv_rfer;
> dstgroup->rsv_excl = srcgroup->rsv_excl;
> + update_limit = false;
>
> qgroup_dirty(fs_info, dstgroup);
> qgroup_dirty(fs_info, srcgroup);
> @@ -3055,6 +3055,11 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
>
> unlock:
> spin_unlock(&fs_info->qgroup_lock);
> + if (update_limit) {
> + err = btrfs_update_quoto_limit(trans, dstgroup, fs_info);
> + if (err)
> + ret = err;
> + }
> if (!ret)
> ret = btrfs_sysfs_add_one_qgroup(fs_info, dstgroup);
> out: