remove warn when bdi is unregisterd in __mark_inode_dirty().

Commit Message

[e-mail] Feb. 29, 2016, 7:03 a.m. UTC

From: zhangaihua <[e-mail]>

When a disk is ejected, we try to do the operation 'ls', it will fails
because device is gone.

Currently we do a WARN() which does not seem to be the right thing.
Convert it to a ratelimited kernel warning.

[ 221.582369] bdi-block not registered
[ 221.625209] ------------[ cut here ]------------
[ 221.680512] WARNING: at fs/fs-writeback.c:2065
[ 221.733632] Modules linked in:
[ 221.770088]
[ 221.787798] CPU: 10 PID: 1454 Comm: ls Not tainted 4.4.0+ #6
[ 221.855604] Hardware name: Hisilicon PhosphorHi1382 EVB (DT)
[ 221.923309] task: ffffffc0d4c74800 ti: ffffffc0d47bc000 task.ti:
ffffffc0d47bc000
[ 222.012896] PC is at __mark_inode_dirty+0x2b0/0x3d0
[ 222.071327] LR is at __mark_inode_dirty+0x2b0/0x3d0
[ 222.129657] pc : [<ffffffc000278490>] lr : [<ffffffc000278490>]
pstate: 80000145
[ 222.218191] sp : ffffffc0d47bfd80
[ 222.257772] x29: ffffffc0d47bfd80 x28: ffffffc0d47bc000
[ 222.321310] x27: ffffffc000ccd000 x26: 000000000000003d
[ 222.384847] x25: 000000000000011d x24: ffffffc00e5d0140
[ 222.448385] x23: 0000000000000000 x22: 0000000000000000
[ 222.511923] x21: ffffffc00e5d00b8 x20: ffffffc00dfc81e0
[ 222.575459] x19: 0000000000000001 x18: 0000000000000000
[ 222.638997] x17: 00000000004b59f0 x16: ffffffc00025a640
[ 222.702535] x15: ffffffffffffffff x14: 0ffffffffffffffe
[ 222.766073] x13: 0000000000000028 x12: 0101010101010101
[ 222.829611] x11: ffffffc0d47bfac0 x10: ffffffc0d47bfac0
[ 222.893149] x9 : ffffffd000000000 x8 : 6b206e6f74207265
[ 222.956685] x7 : 00000000000001f8 x6 : ffffffc0014d7377
[ 223.020223] x5 : 0000000000000005 x4 : 0000000000000000
[ 223.083761] x3 : ffffffc0014d6dda x2 : 0000000000000018
[ 223.147299] x1 : 0000000000000001 x0 : 0000000000000018
[ 223.210836]
[ 223.228543] ---[ end trace 17b671a783f074ad ]---
[ 223.283749] Call trace:
[ 223.312915] [<ffffffc000278490>] __mark_inode_dirty+0x2b0/0x3d0
[ 223.383745] [<ffffffc000262570>] generic_update_time+0x90/0xd0
[ 223.453635] [<ffffffc000264220>] touch_atime+0xa0/0xc0
[ 223.515092] [<ffffffc00025a1b0>] iterate_dir+0x100/0x140
[ 223.578731] [<ffffffc00025a6cc>] SyS_getdents64+0x8c/0x120
[ 223.644355] [<ffffffc000084cb0>] el0_svc_naked+0x24/0x28

Signed-off-by: [e-mail]
---
 fs/fs-writeback.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

Comments

Dave Chinner March 1, 2016, 9:34 p.m. UTC | #1

On Mon, Feb 29, 2016 at 03:03:54PM +0800, [e-mail] wrote:
> From: zhangaihua <[e-mail]>
> 
> When a disk is ejected, we try to do the operation 'ls', it will fails
> because device is gone.
> 
> Currently we do a WARN() which does not seem to be the right thing.
> Convert it to a ratelimited kernel warning.

I don't think this is a good idea - the stack trace is required to
determine what code has failed to detect that the device has gone
and is trying to continue onwards without handling errors correctly.

IMO, iterate_dir() should be fixed to do the right thing when
errors have occurred....

Cheers,

Dave.

Patch

diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 6915c95..2cb5bca 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -2058,9 +2058,10 @@  void __mark_inode_dirty(struct inode *inode, int flags)
 
 			wb = locked_inode_to_wb_and_lock_list(inode);
 
-			WARN(bdi_cap_writeback_dirty(wb->bdi) &&
-			     !test_bit(WB_registered, &wb->state),
-			     "bdi-%s not registered\n", wb->bdi->name);
+			if (bdi_cap_writeback_dirty(wb->bdi) &&
+				!test_bit(WB_registered, &wb->state))
+				pr_warn_ratelimited("bdi-%s not registered\n",
+							wb->bdi->name);
 
 			inode->dirtied_when = jiffies;
 			if (dirtytime)