* [PATCH] fsck: check inline_dentry i_addr[0] same as inline_data
@ 2017-11-03 6:29 Yunlei He
2017-11-03 9:25 ` Chao Yu
0 siblings, 1 reply; 2+ messages in thread
From: Yunlei He @ 2017-11-03 6:29 UTC (permalink / raw)
To: jaegeuk, yuchao0, linux-f2fs-devel; +Cc: heyunlei, morgan.wang
<4>[ 30.222442s][pid:1,cpu4,init][<ffffff8008357ffc>] f2fs_evict_inode+0x850/0xa34
<4>[ 30.222442s][pid:1,cpu4,init][<ffffff8008209ec4>] evict+0xa0/0x168
<4>[ 30.222442s][pid:1,cpu4,init][<ffffff800820ad34>] iput+0x188/0x220
<4>[ 30.222473s][pid:1,cpu4,init][<ffffff800836d050>] recover_orphan_inodes+0x2b4/0xa80
<4>[ 30.222473s][pid:1,cpu4,init][<ffffff80083648b0>] f2fs_fill_super+0xcf4/0x16a0
<4>[ 30.222473s][pid:1,cpu4,init][<ffffff80081f2478>] mount_bdev+0x198/0x1c8
<4>[ 30.222473s][pid:1,cpu4,init][<ffffff80083608a8>] f2fs_mount+0x14/0x1c
<4>[ 30.222503s][pid:1,cpu4,init][<ffffff80081f2f38>] mount_fs+0x3c/0x15c
<4>[ 30.222503s][pid:1,cpu4,init][<ffffff800820e9f8>] vfs_kern_mount+0x7c/0x16c
<4>[ 30.222503s][pid:1,cpu4,init][<ffffff8008212548>] do_mount+0x214/0xcf8
<4>[ 30.222503s][pid:1,cpu4,init][<ffffff800821338c>] SyS_mount+0xa8/0x164
<4>[ 30.222503s][pid:1,cpu4,init][<ffffff80080831b0>] el0_svc_naked+0x24/0x28
Mount failed with message as blow:
recover_orphan_inode: orphan failed (ino=1265), run fsck to fix
One orphan directory with inline_dentry flag, but i_addr[0] is not zero.
By the way, sit bitmap of i_addr[0] is also invalidate.
Signed-off-by: Yunlei He <heyunlei@huawei.com>
---
fsck/fsck.c | 18 +++++++++++++-----
1 file changed, 13 insertions(+), 5 deletions(-)
diff --git a/fsck/fsck.c b/fsck/fsck.c
index 56a47be..17cf6ff 100644
--- a/fsck/fsck.c
+++ b/fsck/fsck.c
@@ -601,6 +601,7 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
u32 i_links = le32_to_cpu(node_blk->i.i_links);
u64 i_size = le64_to_cpu(node_blk->i.i_size);
u64 i_blocks = le64_to_cpu(node_blk->i.i_blocks);
+ int ofs = get_extra_isize(node_blk);
unsigned char en[F2FS_NAME_LEN + 1];
int namelen;
unsigned int idx = 0;
@@ -660,9 +661,7 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
ftype == F2FS_FT_FIFO || ftype == F2FS_FT_SOCK)
goto check;
- if((node_blk->i.i_inline & F2FS_INLINE_DATA)) {
- int ofs = get_extra_isize(node_blk);
-
+ if ((node_blk->i.i_inline & F2FS_INLINE_DATA)) {
if (le32_to_cpu(node_blk->i.i_addr[ofs]) != 0) {
/* should fix this bug all the time */
FIX_MSG("inline_data has wrong 0'th block = %x",
@@ -685,8 +684,18 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
DBG(3, "ino[0x%x] has inline data!\n", nid);
goto check;
}
- if((node_blk->i.i_inline & F2FS_INLINE_DENTRY)) {
+
+ if ((node_blk->i.i_inline & F2FS_INLINE_DENTRY)) {
DBG(3, "ino[0x%x] has inline dentry!\n", nid);
+ if (le32_to_cpu(node_blk->i.i_addr[ofs]) != 0) {
+ /* should fix this bug all the time */
+ FIX_MSG("inline_dentry has wrong 0'th block = %x",
+ le32_to_cpu(node_blk->i.i_addr[ofs]));
+ node_blk->i.i_addr[ofs] = 0;
+ node_blk->i.i_blocks = cpu_to_le64(*blk_cnt);
+ need_fix = 1;
+ }
+
ret = fsck_chk_inline_dentries(sbi, node_blk, &child);
if (ret < 0) {
/* should fix this bug all the time */
@@ -715,7 +724,6 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
/* check data blocks in inode */
for (idx = 0; idx < ADDRS_PER_INODE(&node_blk->i);
idx++, child.pgofs++) {
- int ofs = get_extra_isize(node_blk);
block_t blkaddr = le32_to_cpu(node_blk->i.i_addr[ofs + idx]);
/* check extent info */
--
1.9.1
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] fsck: check inline_dentry i_addr[0] same as inline_data
2017-11-03 6:29 [PATCH] fsck: check inline_dentry i_addr[0] same as inline_data Yunlei He
@ 2017-11-03 9:25 ` Chao Yu
0 siblings, 0 replies; 2+ messages in thread
From: Chao Yu @ 2017-11-03 9:25 UTC (permalink / raw)
To: Yunlei He, jaegeuk, linux-f2fs-devel; +Cc: morgan.wang
On 2017/11/3 14:29, Yunlei He wrote:
> <4>[ 30.222442s][pid:1,cpu4,init][<ffffff8008357ffc>] f2fs_evict_inode+0x850/0xa34
> <4>[ 30.222442s][pid:1,cpu4,init][<ffffff8008209ec4>] evict+0xa0/0x168
> <4>[ 30.222442s][pid:1,cpu4,init][<ffffff800820ad34>] iput+0x188/0x220
> <4>[ 30.222473s][pid:1,cpu4,init][<ffffff800836d050>] recover_orphan_inodes+0x2b4/0xa80
> <4>[ 30.222473s][pid:1,cpu4,init][<ffffff80083648b0>] f2fs_fill_super+0xcf4/0x16a0
> <4>[ 30.222473s][pid:1,cpu4,init][<ffffff80081f2478>] mount_bdev+0x198/0x1c8
> <4>[ 30.222473s][pid:1,cpu4,init][<ffffff80083608a8>] f2fs_mount+0x14/0x1c
> <4>[ 30.222503s][pid:1,cpu4,init][<ffffff80081f2f38>] mount_fs+0x3c/0x15c
> <4>[ 30.222503s][pid:1,cpu4,init][<ffffff800820e9f8>] vfs_kern_mount+0x7c/0x16c
> <4>[ 30.222503s][pid:1,cpu4,init][<ffffff8008212548>] do_mount+0x214/0xcf8
> <4>[ 30.222503s][pid:1,cpu4,init][<ffffff800821338c>] SyS_mount+0xa8/0x164
> <4>[ 30.222503s][pid:1,cpu4,init][<ffffff80080831b0>] el0_svc_naked+0x24/0x28
>
> Mount failed with message as blow:
> recover_orphan_inode: orphan failed (ino=1265), run fsck to fix
>
> One orphan directory with inline_dentry flag, but i_addr[0] is not zero.
> By the way, sit bitmap of i_addr[0] is also invalidate.
>
> Signed-off-by: Yunlei He <heyunlei@huawei.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Thanks,
> ---
> fsck/fsck.c | 18 +++++++++++++-----
> 1 file changed, 13 insertions(+), 5 deletions(-)
>
> diff --git a/fsck/fsck.c b/fsck/fsck.c
> index 56a47be..17cf6ff 100644
> --- a/fsck/fsck.c
> +++ b/fsck/fsck.c
> @@ -601,6 +601,7 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
> u32 i_links = le32_to_cpu(node_blk->i.i_links);
> u64 i_size = le64_to_cpu(node_blk->i.i_size);
> u64 i_blocks = le64_to_cpu(node_blk->i.i_blocks);
> + int ofs = get_extra_isize(node_blk);
> unsigned char en[F2FS_NAME_LEN + 1];
> int namelen;
> unsigned int idx = 0;
> @@ -660,9 +661,7 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
> ftype == F2FS_FT_FIFO || ftype == F2FS_FT_SOCK)
> goto check;
>
> - if((node_blk->i.i_inline & F2FS_INLINE_DATA)) {
> - int ofs = get_extra_isize(node_blk);
> -
> + if ((node_blk->i.i_inline & F2FS_INLINE_DATA)) {
> if (le32_to_cpu(node_blk->i.i_addr[ofs]) != 0) {
> /* should fix this bug all the time */
> FIX_MSG("inline_data has wrong 0'th block = %x",
> @@ -685,8 +684,18 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
> DBG(3, "ino[0x%x] has inline data!\n", nid);
> goto check;
> }
> - if((node_blk->i.i_inline & F2FS_INLINE_DENTRY)) {
> +
> + if ((node_blk->i.i_inline & F2FS_INLINE_DENTRY)) {
> DBG(3, "ino[0x%x] has inline dentry!\n", nid);
> + if (le32_to_cpu(node_blk->i.i_addr[ofs]) != 0) {
> + /* should fix this bug all the time */
> + FIX_MSG("inline_dentry has wrong 0'th block = %x",
> + le32_to_cpu(node_blk->i.i_addr[ofs]));
> + node_blk->i.i_addr[ofs] = 0;
> + node_blk->i.i_blocks = cpu_to_le64(*blk_cnt);
> + need_fix = 1;
> + }
> +
> ret = fsck_chk_inline_dentries(sbi, node_blk, &child);
> if (ret < 0) {
> /* should fix this bug all the time */
> @@ -715,7 +724,6 @@ void fsck_chk_inode_blk(struct f2fs_sb_info *sbi, u32 nid,
> /* check data blocks in inode */
> for (idx = 0; idx < ADDRS_PER_INODE(&node_blk->i);
> idx++, child.pgofs++) {
> - int ofs = get_extra_isize(node_blk);
> block_t blkaddr = le32_to_cpu(node_blk->i.i_addr[ofs + idx]);
>
> /* check extent info */
>
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2017-11-03 9:26 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-11-03 6:29 [PATCH] fsck: check inline_dentry i_addr[0] same as inline_data Yunlei He
2017-11-03 9:25 ` Chao Yu
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.