[PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu Wenruo]

We have at least two user reports about bad inode generation makes
kernel reject the fs.

According to the creation time, the inode is created by some 2014
kernel.
And the generation member of INODE_ITEM is not updated (unlike the
transid member) so the error persists until latest tree-checker detects.

Even the situation can be fixed by reverting back to older kernel and
copying the offending dir/file to another inode and delete the offending
one, it still should be done by btrfs-progs.

This patchset adds such check and repair ability to btrfs-check, with a
simple test image.

Qu Wenruo (3):
  btrfs-progs: check/lowmem: Add check and repair for invalid inode
    generation
  btrfs-progs: check/original: Add check and repair for invalid inode
    generation
  btrfs-progs: fsck-tests: Add test image for invalid inode generation
    repair

 check/main.c                                  |  50 +++++++++++-
 check/mode-lowmem.c                           |  76 ++++++++++++++++++
 check/mode-original.h                         |   1 +
 .../.lowmem_repairable                        |   0
 .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
 5 files changed, 126 insertions(+), 1 deletion(-)
 create mode 100644 tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
 create mode 100644 tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz

-- 
2.23.0

[PATCH 1/3] btrfs-progs: check/lowmem: Add check and repair for invalid inode generation

[Qu Wenruo]

There are at least two bug reports of kernel tree-checker complaining
about invalid inode generation.

All offending inodes seem to be caused by old kernel around 2014, with
inode generation overflow.

So add such check and repair ability to lowmem mode check first.

This involves:
- Calculate the inode generation upper limit
  If it's an inode from log tree, then the upper limit is
  super_generation + 1, otherwise it's super_generation.

- Check if the inode generation is larger than the upper limit

- Repair by resetting inode generation to current transaction
  generation

Reported-by: Charles Wright <charles.v.wright@gmail.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 check/mode-lowmem.c | 76 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 76 insertions(+)

diff --git a/check/mode-lowmem.c b/check/mode-lowmem.c
index 5f7f101d..7af29ba9 100644
--- a/check/mode-lowmem.c
+++ b/check/mode-lowmem.c
@@ -2472,6 +2472,59 @@ static bool has_orphan_item(struct btrfs_root *root, u64 ino)
 	return false;
 }
 
+static int repair_inode_gen_lowmem(struct btrfs_root *root,
+				   struct btrfs_path *path)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_inode_item *ii;
+	struct btrfs_key key;
+	u64 transid;
+	int ret;
+
+	trans = btrfs_start_transaction(root, 1);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		errno = -ret;
+		error("failed to start transaction for inode gen repair: %m");
+		return ret;
+	}
+	transid = trans->transid;
+	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+	ASSERT(key.type == BTRFS_INODE_ITEM_KEY);
+
+	btrfs_release_path(path);
+
+	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+	if (ret > 0) {
+		ret = -ENOENT;
+		error("no inode item found for ino %llu", key.objectid);
+		goto error;
+	}
+	if (ret < 0) {
+		errno = -ret;
+		error("failed to find inode item for ino %llu: %m",
+		      key.objectid);
+		goto error;
+	}
+	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			    struct btrfs_inode_item);
+	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	ret = btrfs_commit_transaction(trans, root);
+	if (ret < 0) {
+		errno = -ret;
+		error("failed to commit transaction: %m");
+		goto error;
+	}
+	printf("reseting inode generation to %llu for ino %llu\n",
+		transid, key.objectid);
+	return ret;
+
+error:
+	btrfs_abort_transaction(trans, ret);
+	return ret;
+}
+
 /*
  * Check INODE_ITEM and related ITEMs (the same inode number)
  * 1. check link count
@@ -2487,6 +2540,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
 	struct btrfs_inode_item *ii;
 	struct btrfs_key key;
 	struct btrfs_key last_key;
+	struct btrfs_super_block *super = root->fs_info->super_copy;
 	u64 inode_id;
 	u32 mode;
 	u64 flags;
@@ -2497,6 +2551,8 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
 	u64 refs = 0;
 	u64 extent_end = 0;
 	u64 extent_size = 0;
+	u64 generation;
+	u64 gen_uplimit;
 	unsigned int dir;
 	unsigned int nodatasum;
 	bool is_orphan = false;
@@ -2527,6 +2583,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
 	flags = btrfs_inode_flags(node, ii);
 	dir = imode_to_type(mode) == BTRFS_FT_DIR;
 	nlink = btrfs_inode_nlink(node, ii);
+	generation = btrfs_inode_generation(node, ii);
 	nodatasum = btrfs_inode_flags(node, ii) & BTRFS_INODE_NODATASUM;
 
 	if (!is_valid_imode(mode)) {
@@ -2540,6 +2597,25 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
 		}
 	}
 
+	if (btrfs_super_log_root(super) != 0 &&
+	    root->objectid == BTRFS_TREE_LOG_OBJECTID)
+		gen_uplimit = btrfs_super_generation(super) + 1;
+	else
+		gen_uplimit = btrfs_super_generation(super);
+
+	if (generation > gen_uplimit) {
+		error(
+	"invalid inode generation for ino %llu, have %llu expect [0, %llu)",
+		      inode_id, generation, gen_uplimit);
+		if (repair) {
+			ret = repair_inode_gen_lowmem(root, path);
+			if (ret < 0)
+				err |= INVALID_GENERATION;
+		} else {
+			err |= INVALID_GENERATION;
+		}
+
+	}
 	if (S_ISLNK(mode) &&
 	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND)) {
 		err |= INODE_FLAGS_ERROR;
-- 
2.23.0

[PATCH 2/3] btrfs-progs: check/original: Add check and repair for invalid inode generation

[Qu Wenruo]

There are at least two bug reports of kernel tree-checker complaining
about invalid inode generation.

All offending inodes seem to be caused by old kernel around 2014, with
inode generation overflow.

So add such check and repair ability to lowmem mode check first.

This involves:
- Calculate the inode generation upper limit
  Unlike the original mode context, we don't have anyway to determine if
  this inode belongs to log tree.
  So we use super_generation + 1 as upper limit, just like what we did
  in kernel tree checker.

- Check if the inode generation is larger than the upper limit

- Repair by resetting inode generation to current transaction
  generation
  The difference is, in original mode, we have a common trans handle for
  all repair and reset path for each repair.

Reported-by: Charles Wright <charles.v.wright@gmail.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 check/main.c          | 50 ++++++++++++++++++++++++++++++++++++++++++-
 check/mode-original.h |  1 +
 2 files changed, 50 insertions(+), 1 deletion(-)

diff --git a/check/main.c b/check/main.c
index c2d0f394..018707c8 100644
--- a/check/main.c
+++ b/check/main.c
@@ -604,6 +604,8 @@ static void print_inode_error(struct btrfs_root *root, struct inode_record *rec)
 	if (errors & I_ERR_INVALID_IMODE)
 		fprintf(stderr, ", invalid inode mode bit 0%o",
 			rec->imode & ~07777);
+	if (errors & I_ERR_INVALID_GEN)
+		fprintf(stderr, ", invalid inode generation");
 	fprintf(stderr, "\n");
 
 	/* Print the holes if needed */
@@ -857,6 +859,7 @@ static int process_inode_item(struct extent_buffer *eb,
 {
 	struct inode_record *rec;
 	struct btrfs_inode_item *item;
+	u64 gen_uplimit;
 	u64 flags;
 
 	rec = active_node->current;
@@ -879,6 +882,13 @@ static int process_inode_item(struct extent_buffer *eb,
 	if (S_ISLNK(rec->imode) &&
 	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND))
 		rec->errors |= I_ERR_ODD_INODE_FLAGS;
+	/*
+	 * We don't have accurate root info to determine the correct
+	 * inode generation uplimit, use super_generation + 1 anyway
+	 */
+	gen_uplimit = btrfs_super_generation(global_info->super_copy) + 1;
+	if (btrfs_inode_generation(eb, item) > gen_uplimit)
+		rec->errors |= I_ERR_INVALID_GEN;
 	maybe_free_inode_rec(&active_node->inode_cache, rec);
 	return 0;
 }
@@ -2774,6 +2784,41 @@ static int repair_imode_original(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
+static int repair_inode_gen_original(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root,
+				     struct btrfs_path *path,
+				     struct inode_record *rec)
+{
+	struct btrfs_inode_item *ii;
+	struct btrfs_key key;
+	int ret;
+
+	key.objectid = rec->ino;
+	key.offset = 0;
+	key.type = BTRFS_INODE_ITEM_KEY;
+
+	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+	if (ret > 0) {
+		ret = -ENOENT;
+		error("no inode item found for ino %llu", rec->ino);
+		return ret;
+	}
+	if (ret < 0) {
+		errno = -ret;
+		error("failed to search inode item for ino %llu: %m", rec->ino);
+		return ret;
+	}
+	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			    struct btrfs_inode_item);
+	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	btrfs_release_path(path);
+	printf("resetting inode generation to %llu for ino %llu\n",
+		trans->transid, rec->ino);
+	rec->errors &= ~I_ERR_INVALID_GEN;
+	return 0;
+}
+
 static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
 {
 	struct btrfs_trans_handle *trans;
@@ -2794,7 +2839,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
 			     I_ERR_FILE_NBYTES_WRONG |
 			     I_ERR_INLINE_RAM_BYTES_WRONG |
 			     I_ERR_MISMATCH_DIR_HASH |
-			     I_ERR_UNALIGNED_EXTENT_REC)))
+			     I_ERR_UNALIGNED_EXTENT_REC |
+			     I_ERR_INVALID_GEN)))
 		return rec->errors;
 
 	/*
@@ -2829,6 +2875,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
 		ret = repair_unaligned_extent_recs(trans, root, &path, rec);
 	if (!ret && rec->errors & I_ERR_INVALID_IMODE)
 		ret = repair_imode_original(trans, root, &path, rec);
+	if (!ret && rec->errors & I_ERR_INVALID_GEN)
+		ret = repair_inode_gen_original(trans, root, &path, rec);
 	btrfs_commit_transaction(trans, root);
 	btrfs_release_path(&path);
 	return ret;
diff --git a/check/mode-original.h b/check/mode-original.h
index 78d6bb30..b075a95c 100644
--- a/check/mode-original.h
+++ b/check/mode-original.h
@@ -185,6 +185,7 @@ struct unaligned_extent_rec_t {
 #define I_ERR_INLINE_RAM_BYTES_WRONG	(1 << 17)
 #define I_ERR_MISMATCH_DIR_HASH		(1 << 18)
 #define I_ERR_INVALID_IMODE		(1 << 19)
+#define I_ERR_INVALID_GEN		(1 << 20)
 
 struct inode_record {
 	struct list_head backrefs;
-- 
2.23.0

[PATCH 3/3] btrfs-progs: fsck-tests: Add test image for invalid inode generation repair

[Qu Wenruo]

The image contains one inode item with invalid generation.
The image can be crafted by "btrfs-corrupt-block -i 257 -f generation".
It should emulate the bad inode generation caused by older kernel around
2014.

The image is repairable for both original and lowmem mode.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 .../043-bad-inode-generation/.lowmem_repairable  |   0
 .../bad_inode_geneartion.img.xz                  | Bin 0 -> 2012 bytes
 2 files changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
 create mode 100644 tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz

diff --git a/tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable b/tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
new file mode 100644
index 00000000..e69de29b
diff --git a/tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz b/tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
new file mode 100644
index 0000000000000000000000000000000000000000..d7302fcbabc653885ae1fac438183d7a9908a2f1
GIT binary patch
literal 2012
zcmV<22P63XH+ooF000E$*0e?f03iV!0000G&sfah3;ze3T>wRyj;C3^v%$$4d1r37
zhA1?^g@qU6X{4JoZYUD&m)J~cEzT$F8&JddXW->df79?XOzszXC^i=3jv5zdJ#gxs
z9sQ15EP30<h-K;{hVQDH?$vha-HC@MKe-p3;nw!aTGSk1iK3$jhkg^mmfi38LdMcl
zI<0I=(4xt|g|iQ|g);e0j!pQZ26;x3+R`;xzoVds&J=hBAw4B8L-P|Ve!$)LX@XLr
zpzF|j3*&CLczbqE9v1ss1PkPi%^0v(Dlm5pYr2bQv{LZKvdNN5!i)OgFjV9&wH^qU
z$RwlSTsQ2*=LfG2wb84ro>7fR>-UpptOL6$V1qyGH7H=dX29pcbb>H?Kn*GkWCd+>
z4gz!Y>ckaFgrgJeuG6M51wqhmn{kbaDVJg3Ld?lgrBdfj3zsD?PK#AyrUSv9Tph@s
zlWPn@3zTdm{~HMt@#SSDZQo9pza_F;Uft;~oZgvZAnors__Ju}&Qp3w_XQ)tsS2T|
zq$FIL*2A$z^pWDZ*1|!M(O+5AX(EZJ$>dBH`%1hzGX?o=*9kma$y@8m;CS@I{;nWI
z*j}3J7}<P7H+Q_(*~Z?Dn4(YdJ7u&w3CIuVU<B=LH^@^ueAZn4p^hrwCEExS1qvnu
z4q4*GjEao!lqtg??)^V&FT@3f935LlH<Zm7um2i^^>iP+f7~@z9j`a%V<Q{xAs~eT
zLANvc3&|D0;`ED|^@hYz5<#z>r;Y*M<h-VEzVXiL7Mh6+H6<T-L&2+)TQIS?YO`=;
z1+t^Sp;BfP9zw9oIlv>{DE&w$@uL;WPHEno3DACkV>>P=M{Ff3#!`V7d;rR>!$!8m
zzr2L&o3Z*+c<OlucJg#(xtoceTmW4|=pQRoTY`fh<wo$+#w#exT&A73=&D}upU1>u
zeHG)3b0=ql#0j_4)Gl@YHB@!`y9GEG1RvAk6KQ^YKcMM<dkd74f3yf=8L+9fzadT9
z^R}0|n8E?#M8L$TjA--=%23~K=Z?jLDgWEB%=D2u@^?NDFx_kfG5$FBzvx@Ex`6E;
z$Bb%+a#%_FLpS!cB15&Hl(BJ1YJsrNI!=5S`G1G8>rE&8HfO=EEN1E(wLnK-l_6VQ
zUEzt^?TcZ$wp0ul`k`yBMI3y4xfj}%wEPc%>_>?=JzO{|(p?yTYi(E~Ri(3d4PLj|
z-q_V+l1HR|29?y>SZ#C;mEzg;X?sBSb`OI<^BrkUUO!z7;*qFF<o2;dkg|U<S6r~4
zT3tckU+9<pgSwBl2%ZTP%Aj*H9c^EFr68d1;A(bh?%cR!GRbY2!(teY{(>JEH3r(k
z!AXLNe~a4|dsq;NVVL)uHr^WPWU_rXL-F}@@>i&G@F%st)hh9}CE+-jOR(h79xHk6
z^dF)hH#&Z+_Zm95ii!A=V@&q*g58Rx(*t2p*X(59-0l|_l8}Tc_=7*sX$P+QbM+cU
zD9khoX3gK0U$`(inhSdoZ$DF>nS7fUp98+}-M;HFDb*iCz-L!tXG+#5T6AIEtKcQv
zhvAhr@`FA8s`s@OM{J||gVAGFT&>`tX@G+Cg=Ei-+0E(NrO;(56yJ<mn87((;{&<$
z!vP>zdmtUKI8O1Kf)!vjac5Mm>@t~X+|<K2k7pmf2h^whU`9F^M@g6Vf`aSTg#7QK
z2(a&Zf%OC!HmVf#sH&uNw_mGH@-yi@660h)RhD+9fo&z{++a?!tQp^6O`9n0do-`c
zi59G~u!yh#9U1zupeXJzuNfvTSa_>wv7i~&1yym3r0)M?og*@PZg_dS41n#{io?=L
zciazGdOvd@Mjk}WI|g@>=qLkIwL0;W+Q`Jz%`|psoEMb<`|X2P4|$@YN5>%uq&X2^
zkyrwTpaXx3X0D2tvu9=WKrQiTKb>Q=rA#!_$5|X8@4P;=6i;P0gxgU}Ui371E42dz
zn>c#uoIz?0HY94QQQ>8fu27(WWks^}FEXr`TwbDak}wd!J&V%r4bv4`30pmS+T9WT
z=CP(|%4;_3elq$P8k7GWSdXE|_%te~?dW~>3V334JW;b6sM=7DecM0c4J54{58)H;
z-uMbZ9#jCD;4(DYlCY|hSzb~klkcDdPB;a34yOjt&_BAjLZAMSiB5n_KF&COtxbes
zEcszYE;#FVX-vDZ#vs(#R~@^8*^}Dym5@sYubU#`M6B)U4M!{M*{y(555${5Rbm4a
z6JvM!4sOI~3}h^vXC`V4yqE^R&0-o@!R}HlW&R;T32p{en5w!+OZCEHJe^Fia)@}X
zFP1X0u%e_f=pquM0wgP!<`;?+KaY6k(S!tV^G%<uo0IZ`#j)Q18sGk;4M^IaGB4N%
z4q&QW$Md|Hjs2CLsf-H?jQ6x^Rq~%ENLB+>Jl`BSY`lb3-&5!g^eQ6<>R-=RsPB-E
z<vAbCA-Wrm(Pc9YM7>13E=;JDLif@t_INz&YKa!;Zw#THD=8~YwVLhT_7j2&I>Eq6
zaQA0+X>(drz2NQzo;VMNU>DoE_n5rZXC*Rc-adER=MD9Y1w21(wlG${buZ6hhdtQK
utt-Hd+V!p60001@#^<1LEIQ)=0k;o;7ytk=B&NBs#Ao{g000001X)^iOycnX

literal 0
HcmV?d00001

-- 
2.23.0

Re: [PATCH 2/3] btrfs-progs: check/original: Add check and repair for invalid inode generation

[Nikolay Borisov]

On 24.09.19 г. 11:11 ч., Qu Wenruo wrote:
> There are at least two bug reports of kernel tree-checker complaining
> about invalid inode generation.
> 
> All offending inodes seem to be caused by old kernel around 2014, with
> inode generation overflow.
> 
> So add such check and repair ability to lowmem mode check first.
> 
> This involves:
> - Calculate the inode generation upper limit
>   Unlike the original mode context, we don't have anyway to determine if

Perhaps you meant "lowmem mode context" since in lowmem you deduce
whether it's a log-tree inode or not.

>   this inode belongs to log tree.
>   So we use super_generation + 1 as upper limit, just like what we did
>   in kernel tree checker.
> 
> - Check if the inode generation is larger than the upper limit
> 
> - Repair by resetting inode generation to current transaction
>   generation
>   The difference is, in original mode, we have a common trans handle for
>   all repair and reset path for each repair.
> 
> Reported-by: Charles Wright <charles.v.wright@gmail.com>
> Signed-off-by: Qu Wenruo <wqu@suse.com>

I tested that code with the image provided and it does work as expected
as well as it's fairly obvious what's happening. So:

Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Tested-by: Nikolay Borisov <nborisov@suse.com>

> ---
>  check/main.c          | 50 ++++++++++++++++++++++++++++++++++++++++++-
>  check/mode-original.h |  1 +
>  2 files changed, 50 insertions(+), 1 deletion(-)
> 
> diff --git a/check/main.c b/check/main.c
> index c2d0f394..018707c8 100644
> --- a/check/main.c
> +++ b/check/main.c
> @@ -604,6 +604,8 @@ static void print_inode_error(struct btrfs_root *root, struct inode_record *rec)
>  	if (errors & I_ERR_INVALID_IMODE)
>  		fprintf(stderr, ", invalid inode mode bit 0%o",
>  			rec->imode & ~07777);
> +	if (errors & I_ERR_INVALID_GEN)
> +		fprintf(stderr, ", invalid inode generation");
>  	fprintf(stderr, "\n");
>  
>  	/* Print the holes if needed */
> @@ -857,6 +859,7 @@ static int process_inode_item(struct extent_buffer *eb,
>  {
>  	struct inode_record *rec;
>  	struct btrfs_inode_item *item;
> +	u64 gen_uplimit;
>  	u64 flags;
>  
>  	rec = active_node->current;
> @@ -879,6 +882,13 @@ static int process_inode_item(struct extent_buffer *eb,
>  	if (S_ISLNK(rec->imode) &&
>  	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND))
>  		rec->errors |= I_ERR_ODD_INODE_FLAGS;
> +	/*
> +	 * We don't have accurate root info to determine the correct
> +	 * inode generation uplimit, use super_generation + 1 anyway
> +	 */
> +	gen_uplimit = btrfs_super_generation(global_info->super_copy) + 1;
> +	if (btrfs_inode_generation(eb, item) > gen_uplimit)
> +		rec->errors |= I_ERR_INVALID_GEN;
>  	maybe_free_inode_rec(&active_node->inode_cache, rec);
>  	return 0;
>  }
> @@ -2774,6 +2784,41 @@ static int repair_imode_original(struct btrfs_trans_handle *trans,
>  	return ret;
>  }
>  
> +static int repair_inode_gen_original(struct btrfs_trans_handle *trans,
> +				     struct btrfs_root *root,
> +				     struct btrfs_path *path,
> +				     struct inode_record *rec)
> +{
> +	struct btrfs_inode_item *ii;
> +	struct btrfs_key key;
> +	int ret;
> +
> +	key.objectid = rec->ino;
> +	key.offset = 0;
> +	key.type = BTRFS_INODE_ITEM_KEY;
> +
> +	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
> +	if (ret > 0) {
> +		ret = -ENOENT;
> +		error("no inode item found for ino %llu", rec->ino);
> +		return ret;
> +	}
> +	if (ret < 0) {
> +		errno = -ret;
> +		error("failed to search inode item for ino %llu: %m", rec->ino);
> +		return ret;
> +	}
> +	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
> +			    struct btrfs_inode_item);
> +	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
> +	btrfs_mark_buffer_dirty(path->nodes[0]);
> +	btrfs_release_path(path);
> +	printf("resetting inode generation to %llu for ino %llu\n",
> +		trans->transid, rec->ino);
> +	rec->errors &= ~I_ERR_INVALID_GEN;
> +	return 0;
> +}
> +
>  static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>  {
>  	struct btrfs_trans_handle *trans;
> @@ -2794,7 +2839,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>  			     I_ERR_FILE_NBYTES_WRONG |
>  			     I_ERR_INLINE_RAM_BYTES_WRONG |
>  			     I_ERR_MISMATCH_DIR_HASH |
> -			     I_ERR_UNALIGNED_EXTENT_REC)))
> +			     I_ERR_UNALIGNED_EXTENT_REC |
> +			     I_ERR_INVALID_GEN)))
>  		return rec->errors;
>  
>  	/*
> @@ -2829,6 +2875,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>  		ret = repair_unaligned_extent_recs(trans, root, &path, rec);
>  	if (!ret && rec->errors & I_ERR_INVALID_IMODE)
>  		ret = repair_imode_original(trans, root, &path, rec);
> +	if (!ret && rec->errors & I_ERR_INVALID_GEN)
> +		ret = repair_inode_gen_original(trans, root, &path, rec);
>  	btrfs_commit_transaction(trans, root);
>  	btrfs_release_path(&path);
>  	return ret;
> diff --git a/check/mode-original.h b/check/mode-original.h
> index 78d6bb30..b075a95c 100644
> --- a/check/mode-original.h
> +++ b/check/mode-original.h
> @@ -185,6 +185,7 @@ struct unaligned_extent_rec_t {
>  #define I_ERR_INLINE_RAM_BYTES_WRONG	(1 << 17)
>  #define I_ERR_MISMATCH_DIR_HASH		(1 << 18)
>  #define I_ERR_INVALID_IMODE		(1 << 19)
> +#define I_ERR_INVALID_GEN		(1 << 20)
>  
>  struct inode_record {
>  	struct list_head backrefs;
> 

Re: [PATCH 2/3] btrfs-progs: check/original: Add check and repair for invalid inode generation

[Qu Wenruo]

On 2019/9/30 下午4:41, Nikolay Borisov wrote:
>
>
> On 24.09.19 г. 11:11 ч., Qu Wenruo wrote:
>> There are at least two bug reports of kernel tree-checker complaining
>> about invalid inode generation.
>>
>> All offending inodes seem to be caused by old kernel around 2014, with
>> inode generation overflow.
>>
>> So add such check and repair ability to lowmem mode check first.
>>
>> This involves:
>> - Calculate the inode generation upper limit
>>   Unlike the original mode context, we don't have anyway to determine if
>
> Perhaps you meant "lowmem mode context" since in lowmem you deduce
> whether it's a log-tree inode or not.

Oh, right.

Would you mind to fix this in commit time, David?

>
>>   this inode belongs to log tree.
>>   So we use super_generation + 1 as upper limit, just like what we did
>>   in kernel tree checker.
>>
>> - Check if the inode generation is larger than the upper limit
>>
>> - Repair by resetting inode generation to current transaction
>>   generation
>>   The difference is, in original mode, we have a common trans handle for
>>   all repair and reset path for each repair.
>>
>> Reported-by: Charles Wright <charles.v.wright@gmail.com>
>> Signed-off-by: Qu Wenruo <wqu@suse.com>
>
> I tested that code with the image provided and it does work as expected
> as well as it's fairly obvious what's happening. So:
>
> Reviewed-by: Nikolay Borisov <nborisov@suse.com>
> Tested-by: Nikolay Borisov <nborisov@suse.com>

Thanks for the review and test.

Thanks,
Qu
>
>> ---
>>  check/main.c          | 50 ++++++++++++++++++++++++++++++++++++++++++-
>>  check/mode-original.h |  1 +
>>  2 files changed, 50 insertions(+), 1 deletion(-)
>>
>> diff --git a/check/main.c b/check/main.c
>> index c2d0f394..018707c8 100644
>> --- a/check/main.c
>> +++ b/check/main.c
>> @@ -604,6 +604,8 @@ static void print_inode_error(struct btrfs_root *root, struct inode_record *rec)
>>  	if (errors & I_ERR_INVALID_IMODE)
>>  		fprintf(stderr, ", invalid inode mode bit 0%o",
>>  			rec->imode & ~07777);
>> +	if (errors & I_ERR_INVALID_GEN)
>> +		fprintf(stderr, ", invalid inode generation");
>>  	fprintf(stderr, "\n");
>>
>>  	/* Print the holes if needed */
>> @@ -857,6 +859,7 @@ static int process_inode_item(struct extent_buffer *eb,
>>  {
>>  	struct inode_record *rec;
>>  	struct btrfs_inode_item *item;
>> +	u64 gen_uplimit;
>>  	u64 flags;
>>
>>  	rec = active_node->current;
>> @@ -879,6 +882,13 @@ static int process_inode_item(struct extent_buffer *eb,
>>  	if (S_ISLNK(rec->imode) &&
>>  	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND))
>>  		rec->errors |= I_ERR_ODD_INODE_FLAGS;
>> +	/*
>> +	 * We don't have accurate root info to determine the correct
>> +	 * inode generation uplimit, use super_generation + 1 anyway
>> +	 */
>> +	gen_uplimit = btrfs_super_generation(global_info->super_copy) + 1;
>> +	if (btrfs_inode_generation(eb, item) > gen_uplimit)
>> +		rec->errors |= I_ERR_INVALID_GEN;
>>  	maybe_free_inode_rec(&active_node->inode_cache, rec);
>>  	return 0;
>>  }
>> @@ -2774,6 +2784,41 @@ static int repair_imode_original(struct btrfs_trans_handle *trans,
>>  	return ret;
>>  }
>>
>> +static int repair_inode_gen_original(struct btrfs_trans_handle *trans,
>> +				     struct btrfs_root *root,
>> +				     struct btrfs_path *path,
>> +				     struct inode_record *rec)
>> +{
>> +	struct btrfs_inode_item *ii;
>> +	struct btrfs_key key;
>> +	int ret;
>> +
>> +	key.objectid = rec->ino;
>> +	key.offset = 0;
>> +	key.type = BTRFS_INODE_ITEM_KEY;
>> +
>> +	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
>> +	if (ret > 0) {
>> +		ret = -ENOENT;
>> +		error("no inode item found for ino %llu", rec->ino);
>> +		return ret;
>> +	}
>> +	if (ret < 0) {
>> +		errno = -ret;
>> +		error("failed to search inode item for ino %llu: %m", rec->ino);
>> +		return ret;
>> +	}
>> +	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
>> +			    struct btrfs_inode_item);
>> +	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
>> +	btrfs_mark_buffer_dirty(path->nodes[0]);
>> +	btrfs_release_path(path);
>> +	printf("resetting inode generation to %llu for ino %llu\n",
>> +		trans->transid, rec->ino);
>> +	rec->errors &= ~I_ERR_INVALID_GEN;
>> +	return 0;
>> +}
>> +
>>  static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>>  {
>>  	struct btrfs_trans_handle *trans;
>> @@ -2794,7 +2839,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>>  			     I_ERR_FILE_NBYTES_WRONG |
>>  			     I_ERR_INLINE_RAM_BYTES_WRONG |
>>  			     I_ERR_MISMATCH_DIR_HASH |
>> -			     I_ERR_UNALIGNED_EXTENT_REC)))
>> +			     I_ERR_UNALIGNED_EXTENT_REC |
>> +			     I_ERR_INVALID_GEN)))
>>  		return rec->errors;
>>
>>  	/*
>> @@ -2829,6 +2875,8 @@ static int try_repair_inode(struct btrfs_root *root, struct inode_record *rec)
>>  		ret = repair_unaligned_extent_recs(trans, root, &path, rec);
>>  	if (!ret && rec->errors & I_ERR_INVALID_IMODE)
>>  		ret = repair_imode_original(trans, root, &path, rec);
>> +	if (!ret && rec->errors & I_ERR_INVALID_GEN)
>> +		ret = repair_inode_gen_original(trans, root, &path, rec);
>>  	btrfs_commit_transaction(trans, root);
>>  	btrfs_release_path(&path);
>>  	return ret;
>> diff --git a/check/mode-original.h b/check/mode-original.h
>> index 78d6bb30..b075a95c 100644
>> --- a/check/mode-original.h
>> +++ b/check/mode-original.h
>> @@ -185,6 +185,7 @@ struct unaligned_extent_rec_t {
>>  #define I_ERR_INLINE_RAM_BYTES_WRONG	(1 << 17)
>>  #define I_ERR_MISMATCH_DIR_HASH		(1 << 18)
>>  #define I_ERR_INVALID_IMODE		(1 << 19)
>> +#define I_ERR_INVALID_GEN		(1 << 20)
>>
>>  struct inode_record {
>>  	struct list_head backrefs;
>>

Re: [PATCH 1/3] btrfs-progs: check/lowmem: Add check and repair for invalid inode generation

[Nikolay Borisov]

On 24.09.19 г. 11:11 ч., Qu Wenruo wrote:
> There are at least two bug reports of kernel tree-checker complaining
> about invalid inode generation.
> 
> All offending inodes seem to be caused by old kernel around 2014, with
> inode generation overflow.
> 
> So add such check and repair ability to lowmem mode check first.
> 
> This involves:
> - Calculate the inode generation upper limit
>   If it's an inode from log tree, then the upper limit is
>   super_generation + 1, otherwise it's super_generation.
> 
> - Check if the inode generation is larger than the upper limit
> 
> - Repair by resetting inode generation to current transaction
>   generation
> 
> Reported-by: Charles Wright <charles.v.wright@gmail.com>
> Signed-off-by: Qu Wenruo <wqu@suse.com>

Tested-by: Nikolay Borisov <nborisov@suse.com>

There is one small nit with the assert once rectified you can add:

Reviewed-by: Nikolay Borisov <nborisov@suse.com>

> ---
>  check/mode-lowmem.c | 76 +++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 76 insertions(+)
> 
> diff --git a/check/mode-lowmem.c b/check/mode-lowmem.c
> index 5f7f101d..7af29ba9 100644
> --- a/check/mode-lowmem.c
> +++ b/check/mode-lowmem.c
> @@ -2472,6 +2472,59 @@ static bool has_orphan_item(struct btrfs_root *root, u64 ino)
>  	return false;
>  }
>  
> +static int repair_inode_gen_lowmem(struct btrfs_root *root,
> +				   struct btrfs_path *path)
> +{
> +	struct btrfs_trans_handle *trans;
> +	struct btrfs_inode_item *ii;
> +	struct btrfs_key key;
> +	u64 transid;
> +	int ret;
> +
> +	trans = btrfs_start_transaction(root, 1);
> +	if (IS_ERR(trans)) {
> +		ret = PTR_ERR(trans);
> +		errno = -ret;
> +		error("failed to start transaction for inode gen repair: %m");
> +		return ret;
> +	}
> +	transid = trans->transid;

> +	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
> +	ASSERT(key.type == BTRFS_INODE_ITEM_KEY);

nit: This function's sole caller, check_inode_item, is guaranteed to be
called with a path pointing to BTRFS_INODE_ITEM_KEY thanks to the logic
in the 'for' loop in process_one_leaf. This renders the assert
redundant. At the very least I think it should be moved to
check_inode_item.

> +
> +	btrfs_release_path(path);
> +
> +	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
> +	if (ret > 0) {
> +		ret = -ENOENT;
> +		error("no inode item found for ino %llu", key.objectid);
> +		goto error;
> +	}
> +	if (ret < 0) {
> +		errno = -ret;
> +		error("failed to find inode item for ino %llu: %m",
> +		      key.objectid);
> +		goto error;
> +	}
> +	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
> +			    struct btrfs_inode_item);
> +	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
> +	btrfs_mark_buffer_dirty(path->nodes[0]);
> +	ret = btrfs_commit_transaction(trans, root);
> +	if (ret < 0) {
> +		errno = -ret;
> +		error("failed to commit transaction: %m");
> +		goto error;
> +	}
> +	printf("reseting inode generation to %llu for ino %llu\n",
> +		transid, key.objectid);
> +	return ret;
> +
> +error:
> +	btrfs_abort_transaction(trans, ret);
> +	return ret;
> +}
> +
>  /*
>   * Check INODE_ITEM and related ITEMs (the same inode number)
>   * 1. check link count
> @@ -2487,6 +2540,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>  	struct btrfs_inode_item *ii;
>  	struct btrfs_key key;
>  	struct btrfs_key last_key;
> +	struct btrfs_super_block *super = root->fs_info->super_copy;
>  	u64 inode_id;
>  	u32 mode;
>  	u64 flags;
> @@ -2497,6 +2551,8 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>  	u64 refs = 0;
>  	u64 extent_end = 0;
>  	u64 extent_size = 0;
> +	u64 generation;
> +	u64 gen_uplimit;
>  	unsigned int dir;
>  	unsigned int nodatasum;
>  	bool is_orphan = false;
> @@ -2527,6 +2583,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>  	flags = btrfs_inode_flags(node, ii);
>  	dir = imode_to_type(mode) == BTRFS_FT_DIR;
>  	nlink = btrfs_inode_nlink(node, ii);
> +	generation = btrfs_inode_generation(node, ii);
>  	nodatasum = btrfs_inode_flags(node, ii) & BTRFS_INODE_NODATASUM;
>  
>  	if (!is_valid_imode(mode)) {
> @@ -2540,6 +2597,25 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>  		}
>  	}
>  
> +	if (btrfs_super_log_root(super) != 0 &&
> +	    root->objectid == BTRFS_TREE_LOG_OBJECTID)
> +		gen_uplimit = btrfs_super_generation(super) + 1;
> +	else
> +		gen_uplimit = btrfs_super_generation(super);
> +
> +	if (generation > gen_uplimit) {
> +		error(
> +	"invalid inode generation for ino %llu, have %llu expect [0, %llu)",
> +		      inode_id, generation, gen_uplimit);
> +		if (repair) {
> +			ret = repair_inode_gen_lowmem(root, path);
> +			if (ret < 0)
> +				err |= INVALID_GENERATION;
> +		} else {
> +			err |= INVALID_GENERATION;
> +		}
> +
> +	}
>  	if (S_ISLNK(mode) &&
>  	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND)) {
>  		err |= INODE_FLAGS_ERROR;
> 

Re: [PATCH 1/3] btrfs-progs: check/lowmem: Add check and repair for invalid inode generation

[Qu Wenruo]

On 2019/9/30 下午7:36, Nikolay Borisov wrote:
>
>
> On 24.09.19 г. 11:11 ч., Qu Wenruo wrote:
>> There are at least two bug reports of kernel tree-checker complaining
>> about invalid inode generation.
>>
>> All offending inodes seem to be caused by old kernel around 2014, with
>> inode generation overflow.
>>
>> So add such check and repair ability to lowmem mode check first.
>>
>> This involves:
>> - Calculate the inode generation upper limit
>>   If it's an inode from log tree, then the upper limit is
>>   super_generation + 1, otherwise it's super_generation.
>>
>> - Check if the inode generation is larger than the upper limit
>>
>> - Repair by resetting inode generation to current transaction
>>   generation
>>
>> Reported-by: Charles Wright <charles.v.wright@gmail.com>
>> Signed-off-by: Qu Wenruo <wqu@suse.com>
>
> Tested-by: Nikolay Borisov <nborisov@suse.com>
>
> There is one small nit with the assert once rectified you can add:
>
> Reviewed-by: Nikolay Borisov <nborisov@suse.com>
>
>> ---
>>  check/mode-lowmem.c | 76 +++++++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 76 insertions(+)
>>
>> diff --git a/check/mode-lowmem.c b/check/mode-lowmem.c
>> index 5f7f101d..7af29ba9 100644
>> --- a/check/mode-lowmem.c
>> +++ b/check/mode-lowmem.c
>> @@ -2472,6 +2472,59 @@ static bool has_orphan_item(struct btrfs_root *root, u64 ino)
>>  	return false;
>>  }
>>
>> +static int repair_inode_gen_lowmem(struct btrfs_root *root,
>> +				   struct btrfs_path *path)
>> +{
>> +	struct btrfs_trans_handle *trans;
>> +	struct btrfs_inode_item *ii;
>> +	struct btrfs_key key;
>> +	u64 transid;
>> +	int ret;
>> +
>> +	trans = btrfs_start_transaction(root, 1);
>> +	if (IS_ERR(trans)) {
>> +		ret = PTR_ERR(trans);
>> +		errno = -ret;
>> +		error("failed to start transaction for inode gen repair: %m");
>> +		return ret;
>> +	}
>> +	transid = trans->transid;
>
>> +	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
>> +	ASSERT(key.type == BTRFS_INODE_ITEM_KEY);
>
> nit: This function's sole caller, check_inode_item, is guaranteed to be
> called with a path pointing to BTRFS_INODE_ITEM_KEY thanks to the logic
> in the 'for' loop in process_one_leaf. This renders the assert
> redundant. At the very least I think it should be moved to
> check_inode_item.

Yes, the ASSERT() doesn't make much sense by itself.

However I still believe it won't be a problem.

It's compiler's job to remove such dead ASSERT(), but for human reader,
I still believe this ASSERT() could still make sense, especially when
the caller or callee can get more and more complex.

Thanks,
Qu

>
>> +
>> +	btrfs_release_path(path);
>> +
>> +	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
>> +	if (ret > 0) {
>> +		ret = -ENOENT;
>> +		error("no inode item found for ino %llu", key.objectid);
>> +		goto error;
>> +	}
>> +	if (ret < 0) {
>> +		errno = -ret;
>> +		error("failed to find inode item for ino %llu: %m",
>> +		      key.objectid);
>> +		goto error;
>> +	}
>> +	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
>> +			    struct btrfs_inode_item);
>> +	btrfs_set_inode_generation(path->nodes[0], ii, trans->transid);
>> +	btrfs_mark_buffer_dirty(path->nodes[0]);
>> +	ret = btrfs_commit_transaction(trans, root);
>> +	if (ret < 0) {
>> +		errno = -ret;
>> +		error("failed to commit transaction: %m");
>> +		goto error;
>> +	}
>> +	printf("reseting inode generation to %llu for ino %llu\n",
>> +		transid, key.objectid);
>> +	return ret;
>> +
>> +error:
>> +	btrfs_abort_transaction(trans, ret);
>> +	return ret;
>> +}
>> +
>>  /*
>>   * Check INODE_ITEM and related ITEMs (the same inode number)
>>   * 1. check link count
>> @@ -2487,6 +2540,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>>  	struct btrfs_inode_item *ii;
>>  	struct btrfs_key key;
>>  	struct btrfs_key last_key;
>> +	struct btrfs_super_block *super = root->fs_info->super_copy;
>>  	u64 inode_id;
>>  	u32 mode;
>>  	u64 flags;
>> @@ -2497,6 +2551,8 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>>  	u64 refs = 0;
>>  	u64 extent_end = 0;
>>  	u64 extent_size = 0;
>> +	u64 generation;
>> +	u64 gen_uplimit;
>>  	unsigned int dir;
>>  	unsigned int nodatasum;
>>  	bool is_orphan = false;
>> @@ -2527,6 +2583,7 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>>  	flags = btrfs_inode_flags(node, ii);
>>  	dir = imode_to_type(mode) == BTRFS_FT_DIR;
>>  	nlink = btrfs_inode_nlink(node, ii);
>> +	generation = btrfs_inode_generation(node, ii);
>>  	nodatasum = btrfs_inode_flags(node, ii) & BTRFS_INODE_NODATASUM;
>>
>>  	if (!is_valid_imode(mode)) {
>> @@ -2540,6 +2597,25 @@ static int check_inode_item(struct btrfs_root *root, struct btrfs_path *path)
>>  		}
>>  	}
>>
>> +	if (btrfs_super_log_root(super) != 0 &&
>> +	    root->objectid == BTRFS_TREE_LOG_OBJECTID)
>> +		gen_uplimit = btrfs_super_generation(super) + 1;
>> +	else
>> +		gen_uplimit = btrfs_super_generation(super);
>> +
>> +	if (generation > gen_uplimit) {
>> +		error(
>> +	"invalid inode generation for ino %llu, have %llu expect [0, %llu)",
>> +		      inode_id, generation, gen_uplimit);
>> +		if (repair) {
>> +			ret = repair_inode_gen_lowmem(root, path);
>> +			if (ret < 0)
>> +				err |= INVALID_GENERATION;
>> +		} else {
>> +			err |= INVALID_GENERATION;
>> +		}
>> +
>> +	}
>>  	if (S_ISLNK(mode) &&
>>  	    flags & (BTRFS_INODE_IMMUTABLE | BTRFS_INODE_APPEND)) {
>>  		err |= INODE_FLAGS_ERROR;
>>

Re: [PATCH 1/3] btrfs-progs: check/lowmem: Add check and repair for invalid inode generation

[Nikolay Borisov]

On 30.09.19 г. 15:24 ч., Qu Wenruo wrote:
> Yes, the ASSERT() doesn't make much sense by itself.
> 
> However I still believe it won't be a problem.

It won't be a problem but it feels wrong to have this assert this deep
into the call chain. IMO It should be put where it can trigger at the
earliest which seems to be in check_inode_item. That function assumes
it's working with an inode item and goes to dereference inode members so
if the type is wrong we'd crash there instead of in repair_inode_gen_lowmem.

> 
> It's compiler's job to remove such dead ASSERT(), but for human reader,
> I still believe this ASSERT() could still make sense, especially when
> the caller or callee can get more and more complex.
> 
> Thanks,

Re: [PATCH 1/3] btrfs-progs: check/lowmem: Add check and repair for invalid inode generation

[Qu Wenruo]

On 2019/9/30 下午9:34, Nikolay Borisov wrote:
>
>
> On 30.09.19 г. 15:24 ч., Qu Wenruo wrote:
>> Yes, the ASSERT() doesn't make much sense by itself.
>>
>> However I still believe it won't be a problem.
>
> It won't be a problem but it feels wrong to have this assert this deep
> into the call chain. IMO It should be put where it can trigger at the
> earliest which seems to be in check_inode_item. That function assumes
> it's working with an inode item and goes to dereference inode members so
> if the type is wrong we'd crash there instead of in repair_inode_gen_lowmem.

This is going to be a preference thing.

To my taste, if possible, I would put ASSERT() in every function with
more than 24 lines, as an alternative to comment, to show the
prerequisite or the assumption.

Thanks,
Qu

>
>>
>> It's compiler's job to remove such dead ASSERT(), but for human reader,
>> I still believe this ASSERT() could still make sense, especially when
>> the caller or callee can get more and more complex.
>>
>> Thanks,

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 24-09-2019 om 10:11 schreef Qu Wenruo:
> We have at least two user reports about bad inode generation makes
> kernel reject the fs.

May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so 
kernel went from 5.0 -> 5.3 (but I was using 4.15 too).

Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@

In initramfs I can try to mount but get something like
btrfs critical corrupt leaf invalid inode generation open_ctree failed

Booting old kernel works just as before, no errors.

> According to the creation time, the inode is created by some 2014
> kernel.

How do I get the creation time?

> And the generation member of INODE_ITEM is not updated (unlike the
> transid member) so the error persists until latest tree-checker detects.
> 
> Even the situation can be fixed by reverting back to older kernel and
> copying the offending dir/file to another inode and delete the offending
> one, it still should be done by btrfs-progs.
> 
How to find the offending dir/file from the command line manually?

> This patchset adds such check and repair ability to btrfs-check, with a
> simple test image.
> 
> Qu Wenruo (3):
>    btrfs-progs: check/lowmem: Add check and repair for invalid inode
>      generation
>    btrfs-progs: check/original: Add check and repair for invalid inode
>      generation
>    btrfs-progs: fsck-tests: Add test image for invalid inode generation
>      repair
> 
>   check/main.c                                  |  50 +++++++++++-
>   check/mode-lowmem.c                           |  76 ++++++++++++++++++
>   check/mode-original.h                         |   1 +
>   .../.lowmem_repairable                        |   0
>   .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>   5 files changed, 126 insertions(+), 1 deletion(-)
>   create mode 100644 tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>   create mode 100644 tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu WenRuo]

On 2019/10/19 上午4:32, Ferry Toth wrote:
> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>> We have at least two user reports about bad inode generation makes
>> kernel reject the fs.
> 
> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
> 
> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
> 
> In initramfs I can try to mount but get something like
> btrfs critical corrupt leaf invalid inode generation open_ctree failed
> 
> Booting old kernel works just as before, no errors.
> 
>> According to the creation time, the inode is created by some 2014
>> kernel.
> 
> How do I get the creation time?

# btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
> 
>> And the generation member of INODE_ITEM is not updated (unlike the
>> transid member) so the error persists until latest tree-checker detects.
>>
>> Even the situation can be fixed by reverting back to older kernel and
>> copying the offending dir/file to another inode and delete the offending
>> one, it still should be done by btrfs-progs.
>>
> How to find the offending dir/file from the command line manually?

# find <mount point> -inum <inode number>

Thanks,
Qu

> 
>> This patchset adds such check and repair ability to btrfs-check, with a
>> simple test image.
>>
>> Qu Wenruo (3):
>>    btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>      generation
>>    btrfs-progs: check/original: Add check and repair for invalid inode
>>      generation
>>    btrfs-progs: fsck-tests: Add test image for invalid inode generation
>>      repair
>>
>>   check/main.c                                  |  50 +++++++++++-
>>   check/mode-lowmem.c                           |  76 ++++++++++++++++++
>>   check/mode-original.h                         |   1 +
>>   .../.lowmem_repairable                        |   0
>>   .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>   5 files changed, 126 insertions(+), 1 deletion(-)
>>   create mode 100644
>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>   create mode 100644
>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Hi,

Op 19-10-2019 om 01:50 schreef Qu WenRuo:
> 
> 
> On 2019/10/19 上午4:32, Ferry Toth wrote:
>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>> We have at least two user reports about bad inode generation makes
>>> kernel reject the fs.
>>
>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>
>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
>>
>> In initramfs I can try to mount but get something like
>> btrfs critical corrupt leaf invalid inode generation open_ctree failed
>>
>> Booting old kernel works just as before, no errors.
>>
>>> According to the creation time, the inode is created by some 2014
>>> kernel.
>>
>> How do I get the creation time?
> 
> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>

I just went back to the office to reboot to 5.3 and check the creation 
times and found they were 2013 - 2014.

>>
>>> And the generation member of INODE_ITEM is not updated (unlike the
>>> transid member) so the error persists until latest tree-checker detects.
>>>
>>> Even the situation can be fixed by reverting back to older kernel and
>>> copying the offending dir/file to another inode and delete the offending
>>> one, it still should be done by btrfs-progs.
>>>
>> How to find the offending dir/file from the command line manually?
> 
> # find <mount point> -inum <inode number>

This works, thanks.

But appears unpractical. After fix 2 files and reboot, I found 4 more, 
then 16, then I gave up.

> Thanks,
> Qu
> 
>>
>>> This patchset adds such check and repair ability to btrfs-check, with a
>>> simple test image.
>>>
>>> Qu Wenruo (3):
>>>     btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>>       generation
>>>     btrfs-progs: check/original: Add check and repair for invalid inode
>>>       generation
>>>     btrfs-progs: fsck-tests: Add test image for invalid inode generation
>>>       repair
>>>
>>>    check/main.c                                  |  50 +++++++++++-
>>>    check/mode-lowmem.c                           |  76 ++++++++++++++++++
>>>    check/mode-original.h                         |   1 +
>>>    .../.lowmem_repairable                        |   0
>>>    .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>>    5 files changed, 126 insertions(+), 1 deletion(-)
>>>    create mode 100644
>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>    create mode 100644
>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>
>>

I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my 
mounted rootfs with linux 5.0 and captured the log (~1800 lines 209 
errors).

I'm not sure if using the v5.0 kernel and/or checking mounted distorts 
the results? Else I'm going to need a live usb with a v5.3 kernel and 
v5.3 btrfs-progs.

If you like I can share the log. Let me know.

This issue can potentially cause a lot of grief. Our company server runs 
Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with 
~100 snapshots. I guess the problematic inodes need to be fixed on each 
snapshot prior to upgrading to 20.04 LTS (which might be on kernel ~5.6)?

Do I understand correctly that this FTB is caused by more strict 
checking of the fs by the kernel, while the tools to fix the detected 
corruptions are not yet released?

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu Wenruo]

[-- Attachment #1.1: Type: text/plain, Size: 4057 bytes --]



On 2019/10/20 上午12:24, Ferry Toth wrote:
> Hi,
> 
> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>
>>
>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>> We have at least two user reports about bad inode generation makes
>>>> kernel reject the fs.
>>>
>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>
>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
>>>
>>> In initramfs I can try to mount but get something like
>>> btrfs critical corrupt leaf invalid inode generation open_ctree failed
>>>
>>> Booting old kernel works just as before, no errors.
>>>
>>>> According to the creation time, the inode is created by some 2014
>>>> kernel.
>>>
>>> How do I get the creation time?
>>
>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
> 
> I just went back to the office to reboot to 5.3 and check the creation
> times and found they were 2013 - 2014.
> 
>>>
>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>> transid member) so the error persists until latest tree-checker
>>>> detects.
>>>>
>>>> Even the situation can be fixed by reverting back to older kernel and
>>>> copying the offending dir/file to another inode and delete the
>>>> offending
>>>> one, it still should be done by btrfs-progs.
>>>>
>>> How to find the offending dir/file from the command line manually?
>>
>> # find <mount point> -inum <inode number>
> 
> This works, thanks.
> 
> But appears unpractical. After fix 2 files and reboot, I found 4 more,
> then 16, then I gave up.
> 
>> Thanks,
>> Qu
>>
>>>
>>>> This patchset adds such check and repair ability to btrfs-check, with a
>>>> simple test image.
>>>>
>>>> Qu Wenruo (3):
>>>>     btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>>>       generation
>>>>     btrfs-progs: check/original: Add check and repair for invalid inode
>>>>       generation
>>>>     btrfs-progs: fsck-tests: Add test image for invalid inode
>>>> generation
>>>>       repair
>>>>
>>>>    check/main.c                                  |  50 +++++++++++-
>>>>    check/mode-lowmem.c                           |  76
>>>> ++++++++++++++++++
>>>>    check/mode-original.h                         |   1 +
>>>>    .../.lowmem_repairable                        |   0
>>>>    .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>>>    5 files changed, 126 insertions(+), 1 deletion(-)
>>>>    create mode 100644
>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>    create mode 100644
>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>
>>>
> 
> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
> errors).

It's really not recommended to run btrfs check, especially repair on the
mounted fs, unless it's RO.

A new transaction from kernel can easily screw up the repaired fs.
> 
> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
> the results? Else I'm going to need a live usb with a v5.3 kernel and
> v5.3 btrfs-progs.
> 
> If you like I can share the log. Let me know.
> 
> This issue can potentially cause a lot of grief. Our company server runs
> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
> ~100 snapshots. I guess the problematic inodes need to be fixed on each
> snapshot prior to upgrading to 20.04 LTS (which might be on kernel ~5.6)?

Yes.

> 
> Do I understand correctly that this FTB is caused by more strict
> checking of the fs by the kernel, while the tools to fix the detected
> corruptions are not yet released?

Yes.

Thanks,
Qu


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu Wenruo]

[-- Attachment #1.1: Type: text/plain, Size: 4461 bytes --]



On 2019/10/20 上午8:26, Qu Wenruo wrote:
> 
> 
> On 2019/10/20 上午12:24, Ferry Toth wrote:
>> Hi,
>>
>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>
>>>
>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>> We have at least two user reports about bad inode generation makes
>>>>> kernel reject the fs.
>>>>
>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>
>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
>>>>
>>>> In initramfs I can try to mount but get something like
>>>> btrfs critical corrupt leaf invalid inode generation open_ctree failed
>>>>
>>>> Booting old kernel works just as before, no errors.
>>>>
>>>>> According to the creation time, the inode is created by some 2014
>>>>> kernel.
>>>>
>>>> How do I get the creation time?
>>>
>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
>>
>> I just went back to the office to reboot to 5.3 and check the creation
>> times and found they were 2013 - 2014.
>>
>>>>
>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>> transid member) so the error persists until latest tree-checker
>>>>> detects.
>>>>>
>>>>> Even the situation can be fixed by reverting back to older kernel and
>>>>> copying the offending dir/file to another inode and delete the
>>>>> offending
>>>>> one, it still should be done by btrfs-progs.
>>>>>
>>>> How to find the offending dir/file from the command line manually?
>>>
>>> # find <mount point> -inum <inode number>
>>
>> This works, thanks.
>>
>> But appears unpractical. After fix 2 files and reboot, I found 4 more,
>> then 16, then I gave up.

Another solution is use "find" to locate the files with creation time
before 2015, and copy them to a new file, then replace the old file with
the new file.

It would be much safer than btrfs check --repair.

Thanks,
Qu


>>
>>> Thanks,
>>> Qu
>>>
>>>>
>>>>> This patchset adds such check and repair ability to btrfs-check, with a
>>>>> simple test image.
>>>>>
>>>>> Qu Wenruo (3):
>>>>>     btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>>>>       generation
>>>>>     btrfs-progs: check/original: Add check and repair for invalid inode
>>>>>       generation
>>>>>     btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>> generation
>>>>>       repair
>>>>>
>>>>>    check/main.c                                  |  50 +++++++++++-
>>>>>    check/mode-lowmem.c                           |  76
>>>>> ++++++++++++++++++
>>>>>    check/mode-original.h                         |   1 +
>>>>>    .../.lowmem_repairable                        |   0
>>>>>    .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>>>>    5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>    create mode 100644
>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>    create mode 100644
>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>
>>>>
>>
>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>> errors).
> 
> It's really not recommended to run btrfs check, especially repair on the
> mounted fs, unless it's RO.
> 
> A new transaction from kernel can easily screw up the repaired fs.
>>
>> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>> v5.3 btrfs-progs.
>>
>> If you like I can share the log. Let me know.
>>
>> This issue can potentially cause a lot of grief. Our company server runs
>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>> ~100 snapshots. I guess the problematic inodes need to be fixed on each
>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel ~5.6)?
> 
> Yes.
> 
>>
>> Do I understand correctly that this FTB is caused by more strict
>> checking of the fs by the kernel, while the tools to fix the detected
>> corruptions are not yet released?
> 
> Yes.
> 
> Thanks,
> Qu
> 


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 20-10-2019 om 02:26 schreef Qu Wenruo:
> 
> 
> On 2019/10/20 上午12:24, Ferry Toth wrote:
>> Hi,
>>
>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>
>>>
>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>> We have at least two user reports about bad inode generation makes
>>>>> kernel reject the fs.
>>>>
>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>
>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
>>>>
>>>> In initramfs I can try to mount but get something like
>>>> btrfs critical corrupt leaf invalid inode generation open_ctree failed
>>>>
>>>> Booting old kernel works just as before, no errors.
>>>>
>>>>> According to the creation time, the inode is created by some 2014
>>>>> kernel.
>>>>
>>>> How do I get the creation time?
>>>
>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
>>
>> I just went back to the office to reboot to 5.3 and check the creation
>> times and found they were 2013 - 2014.
>>
>>>>
>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>> transid member) so the error persists until latest tree-checker
>>>>> detects.
>>>>>
>>>>> Even the situation can be fixed by reverting back to older kernel and
>>>>> copying the offending dir/file to another inode and delete the
>>>>> offending
>>>>> one, it still should be done by btrfs-progs.
>>>>>
>>>> How to find the offending dir/file from the command line manually?
>>>
>>> # find <mount point> -inum <inode number>
>>
>> This works, thanks.
>>
>> But appears unpractical. After fix 2 files and reboot, I found 4 more,
>> then 16, then I gave up.
>>
>>> Thanks,
>>> Qu
>>>
>>>>
>>>>> This patchset adds such check and repair ability to btrfs-check, with a
>>>>> simple test image.
>>>>>
>>>>> Qu Wenruo (3):
>>>>>      btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>>>>        generation
>>>>>      btrfs-progs: check/original: Add check and repair for invalid inode
>>>>>        generation
>>>>>      btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>> generation
>>>>>        repair
>>>>>
>>>>>     check/main.c                                  |  50 +++++++++++-
>>>>>     check/mode-lowmem.c                           |  76
>>>>> ++++++++++++++++++
>>>>>     check/mode-original.h                         |   1 +
>>>>>     .../.lowmem_repairable                        |   0
>>>>>     .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>>>>     5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>     create mode 100644
>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>     create mode 100644
>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>
>>>>
>>
>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>> errors).
> 
> It's really not recommended to run btrfs check, especially repair on the
> mounted fs, unless it's RO.

Yes, I know. As the fs is mounted btrfs refuses to repair it.
So I only forced to skip the mount check, but no repair has been done, 
only the check.

> A new transaction from kernel can easily screw up the repaired fs.
>>
>> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>> v5.3 btrfs-progs.
>>
>> If you like I can share the log. Let me know.
>>
>> This issue can potentially cause a lot of grief. Our company server runs
>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>> ~100 snapshots. I guess the problematic inodes need to be fixed on each
>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel ~5.6)?
> 
> Yes.
> 
>>
>> Do I understand correctly that this FTB is caused by more strict
>> checking of the fs by the kernel, while the tools to fix the detected
>> corruptions are not yet released?
> 
> Yes.
> 
> Thanks,
> Qu
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 20-10-2019 om 02:51 schreef Qu Wenruo:
> 
> 
> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>
>>
>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>> Hi,
>>>
>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>
>>>>
>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>> We have at least two user reports about bad inode generation makes
>>>>>> kernel reject the fs.
>>>>>
>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>
>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and / on /@
>>>>>
>>>>> In initramfs I can try to mount but get something like
>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree failed
>>>>>
>>>>> Booting old kernel works just as before, no errors.
>>>>>
>>>>>> According to the creation time, the inode is created by some 2014
>>>>>> kernel.
>>>>>
>>>>> How do I get the creation time?
>>>>
>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
>>>
>>> I just went back to the office to reboot to 5.3 and check the creation
>>> times and found they were 2013 - 2014.
>>>
>>>>>
>>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>>> transid member) so the error persists until latest tree-checker
>>>>>> detects.
>>>>>>
>>>>>> Even the situation can be fixed by reverting back to older kernel and
>>>>>> copying the offending dir/file to another inode and delete the
>>>>>> offending
>>>>>> one, it still should be done by btrfs-progs.
>>>>>>
>>>>> How to find the offending dir/file from the command line manually?
>>>>
>>>> # find <mount point> -inum <inode number>
>>>
>>> This works, thanks.
>>>
>>> But appears unpractical. After fix 2 files and reboot, I found 4 more,
>>> then 16, then I gave up.
> 
> Another solution is use "find" to locate the files with creation time
> before 2015, and copy them to a new file, then replace the old file with
> the new file.

Hmm. But how do I "find" by creation time (otime)? Do you have a 
suggestion for this?

> It would be much safer than btrfs check --repair.
> 
> Thanks,
> Qu
> 
> 
>>>
>>>> Thanks,
>>>> Qu
>>>>
>>>>>
>>>>>> This patchset adds such check and repair ability to btrfs-check, with a
>>>>>> simple test image.
>>>>>>
>>>>>> Qu Wenruo (3):
>>>>>>      btrfs-progs: check/lowmem: Add check and repair for invalid inode
>>>>>>        generation
>>>>>>      btrfs-progs: check/original: Add check and repair for invalid inode
>>>>>>        generation
>>>>>>      btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>> generation
>>>>>>        repair
>>>>>>
>>>>>>     check/main.c                                  |  50 +++++++++++-
>>>>>>     check/mode-lowmem.c                           |  76
>>>>>> ++++++++++++++++++
>>>>>>     check/mode-original.h                         |   1 +
>>>>>>     .../.lowmem_repairable                        |   0
>>>>>>     .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012 bytes
>>>>>>     5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>     create mode 100644
>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>     create mode 100644
>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>>
>>>>>
>>>
>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>> errors).
>>
>> It's really not recommended to run btrfs check, especially repair on the
>> mounted fs, unless it's RO.
>>
>> A new transaction from kernel can easily screw up the repaired fs.
>>>
>>> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
>>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>>> v5.3 btrfs-progs.
>>>
>>> If you like I can share the log. Let me know.
>>>
>>> This issue can potentially cause a lot of grief. Our company server runs
>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>>> ~100 snapshots. I guess the problematic inodes need to be fixed on each
>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel ~5.6)?
>>
>> Yes.
>>
>>>
>>> Do I understand correctly that this FTB is caused by more strict
>>> checking of the fs by the kernel, while the tools to fix the detected
>>> corruptions are not yet released?
>>
>> Yes.
>>
>> Thanks,
>> Qu
>>
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu WenRuo]

On 2019/10/20 下午9:04, Ferry Toth wrote:
> Op 20-10-2019 om 02:51 schreef Qu Wenruo:
>>
>>
>> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>>
>>>
>>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>>> Hi,
>>>>
>>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>>
>>>>>
>>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>>> We have at least two user reports about bad inode generation makes
>>>>>>> kernel reject the fs.
>>>>>>
>>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>>
>>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and /
>>>>>> on /@
>>>>>>
>>>>>> In initramfs I can try to mount but get something like
>>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree
>>>>>> failed
>>>>>>
>>>>>> Booting old kernel works just as before, no errors.
>>>>>>
>>>>>>> According to the creation time, the inode is created by some 2014
>>>>>>> kernel.
>>>>>>
>>>>>> How do I get the creation time?
>>>>>
>>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
>>>>
>>>> I just went back to the office to reboot to 5.3 and check the creation
>>>> times and found they were 2013 - 2014.
>>>>
>>>>>>
>>>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>>>> transid member) so the error persists until latest tree-checker
>>>>>>> detects.
>>>>>>>
>>>>>>> Even the situation can be fixed by reverting back to older kernel
>>>>>>> and
>>>>>>> copying the offending dir/file to another inode and delete the
>>>>>>> offending
>>>>>>> one, it still should be done by btrfs-progs.
>>>>>>>
>>>>>> How to find the offending dir/file from the command line manually?
>>>>>
>>>>> # find <mount point> -inum <inode number>
>>>>
>>>> This works, thanks.
>>>>
>>>> But appears unpractical. After fix 2 files and reboot, I found 4 more,
>>>> then 16, then I gave up.
>>
>> Another solution is use "find" to locate the files with creation time
>> before 2015, and copy them to a new file, then replace the old file with
>> the new file.
> 
> Hmm. But how do I "find" by creation time (otime)? Do you have a
> suggestion for this?

$ touch -t 201501010000 /tmp/sample
$ find <mnt> -not -cnewer /tmp/sample

If you want, you can add -exec to that find, but I'd only add that after
confirming the execution load is verified.

Thanks,
Qu

> 
>> It would be much safer than btrfs check --repair.
>>
>> Thanks,
>> Qu
>>
>>
>>>>
>>>>> Thanks,
>>>>> Qu
>>>>>
>>>>>>
>>>>>>> This patchset adds such check and repair ability to btrfs-check,
>>>>>>> with a
>>>>>>> simple test image.
>>>>>>>
>>>>>>> Qu Wenruo (3):
>>>>>>>      btrfs-progs: check/lowmem: Add check and repair for invalid
>>>>>>> inode
>>>>>>>        generation
>>>>>>>      btrfs-progs: check/original: Add check and repair for
>>>>>>> invalid inode
>>>>>>>        generation
>>>>>>>      btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>>> generation
>>>>>>>        repair
>>>>>>>
>>>>>>>     check/main.c                                  |  50 +++++++++++-
>>>>>>>     check/mode-lowmem.c                           |  76
>>>>>>> ++++++++++++++++++
>>>>>>>     check/mode-original.h                         |   1 +
>>>>>>>     .../.lowmem_repairable                        |   0
>>>>>>>     .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012
>>>>>>> bytes
>>>>>>>     5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>>     create mode 100644
>>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>>     create mode 100644
>>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>>>
>>>>>>>
>>>>>>
>>>>
>>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>>> errors).
>>>
>>> It's really not recommended to run btrfs check, especially repair on the
>>> mounted fs, unless it's RO.
>>>
>>> A new transaction from kernel can easily screw up the repaired fs.
>>>>
>>>> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
>>>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>>>> v5.3 btrfs-progs.
>>>>
>>>> If you like I can share the log. Let me know.
>>>>
>>>> This issue can potentially cause a lot of grief. Our company server
>>>> runs
>>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>>>> ~100 snapshots. I guess the problematic inodes need to be fixed on each
>>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel
>>>> ~5.6)?
>>>
>>> Yes.
>>>
>>>>
>>>> Do I understand correctly that this FTB is caused by more strict
>>>> checking of the fs by the kernel, while the tools to fix the detected
>>>> corruptions are not yet released?
>>>
>>> Yes.
>>>
>>> Thanks,
>>> Qu
>>>
>>
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 20-10-2019 om 15:15 schreef Qu WenRuo:
> 
> 
> On 2019/10/20 下午9:04, Ferry Toth wrote:
>> Op 20-10-2019 om 02:51 schreef Qu Wenruo:
>>>
>>>
>>> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>>>
>>>>
>>>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>>>> Hi,
>>>>>
>>>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>>>
>>>>>>
>>>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>>>> We have at least two user reports about bad inode generation makes
>>>>>>>> kernel reject the fs.
>>>>>>>
>>>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>>>
>>>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and /
>>>>>>> on /@
>>>>>>>
>>>>>>> In initramfs I can try to mount but get something like
>>>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree
>>>>>>> failed
>>>>>>>
>>>>>>> Booting old kernel works just as before, no errors.
>>>>>>>
>>>>>>>> According to the creation time, the inode is created by some 2014
>>>>>>>> kernel.
>>>>>>>
>>>>>>> How do I get the creation time?
>>>>>>
>>>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your device>
>>>>>
>>>>> I just went back to the office to reboot to 5.3 and check the creation
>>>>> times and found they were 2013 - 2014.
>>>>>
>>>>>>>
>>>>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>>>>> transid member) so the error persists until latest tree-checker
>>>>>>>> detects.
>>>>>>>>
>>>>>>>> Even the situation can be fixed by reverting back to older kernel
>>>>>>>> and
>>>>>>>> copying the offending dir/file to another inode and delete the
>>>>>>>> offending
>>>>>>>> one, it still should be done by btrfs-progs.
>>>>>>>>
>>>>>>> How to find the offending dir/file from the command line manually?
>>>>>>
>>>>>> # find <mount point> -inum <inode number>
>>>>>
>>>>> This works, thanks.
>>>>>
>>>>> But appears unpractical. After fix 2 files and reboot, I found 4 more,
>>>>> then 16, then I gave up.
>>>
>>> Another solution is use "find" to locate the files with creation time
>>> before 2015, and copy them to a new file, then replace the old file with
>>> the new file.
>>
>> Hmm. But how do I "find" by creation time (otime)? Do you have a
>> suggestion for this?
> 
> $ touch -t 201501010000 /tmp/sample
> $ find <mnt> -not -cnewer /tmp/sample

AFAIK this compares file modified date with status changed date. So, no 
search for creation date.

And stat /tmp/sample (sorry dutch lang output):

ferry@ferry-quad:~$ stat /tmp/sample
   Bestand: /tmp/sample
   Grootte: 0            Blokken: 0            IO-blok: 4096   leeg 
normaal bestand
Apparaat: 1bh/27d   Inode: 62005381     Koppelingen: 1
Toegang: (0664/-rw-rw-r--)   UID: ( 1001/   ferry)   GID: ( 1001/   ferry)
Toegang:   2015-01-01 00:00:00.000000000 +0100
Gewijzigd: 2015-01-01 00:00:00.000000000 +0100
Veranderd: 2019-10-20 15:20:50.366163766 +0200
Ontstaan:  -


> If you want, you can add -exec to that find, but I'd only add that after
> confirming the execution load is verified.
> 
> Thanks,
> Qu
> 
>>
>>> It would be much safer than btrfs check --repair.
>>>
>>> Thanks,
>>> Qu
>>>
>>>
>>>>>
>>>>>> Thanks,
>>>>>> Qu
>>>>>>
>>>>>>>
>>>>>>>> This patchset adds such check and repair ability to btrfs-check,
>>>>>>>> with a
>>>>>>>> simple test image.
>>>>>>>>
>>>>>>>> Qu Wenruo (3):
>>>>>>>>       btrfs-progs: check/lowmem: Add check and repair for invalid
>>>>>>>> inode
>>>>>>>>         generation
>>>>>>>>       btrfs-progs: check/original: Add check and repair for
>>>>>>>> invalid inode
>>>>>>>>         generation
>>>>>>>>       btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>>>> generation
>>>>>>>>         repair
>>>>>>>>
>>>>>>>>      check/main.c                                  |  50 +++++++++++-
>>>>>>>>      check/mode-lowmem.c                           |  76
>>>>>>>> ++++++++++++++++++
>>>>>>>>      check/mode-original.h                         |   1 +
>>>>>>>>      .../.lowmem_repairable                        |   0
>>>>>>>>      .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012
>>>>>>>> bytes
>>>>>>>>      5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>>>      create mode 100644
>>>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>>>      create mode 100644
>>>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>
>>>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>>>> errors).
>>>>
>>>> It's really not recommended to run btrfs check, especially repair on the
>>>> mounted fs, unless it's RO.
>>>>
>>>> A new transaction from kernel can easily screw up the repaired fs.
>>>>>
>>>>> I'm not sure if using the v5.0 kernel and/or checking mounted distorts
>>>>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>>>>> v5.3 btrfs-progs.
>>>>>
>>>>> If you like I can share the log. Let me know.
>>>>>
>>>>> This issue can potentially cause a lot of grief. Our company server
>>>>> runs
>>>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>>>>> ~100 snapshots. I guess the problematic inodes need to be fixed on each
>>>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel
>>>>> ~5.6)?
>>>>
>>>> Yes.
>>>>
>>>>>
>>>>> Do I understand correctly that this FTB is caused by more strict
>>>>> checking of the fs by the kernel, while the tools to fix the detected
>>>>> corruptions are not yet released?
>>>>
>>>> Yes.
>>>>
>>>> Thanks,
>>>> Qu
>>>>
>>>
>>

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Qu Wenruo]

[-- Attachment #1.1: Type: text/plain, Size: 6844 bytes --]



On 2019/10/20 下午9:29, Ferry Toth wrote:
> Op 20-10-2019 om 15:15 schreef Qu WenRuo:
>>
>>
>> On 2019/10/20 下午9:04, Ferry Toth wrote:
>>> Op 20-10-2019 om 02:51 schreef Qu Wenruo:
>>>>
>>>>
>>>> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>>>>
>>>>>
>>>>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>>>>
>>>>>>>
>>>>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>>>>> We have at least two user reports about bad inode generation makes
>>>>>>>>> kernel reject the fs.
>>>>>>>>
>>>>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>>>>
>>>>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and /
>>>>>>>> on /@
>>>>>>>>
>>>>>>>> In initramfs I can try to mount but get something like
>>>>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree
>>>>>>>> failed
>>>>>>>>
>>>>>>>> Booting old kernel works just as before, no errors.
>>>>>>>>
>>>>>>>>> According to the creation time, the inode is created by some 2014
>>>>>>>>> kernel.
>>>>>>>>
>>>>>>>> How do I get the creation time?
>>>>>>>
>>>>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your
>>>>>>> device>
>>>>>>
>>>>>> I just went back to the office to reboot to 5.3 and check the
>>>>>> creation
>>>>>> times and found they were 2013 - 2014.
>>>>>>
>>>>>>>>
>>>>>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>>>>>> transid member) so the error persists until latest tree-checker
>>>>>>>>> detects.
>>>>>>>>>
>>>>>>>>> Even the situation can be fixed by reverting back to older kernel
>>>>>>>>> and
>>>>>>>>> copying the offending dir/file to another inode and delete the
>>>>>>>>> offending
>>>>>>>>> one, it still should be done by btrfs-progs.
>>>>>>>>>
>>>>>>>> How to find the offending dir/file from the command line manually?
>>>>>>>
>>>>>>> # find <mount point> -inum <inode number>
>>>>>>
>>>>>> This works, thanks.
>>>>>>
>>>>>> But appears unpractical. After fix 2 files and reboot, I found 4
>>>>>> more,
>>>>>> then 16, then I gave up.
>>>>
>>>> Another solution is use "find" to locate the files with creation time
>>>> before 2015, and copy them to a new file, then replace the old file
>>>> with
>>>> the new file.
>>>
>>> Hmm. But how do I "find" by creation time (otime)? Do you have a
>>> suggestion for this?
>>
>> $ touch -t 201501010000 /tmp/sample
>> $ find <mnt> -not -cnewer /tmp/sample
> 
> AFAIK this compares file modified date with status changed date. So, no
> search for creation date.
> 
> And stat /tmp/sample (sorry dutch lang output):
> 
> ferry@ferry-quad:~$ stat /tmp/sample
>   Bestand: /tmp/sample
>   Grootte: 0            Blokken: 0            IO-blok: 4096   leeg
> normaal bestand
> Apparaat: 1bh/27d   Inode: 62005381     Koppelingen: 1
> Toegang: (0664/-rw-rw-r--)   UID: ( 1001/   ferry)   GID: ( 1001/   ferry)
> Toegang:   2015-01-01 00:00:00.000000000 +0100
> Gewijzigd: 2015-01-01 00:00:00.000000000 +0100
> Veranderd: 2019-10-20 15:20:50.366163766 +0200
> Ontstaan:  -

My bad, always got confused by o/a/c/mtime, as c really looks like *c*
reation, so I always got confused between ctime and otime.

Then considering not all fs supports otime, find doesn't support that.
I guess it's only possible by other tools....

BTW, did you find any patterns in those existing offending inodes?
I guess it would be faster than finding a tool supporting otime search.

Thanks,
Qu

> 
> 
>> If you want, you can add -exec to that find, but I'd only add that after
>> confirming the execution load is verified.
>>
>> Thanks,
>> Qu
>>
>>>
>>>> It would be much safer than btrfs check --repair.
>>>>
>>>> Thanks,
>>>> Qu
>>>>
>>>>
>>>>>>
>>>>>>> Thanks,
>>>>>>> Qu
>>>>>>>
>>>>>>>>
>>>>>>>>> This patchset adds such check and repair ability to btrfs-check,
>>>>>>>>> with a
>>>>>>>>> simple test image.
>>>>>>>>>
>>>>>>>>> Qu Wenruo (3):
>>>>>>>>>       btrfs-progs: check/lowmem: Add check and repair for invalid
>>>>>>>>> inode
>>>>>>>>>         generation
>>>>>>>>>       btrfs-progs: check/original: Add check and repair for
>>>>>>>>> invalid inode
>>>>>>>>>         generation
>>>>>>>>>       btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>>>>> generation
>>>>>>>>>         repair
>>>>>>>>>
>>>>>>>>>      check/main.c                                  |  50
>>>>>>>>> +++++++++++-
>>>>>>>>>      check/mode-lowmem.c                           |  76
>>>>>>>>> ++++++++++++++++++
>>>>>>>>>      check/mode-original.h                         |   1 +
>>>>>>>>>      .../.lowmem_repairable                        |   0
>>>>>>>>>      .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012
>>>>>>>>> bytes
>>>>>>>>>      5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>>>>      create mode 100644
>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>>>>      create mode 100644
>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>>>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>>>>> errors).
>>>>>
>>>>> It's really not recommended to run btrfs check, especially repair
>>>>> on the
>>>>> mounted fs, unless it's RO.
>>>>>
>>>>> A new transaction from kernel can easily screw up the repaired fs.
>>>>>>
>>>>>> I'm not sure if using the v5.0 kernel and/or checking mounted
>>>>>> distorts
>>>>>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>>>>>> v5.3 btrfs-progs.
>>>>>>
>>>>>> If you like I can share the log. Let me know.
>>>>>>
>>>>>> This issue can potentially cause a lot of grief. Our company server
>>>>>> runs
>>>>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>>>>>> ~100 snapshots. I guess the problematic inodes need to be fixed on
>>>>>> each
>>>>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel
>>>>>> ~5.6)?
>>>>>
>>>>> Yes.
>>>>>
>>>>>>
>>>>>> Do I understand correctly that this FTB is caused by more strict
>>>>>> checking of the fs by the kernel, while the tools to fix the detected
>>>>>> corruptions are not yet released?
>>>>>
>>>>> Yes.
>>>>>
>>>>> Thanks,
>>>>> Qu
>>>>>
>>>>
>>>
> 


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 20-10-2019 om 16:11 schreef Qu Wenruo:
> 
> 
> On 2019/10/20 下午9:29, Ferry Toth wrote:
>> Op 20-10-2019 om 15:15 schreef Qu WenRuo:
>>>
>>>
>>> On 2019/10/20 下午9:04, Ferry Toth wrote:
>>>> Op 20-10-2019 om 02:51 schreef Qu Wenruo:
>>>>>
>>>>>
>>>>> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>>>>>
>>>>>>
>>>>>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>>>>>
>>>>>>>>
>>>>>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>>>>>> We have at least two user reports about bad inode generation makes
>>>>>>>>>> kernel reject the fs.
>>>>>>>>>
>>>>>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 19.10 so
>>>>>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>>>>>
>>>>>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and /
>>>>>>>>> on /@
>>>>>>>>>
>>>>>>>>> In initramfs I can try to mount but get something like
>>>>>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree
>>>>>>>>> failed
>>>>>>>>>
>>>>>>>>> Booting old kernel works just as before, no errors.
>>>>>>>>>
>>>>>>>>>> According to the creation time, the inode is created by some 2014
>>>>>>>>>> kernel.
>>>>>>>>>
>>>>>>>>> How do I get the creation time?
>>>>>>>>
>>>>>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your
>>>>>>>> device>
>>>>>>>
>>>>>>> I just went back to the office to reboot to 5.3 and check the
>>>>>>> creation
>>>>>>> times and found they were 2013 - 2014.
>>>>>>>
>>>>>>>>>
>>>>>>>>>> And the generation member of INODE_ITEM is not updated (unlike the
>>>>>>>>>> transid member) so the error persists until latest tree-checker
>>>>>>>>>> detects.
>>>>>>>>>>
>>>>>>>>>> Even the situation can be fixed by reverting back to older kernel
>>>>>>>>>> and
>>>>>>>>>> copying the offending dir/file to another inode and delete the
>>>>>>>>>> offending
>>>>>>>>>> one, it still should be done by btrfs-progs.
>>>>>>>>>>
>>>>>>>>> How to find the offending dir/file from the command line manually?
>>>>>>>>
>>>>>>>> # find <mount point> -inum <inode number>
>>>>>>>
>>>>>>> This works, thanks.
>>>>>>>
>>>>>>> But appears unpractical. After fix 2 files and reboot, I found 4
>>>>>>> more,
>>>>>>> then 16, then I gave up.
>>>>>
>>>>> Another solution is use "find" to locate the files with creation time
>>>>> before 2015, and copy them to a new file, then replace the old file
>>>>> with
>>>>> the new file.
>>>>
>>>> Hmm. But how do I "find" by creation time (otime)? Do you have a
>>>> suggestion for this?
>>>
>>> $ touch -t 201501010000 /tmp/sample
>>> $ find <mnt> -not -cnewer /tmp/sample
>>
>> AFAIK this compares file modified date with status changed date. So, no
>> search for creation date.
>>
>> And stat /tmp/sample (sorry dutch lang output):
>>
>> ferry@ferry-quad:~$ stat /tmp/sample
>>    Bestand: /tmp/sample
>>    Grootte: 0            Blokken: 0            IO-blok: 4096   leeg
>> normaal bestand
>> Apparaat: 1bh/27d   Inode: 62005381     Koppelingen: 1
>> Toegang: (0664/-rw-rw-r--)   UID: ( 1001/   ferry)   GID: ( 1001/   ferry)
>> Toegang:   2015-01-01 00:00:00.000000000 +0100
>> Gewijzigd: 2015-01-01 00:00:00.000000000 +0100
>> Veranderd: 2019-10-20 15:20:50.366163766 +0200
>> Ontstaan:  -
> 
> My bad, always got confused by o/a/c/mtime, as c really looks like *c*
> reation, so I always got confused between ctime and otime.
> 
> Then considering not all fs supports otime, find doesn't support that.
> I guess it's only possible by other tools....

New stat will support it, but not in Ubuntu 19.10. We did find this:
https://github.com/torvalds/linux/blob/master/samples/vfs/test-statx.c

and are trying to work out a script that will gzip all files with 
creation < 2015. Then we can unzip again and overwrite.

> BTW, did you find any patterns in those existing offending inodes?
> I guess it would be faster than finding a tool supporting otime search.

I didn't see any logic. A mix of logs, cached files, journal files etc. 
+ a .kde directory in the /

> Thanks,
> Qu
> 
>>
>>
>>> If you want, you can add -exec to that find, but I'd only add that after
>>> confirming the execution load is verified.
>>>
>>> Thanks,
>>> Qu
>>>
>>>>
>>>>> It would be much safer than btrfs check --repair.
>>>>>
>>>>> Thanks,
>>>>> Qu
>>>>>
>>>>>
>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Qu
>>>>>>>>
>>>>>>>>>
>>>>>>>>>> This patchset adds such check and repair ability to btrfs-check,
>>>>>>>>>> with a
>>>>>>>>>> simple test image.
>>>>>>>>>>
>>>>>>>>>> Qu Wenruo (3):
>>>>>>>>>>        btrfs-progs: check/lowmem: Add check and repair for invalid
>>>>>>>>>> inode
>>>>>>>>>>          generation
>>>>>>>>>>        btrfs-progs: check/original: Add check and repair for
>>>>>>>>>> invalid inode
>>>>>>>>>>          generation
>>>>>>>>>>        btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>>>>>> generation
>>>>>>>>>>          repair
>>>>>>>>>>
>>>>>>>>>>       check/main.c                                  |  50
>>>>>>>>>> +++++++++++-
>>>>>>>>>>       check/mode-lowmem.c                           |  76
>>>>>>>>>> ++++++++++++++++++
>>>>>>>>>>       check/mode-original.h                         |   1 +
>>>>>>>>>>       .../.lowmem_repairable                        |   0
>>>>>>>>>>       .../bad_inode_geneartion.img.xz               | Bin 0 -> 2012
>>>>>>>>>> bytes
>>>>>>>>>>       5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>>>>>       create mode 100644
>>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>>>>>       create mode 100644
>>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>
>>>>>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>>>>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>>>>>> errors).
>>>>>>
>>>>>> It's really not recommended to run btrfs check, especially repair
>>>>>> on the
>>>>>> mounted fs, unless it's RO.
>>>>>>
>>>>>> A new transaction from kernel can easily screw up the repaired fs.
>>>>>>>
>>>>>>> I'm not sure if using the v5.0 kernel and/or checking mounted
>>>>>>> distorts
>>>>>>> the results? Else I'm going to need a live usb with a v5.3 kernel and
>>>>>>> v5.3 btrfs-progs.
>>>>>>>
>>>>>>> If you like I can share the log. Let me know.
>>>>>>>
>>>>>>> This issue can potentially cause a lot of grief. Our company server
>>>>>>> runs
>>>>>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs with
>>>>>>> ~100 snapshots. I guess the problematic inodes need to be fixed on
>>>>>>> each
>>>>>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel
>>>>>>> ~5.6)?
>>>>>>
>>>>>> Yes.
>>>>>>
>>>>>>>
>>>>>>> Do I understand correctly that this FTB is caused by more strict
>>>>>>> checking of the fs by the kernel, while the tools to fix the detected
>>>>>>> corruptions are not yet released?
>>>>>>
>>>>>> Yes.
>>>>>>
>>>>>> Thanks,
>>>>>> Qu
>>>>>>
>>>>>
>>>>
>>
> 

Re: [PATCH 0/3] btrfs-progs: Add check and repair for invalid inode generation

[Ferry Toth]

Op 20-10-2019 om 16:24 schreef Ferry Toth:
> Op 20-10-2019 om 16:11 schreef Qu Wenruo:
>>
>>
>> On 2019/10/20 下午9:29, Ferry Toth wrote:
>>> Op 20-10-2019 om 15:15 schreef Qu WenRuo:
>>>>
>>>>
>>>> On 2019/10/20 下午9:04, Ferry Toth wrote:
>>>>> Op 20-10-2019 om 02:51 schreef Qu Wenruo:
>>>>>>
>>>>>>
>>>>>> On 2019/10/20 上午8:26, Qu Wenruo wrote:
>>>>>>>
>>>>>>>
>>>>>>> On 2019/10/20 上午12:24, Ferry Toth wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Op 19-10-2019 om 01:50 schreef Qu WenRuo:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 2019/10/19 上午4:32, Ferry Toth wrote:
>>>>>>>>>> Op 24-09-2019 om 10:11 schreef Qu Wenruo:
>>>>>>>>>>> We have at least two user reports about bad inode generation 
>>>>>>>>>>> makes
>>>>>>>>>>> kernel reject the fs.
>>>>>>>>>>
>>>>>>>>>> May I add my report? I just upgraded Ubuntu from 19.04 -> 
>>>>>>>>>> 19.10 so
>>>>>>>>>> kernel went from 5.0 -> 5.3 (but I was using 4.15 too).
>>>>>>>>>>
>>>>>>>>>> Booting 5.3 leaves me in initramfs as I have /boot on @boot and /
>>>>>>>>>> on /@
>>>>>>>>>>
>>>>>>>>>> In initramfs I can try to mount but get something like
>>>>>>>>>> btrfs critical corrupt leaf invalid inode generation open_ctree
>>>>>>>>>> failed
>>>>>>>>>>
>>>>>>>>>> Booting old kernel works just as before, no errors.
>>>>>>>>>>
>>>>>>>>>>> According to the creation time, the inode is created by some 
>>>>>>>>>>> 2014
>>>>>>>>>>> kernel.
>>>>>>>>>>
>>>>>>>>>> How do I get the creation time?
>>>>>>>>>
>>>>>>>>> # btrfs ins dump-tree -b <the bytenr reported by kernel> <your
>>>>>>>>> device>
>>>>>>>>
>>>>>>>> I just went back to the office to reboot to 5.3 and check the
>>>>>>>> creation
>>>>>>>> times and found they were 2013 - 2014.
>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> And the generation member of INODE_ITEM is not updated 
>>>>>>>>>>> (unlike the
>>>>>>>>>>> transid member) so the error persists until latest tree-checker
>>>>>>>>>>> detects.
>>>>>>>>>>>
>>>>>>>>>>> Even the situation can be fixed by reverting back to older 
>>>>>>>>>>> kernel
>>>>>>>>>>> and
>>>>>>>>>>> copying the offending dir/file to another inode and delete the
>>>>>>>>>>> offending
>>>>>>>>>>> one, it still should be done by btrfs-progs.
>>>>>>>>>>>
>>>>>>>>>> How to find the offending dir/file from the command line 
>>>>>>>>>> manually?
>>>>>>>>>
>>>>>>>>> # find <mount point> -inum <inode number>
>>>>>>>>
>>>>>>>> This works, thanks.
>>>>>>>>
>>>>>>>> But appears unpractical. After fix 2 files and reboot, I found 4
>>>>>>>> more,
>>>>>>>> then 16, then I gave up.
>>>>>>
>>>>>> Another solution is use "find" to locate the files with creation time
>>>>>> before 2015, and copy them to a new file, then replace the old file
>>>>>> with
>>>>>> the new file.
>>>>>
>>>>> Hmm. But how do I "find" by creation time (otime)? Do you have a
>>>>> suggestion for this?
>>>>
>>>> $ touch -t 201501010000 /tmp/sample
>>>> $ find <mnt> -not -cnewer /tmp/sample
>>>
>>> AFAIK this compares file modified date with status changed date. So, no
>>> search for creation date.
>>>
>>> And stat /tmp/sample (sorry dutch lang output):
>>>
>>> ferry@ferry-quad:~$ stat /tmp/sample
>>>    Bestand: /tmp/sample
>>>    Grootte: 0            Blokken: 0            IO-blok: 4096   leeg
>>> normaal bestand
>>> Apparaat: 1bh/27d   Inode: 62005381     Koppelingen: 1
>>> Toegang: (0664/-rw-rw-r--)   UID: ( 1001/   ferry)   GID: ( 1001/   
>>> ferry)
>>> Toegang:   2015-01-01 00:00:00.000000000 +0100
>>> Gewijzigd: 2015-01-01 00:00:00.000000000 +0100
>>> Veranderd: 2019-10-20 15:20:50.366163766 +0200
>>> Ontstaan:  -
>>
>> My bad, always got confused by o/a/c/mtime, as c really looks like *c*
>> reation, so I always got confused between ctime and otime.
>>
>> Then considering not all fs supports otime, find doesn't support that.
>> I guess it's only possible by other tools....
> 
> New stat will support it, but not in Ubuntu 19.10. We did find this:
> https://github.com/torvalds/linux/blob/master/samples/vfs/test-statx.c
> 
> and are trying to work out a script that will gzip all files with 
> creation < 2015. Then we can unzip again and overwrite.

So we combined find + test-statx + awk to produce a list of files with 
creation date 1970 that are then fed to tar. Finally we untar that to 
overwrite the original files.

I found the following:
- the problematic files are both ordinary and symlinks
- they all seem to have the same inode generation
- all had no creation date (so were set to jan 1970)
- not all files without creation date were problematic (only ~10%)
- modified dates were 2013/2014 afaikt

The procedure keeps the modified date, but sets creation date to today, 
creates a new inode and fixes the error.

>> BTW, did you find any patterns in those existing offending inodes?
>> I guess it would be faster than finding a tool supporting otime search.
> 
> I didn't see any logic. A mix of logs, cached files, journal files etc. 
> + a .kde directory in the /
> 
>> Thanks,
>> Qu
>>
>>>
>>>
>>>> If you want, you can add -exec to that find, but I'd only add that 
>>>> after
>>>> confirming the execution load is verified.
>>>>
>>>> Thanks,
>>>> Qu
>>>>
>>>>>
>>>>>> It would be much safer than btrfs check --repair.
>>>>>>
>>>>>> Thanks,
>>>>>> Qu
>>>>>>
>>>>>>
>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> Qu
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> This patchset adds such check and repair ability to btrfs-check,
>>>>>>>>>>> with a
>>>>>>>>>>> simple test image.
>>>>>>>>>>>
>>>>>>>>>>> Qu Wenruo (3):
>>>>>>>>>>>        btrfs-progs: check/lowmem: Add check and repair for 
>>>>>>>>>>> invalid
>>>>>>>>>>> inode
>>>>>>>>>>>          generation
>>>>>>>>>>>        btrfs-progs: check/original: Add check and repair for
>>>>>>>>>>> invalid inode
>>>>>>>>>>>          generation
>>>>>>>>>>>        btrfs-progs: fsck-tests: Add test image for invalid inode
>>>>>>>>>>> generation
>>>>>>>>>>>          repair
>>>>>>>>>>>
>>>>>>>>>>>       check/main.c                                  |  50
>>>>>>>>>>> +++++++++++-
>>>>>>>>>>>       check/mode-lowmem.c                           |  76
>>>>>>>>>>> ++++++++++++++++++
>>>>>>>>>>>       check/mode-original.h                         |   1 +
>>>>>>>>>>>       .../.lowmem_repairable                        |   0
>>>>>>>>>>>       .../bad_inode_geneartion.img.xz               | Bin 0 
>>>>>>>>>>> -> 2012
>>>>>>>>>>> bytes
>>>>>>>>>>>       5 files changed, 126 insertions(+), 1 deletion(-)
>>>>>>>>>>>       create mode 100644
>>>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/.lowmem_repairable
>>>>>>>>>>>       create mode 100644
>>>>>>>>>>> tests/fsck-tests/043-bad-inode-generation/bad_inode_geneartion.img.xz 
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>
>>>>>>>> I checked out and built v5.3-rc1 of btrfs-progs. Then ran it on my
>>>>>>>> mounted rootfs with linux 5.0 and captured the log (~1800 lines 209
>>>>>>>> errors).
>>>>>>>
>>>>>>> It's really not recommended to run btrfs check, especially repair
>>>>>>> on the
>>>>>>> mounted fs, unless it's RO.
>>>>>>>
>>>>>>> A new transaction from kernel can easily screw up the repaired fs.
>>>>>>>>
>>>>>>>> I'm not sure if using the v5.0 kernel and/or checking mounted
>>>>>>>> distorts
>>>>>>>> the results? Else I'm going to need a live usb with a v5.3 
>>>>>>>> kernel and
>>>>>>>> v5.3 btrfs-progs.
>>>>>>>>
>>>>>>>> If you like I can share the log. Let me know.
>>>>>>>>
>>>>>>>> This issue can potentially cause a lot of grief. Our company server
>>>>>>>> runs
>>>>>>>> Ubuntu LTS (18.04.02) with a 4.15 kernel on a btrfs boot/rootfs 
>>>>>>>> with
>>>>>>>> ~100 snapshots. I guess the problematic inodes need to be fixed on
>>>>>>>> each
>>>>>>>> snapshot prior to upgrading to 20.04 LTS (which might be on kernel
>>>>>>>> ~5.6)?
>>>>>>>
>>>>>>> Yes.
>>>>>>>
>>>>>>>>
>>>>>>>> Do I understand correctly that this FTB is caused by more strict
>>>>>>>> checking of the fs by the kernel, while the tools to fix the 
>>>>>>>> detected
>>>>>>>> corruptions are not yet released?
>>>>>>>
>>>>>>> Yes.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Qu
>>>>>>>
>>>>>>
>>>>>
>>>
>>
>