f2fs: fix to detect obsolete inner inode during fill_super()

PATCH In review Chao Yu 2 participants 6 messages Sep 8, 2022 – Sep 16, 2022

Subsystems (2)

Stable branch F2FS file system

lore

Share & tools

lore

Chao Yu

author maintainer

2022-09-08 10:53 UTC #0 lore raw

PATCH 1/1 f2fs: fix to detect obsolete inner inode during fill_super()

Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.

Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000

Mem abort info:
ESR = 0x0000000086000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000 [0000000000000000] pgd=0000000000000000, p4d=0000000000000000

Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748

sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8

Call trace:
 0x0
 set_page_dirty+0x38/0xbc mm/folio-compat.c:62
 f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
 do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
 f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679

The root cause is, quoted from Jaegeuk:

It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.

This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.

Cc [email protected]

Reported-by [email protected]

Signed-off-by Jaegeuk Kim <[email protected]>

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/inode.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

fs/f2fs/inode.c +11 -0

index ccb29034af59..df1a82fbfaf2 100644
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
 	struct inode *inode;
 	int ret = 0;
 
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);
+		}
+	}
+
 	inode = iget_locked(sb, ino);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);

Signature · 1 line

2.25.1

Jaegeuk Kim

maintainer

2022-09-12 15:38 UTC #1 lore raw

Quoted 61 lines · Chao Yu

Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Mem abort info:
ESR = 0x0000000086000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000 [0000000000000000] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748
sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8
Call trace:
 0x0
 set_page_dirty+0x38/0xbc mm/folio-compat.c:62
 f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
 do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
 f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679
The root cause is, quoted from Jaegeuk:
It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.
This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.
Cc [email protected]
Reported-by [email protected]
Signed-off-by Jaegeuk Kim <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
 fs/f2fs/inode.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
fs/f2fs/inode.c +0 -0
index ccb29034af59..df1a82fbfaf2 100644
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
 	struct inode *inode;

fs/f2fs/inode.c struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino) · line 503

Quoted 10 lines

fs/f2fs/inode.c +8 -0

 	int ret = 0;
 
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);

Well, this does not indicate f2fs is corrupted. I'd rather expect to fix reiserfs instead of f2fs workaround which hides the bug.

Quoted 8 lines

fs/f2fs/inode.c +3 -0

+		}
+	}
+
 	inode = iget_locked(sb, ino);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);

Signature · 1 line

2.25.1

Chao Yu

author maintainer

2022-09-13 02:33 UTC #2 lore raw

Quoted 64 lines · Jaegeuk Kim

Quoted 63 lines · Chao Yu
Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Mem abort info:
   ESR = 0x0000000086000004
   EC = 0x21: IABT (current EL), IL = 32 bits
   SET = 0, FnV = 0
   EA = 0, S1PTW = 0
   FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000
[0000000000000000] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748
sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8
Call trace:
  0x0
  set_page_dirty+0x38/0xbc mm/folio-compat.c:62
  f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
  do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
  f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679
The root cause is, quoted from Jaegeuk:
It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.
This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.
Cc [email protected]
Reported-by [email protected]
Signed-off-by Jaegeuk Kim <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
  fs/f2fs/inode.c | 11 +++++++++++
  1 file changed, 11 insertions(+)
fs/f2fs/inode.c +0 -0
index ccb29034af59..df1a82fbfaf2 100644
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
  	struct inode *inode;
  	int ret = 0;
 

Quoted 11 lines

fs/f2fs/inode.c struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino) · line 503
Quoted 8 lines
fs/f2fs/inode.c +8 -0
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);
Well, this does not indicate f2fs is corrupted. I'd rather expect to fix reiserfs instead of f2fs workaround which hides the bug.

Well, is there a fixing patch for reiserfs? If not, how about applying this patch first, later, we can revert it after reiserfs has been fixed.

Thanks,

Quoted 9 lines

fs/f2fs/inode.c +3 -0

+		}
+	}
+
  	inode = iget_locked(sb, ino);
  	if (!inode)
  		return ERR_PTR(-ENOMEM);

Signature · 1 line

2.25.1

Jaegeuk Kim

maintainer

2022-09-13 06:04 UTC #3 lore raw

Quoted 76 lines · Chao Yu

Quoted 66 lines · Jaegeuk Kim
Quoted 65 lines · Chao Yu
Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Mem abort info:
   ESR = 0x0000000086000004
   EC = 0x21: IABT (current EL), IL = 32 bits
   SET = 0, FnV = 0
   EA = 0, S1PTW = 0
   FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000
[0000000000000000] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748
sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8
Call trace:
  0x0
  set_page_dirty+0x38/0xbc mm/folio-compat.c:62
  f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
  do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
  f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679
The root cause is, quoted from Jaegeuk:
It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.
This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.
Cc [email protected]
Reported-by [email protected]
Signed-off-by Jaegeuk Kim <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
  fs/f2fs/inode.c | 11 +++++++++++
  1 file changed, 11 insertions(+)
fs/f2fs/inode.c +3 -0
index ccb29034af59..df1a82fbfaf2 100644
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
  	struct inode *inode;
  	int ret = 0;
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
Quoted 8 lines
fs/f2fs/inode.c struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino) · line 503
Quoted 5 lines
fs/f2fs/inode.c +5 -0
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);
Well, this does not indicate f2fs is corrupted. I'd rather expect to fix reiserfs instead of f2fs workaround which hides the bug.

Quoted 2 lines

Well, is there a fixing patch for reiserfs? If not, how about applying this patch first, later, we can revert it after reiserfs has been fixed.

I don't feel this is a right way to deal with that. If we think it'd be worth checking any stale inode object during f2fs_fill_super, we'd better check any cached inode given superblock pointer rather than our inner inodes only.

Quoted 12 lines

Thanks,
Quoted 9 lines
Quoted 9 lines
fs/f2fs/inode.c +3 -0
+		}
+	}
+
  	inode = iget_locked(sb, ino);
  	if (!inode)
  		return ERR_PTR(-ENOMEM);
Signature · 1 line
2.25.1

Chao Yu

author maintainer

2022-09-14 15:54 UTC #4 lore raw

Quoted 77 lines · Jaegeuk Kim

Quoted 76 lines · Chao Yu
Quoted 74 lines · Jaegeuk Kim
Quoted 60 lines · Chao Yu
Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Mem abort info:
    ESR = 0x0000000086000004
    EC = 0x21: IABT (current EL), IL = 32 bits
    SET = 0, FnV = 0
    EA = 0, S1PTW = 0
    FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000
[0000000000000000] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748
sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8
Call trace:
   0x0
   set_page_dirty+0x38/0xbc mm/folio-compat.c:62
   f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
   do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
   f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679
The root cause is, quoted from Jaegeuk:
It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.
This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.
Cc [email protected]
Reported-by [email protected]
Signed-off-by Jaegeuk Kim <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
   fs/f2fs/inode.c | 11 +++++++++++
   1 file changed, 11 insertions(+)
fs/f2fs/inode.c +0 -0
index ccb29034af59..df1a82fbfaf2 100644
fs/f2fs/inode.c struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino) · line 502
Quoted 10 lines
fs/f2fs/inode.c +8 -0
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
   	struct inode *inode;
   	int ret = 0;
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);
Well, this does not indicate f2fs is corrupted. I'd rather expect to fix reiserfs instead of f2fs workaround which hides the bug.

Quoted 6 lines

Quoted 2 lines
Well, is there a fixing patch for reiserfs? If not, how about applying this patch first, later, we can revert it after reiserfs has been fixed.
I don't feel this is a right way to deal with that. If we think it'd be worth checking any stale inode object during f2fs_fill_super, we'd better check any cached inode given superblock pointer rather than our inner inodes only.

Well, something like this?

f2fs_fill_super()

	for (ino = root_ino; ino < max_nid; ino++) {
		inode = iget_locked(sb, ino);
		if (!inode)
			continue;
		iput(inode);
		ret = -EFSCORRUPTED;
		goto error_handling;
	}

Quoted 13 lines

Quoted 13 lines
Thanks,
Quoted 9 lines
Quoted 9 lines
fs/f2fs/inode.c +3 -0
+		}
+	}
+
   	inode = iget_locked(sb, ino);
   	if (!inode)
   		return ERR_PTR(-ENOMEM);
Signature · 1 line
2.25.1

Jaegeuk Kim

maintainer

2022-09-16 18:21 UTC #5 lore raw

Quoted 78 lines · Chao Yu

Quoted 77 lines · Jaegeuk Kim
Quoted 76 lines · Chao Yu
Quoted 74 lines · Jaegeuk Kim
Quoted 60 lines · Chao Yu
Sometimes we can get a cached meta_inode which has no aops yet. Let's set it all the time to fix the below panic.
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Mem abort info:
    ESR = 0x0000000086000004
    EC = 0x21: IABT (current EL), IL = 32 bits
    SET = 0, FnV = 0
    EA = 0, S1PTW = 0
    FSC = 0x04: level 0 translation fault
user pgtable: 4k pages, 48-bit VAs, pgdp=0000000109ee4000
[0000000000000000] pgd=0000000000000000, p4d=0000000000000000
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 1 PID: 3045 Comm: syz-executor330 Not tainted 6.0.0-rc2-syzkaller-16455-ga41a877bc12d #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : 0x0
lr : folio_mark_dirty+0xbc/0x208 mm/page-writeback.c:2748
sp : ffff800012783970
x29: ffff800012783970 x28: 0000000000000000 x27: ffff800012783b08
x26: 0000000000000001 x25: 0000000000000400 x24: 0000000000000001
x23: ffff0000c736e000 x22: 0000000000000045 x21: 05ffc00000000015
x20: ffff0000ca7403b8 x19: fffffc00032ec600 x18: 0000000000000181
x17: ffff80000c04d6bc x16: ffff80000dbb8658 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ff808000083e9814 x10: 0000000000000000 x9 : ffff8000083e9814 x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff0000cbb19000 x4 : ffff0000cb3d2000 x3 : ffff0000cbb18f80 x2 : fffffffffffffff0 x1 : fffffc00032ec600 x0 : ffff0000ca7403b8
Call trace:
   0x0
   set_page_dirty+0x38/0xbc mm/folio-compat.c:62
   f2fs_update_meta_page+0x80/0xa8 fs/f2fs/segment.c:2369
   do_checkpoint+0x794/0xea8 fs/f2fs/checkpoint.c:1522
   f2fs_write_checkpoint+0x3b8/0x568 fs/f2fs/checkpoint.c:1679
The root cause is, quoted from Jaegeuk:
It turned out there is a bug in reiserfs which doesn't free the root inode (ino=2). That leads f2fs to find an ino=2 with the previous superblock point used by reiserfs. That stale inode has no valid mapping that f2fs can use, result in kernel panic.
This patch adds sanity check in f2fs_iget() to avoid finding stale inode during inner inode initialization.
Cc [email protected]
Reported-by [email protected]
Signed-off-by Jaegeuk Kim <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
   fs/f2fs/inode.c | 11 +++++++++++
   1 file changed, 11 insertions(+)
fs/f2fs/inode.c +0 -0
index ccb29034af59..df1a82fbfaf2 100644
fs/f2fs/inode.c struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino) · line 502
Quoted 10 lines
fs/f2fs/inode.c +8 -0
@@ -493,6 +493,17 @@ struct inode *f2fs_iget_inner(struct super_block *sb, unsigned long ino)
   	struct inode *inode;
   	int ret = 0;
+	if (ino == F2FS_NODE_INO(sbi) || ino == F2FS_META_INO(sbi) ||
+					ino == F2FS_COMPRESS_INO(sbi)) {
+		inode = ilookup(sb, ino);
+		if (inode) {
+			iput(inode);
+			f2fs_err(sbi, "there is obsoleted inner inode %lu cached in hash table",
+					ino);
+			return ERR_PTR(-EFSCORRUPTED);
Well, this does not indicate f2fs is corrupted. I'd rather expect to fix reiserfs instead of f2fs workaround which hides the bug.

Quoted 8 lines

Quoted 6 lines
Quoted 2 lines
Well, is there a fixing patch for reiserfs? If not, how about applying this patch first, later, we can revert it after reiserfs has been fixed.
I don't feel this is a right way to deal with that. If we think it'd be worth checking any stale inode object during f2fs_fill_super, we'd better check any cached inode given superblock pointer rather than our inner inodes only.
Well, something like this?

Any chance to get a better way?

Quoted 25 lines

f2fs_fill_super()

	for (ino = root_ino; ino < max_nid; ino++) {
		inode = iget_locked(sb, ino);
		if (!inode)
			continue;
		iput(inode);
		ret = -EFSCORRUPTED;
		goto error_handling;
	}

Quoted 13 lines

Quoted 13 lines
Thanks,
Quoted 9 lines
Quoted 9 lines
fs/f2fs/inode.c +3 -0
+		}
+	}
+
   	inode = iget_locked(sb, ino);
   	if (!inode)
   		return ERR_PTR(-ENOMEM);
Signature · 1 line
2.25.1