Skip to contentSkip to conversation

ipv4: Fix refcount warning for new fib_info

PATCH In review David Ahern 5 participants 5 messages Aug 2, 2021 – Aug 3, 2021

Subsystems (3)
lore
Share & tools
lore

David Ahern

author maintainer
#0 lore raw

PATCH 1/1 ipv4: Fix refcount warning for new fib_info

Ioana reported a refcount warning when booting over NFS:

[    5.042532] ------------[ cut here ]------------
[    5.047184] refcount_t: addition on 0; use-after-free.
[    5.052324] WARNING: CPU: 7 PID: 1 at lib/refcount.c:25 refcount_warn_saturate+0xa4/0x150
...
[    5.167201] Call trace:
[    5.169635]  refcount_warn_saturate+0xa4/0x150
[    5.174067]  fib_create_info+0xc00/0xc90
[    5.177982]  fib_table_insert+0x8c/0x620
[    5.181893]  fib_magic.isra.0+0x110/0x11c
[    5.185891]  fib_add_ifaddr+0xb8/0x190
[    5.189629]  fib_inetaddr_event+0x8c/0x140

fib_treeref needs to be set after kzalloc. The old code had a ++ which led to the confusion when the int was replaced by a refcount_t.

Fixes 79976892f7ea ("net: convert fib_treeref from int to refcount_t")
Cc Yajun Deng <[email protected]>
---
 net/ipv4/fib_semantics.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/ipv4/fib_semantics.c +1 -1
index fa19f4cdf3a4..f29feb7772da 100644
@@ -1551,7 +1551,7 @@ struct fib_info *fib_create_info(struct fib_config *cfg,
 		return ofi;
 	}
 
-	refcount_inc(&fi->fib_treeref);
+	refcount_set(&fi->fib_treeref, 1);
 	refcount_set(&fi->fib_clntref, 1);
 	spin_lock_bh(&fib_info_lock);
 	hlist_add_head(&fi->fib_hash,
Signature · 1 line
2.24.3 (Apple Git-128)
Tested-by on v1 1/1
#1 lore raw

Hi David,

Quoted 14 lines · David Ahern

Ioana reported a refcount warning when booting over NFS:

[    5.042532] ------------[ cut here ]------------
[    5.047184] refcount_t: addition on 0; use-after-free.
[    5.052324] WARNING: CPU: 7 PID: 1 at lib/refcount.c:25 refcount_warn_saturate+0xa4/0x150
...
[    5.167201] Call trace:
[    5.169635]  refcount_warn_saturate+0xa4/0x150
[    5.174067]  fib_create_info+0xc00/0xc90
[    5.177982]  fib_table_insert+0x8c/0x620
[    5.181893]  fib_magic.isra.0+0x110/0x11c
[    5.185891]  fib_add_ifaddr+0xb8/0x190
[    5.189629]  fib_inetaddr_event+0x8c/0x140
Quoted 2 lines

fib_treeref needs to be set after kzalloc. The old code had a ++ which led to the confusion when the int was replaced by a refcount_t.

Thank you for the patch!

My CI was also complaining of not being able to run kernel selftests [1]. Your patch fixes the issue, thanks!

Cheers,
Matt

[1] https://cirrus-ci.com/task/5688032394215424

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
#2 lore raw
Quoted 20 lines · Matthieu Baerts

Hi David,

Quoted 14 lines · David Ahern

Ioana reported a refcount warning when booting over NFS:

[    5.042532] ------------[ cut here ]------------
[    5.047184] refcount_t: addition on 0; use-after-free.
[    5.052324] WARNING: CPU: 7 PID: 1 at lib/refcount.c:25 refcount_warn_saturate+0xa4/0x150
...
[    5.167201] Call trace:
[    5.169635]  refcount_warn_saturate+0xa4/0x150
[    5.174067]  fib_create_info+0xc00/0xc90
[    5.177982]  fib_table_insert+0x8c/0x620
[    5.181893]  fib_magic.isra.0+0x110/0x11c
[    5.185891]  fib_add_ifaddr+0xb8/0x190
[    5.189629]  fib_inetaddr_event+0x8c/0x140
Quoted 2 lines

fib_treeref needs to be set after kzalloc. The old code had a ++ which led to the confusion when the int was replaced by a refcount_t.

Quoted 7 lines

Thank you for the patch!

My CI was also complaining of not being able to run kernel selftests [1]. Your patch fixes the issue, thanks!

Given how easily it is to trigger the warning, I get the impression the original was an untested patch.

maintainer
#3 lore raw
Quoted 23 lines · On Mon, 2 Aug 2021 12:04:01 -0600 David Ahern wrote:
Quoted 22 lines · Matthieu Baerts

Hi David,

Quoted 14 lines · David Ahern

Ioana reported a refcount warning when booting over NFS:

[    5.042532] ------------[ cut here ]------------
[    5.047184] refcount_t: addition on 0; use-after-free.
[    5.052324] WARNING: CPU: 7 PID: 1 at lib/refcount.c:25 refcount_warn_saturate+0xa4/0x150
...
[    5.167201] Call trace:
[    5.169635]  refcount_warn_saturate+0xa4/0x150
[    5.174067]  fib_create_info+0xc00/0xc90
[    5.177982]  fib_table_insert+0x8c/0x620
[    5.181893]  fib_magic.isra.0+0x110/0x11c
[    5.185891]  fib_add_ifaddr+0xb8/0x190
[    5.189629]  fib_inetaddr_event+0x8c/0x140
Quoted 2 lines

fib_treeref needs to be set after kzalloc. The old code had a ++ which led to the confusion when the int was replaced by a refcount_t.

Thank you for the patch!

Quoted 8 lines
Quoted 5 lines

My CI was also complaining of not being able to run kernel selftests [1]. Your patch fixes the issue, thanks!

Given how easily it is to trigger the warning, I get the impression the original was an untested patch.

Yeah :( In hindsight any refcount patch which doesn't contain a refcount_set() is suspicious. Thanks for the quick fix, applied!

#4 lore raw
Quoted 28 lines · Jakub Kicinski
Quoted 26 lines · On Mon, 2 Aug 2021 12:04:01 -0600 David Ahern wrote:

On 8/2/21 11:58 AM, Matthieu Baerts wrote:
Hi David,

On 02/08/2021 18:02, David Ahern wrote:
Ioana reported a refcount warning when booting over NFS:

[ 5.042532] ------------[ cut here ]------------
[ 5.047184] refcount_t: addition on 0; use-after-free.
[ 5.052324] WARNING: CPU: 7 PID: 1 at lib/refcount.c:25 refcount_warn_saturate+0xa4/0x150
...
[ 5.167201] Call trace:
[ 5.169635] refcount_warn_saturate+0xa4/0x150
[ 5.174067] fib_create_info+0xc00/0xc90
[ 5.177982] fib_table_insert+0x8c/0x620
[ 5.181893] fib_magic.isra.0+0x110/0x11c
[ 5.185891] fib_add_ifaddr+0xb8/0x190
[ 5.189629] fib_inetaddr_event+0x8c/0x140

fib_treeref needs to be set after kzalloc. The old code had a ++ which led to the confusion when the int was replaced by a refcount_t.

Thank you for the patch!

My CI was also complaining of not being able to run kernel selftests [1]. Your patch fixes the issue, thanks!

Quoted 7 lines
Quoted 4 lines

Given how easily it is to trigger the warning, I get the impression the original was an untested patch.

Yeah :( In hindsight any refcount patch which doesn't contain a refcount_set() is suspicious. Thanks for the quick fix, applied!

Sorry for that, there is another patch needed to apply for the same reason. I just submitted it.