2019-07-29 02:31:00

by Jia-Ju Bai

Subject: [PATCH] target: iscsi: iscsi_target_tpg: Fix a possible null-pointer dereference in iscsit_tpg_add_network_portal()

In iscsit_tpg_add_network_portal(), there is an if statement on line 496
to check whether tpg->tpg_tiqn is NULL:
if (tpg->tpg_tiqn)

When tpg->tpg_tiqn is NULL, it is used on line 508:
pr_debug(..., tpg->tpg_tiqn->tiqn, ...);

Thus, a possible null-pointer dereference may occur.

To fix this bug, tpg->tpg_tiqn is checked before being used.

This bug is found by a static analysis tool STCheck written by us.

Signed-off-by: Jia-Ju Bai <[email protected]>
---
drivers/target/iscsi/iscsi_target_tpg.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/target/iscsi/iscsi_target_tpg.c b/drivers/target/iscsi/iscsi_target_tpg.c
index 8075f60fd02c..bf97be36ec1f 100644
--- a/drivers/target/iscsi/iscsi_target_tpg.c
+++ b/drivers/target/iscsi/iscsi_target_tpg.c
@@ -505,9 +505,11 @@ struct iscsi_tpg_np *iscsit_tpg_add_network_portal(
spin_unlock(&tpg_np_parent->tpg_np_parent_lock);
}

- pr_debug("CORE[%s] - Added Network Portal: %pISpc,%hu on %s\n",
- tpg->tpg_tiqn->tiqn, &np->np_sockaddr, tpg->tpgt,
- np->np_transport->name);
+ if (tpg->tpg_tiqn) {
+ pr_debug("CORE[%s] - Added Network Portal: %pISpc,%hu on %s\n",
+ tpg->tpg_tiqn->tiqn, &np->np_sockaddr, tpg->tpgt,
+ np->np_transport->name);
+ }

return tpg_np;
}
--
2.17.0

2019-07-29 18:13:06

by James Bottomley

Subject: Re: [PATCH] target: iscsi: iscsi_target_tpg: Fix a possible null-pointer dereference in iscsit_tpg_add_network_portal()

On Mon, 2019-07-29 at 10:29 +0800, Jia-Ju Bai wrote:
> In iscsit_tpg_add_network_portal(), there is an if statement on line
> 496
> to check whether tpg->tpg_tiqn is NULL:
> if (tpg->tpg_tiqn)
>
> When tpg->tpg_tiqn is NULL, it is used on line 508:
> pr_debug(..., tpg->tpg_tiqn->tiqn, ...);
>
> Thus, a possible null-pointer dereference may occur.
>
> To fix this bug, tpg->tpg_tiqn is checked before being used.
>
> This bug is found by a static analysis tool STCheck written by us.

I don't really think this is helpful. The first question is, is the
implied might be NULL check correct? The tpg_tiqn is always set by a
non-dummy driver and I think network configuration is only done for the
non dummy driver, so I suspect the NULL check is wrong. Secondly even
if the NULL check were correct, I think there's still a need for some
debugging output, so the proposed patch also looks wrong.

James