[PATCH v3] thunderbolt: Fix to check the return value of kmemdup

[PATCH v3] thunderbolt: Fix to check the return value of kmemdup

[Aditya Pakki]

uuid in add_switch is allocted via kmemdup which can fail. The patch
logs the error and cleans up the allocated memory for switch.

Signed-off-by: Aditya Pakki <pakki001@umn.edu>

---
v2: replace WARN_ONCE with tb_sw_warn, as suggested by Mika
v1: Change warn_once to return after clean up.
---
 drivers/thunderbolt/icm.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/thunderbolt/icm.c b/drivers/thunderbolt/icm.c
index e3fc920af682..8b7f9131e9d1 100644
--- a/drivers/thunderbolt/icm.c
+++ b/drivers/thunderbolt/icm.c
@@ -473,6 +473,11 @@ static void add_switch(struct tb_switch *parent_sw, u64 route,
 		goto out;
 
 	sw->uuid = kmemdup(uuid, sizeof(*uuid), GFP_KERNEL);
+	if (!sw->uuid) {
+		tb_sw_warn(sw, "cannot allocate memory for switch\n");
+		tb_switch_put(sw);
+		goto out;
+	}
 	sw->connection_id = connection_id;
 	sw->connection_key = connection_key;
 	sw->link = link;
-- 
2.17.1

Re: [PATCH v3] thunderbolt: Fix to check the return value of kmemdup

[Mukesh Ojha]

On 3/26/2019 2:55 AM, Aditya Pakki wrote:
> uuid in add_switch is allocted via kmemdup which can fail. The patch
> logs the error and cleans up the allocated memory for switch.
>
> Signed-off-by: Aditya Pakki <pakki001@umn.edu>

Reviewed-by: Mukesh Ojha <mojha@codeaurora.org>

-Mukesh

>
> ---
> v2: replace WARN_ONCE with tb_sw_warn, as suggested by Mika
> v1: Change warn_once to return after clean up.
> ---
>   drivers/thunderbolt/icm.c | 5 +++++
>   1 file changed, 5 insertions(+)
>
> diff --git a/drivers/thunderbolt/icm.c b/drivers/thunderbolt/icm.c
> index e3fc920af682..8b7f9131e9d1 100644
> --- a/drivers/thunderbolt/icm.c
> +++ b/drivers/thunderbolt/icm.c
> @@ -473,6 +473,11 @@ static void add_switch(struct tb_switch *parent_sw, u64 route,
>   		goto out;
>   
>   	sw->uuid = kmemdup(uuid, sizeof(*uuid), GFP_KERNEL);
> +	if (!sw->uuid) {
> +		tb_sw_warn(sw, "cannot allocate memory for switch\n");
> +		tb_switch_put(sw);
> +		goto out;
> +	}
>   	sw->connection_id = connection_id;
>   	sw->connection_key = connection_key;
>   	sw->link = link;

Re: [PATCH v3] thunderbolt: Fix to check the return value of kmemdup

[Mika Westerberg]

On Mon, Mar 25, 2019 at 04:25:22PM -0500, Aditya Pakki wrote:
> uuid in add_switch is allocted via kmemdup which can fail. The patch
> logs the error and cleans up the allocated memory for switch.
> 
> Signed-off-by: Aditya Pakki <pakki001@umn.edu>

Applied, thanks!

Re: [v3] thunderbolt: Fix to check the return value of kmemdup

[Markus Elfring]

> uuid in add_switch is allocted via kmemdup which can fail.

I have tried another script for the semantic patch language out.
This source code analysis approach points out that the implementation
of the function “icm_handle_event” contains still an unchecked call
of the function “kmemdup”.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/thunderbolt/icm.c?id=3cdb9446a117d5d63af823bde6fe6babc312e77b#n1627
https://elixir.bootlin.com/linux/v5.4-rc2/source/drivers/thunderbolt/icm.c#L1627

How do you think about to improve it?

Regards,
Markus

Re: [v3] thunderbolt: Fix to check the return value of kmemdup

[Mika Westerberg]

On Fri, Oct 11, 2019 at 03:00:13PM +0200, Markus Elfring wrote:
> > uuid in add_switch is allocted via kmemdup which can fail.
> 
> I have tried another script for the semantic patch language out.
> This source code analysis approach points out that the implementation
> of the function “icm_handle_event” contains still an unchecked call
> of the function “kmemdup”.
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/thunderbolt/icm.c?id=3cdb9446a117d5d63af823bde6fe6babc312e77b#n1627
> https://elixir.bootlin.com/linux/v5.4-rc2/source/drivers/thunderbolt/icm.c#L1627

Right it misses that.

> How do you think about to improve it?

Feel free to send a patch fixing it ;-) Or I can do that myself.

Re: [v3] thunderbolt: Fix to check the return value of kmemdup

[Markus Elfring]

>> This source code analysis approach points out that the implementation
>> of the function “icm_handle_event” contains still an unchecked call
>> of the function “kmemdup”.
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/thunderbolt/icm.c?id=3cdb9446a117d5d63af823bde6fe6babc312e77b#n1627
>> https://elixir.bootlin.com/linux/v5.4-rc2/source/drivers/thunderbolt/icm.c#L1627
>
> Right it misses that.

Thanks for your quick feedback.


> Feel free to send a patch fixing it ;-) Or I can do that myself.

Would you like to reconsider also the addition of the function call
“tb_sw_warn(sw, "cannot allocate memory for switch\n")”?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/coding-style.rst?id=9e208aa06c2109b45eec6be049a8e47034748c20#n878

Regards,
Markus

Re: [v3] thunderbolt: Fix to check the return value of kmemdup

[Mika Westerberg]

On Fri, Oct 11, 2019 at 04:13:22PM +0200, Markus Elfring wrote:
> Would you like to reconsider also the addition of the function call
> “tb_sw_warn(sw, "cannot allocate memory for switch\n")”?

For that I already have a patch as part of my USB4 support v2 series.