diff mbox series

soundwire: qcom: fix storing port config out-of-bounds

Message ID 20230601102525.609627-1-krzysztof.kozlowski@linaro.org
State Accepted
Commit 490937d479abe5f6584e69b96df066bc87be92e9
Headers show
Series soundwire: qcom: fix storing port config out-of-bounds | expand

Commit Message

Krzysztof Kozlowski June 1, 2023, 10:25 a.m. UTC
The 'qcom_swrm_ctrl->pconfig' has size of QCOM_SDW_MAX_PORTS (14),
however we index it starting from 1, not 0, to match real port numbers.
This can lead to writing port config past 'pconfig' bounds and
overwriting next member of 'qcom_swrm_ctrl' struct.  Reported also by
smatch:

  drivers/soundwire/qcom.c:1269 qcom_swrm_get_port_config() error: buffer overflow 'ctrl->pconfig' 14 <= 14

Fixes: 9916c02ccd74 ("soundwire: qcom: cleanup internal port config indexing")
Cc: <stable@vger.kernel.org>
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Link: https://lore.kernel.org/r/202305201301.sCJ8UDKV-lkp@intel.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
---
 drivers/soundwire/qcom.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Vinod Koul June 7, 2023, 1:10 p.m. UTC | #1
On 01-06-23, 12:25, Krzysztof Kozlowski wrote:
> The 'qcom_swrm_ctrl->pconfig' has size of QCOM_SDW_MAX_PORTS (14),
> however we index it starting from 1, not 0, to match real port numbers.
> This can lead to writing port config past 'pconfig' bounds and
> overwriting next member of 'qcom_swrm_ctrl' struct.  Reported also by
> smatch:
> 
>   drivers/soundwire/qcom.c:1269 qcom_swrm_get_port_config() error: buffer overflow 'ctrl->pconfig' 14 <= 14
> 
> Fixes: 9916c02ccd74 ("soundwire: qcom: cleanup internal port config indexing")
> Cc: <stable@vger.kernel.org>
> Reported-by: kernel test robot <lkp@intel.com>
> Reported-by: Dan Carpenter <error27@gmail.com>
> Link: https://lore.kernel.org/r/202305201301.sCJ8UDKV-lkp@intel.com/
> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> ---
>  drivers/soundwire/qcom.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
> index 7cb1b7eba814..88a772075907 100644
> --- a/drivers/soundwire/qcom.c
> +++ b/drivers/soundwire/qcom.c
> @@ -202,7 +202,8 @@ struct qcom_swrm_ctrl {
>  	u32 intr_mask;
>  	u8 rcmd_id;
>  	u8 wcmd_id;
> -	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS];
> +	/* Port numbers are 1 - 14 */
> +	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS + 1];

Better use SDW_MAX_PORTS ?

>  	struct sdw_stream_runtime *sruntime[SWRM_MAX_DAIS];
>  	enum sdw_slave_status status[SDW_MAX_DEVICES + 1];
>  	int (*reg_read)(struct qcom_swrm_ctrl *ctrl, int reg, u32 *val);
> -- 
> 2.34.1
Krzysztof Kozlowski June 7, 2023, 1:51 p.m. UTC | #2
On 07/06/2023 15:10, Vinod Koul wrote:
> On 01-06-23, 12:25, Krzysztof Kozlowski wrote:
>> The 'qcom_swrm_ctrl->pconfig' has size of QCOM_SDW_MAX_PORTS (14),
>> however we index it starting from 1, not 0, to match real port numbers.
>> This can lead to writing port config past 'pconfig' bounds and
>> overwriting next member of 'qcom_swrm_ctrl' struct.  Reported also by
>> smatch:
>>
>>   drivers/soundwire/qcom.c:1269 qcom_swrm_get_port_config() error: buffer overflow 'ctrl->pconfig' 14 <= 14
>>
>> Fixes: 9916c02ccd74 ("soundwire: qcom: cleanup internal port config indexing")
>> Cc: <stable@vger.kernel.org>
>> Reported-by: kernel test robot <lkp@intel.com>
>> Reported-by: Dan Carpenter <error27@gmail.com>
>> Link: https://lore.kernel.org/r/202305201301.sCJ8UDKV-lkp@intel.com/
>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>> ---
>>  drivers/soundwire/qcom.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
>> index 7cb1b7eba814..88a772075907 100644
>> --- a/drivers/soundwire/qcom.c
>> +++ b/drivers/soundwire/qcom.c
>> @@ -202,7 +202,8 @@ struct qcom_swrm_ctrl {
>>  	u32 intr_mask;
>>  	u8 rcmd_id;
>>  	u8 wcmd_id;
>> -	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS];
>> +	/* Port numbers are 1 - 14 */
>> +	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS + 1];
> 
> Better use SDW_MAX_PORTS ?

That's interesting idea, but except of value, is the meaning actually
the same? Driver claims that port 0 is masked and max number of ports is
14. Therefore it uses in all places constant QCOM_SDW_MAX_PORTS. We need
here +1, only because we index from 1, not 0, but we still index over
QCOM_SDW_MAX_PORTS, not SDW_MAX_PORTS. Wouldn't it be also confusing to
use here SDW_MAX_PORTS but then index over something else?

Best regards,
Krzysztof
Pierre-Louis Bossart June 7, 2023, 2:51 p.m. UTC | #3
On 6/7/23 08:51, Krzysztof Kozlowski wrote:
> On 07/06/2023 15:10, Vinod Koul wrote:
>> On 01-06-23, 12:25, Krzysztof Kozlowski wrote:
>>> The 'qcom_swrm_ctrl->pconfig' has size of QCOM_SDW_MAX_PORTS (14),
>>> however we index it starting from 1, not 0, to match real port numbers.
>>> This can lead to writing port config past 'pconfig' bounds and
>>> overwriting next member of 'qcom_swrm_ctrl' struct.  Reported also by
>>> smatch:
>>>
>>>   drivers/soundwire/qcom.c:1269 qcom_swrm_get_port_config() error: buffer overflow 'ctrl->pconfig' 14 <= 14
>>>
>>> Fixes: 9916c02ccd74 ("soundwire: qcom: cleanup internal port config indexing")
>>> Cc: <stable@vger.kernel.org>
>>> Reported-by: kernel test robot <lkp@intel.com>
>>> Reported-by: Dan Carpenter <error27@gmail.com>
>>> Link: https://lore.kernel.org/r/202305201301.sCJ8UDKV-lkp@intel.com/
>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>> ---
>>>  drivers/soundwire/qcom.c | 3 ++-
>>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
>>> index 7cb1b7eba814..88a772075907 100644
>>> --- a/drivers/soundwire/qcom.c
>>> +++ b/drivers/soundwire/qcom.c
>>> @@ -202,7 +202,8 @@ struct qcom_swrm_ctrl {
>>>  	u32 intr_mask;
>>>  	u8 rcmd_id;
>>>  	u8 wcmd_id;
>>> -	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS];
>>> +	/* Port numbers are 1 - 14 */
>>> +	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS + 1];
>>
>> Better use SDW_MAX_PORTS ?
> 
> That's interesting idea, but except of value, is the meaning actually
> the same? Driver claims that port 0 is masked and max number of ports is
> 14. Therefore it uses in all places constant QCOM_SDW_MAX_PORTS. We need
> here +1, only because we index from 1, not 0, but we still index over
> QCOM_SDW_MAX_PORTS, not SDW_MAX_PORTS. Wouldn't it be also confusing to
> use here SDW_MAX_PORTS but then index over something else?

SDW_MAX_PORTS only applies for the peripheral. DP0 is reserved for
non-audio/Bulk request, DP15 is an alias for "all ports"

There's nothing in the spec that restricts the ports on the manager
side, be it to dedicate Port0 or Port15 to a specific purpose or even
the number of ports.

I would recommend using a vendor-specific definition rather than
overloading a peripheral specification requirement.
diff mbox series

Patch

diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
index 7cb1b7eba814..88a772075907 100644
--- a/drivers/soundwire/qcom.c
+++ b/drivers/soundwire/qcom.c
@@ -202,7 +202,8 @@  struct qcom_swrm_ctrl {
 	u32 intr_mask;
 	u8 rcmd_id;
 	u8 wcmd_id;
-	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS];
+	/* Port numbers are 1 - 14 */
+	struct qcom_swrm_port_config pconfig[QCOM_SDW_MAX_PORTS + 1];
 	struct sdw_stream_runtime *sruntime[SWRM_MAX_DAIS];
 	enum sdw_slave_status status[SDW_MAX_DEVICES + 1];
 	int (*reg_read)(struct qcom_swrm_ctrl *ctrl, int reg, u32 *val);