[PATCH] drm/bridge: ti-sn65dsi86: Avoid possible buffer overflow

[PATCH] drm/bridge: ti-sn65dsi86: Avoid possible buffer overflow

[Su Hui]

Smatch error:buffer overflow 'ti_sn_bridge_refclk_lut' 5 <= 5.

Fixes: cea86c5bb442 ("drm/bridge: ti-sn65dsi86: Implement the pwm_chip")
Signed-off-by: Su Hui <suhui@nfschina.com>
---
 drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
index 7a748785c545..952aae4221e7 100644
--- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c
+++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
@@ -305,7 +305,8 @@ static void ti_sn_bridge_set_refclk_freq(struct ti_sn65dsi86 *pdata)
 	 * The PWM refclk is based on the value written to SN_DPPLL_SRC_REG,
 	 * regardless of its actual sourcing.
 	 */
-	pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
+	if (i < refclk_lut_size)
+		pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
 }
 
 static void ti_sn65dsi86_enable_comms(struct ti_sn65dsi86 *pdata)
-- 
2.30.2

Re: [PATCH] drm/bridge: ti-sn65dsi86: Avoid possible buffer overflow

[Doug Anderson]

Hi,

On Tue, Jun 6, 2023 at 12:56 AM Su Hui <suhui@nfschina.com> wrote:
>
> Smatch error:buffer overflow 'ti_sn_bridge_refclk_lut' 5 <= 5.
>
> Fixes: cea86c5bb442 ("drm/bridge: ti-sn65dsi86: Implement the pwm_chip")
> Signed-off-by: Su Hui <suhui@nfschina.com>
> ---
>  drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
> index 7a748785c545..952aae4221e7 100644
> --- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c
> +++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
> @@ -305,7 +305,8 @@ static void ti_sn_bridge_set_refclk_freq(struct ti_sn65dsi86 *pdata)
>          * The PWM refclk is based on the value written to SN_DPPLL_SRC_REG,
>          * regardless of its actual sourcing.
>          */
> -       pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
> +       if (i < refclk_lut_size)
> +               pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];

I don't think this is quite the right fix. I don't think we can just
skip assigning "pdata->pwm_refclk_freq". In general I think we're in
pretty bad shape if we ever fail to match a refclk from the table and
I'm not quite sure how the bridge chip could work at all in this case.
Probably that at least deserves a warning message in the logs. There's
no place to return an error though, so I guess the warning is the best
we can do and then we can do our best to do something reasonable.

In this case, I think "reasonable" might be that if the for loop exits
and "i == refclk_lut_size" that we should set "i" to 1. According to
the datasheet [1] setting a value of 5 (which the existing code does)
is the same as setting a value of 1 (the default) and if it's 1 then
we'll be able to look this up in the table.


[1] https://www.ti.com/lit/gpn/sn65dsi86

-Doug

Re: [PATCH] drm/bridge: ti-sn65dsi86: Avoid possible buffer overflow

[Longsuhui]

Hi,

On 2023/6/6 23:28, Doug Anderson wrote:
> Hi,
>
> On Tue, Jun 6, 2023 at 12:56 AM Su Hui <suhui@nfschina.com> wrote:
>> Smatch error:buffer overflow 'ti_sn_bridge_refclk_lut' 5 <= 5.
>>
>> Fixes: cea86c5bb442 ("drm/bridge: ti-sn65dsi86: Implement the pwm_chip")
>> Signed-off-by: Su Hui <suhui@nfschina.com>
>> ---
>>   drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> index 7a748785c545..952aae4221e7 100644
>> --- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> +++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> @@ -305,7 +305,8 @@ static void ti_sn_bridge_set_refclk_freq(struct ti_sn65dsi86 *pdata)
>>           * The PWM refclk is based on the value written to SN_DPPLL_SRC_REG,
>>           * regardless of its actual sourcing.
>>           */
>> -       pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
>> +       if (i < refclk_lut_size)
>> +               pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
> I don't think this is quite the right fix. I don't think we can just
> skip assigning "pdata->pwm_refclk_freq". In general I think we're in
> pretty bad shape if we ever fail to match a refclk from the table and
> I'm not quite sure how the bridge chip could work at all in this case.
> Probably that at least deserves a warning message in the logs. There's
> no place to return an error though, so I guess the warning is the best
> we can do and then we can do our best to do something reasonable.
>
> In this case, I think "reasonable" might be that if the for loop exits
> and "i == refclk_lut_size" that we should set "i" to 1. According to
> the datasheet [1] setting a value of 5 (which the existing code does)
> is the same as setting a value of 1 (the default) and if it's 1 then
> we'll be able to look this up in the table.
I think you are right. And " if ( i >= refclk_lut_size) i=1" is a 
suitable change.
I will send patch v2 a litter latter.
Thanks for your suggestion.

Su Hui

>
> [1] https://www.ti.com/lit/gpn/sn65dsi86
>
> -Doug

Re: [PATCH] drm/bridge: ti-sn65dsi86: Avoid possible buffer overflow

[Su Hui]

Hi,

On 2023/6/6 23:28, Doug Anderson wrote:
> Hi,
>
> On Tue, Jun 6, 2023 at 12:56 AM Su Hui <suhui@nfschina.com> wrote:
>> Smatch error:buffer overflow 'ti_sn_bridge_refclk_lut' 5 <= 5.
>>
>> Fixes: cea86c5bb442 ("drm/bridge: ti-sn65dsi86: Implement the pwm_chip")
>> Signed-off-by: Su Hui <suhui@nfschina.com>
>> ---
>>   drivers/gpu/drm/bridge/ti-sn65dsi86.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> index 7a748785c545..952aae4221e7 100644
>> --- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> +++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c
>> @@ -305,7 +305,8 @@ static void ti_sn_bridge_set_refclk_freq(struct ti_sn65dsi86 *pdata)
>>           * The PWM refclk is based on the value written to SN_DPPLL_SRC_REG,
>>           * regardless of its actual sourcing.
>>           */
>> -       pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
>> +       if (i < refclk_lut_size)
>> +               pdata->pwm_refclk_freq = ti_sn_bridge_refclk_lut[i];
> I don't think this is quite the right fix. I don't think we can just
> skip assigning "pdata->pwm_refclk_freq". In general I think we're in
> pretty bad shape if we ever fail to match a refclk from the table and
> I'm not quite sure how the bridge chip could work at all in this case.
> Probably that at least deserves a warning message in the logs. There's
> no place to return an error though, so I guess the warning is the best
> we can do and then we can do our best to do something reasonable.
>
> In this case, I think "reasonable" might be that if the for loop exits
> and "i == refclk_lut_size" that we should set "i" to 1. According to
> the datasheet [1] setting a value of 5 (which the existing code does)
> is the same as setting a value of 1 (the default) and if it's 1 then
> we'll be able to look this up in the table.
I think you are right, set i to 1 if i >= refclk_lut_size. I will resend 
this patch soon.
Thanks for your reply!

Su Hui

>
> [1] https://www.ti.com/lit/gpn/sn65dsi86
>
> -Doug