This patch prevents from registering thermal entries and letting the
driver misbehave if efuse data is invalid. A device is not properly
calibrated if the golden temperature is zero.
Signed-off-by: Julien Panis <[email protected]>
---
Guard against invalid calibration data, following this discussion:
https://lore.kernel.org/all/[email protected]/
---
drivers/thermal/mediatek/lvts_thermal.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/thermal/mediatek/lvts_thermal.c b/drivers/thermal/mediatek/lvts_thermal.c
index 0bb3a495b56e..185d5a32711f 100644
--- a/drivers/thermal/mediatek/lvts_thermal.c
+++ b/drivers/thermal/mediatek/lvts_thermal.c
@@ -769,7 +769,11 @@ static int lvts_golden_temp_init(struct device *dev, u8 *calib,
*/
gt = (((u32 *)calib)[0] >> lvts_data->gt_calib_bit_offset) & 0xff;
- if (gt && gt < LVTS_GOLDEN_TEMP_MAX)
+ /* A zero value for gt means that device has invalid efuse data */
+ if (!gt)
+ return -ENODATA;
+
+ if (gt < LVTS_GOLDEN_TEMP_MAX)
golden_temp = gt;
golden_temp_offset = golden_temp * 500 + lvts_data->temp_offset;
---
base-commit: 632483ea8004edfadd035de36e1ab2c7c4f53158
change-id: 20240603-mtk-thermal-calib-check-ba2ec24a1c32
Best regards,
--
Julien Panis <[email protected]>
On Mon, 3 Jun 2024, Julien Panis wrote:
> This patch prevents from registering thermal entries and letting the
> driver misbehave if efuse data is invalid. A device is not properly
> calibrated if the golden temperature is zero.
>
> Signed-off-by: Julien Panis <[email protected]>
Reviewed-by: Nicolas Pitre <[email protected]>
> ---
> Guard against invalid calibration data, following this discussion:
> https://lore.kernel.org/all/[email protected]/
> ---
> drivers/thermal/mediatek/lvts_thermal.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/thermal/mediatek/lvts_thermal.c b/drivers/thermal/mediatek/lvts_thermal.c
> index 0bb3a495b56e..185d5a32711f 100644
> --- a/drivers/thermal/mediatek/lvts_thermal.c
> +++ b/drivers/thermal/mediatek/lvts_thermal.c
> @@ -769,7 +769,11 @@ static int lvts_golden_temp_init(struct device *dev, u8 *calib,
> */
> gt = (((u32 *)calib)[0] >> lvts_data->gt_calib_bit_offset) & 0xff;
>
> - if (gt && gt < LVTS_GOLDEN_TEMP_MAX)
> + /* A zero value for gt means that device has invalid efuse data */
> + if (!gt)
> + return -ENODATA;
> +
> + if (gt < LVTS_GOLDEN_TEMP_MAX)
> golden_temp = gt;
>
> golden_temp_offset = golden_temp * 500 + lvts_data->temp_offset;
>
> ---
> base-commit: 632483ea8004edfadd035de36e1ab2c7c4f53158
> change-id: 20240603-mtk-thermal-calib-check-ba2ec24a1c32
>
> Best regards,
> --
> Julien Panis <[email protected]>
>
>
Il 03/06/24 14:06, Julien Panis ha scritto:
> This patch prevents from registering thermal entries and letting the
> driver misbehave if efuse data is invalid. A device is not properly
> calibrated if the golden temperature is zero.
>
> Signed-off-by: Julien Panis <[email protected]>
Fixes tag, please.
Regards,
Angelo
> ---
> Guard against invalid calibration data, following this discussion:
> https://lore.kernel.org/all/[email protected]/
> ---
> drivers/thermal/mediatek/lvts_thermal.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/thermal/mediatek/lvts_thermal.c b/drivers/thermal/mediatek/lvts_thermal.c
> index 0bb3a495b56e..185d5a32711f 100644
> --- a/drivers/thermal/mediatek/lvts_thermal.c
> +++ b/drivers/thermal/mediatek/lvts_thermal.c
> @@ -769,7 +769,11 @@ static int lvts_golden_temp_init(struct device *dev, u8 *calib,
> */
> gt = (((u32 *)calib)[0] >> lvts_data->gt_calib_bit_offset) & 0xff;
>
> - if (gt && gt < LVTS_GOLDEN_TEMP_MAX)
> + /* A zero value for gt means that device has invalid efuse data */
> + if (!gt)
> + return -ENODATA;
> +
> + if (gt < LVTS_GOLDEN_TEMP_MAX)
> golden_temp = gt;
>
> golden_temp_offset = golden_temp * 500 + lvts_data->temp_offset;
>
> ---
> base-commit: 632483ea8004edfadd035de36e1ab2c7c4f53158
> change-id: 20240603-mtk-thermal-calib-check-ba2ec24a1c32
>
> Best regards,
On Tue, 4 Jun 2024, AngeloGioacchino Del Regno wrote:
> Il 03/06/24 14:06, Julien Panis ha scritto:
> > This patch prevents from registering thermal entries and letting the
> > driver misbehave if efuse data is invalid. A device is not properly
> > calibrated if the golden temperature is zero.
> >
> > Signed-off-by: Julien Panis <[email protected]>
>
> Fixes tag, please.
What should it refer to? This is arguably a robustness enhancement for a
defficiency that has been there all along.
Nicolas
On 03/06/2024 14:06, Julien Panis wrote:
> This patch prevents from registering thermal entries and letting the
> driver misbehave if efuse data is invalid. A device is not properly
> calibrated if the golden temperature is zero.
>
> Signed-off-by: Julien Panis <[email protected]>
> ---
Applied, thanks
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
Il 04/06/24 18:33, Nicolas Pitre ha scritto:
> On Tue, 4 Jun 2024, AngeloGioacchino Del Regno wrote:
>
>> Il 03/06/24 14:06, Julien Panis ha scritto:
>>> This patch prevents from registering thermal entries and letting the
>>> driver misbehave if efuse data is invalid. A device is not properly
>>> calibrated if the golden temperature is zero.
>>>
>>> Signed-off-by: Julien Panis <[email protected]>
>>
>> Fixes tag, please.
>
> What should it refer to? This is arguably a robustness enhancement for a
> defficiency that has been there all along.
>
>
If that's been there since day 1, the fixes tag should be the day 1 commit,
of course.
I see that as an important bugfix and not just a robustness enhancement, as
SoCs with uncalibrated LVTS would *not work*, as the kernel would shut down
the system before even loading the userspace.
Regards,
Angelo