linux-kernel - Re: [PATCH] [thermal] adding check if the thermal firmware is running

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Fri, 17 May 2024 12:20:49 +0200
From: "Wysocki, Rafael J" <rafael.j.wysocki@...el.com>
To: Guilherme Giacomo Simoes <trintaeoitogc@...il.com>,
	<miriam.rachel.korenblit@...el.com>, <kvalo@...nel.org>,
	<daniel.lezcano@...aro.org>, <johannes.berg@...el.com>, <dmantipov@...dex.ru>
CC: <linux-wireless@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] [thermal] adding check if the thermal firmware is running


On 5/16/2024 10:05 PM, Guilherme Giacomo Simoes wrote:
> in the dmesg is showing the message "failed to read out thermal zone"
> as if the temperature read is failed by don't find the thermal zone.
>
> After researching and debugging, I see that this specific error is
> occurrence because the thermal try read the temperature when is started,
> but the firmware is not running yet.
>
> For more legibiliti i create the NOTLOAD error code in the errno-base.h,
> and in the iwl_mvm_tzone_get_temp() on tt.c i check if firmware is
> running and I set the NOTLOAD code for ret variable and goto out.
>
> After this, in the update_temperature() in the thermal_code.c i received
> the return of thermal_zone_get_temp() and check if return is NOTLOAD,
> because if it is, I print the warning message "firmware yet not load"
> and return for caller
>
> The thermal_core.c i think that is generic for any thermal drivers and
> not only used for tt.c of course.
> But if this ipotetic driver not check if firmware is running before read
> the temperature, the thermal_code.c is work as a before this change.
>
> After this change, in my computer I compile and install kernel in /boot
> and in my dmesg the message "failed to read out thermal zone" is not
> show any more. In your place the warning messafe "Firmware yet not
> load" is showing.
>
> I would like to thank you in advance for any contribution, suggestion
> or criticism of my patch suggestion.
> ---
>   drivers/net/wireless/intel/iwlwifi/mvm/tt.c | 10 ++++++++--
>   drivers/thermal/thermal_core.c              | 10 +++++++---
>   include/uapi/asm-generic/errno-base.h       |  1 +
>   3 files changed, 16 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/tt.c b/drivers/net/wireless/intel/iwlwifi/mvm/tt.c
> index 8083c4b2ab6b..dd5725db06d2 100644
> --- a/drivers/net/wireless/intel/iwlwifi/mvm/tt.c
> +++ b/drivers/net/wireless/intel/iwlwifi/mvm/tt.c
> @@ -620,8 +620,14 @@ static int iwl_mvm_tzone_get_temp(struct thermal_zone_device *device,
>   
>   	mutex_lock(&mvm->mutex);
>   
> -	if (!iwl_mvm_firmware_running(mvm) ||
> -	    mvm->fwrt.cur_fw_img != IWL_UCODE_REGULAR) {
> +	const int res = iwl_mvm_firmware_running(mvm);
> +
> +	if (!res) {
> +		ret = -NOTLOAD;
> +		goto out;
> +	}
> +
> +	if (mvm->fwrt.cur_fw_img != IWL_UCODE_REGULAR) {
>   		ret = -ENODATA;
>   		goto out;
>   	}
> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
> index 34a31bc72023..4116d312d4a1 100644
> --- a/drivers/thermal/thermal_core.c
> +++ b/drivers/thermal/thermal_core.c
> @@ -414,10 +414,14 @@ static void handle_thermal_trip(struct thermal_zone_device *tz,
>   
>   static void update_temperature(struct thermal_zone_device *tz)
>   {
> -	int temp, ret;
> -
> -	ret = __thermal_zone_get_temp(tz, &temp);
> +	int temp;
> +	int ret = __thermal_zone_get_temp(tz, &temp);
>   	if (ret) {
> +		if (ret == -NOTLOAD) {
> +			pr_warn("Firmware yet not load");
> +			return;
> +		}
> +

The thermal core doesn't need to be modified for this.

Please print the new message from the driver and you may as well return 
-EAGAIN from it in all cases when the issue is expected to be 
intermittent to prevent the core from printing the (existing) warning 
message.

Thanks!


>   		if (ret != -EAGAIN)
>   			dev_warn(&tz->device,
>   				 "failed to read out thermal zone (%d)\n",
> diff --git a/include/uapi/asm-generic/errno-base.h b/include/uapi/asm-generic/errno-base.h
> index 9653140bff92..8b92c41f7993 100644
> --- a/include/uapi/asm-generic/errno-base.h
> +++ b/include/uapi/asm-generic/errno-base.h
> @@ -36,5 +36,6 @@
>   #define	EPIPE		32	/* Broken pipe */
>   #define	EDOM		33	/* Math argument out of domain of func */
>   #define	ERANGE		34	/* Math result not representable */
> +#define	NOTLOAD		35	/* Firmware yet not load */
>   
>   #endif