linux-kernel - Re: [GIT PULL] Thermal management updates for v4.17-rc1

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, 13 Apr 2018 13:10:21 +0200
From:   Daniel Lezcano <daniel.lezcano@...aro.org>
To:     Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>
Cc:     Zhang Rui <rui.zhang@...el.com>,
        Eduardo Valentin <edubezval@...il.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Linux PM list <linux-pm@...r.kernel.org>,
        "Li, Philip" <philip.li@...el.com>
Subject: Re: [GIT PULL] Thermal management updates for v4.17-rc1

On 13/04/2018 12:41, Bartlomiej Zolnierkiewicz wrote:
> On Friday, April 13, 2018 12:30:04 PM Daniel Lezcano wrote:
>> On 13/04/2018 11:28, Bartlomiej Zolnierkiewicz wrote:
>>
>> [ ... ]
>>
>>>>> It is okay to return 0 because this code-path (the default one) will be
>>>>> never hit by the driver (probe makes sure of it) - the default case is
>>>>> here is just to silence compilation errors..
>>>>
>>>> The init function is making sure cal_type is one or another. Can you fix
>>>> it correctly by replacing the 'switch' by a 'if' instead of adding dead
>>>> branches to please gcc?
>>>>
>>>> if (data->cal_type == TYPE_TWO_POINT_TRIMMING) {
>>>> 	return ...;
>>>> }
>>>>
>>>> return ...;
>>>
>>> I'm not the one that added this switch statement (it has been there since
>>> 2011) and I would be happy to remove it. 
>>
>> Actually the switch statement was fine until the cleanup.
> 
> I don't see how it was fine before as the driver has never used the default
> case (always used TYPE_ONE_POINT_TRIMMING or TYPE_TWO_POINT_TRIMMING).
> 
> Could you please explain this more?

>From commit 480b5bfc16e17ef51ca1c

+++ b/drivers/thermal/samsung/exynos_tmu.c
@@ -260,7 +260,7 @@ static int temp_to_code(struct exynos_tmu_data
*data, u8 temp)
                temp_code = temp + data->temp_error1 -
pdata->first_point_trim;
                break;
        default:
-               temp_code = temp + pdata->default_temp_offset;
+               WARN_ON(1);
                break;
        }

@@ -287,7 +287,7 @@ static int code_to_temp(struct exynos_tmu_data
*data, u16 temp_code)
                temp = temp_code - data->temp_error1 +
pdata->first_point_trim;
                break;
        default:
-               temp = temp_code - pdata->default_temp_offset;
+               WARN_ON(1);
                break;
        }

I'm not saying the code path was fine but from the compiler point of
view, it was. By removing the defaulting temp value there is a code path
gcc sees the temp variable as not initialized.

Your cleanups are relevant.


> --
> Bartlomiej Zolnierkiewicz
> Samsung R&D Institute Poland
> Samsung Electronics
> 


-- 
 <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