linux-kernel - Re: [PATCH v3] platform/x86: Add support for improved performance mode

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] [day] [month] [year] [list]

Date:   Wed, 22 Nov 2023 21:03:57 -0500
From:   "Mark Pearson" <mpearson-lenovo@...ebb.ca>
To:     "Breno Leitao" <leitao@...ian.org>
Cc:     "platform-driver-x86@...r.kernel.org" 
        <platform-driver-x86@...r.kernel.org>,
        "Hans de Goede" <hdegoede@...hat.com>,
        Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>,
        ibm-acpi-devel@...ts.sourceforge.net,
        "Limonciello, Mario" <mario.limonciello@....com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] platform/x86: Add support for improved performance mode

Hi Breno,

Thanks for the review!

On Wed, Nov 22, 2023, at 2:44 PM, Breno Leitao wrote:
> On Mon, Nov 13, 2023 at 11:54:33AM -0500, Mark Pearson wrote:
>> @@ -10355,6 +10361,17 @@ static int dytc_profile_set(struct platform_profile_handler *pprof,
>>  	if (err)
>>  		goto unlock;
>>  
>> +	/* Set TMS mode appropriately (enable for performance), if available */
>> +	if (dytc_ultraperf_cap) {
>> +		int cmd;
>> +
>> +		cmd = DYTC_SET_COMMAND(DYTC_FUNCTION_TMS, DYTC_NOMODE,
>> +				       profile == PLATFORM_PROFILE_PERFORMANCE);
>> +		err = dytc_command(cmd, &output);
>> +		if (err)
>> +			return err;
>
> Aren't you returning holding the 'dytc_mutex' mutex?
>
> From what I understand, in the first line of this function you get the lock,
> and release later, at the exit, so, returning without releasing the lock might
> be dangerous. Here is a summary of how I read this function with your change:
>
>
> 	mutex_lock_interruptible(&dytc_mutex);
> 	...
> 	err = dytc_command(cmd, &output);
> 	if (err)
> 		return err;
>
> unlock:
> 	mutex_unlock(&dytc_mutex);
> 	return err;
>
>
> I think "goto unlock" might solve it.

Yep - you're right. Good catch. 
Will fix in the next revision.

Thank you
Mark