[PATCH] ASoC: Intel: Skylake: Fix available clock counter incrementation

Cezary Rojewski cezary.rojewski at intel.com
Mon Feb 24 18:42:45 CET 2020


On 2020-02-24 17:18, Pierre-Louis Bossart wrote:
> On 2/24/20 6:52 AM, Amadeusz Sławiński wrote:
>> Incrementation of avail_clk_cnt was incorrectly moved to error path. Put
>> it back to success path.
>>

>> diff --git a/sound/soc/intel/skylake/skl-ssp-clk.c 
>> b/sound/soc/intel/skylake/skl-ssp-clk.c
>> index 1c0e5226cb5b..bd43885f3805 100644
>> --- a/sound/soc/intel/skylake/skl-ssp-clk.c
>> +++ b/sound/soc/intel/skylake/skl-ssp-clk.c
>> @@ -384,9 +384,11 @@ static int skl_clk_dev_probe(struct 
>> platform_device *pdev)
>>                   &clks[i], clk_pdata, i);
>>           if (IS_ERR(data->clk[data->avail_clk_cnt])) {
>> -            ret = PTR_ERR(data->clk[data->avail_clk_cnt++]);
>> +            ret = PTR_ERR(data->clk[data->avail_clk_cnt]);
> 
> Are you sure?
> 
> If you start with avail_clk_cnt set to zero, the error handling will 
> decrement and access offset -1
> 
> static void unregister_src_clk(struct skl_clk_data *dclk)
> {
>       while (dclk->avail_clk_cnt--)
>           clkdev_drop(dclk->clk[dclk->avail_clk_cnt]->lookup); <<< oob 
> access with offset -1
> }
> 

Decrementation will occur after while's check evaluation - loop will 
exit, decrementation happens regardless.


More information about the Alsa-devel mailing list