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]
Message-ID: <f0cb97b0-19aa-aff5-0ac4-60406e6e1a97@amd.com>
Date:   Mon, 22 Apr 2019 08:23:29 +0000
From:   "Agrawal, Akshu" <Akshu.Agrawal@....com>
To:     Cheng-yi Chiang <cychiang@...omium.org>
CC:     "moderated list:SOUND - SOC LAYER / DYNAMIC AUDIO POWER MANAGEM..." 
        <alsa-devel@...a-project.org>,
        Support Opensource <support.opensource@...semi.com>,
        open list <linux-kernel@...r.kernel.org>,
        Takashi Iwai <tiwai@...e.com>,
        Liam Girdwood <lgirdwood@...il.com>,
        "djkurtz@...omium.org" <djkurtz@...omium.org>,
        Mark Brown <broonie@...nel.org>,
        "Deucher, Alexander" <Alexander.Deucher@....com>,
        "Adam.Thomson.Opensource@...semi.com" 
        <Adam.Thomson.Opensource@...semi.com>,
        "yuhsuan@...omium.org" <yuhsuan@...omium.org>,
        "tzungbi@...omium.org" <tzungbi@...omium.org>
Subject: Re: [alsa-devel] [PATCH] ASoC: DA7219: Fix failure in hw_params by
 not letting set_rate error out



On 4/22/2019 1:09 PM, Cheng-yi Chiang wrote:
> Hi Akshu,
> 
> 
> On Mon, Apr 22, 2019 at 2:15 PM Agrawal, Akshu <Akshu.Agrawal@....com> wrote:
>>
>> We need to set minimum bclk 64x of wclk as this is hw constraint in one
>> of the component used.
>> Since, clk_set_rate and clk is enabled in machine driver the
>> clk_set_rate in hw_params of da7219 fails and errors out and when it
>> tries to override the value.
>>
>> In cases like these not only clk_set_rate of da7219 codec should fail
>> and not override the value but also should not error out.
>>
>> Signed-off-by: Akshu Agrawal <akshu.agrawal@....com>
>> ---
>>  sound/soc/codecs/da7219.c | 8 +-------
>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>
>> diff --git a/sound/soc/codecs/da7219.c b/sound/soc/codecs/da7219.c
>> index 5f5fa3416af3..a041dbf442af 100644
>> --- a/sound/soc/codecs/da7219.c
>> +++ b/sound/soc/codecs/da7219.c
>> @@ -1621,13 +1621,7 @@ static int da7219_hw_params(struct snd_pcm_substream *substream,
>>
>>                 if (bclk) {
>>                         bclk_rate = frame_size * sr;
>> -                       ret = clk_set_rate(bclk, bclk_rate);
>> -                       if (ret) {
>> -                               dev_err(component->dev,
>> -                                       "Failed to set BCLK rate %lu: %d\n",
>> -                                       bclk_rate, ret);
>> -                               return ret;
>> -                       }
>> +                       clk_set_rate(bclk, bclk_rate);
> 
> I think this might mask the real issue when setting rate fails.

Not doing an error check for clk_set_rate is not rare because
clk_set_rate fails if the clk which you are trying to set is already
enabled. Now we can add a is_prepared check before calling clk_set_rate
but that would be redundant as clk_set_rate does the same check.

> A more fundamental problem might be what are the acceptable ratios
> between bclk and wclk.
> 

acceptable ratios for da7219 from spec are:
00 = 32 BCLKS per WCLK
01 = 64 BCLKS per WCLK
10 = 128 BCLK per WCLK
11 = 256 BCLK per WCLK

da7219 derives this ration from SR for a wclk. The same we do in machine
driver.
Conflict comes only in case of an exception where there is constraint
from a hw component. Else when the audio is played on da7219 codec we
could have let da7219 only set rate and enable clks.

Thanks,
Akshu

> From existing code of da7219 I see there are only two choices of
> frame_size: 32 or 64.
> 
> In da7219_hw_params:
> ...
> if (da7219->master && !da7219->tdm_en) {
>   if ((word_len * DA7219_DAI_CH_NUM_MAX) <= 32)
>      frame_size = 32;
>   else
>      frame_size = 64;
> 
> And in https://lkml.org/lkml/2019/4/17/322
> the choice from machine driver is 64 or higher.
> 
> In cz_da7219_params:
>  ...
>  if (bclk < (wclk * 64))
>   bclk = wclk * 64;
>  return da7219_clk_enable(substream, wclk, bclk);
> 
> I don't know what is the best solution to avoid this conflict.
> Maybe just set a fixed ratio from device property that can best work
> for the combination of codecs on the machine?
> 
> 
> 
> 
> 
>>                 } else {
>>                         ret = da7219_set_bclks_per_wclk(component, frame_size);
>>                         if (ret) {
>> --
>> 2.17.1
>>
>> _______________________________________________
>> Alsa-devel mailing list
>> Alsa-devel@...a-project.org
>> https://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ