From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sowjanya Komatineni Subject: Re: [PATCH v5 12/19] ASoC: tegra: Add initial parent configuration for audio mclk Date: Mon, 6 Jan 2020 08:09:33 -0800 Message-ID: <5921f8e1-359d-4352-a785-8a685cd1a200@nvidia.com> References: <1576880825-15010-1-git-send-email-skomatineni@nvidia.com> <1576880825-15010-13-git-send-email-skomatineni@nvidia.com> <20191225175736.GC27497@sirena.org.uk> <856d8a92-0c24-6722-952c-06b86c706e97@gmail.com> <576fca44-6734-5431-b523-512747a0bf12@gmail.com> <0f956c12-98cf-3466-f3ee-ffc3a23e3e3d@gmail.com> <9e08a1eb-7a9f-0ed0-a2f4-08f1d6f93958@nvidia.com> <359bf92f-2e67-28b2-8412-9e47bb24ffdb@nvidia.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <359bf92f-2e67-28b2-8412-9e47bb24ffdb@nvidia.com> Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org To: Sameer Pujar , Dmitry Osipenko , Mark Brown Cc: thierry.reding@gmail.com, jonathanh@nvidia.com, lgirdwood@gmail.com, perex@perex.cz, tiwai@suse.com, mperttunen@nvidia.com, gregkh@linuxfoundation.org, sboyd@kernel.org, robh+dt@kernel.org, mark.rutland@arm.com, pdeschrijver@nvidia.com, pgaikwad@nvidia.com, josephl@nvidia.com, daniel.lezcano@linaro.org, mmaddireddy@nvidia.com, markz@nvidia.com, devicetree@vger.kernel.org, linux-clk@vger.kernel.org, linux-tegra@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: linux-tegra@vger.kernel.org On 1/5/20 8:21 PM, Sameer Pujar wrote: > > On 1/5/2020 10:33 AM, Sowjanya Komatineni wrote: >> >> On 1/4/20 5:05 PM, Dmitry Osipenko wrote: >>> 04.01.2020 08:49, Sowjanya Komatineni =D0=BF=D0=B8=D1=88=D0=B5=D1=82: >>>> On 1/2/20 8:12 AM, Dmitry Osipenko wrote: >>>>> 02.01.2020 10:03, Sowjanya Komatineni =D0=BF=D0=B8=D1=88=D0=B5=D1=82: >>>>>> On 12/27/19 1:19 PM, Sowjanya Komatineni wrote: >>>>>>> On 12/27/19 6:56 AM, Dmitry Osipenko wrote: >>>>>>>> 25.12.2019 20:57, Mark Brown =D0=BF=D0=B8=D1=88=D0=B5=D1=82: >>>>>>>>> On Mon, Dec 23, 2019 at 12:14:34AM +0300, Dmitry Osipenko wrote: >>>>>>>>>> 21.12.2019 01:26, Sowjanya Komatineni =D0=BF=D0=B8=D1=88=D0=B5= =D1=82: >>>>>>>>>>> Tegra PMC clock clk_out_1 is dedicated for audio mclk from=20 >>>>>>>>>>> Tegra30 >>>>>>>>>>> through Tegra210 and currently Tegra clock driver does=20 >>>>>>>>>>> initial parent >>>>>>>>>>> configuration for audio mclk "clk_out_1" and enables them by=20 >>>>>>>>>>> default. >>>>>>>>> Please delete unneeded context from mails when replying.=C2=A0=20 >>>>>>>>> Doing this >>>>>>>>> makes it much easier to find your reply in the message,=20 >>>>>>>>> helping ensure >>>>>>>>> it won't be missed by people scrolling through the irrelevant=20 >>>>>>>>> quoted >>>>>>>>> material. >>>>>>>> Ok >>>>>>>> >>>>>>>>>>> - clk_disable_unprepare(data->clk_cdev1); >>>>>>>>>>> - clk_disable_unprepare(data->clk_pll_a_out0); >>>>>>>>>>> - clk_disable_unprepare(data->clk_pll_a); >>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 if (__clk_is_enabled(data->clk_cdev1)) >>>>>>>>>>> + clk_disable_unprepare(data->clk_cdev1); >>>>>>>>>> The root of the problem is that you removed clocks enabling from >>>>>>>>>> tegra_asoc_utils_init(). >>>>>>> currently, audio mclk and its parent clocks enabling are from clock >>>>>>> driver init and not from tegra_asoc_utils_init. >>>>>>>>>> I'm not sure why clocks should be disabled during the=20 >>>>>>>>>> rate-changing, >>>>>>>>>> probably this action is not really needed. >>>>>>>>> I know nothing about this particular device but this is not that >>>>>>>>> unusual a restriction for audio hardware, you often can't >>>>>>>>> robustly reconfigure the clocking for a device while it's active >>>>>>>>> due to issues in the hardware.=C2=A0 You often see issues with FI= FOs >>>>>>>>> glitching or state machines getting stuck. This may not be an >>>>>>>>> issue here but if it's something that's documented as a >>>>>>>>> requirement it's probably good to pay attention. >>>>>>>> I don't know details about that hardware either, maybe it is=20 >>>>>>>> simply not >>>>>>>> safe to change PLL_A rate dynamically and then=20 >>>>>>>> CLK_SET_RATE_GATE could >>>>>>>> be used. >>>>>>>> >>>>>>>> If nobody knows for sure, then will be better to keep >>>>>>>> tegra_asoc_utils_set_rate() unchanged. >>>>>>> plla rate change through tegra_asoc_utils_set_rate() happens=20 >>>>>>> only when >>>>>>> there is not active playback or record corresponding to this sound >>>>>>> device. >>>>>>> >>>>>>> So, I don't see reason for disabling clock during rate change=20 >>>>>>> and not >>>>>>> sure why we had this from the beginning. >>>>>>> >>>>>>> Thierry/Sameer, >>>>>>> >>>>>>> Can you please comment? >>>>>>> >>>>>>> Yes, we can use CLK_SET_RATE_GATE for PLLA and remove clock=20 >>>>>>> disabling >>>>>>> before rate change. >>>>>>> >>>>>> PLLA is used for both I2S controller clock and also for audio=20 >>>>>> mclk. I2S >>>>>> driver suspend resume implementations takes care of enabling and >>>>>> disabling I2S clock but audio mclk will be enabled during that=20 >>>>>> time and >>>>>> PLLA disable might not happen. So using CLK_SET_RATE_GATE=20 >>>>>> prevents rate >>>>>> change to happen and as rate change happens only when there is no=20 >>>>>> active >>>>>> audio record/playback, we can perform rate change without=20 >>>>>> disable/enable >>>>>> during rate change. >>>>>> >>>>>> So probably below changes should be good. >>>>>> >>>>>> =C2=A0=C2=A0 * remove asoc_utils_set_rate call from asoc_utils_init = as=20 >>>>>> set_rate >>>>>> =C2=A0=C2=A0=C2=A0=C2=A0 happens during existing hw_params callback = implementations=20 >>>>>> in sound >>>>>> =C2=A0=C2=A0=C2=A0=C2=A0 drivers and there is no need to do rate cha= nge during=20 >>>>>> asoc_utils_init. >>>>>> =C2=A0=C2=A0 * remove disable/enable clocks during rate change in=20 >>>>>> asoc_utils_set_rate. >>>>>> =C2=A0=C2=A0 * add startup and shutdown snd_soc_ops callbacks to do = enable and >>>>>> =C2=A0=C2=A0=C2=A0=C2=A0 disable audio mclk. >>>>>> >>>>> Sounds good, thanks. I'll be happy to review and test it. >>>> Regarding disabling audio mclk during PLLA rate change, no need to >>>> explicitly disable PLLA on asoc utils as clock driver takes care of it >>>> properly during pll rate change. >>>> >>>> But the downstream clock divider hardware can malfunction without >>>> recovery when subject to unstable PLL output during locking, unless >>>> clock is gated. >>>> >>>> So it is recommended to disable downstream clocks during PLL rate=20 >>>> change. >>>> >>>> PLLA downstream clocks are I2S and audio mclk (cdev1/clk1 and extern1 >>>> clocks) and I2S clock is disabled in I2S driver by PM runtime ops. >>> The I2S driver uses asynchronous pm_runtime_put() and thus there is no >>> guarantee that I2S clock is disabled at the time of changing PLLA rate. >>> Could this be a problem? >> Looking into soc_pcm_hw_params, I see dai_link hw_params ops happens=20 >> prior to=C2=A0 platform snd_soc_dai_driver hw_params ops. > > This is true. > >> >> So, PLL rate change thru asoc_utils_set_rate happens during sample=20 >> rate config of dai_link hw_params ops and during this time I2S will=20 >> always be in idle state. > > This is probably not the case, since runtime resume for I2S would have=20 > already enabled the clock for I2S and in turn PLLA. The hw_param()=20 > call would happen later. > We could have used a fixed clock rate for PLLA, but the reason why we=20 > are setting the rate at runtime is, we support sample rates (and=20 > multuples) of 8kHz and 11.025kHz. > Both of these require a different PLLA base rate for downstream clock=20 > dividers to work properly. That is why, I think we have two base rates=20 > for PLLA. > > Even if we want to enable the clocks (for i2s) in hw_param(), this=20 > still may not help. > For example there could be multiple I2S instances, which can use the=20 > same PLLA source. ALSA playback/capture on different I2S can run=20 > independently. > Hence we are not sure if clk_disable_unprepare() in tegra_asoc_utils.c=20 > would actually disable PLLA. Hence I think the problem exists in=20 > current code too. clk_disable_unprepare in aosc_utils_set_rate will not disable PLLA as it=20 will be in use by other consumer (I2S). But clock driver it self takes care of disabling pll output and keeping=20 it in bypass state so its not really like PLLA output is off. So regarding PLLA rate change, we dont have to explicitly disable in=20 audio driver as pll clock driver takes care during rate update, But consumer clocks of PLLA need to be disabled during rate update so=20 existing dividers doesn't cause any malfunction with new rate update. audio mclk can will be in disabled state during rate update thru=20 hw_params with addition of shutdown snd_soc_ops callbacks to disable of=20 audio mclk. But issue is for I2S clock where clock might be in enabled state and=20 this is issue with current I2S driver already. Sameer/Dmitry, Making sure I2S clock is in disabled state during rate update is=20 something need to work thru for proper fix and this is not related to=20 this patch series as this issue exists with current upstream. So, can we take care of this as separate patch out of this series so I=20 can get this series out as this PMC clock changes are needed for=20 upcoming camera drivers. > > We really want to allow user to run any sample rate in the supported=20 > range. However the sample rate is known during hw_param() callback. > > Looking at current discussion, we may have to provide an aternate way=20 > of switching PLLA base rate (may be not in ALSA callbacks) > >> >> Sameer, Please confirm. >> >>>> For audio mclk, need to make sure mclk are disabled during rate=20 >>>> change. >>>> >>>> So below are the changes to audio clocks that will be in next version. >>>> >>>> =C2=A0=C2=A0 * remove tegra_asoc_utils_set_rate call from=20 >>>> tegra_asoc_utils_init as >>>> =C2=A0=C2=A0=C2=A0=C2=A0 tegra_asoc_utils_set_rate happens during hw_p= arams callback. >>>> =C2=A0=C2=A0 * add shutdown snd_soc_ops callbacks to disable of audio = mclk. >>>> =C2=A0=C2=A0 * remove disable audio mclk (cdev1) and plla clocks prior= to rate >>>> =C2=A0=C2=A0=C2=A0=C2=A0 change in tegra_asoc_utils_set_rate (as audio= mclk will always=20 >>>> be in >>>> =C2=A0=C2=A0=C2=A0=C2=A0 disabled state every time hw_param callback g= ets executed) and=20 >>>> keep >>>> =C2=A0=C2=A0=C2=A0=C2=A0 audio mclk enable after the rate change in=20 >>>> tegra_asoc_utils_set_rate. >>>>