| Message ID | 20240221-rk3308-audio-codec-v3-1-dfa34abfcef6@bootlin.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Add support for the internal RK3308 audio codec | expand |
On Mittwoch, 21. Februar 2024 11:22:45 CET Luca Ceresoli wrote: > The sample rates set by the rockchip_i2s_tdm driver in master mode are > inaccurate up to 5% in several cases, due to the driver logic to configure > clocks and a nasty interaction with the Common Clock Framework. > > To understand what happens, here is the relevant section of the clock tree > (slightly simplified), along with the names used in the driver: > > vpll0 _OR_ vpll1 "mclk_root" > clk_i2s2_8ch_tx_src "mclk_parent" > clk_i2s2_8ch_tx_mux > clk_i2s2_8ch_tx "mclk" or "mclk_tx" > > This is what happens when playing back e.g. at 192 kHz using > audio-graph-card (when recording the same applies, only s/tx/rx/): > > 0. at probe, rockchip_i2s_tdm_set_sysclk() stores the passed frequency in > i2s_tdm->mclk_tx_freq (*) which is 50176000, and that is never modified > afterwards > > 1. when playback is started, rockchip_i2s_tdm_hw_params() is called and > does the following two calls > > 2. rockchip_i2s_tdm_calibrate_mclk(): > > 2a. selects mclk_root0 (vpll0) as a parent for mclk_parent > (mclk_tx_src), which is OK because the vpll0 rate is a good for > 192000 (and sumbultiple) rates > > 2b. sets the mclk_root frequency based on ppm calibration computations > > 2c. sets mclk_tx_src to 49152000 (= 256 * 192000), which is also OK as > it is a multiple of the required bit clock > > 3. rockchip_i2s_tdm_set_mclk() > > 3a. calls clk_set_rate() to set the rate of mclk_tx (clk_i2s2_8ch_tx) > to the value of i2s_tdm->mclk_tx_freq (*), i.e. 50176000 which is > not a multiple of the sampling frequency -- this is not OK > > 3a1. clk_set_rate() reacts by reparenting clk_i2s2_8ch_tx_src to > vpll1 -- this is not OK because the default vpll1 rate can be > divided to get 44.1 kHz and related rates, not 192 kHz > > The result is that the driver does a lot of ad-hoc decisions about clocks > and ends up in using the wrong parent at an unoptimal rate. > > Step 0 is one part of the problem: unless the card driver calls set_sysclk > at each stream start, whatever rate is set in mclk_tx_freq during boot will > be taken and used until reboot. Moreover the driver does not care if its > value is not a multiple of any audio frequency. > > Another part of the problem is that the whole reparenting and clock rate > setting logic is conflicting with the CCF algorithms to achieve largely the > same goal: selecting the best parent and setting the closest clock > rate. And it turns out that only calling once clk_set_rate() on > clk_i2s2_8ch_tx picks the correct vpll and sets the correct rate. > > The fix is based on removing the custom logic in the driver to select the > parent and set the various clocks, and just let the Clock Framework do it > all. As a side effect, the set_sysclk() op becomes useless because we now > let the CCF compute the appropriate value for the sampling rate. It also > implies that the whole calibration logic is now dead code and so it is > removed along with the "PCM Clock Compensation in PPM" kcontrol, which has > always been broken anyway. The handling of the 4 optional clocks also > becomes dead code and is removed. > > The actual rates have been tested playing 30 seconds of audio at various > sampling rates before and after this change using sox: > > time play -r <sample_rate> -n synth 30 sine 950 gain -3 > > The time reported in the table below is the 'real' value reported by the > 'time' command in the above command line. > > rate before after > --------- ------ ------ > 8000 Hz 30.60s 30.63s > 11025 Hz 30.45s 30.51s > 16000 Hz 30.47s 30.50s > 22050 Hz 30.78s 30.41s > 32000 Hz 31.02s 30.43s > 44100 Hz 30.78s 30.41s > 48000 Hz 29.81s 30.45s > 88200 Hz 30.78s 30.41s > 96000 Hz 29.79s 30.42s > 176400 Hz 27.40s 30.41s > 192000 Hz 29.79s 30.42s > > While the tests are running the clock tree confirms that: > > * without the patch, vpll1 is always used and clk_i2s2_8ch_tx always > produces 50176000 Hz, which cannot be divided for most audio rates > except the slowest ones, generating inaccurate rates > * with the patch: > - for 192000 Hz vpll0 is used > - for 176400 Hz vpll1 is used > - clk_i2s2_8ch_tx always produces (256 * <rate>) Hz > > Fixes: 081068fd6414 ("ASoC: rockchip: add support for i2s-tdm controller") > Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> > > --- > > I don't know why such a complex yet incorrect logic is present, but it is > possibly a remnant of the downstream 4.19 driver > (https://github.com/rockchip-linux/kernel/blob/develop-4.19/sound/soc/rockchip/rockchip_i2s_tdm.c). > It might be that the CCF was not yet capable of optimally reparenting back > in 4.19, so it did make sense back then. > > <snip> Hello, thank you for your patch series fixing this driver. Since I am the person currently listed as the maintainer of it and was the person who originally upstreamed it, I think I can provide some insight. The mclk calculation and reparenting code indeed originates from the downstream driver. As for why that driver did these calculations itself, I do not know, but I do know why I did not remove them and rely on the CCF to do the reparenting and such for me. For starters, I did not know that the DAI and CCF code have functionality to automatically reparent clocks based on the audio rate. The reason I did not know this is because after looking at the official kernel documentation at [1], I did not see any mention of this. Similarly, the documentation for the CCF at [2] also makes no mention of this feature. In general, guidance for how to properly write a DAI driver in the kernel documentation is sparse, though I am extremely grateful to broonie for his patience in answering some of my questions I had at the time. The input of any of the reviewers at the time, including that of Rockchip's engineer who wrote the downstream driver, made it a lot better than it initially was. Had somebody brought up the existence of this functionality during the reviews of this driver's original submission, I would of course have fixed this. I do not blame the reviewers or maintainers, as they have enough on their plate as-is and don't owe me anything. But I am trying to explain the circumstances here that lead to mclk being slightly broken for this driver. Naturally, now that I know that this functionality does exist, I can do a web search (since in-tree documentation search was fruitless) for "linux common clock framework reparenting", which as its second result brings up a PDF of the slides of a talk by one of your co-workers.[3] The slides do not contain the string "reparent", though they do tell me that I am able to purchase Linux kernel development training from your employer. What I'm trying to say here is that I still don't know how the DAI reparenting mclk through the CCF works, and I'd probably need to dig into the implementation to even know that it does this. Furthermore, your employer has an economic incentive to keep this information out of the in-tree documentation. I hope that clears up any confusion as to "why such a complex yet incorrect logic is present", and what ultimately lead you to having to write this patch. As for the patch itself, I'll hopefully get around to testing it this week on an RK3566 board, as your cover letter does not list the platforms this was tested for regressions on, so I assume you're just using RK3308. The RK356x has some linked clock things going on which mainline does not yet model properly as far as I know (which is why some video related clocks are marked as critical, and why things like [4] are needed for MIPI-DSI output to work. So any changes that might fiddle with the parent clocks I'd like to validate. If I don't get around to it in a timely manner however, I'm fine with this patch as-is. Regards, Nicolas Frattaroli [1]: https://docs.kernel.org/sound/index.html [2]: https://docs.kernel.org/driver-api/clk.html [3]: https://elinux.org/images/b/b8/Elc2013_Clement.pdf [4]: https://github.com/CounterPillow/overlay-examples/blob/main/quartz64a/pine64-lcd-panel.dts#L56-L62
Hello Nicolas, On Mon, 04 Mar 2024 19:40:52 +0100 Nicolas Frattaroli <frattaroli.nicolas@gmail.com> wrote: [...] > Hello, > > thank you for your patch series fixing this driver. Since I am the person > currently listed as the maintainer of it and was the person who originally > upstreamed it, I think I can provide some insight. > > The mclk calculation and reparenting code indeed originates from the > downstream driver. As for why that driver did these calculations itself, > I do not know, but I do know why I did not remove them and rely on the > CCF to do the reparenting and such for me. > > For starters, I did not know that the DAI and CCF code have functionality > to automatically reparent clocks based on the audio rate. The reason I did > not know this is because after looking at the official kernel documentation > at [1], I did not see any mention of this. > > Similarly, the documentation for the CCF at [2] also makes no mention of > this feature. In general, guidance for how to properly write a DAI driver > in the kernel documentation is sparse, though I am extremely grateful to > broonie for his patience in answering some of my questions I had at the > time. The input of any of the reviewers at the time, including that of > Rockchip's engineer who wrote the downstream driver, made it a lot better > than it initially was. > > Had somebody brought up the existence of this functionality during the > reviews of this driver's original submission, I would of course have > fixed this. I do not blame the reviewers or maintainers, as they have > enough on their plate as-is and don't owe me anything. But I am trying > to explain the circumstances here that lead to mclk being slightly broken > for this driver. > > Naturally, now that I know that this functionality does exist, I can do > a web search (since in-tree documentation search was fruitless) for > "linux common clock framework reparenting", which as its second result > brings up a PDF of the slides of a talk by one of your co-workers.[3] The > slides do not contain the string "reparent", though they do tell me that > I am able to purchase Linux kernel development training from your > employer. > > What I'm trying to say here is that I still don't know how the DAI > reparenting mclk through the CCF works, and I'd probably need to dig > into the implementation to even know that it does this. Furthermore, > your employer has an economic incentive to keep this information out > of the in-tree documentation. I am very surprised by your opinion about our contributions, as we are trying to do exactly the opposite, being active members of the community and giving back in various ways. We are trying to contribute documentation and share knowledge as well: we release all of our training materials as Creative Commons (source code included), we contribute to the Buildroot and Yocto documentation and one of our colleagues is the maintainer of the Yocto Project documentation. We also wrote and/or maintain "practical working code examples" such as the simplest-yocto-setup [1] we recently published and the zynqmp-pmufw-builder [2], and published some video tutorials (I remember about one on device tree that seems very appreciated). [1] https://github.com/bootlin/simplest-yocto-setup/ [2] https://github.com/lucaceresoli/zynqmp-pmufw-builder/ We also contributed to the Documentation directory of the Linux kernel tree, as git can show: git shortlog --no-merges -s \ --author=free-electrons.com --author=bootlin.com \ -- Documentation/ ':!Documentation/devicetree/bindings/' Sure, we could do better, but I think we are definitely doing something. About the CCF specifically, it is surely one of the areas of the kernel that has insufficient documentation. As a matter of fact, while some areas are very well documented, for others the only real documentation is the code. I don't like this fact, some contributions certainly help but documenting a complex subsystem written by others is definitely a very relevant effort. However I don't think the sound documentation should describe clock reparenting: it is really a CCF feature that is available to any in-kernel clock user, i.e. pretty much every subsystem. It should definitely be described in the CCF docs, sure. > I hope that clears up any confusion as to "why such a complex yet > incorrect logic is present", and what ultimately lead you to having > to write this patch. Thanks for taking time to explain. This is very useful as there might have been reasons I didn't suspect for the code being as it is. The fact that you based your work on the downstream vendor kernel (as I did for the codec driver in this series) is probably the reason. Not only because the code quality of vendor kernels is sometimes considerably lower than the mainline kernel, but also because that driver might have been written when the reparenting logic in the CCF was not yet as complete and effective as it is today, so chances are the manual reparenting was correct initially. Also, the incorrect sampling rate is barely noticeable at most sampling rates up to 96 kHz, so I'm not blaming anybody for having introduced a bug that is not visible in all use cases. The only way to not introduce any bugs is not writing any code. > As for the patch itself, I'll hopefully get around to testing it this > week on an RK3566 board, as your cover letter does not list the platforms > this was tested for regressions on, so I assume you're just using RK3308. Your testing would be very useful! And yes, I developed and tested my patch on RK3308. Good point, added now to my commit log so it will appear in v4. > The RK356x has some linked clock things going on which mainline does not > yet model properly as far as I know (which is why some video related clocks > are marked as critical, and why things like [4] are needed for MIPI-DSI > output to work. So any changes that might fiddle with the parent clocks > I'd like to validate. Right, video clocks are tricky to get right and I've often seen them require manual hard-coding in the device tree, unfortunately. > If I don't get around to it in a timely manner > however, I'm fine with this patch as-is. Cool, thanks. Best regards, Luca
diff --git a/sound/soc/rockchip/rockchip_i2s_tdm.c b/sound/soc/rockchip/rockchip_i2s_tdm.c index 860e66ec85e8..9fa020ef7eab 100644 --- a/sound/soc/rockchip/rockchip_i2s_tdm.c +++ b/sound/soc/rockchip/rockchip_i2s_tdm.c @@ -25,8 +25,6 @@ #define DEFAULT_MCLK_FS 256 #define CH_GRP_MAX 4 /* The max channel 8 / 2 */ #define MULTIPLEX_CH_MAX 10 -#define CLK_PPM_MIN -1000 -#define CLK_PPM_MAX 1000 #define TRCM_TXRX 0 #define TRCM_TX 1 @@ -53,20 +51,6 @@ struct rk_i2s_tdm_dev { struct clk *hclk; struct clk *mclk_tx; struct clk *mclk_rx; - /* The mclk_tx_src is parent of mclk_tx */ - struct clk *mclk_tx_src; - /* The mclk_rx_src is parent of mclk_rx */ - struct clk *mclk_rx_src; - /* - * The mclk_root0 and mclk_root1 are root parent and supplies for - * the different FS. - * - * e.g: - * mclk_root0 is VPLL0, used for FS=48000Hz - * mclk_root1 is VPLL1, used for FS=44100Hz - */ - struct clk *mclk_root0; - struct clk *mclk_root1; struct regmap *regmap; struct regmap *grf; struct snd_dmaengine_dai_dma_data capture_dma_data; @@ -76,19 +60,11 @@ struct rk_i2s_tdm_dev { const struct rk_i2s_soc_data *soc_data; bool is_master_mode; bool io_multiplex; - bool mclk_calibrate; bool tdm_mode; - unsigned int mclk_rx_freq; - unsigned int mclk_tx_freq; - unsigned int mclk_root0_freq; - unsigned int mclk_root1_freq; - unsigned int mclk_root0_initial_freq; - unsigned int mclk_root1_initial_freq; unsigned int frame_width; unsigned int clk_trcm; unsigned int i2s_sdis[CH_GRP_MAX]; unsigned int i2s_sdos[CH_GRP_MAX]; - int clk_ppm; int refcount; spinlock_t lock; /* xfer lock */ bool has_playback; @@ -114,12 +90,6 @@ static void i2s_tdm_disable_unprepare_mclk(struct rk_i2s_tdm_dev *i2s_tdm) { clk_disable_unprepare(i2s_tdm->mclk_tx); clk_disable_unprepare(i2s_tdm->mclk_rx); - if (i2s_tdm->mclk_calibrate) { - clk_disable_unprepare(i2s_tdm->mclk_tx_src); - clk_disable_unprepare(i2s_tdm->mclk_rx_src); - clk_disable_unprepare(i2s_tdm->mclk_root0); - clk_disable_unprepare(i2s_tdm->mclk_root1); - } } /** @@ -142,29 +112,9 @@ static int i2s_tdm_prepare_enable_mclk(struct rk_i2s_tdm_dev *i2s_tdm) ret = clk_prepare_enable(i2s_tdm->mclk_rx); if (ret) goto err_mclk_rx; - if (i2s_tdm->mclk_calibrate) { - ret = clk_prepare_enable(i2s_tdm->mclk_tx_src); - if (ret) - goto err_mclk_rx; - ret = clk_prepare_enable(i2s_tdm->mclk_rx_src); - if (ret) - goto err_mclk_rx_src; - ret = clk_prepare_enable(i2s_tdm->mclk_root0); - if (ret) - goto err_mclk_root0; - ret = clk_prepare_enable(i2s_tdm->mclk_root1); - if (ret) - goto err_mclk_root1; - } return 0; -err_mclk_root1: - clk_disable_unprepare(i2s_tdm->mclk_root0); -err_mclk_root0: - clk_disable_unprepare(i2s_tdm->mclk_rx_src); -err_mclk_rx_src: - clk_disable_unprepare(i2s_tdm->mclk_tx_src); err_mclk_rx: clk_disable_unprepare(i2s_tdm->mclk_tx); err_mclk_tx: @@ -564,159 +514,6 @@ static void rockchip_i2s_tdm_xfer_resume(struct snd_pcm_substream *substream, I2S_XFER_RXS_START); } -static int rockchip_i2s_tdm_clk_set_rate(struct rk_i2s_tdm_dev *i2s_tdm, - struct clk *clk, unsigned long rate, - int ppm) -{ - unsigned long rate_target; - int delta, ret; - - if (ppm == i2s_tdm->clk_ppm) - return 0; - - if (ppm < 0) - delta = -1; - else - delta = 1; - - delta *= (int)div64_u64((u64)rate * (u64)abs(ppm) + 500000, - 1000000); - - rate_target = rate + delta; - - if (!rate_target) - return -EINVAL; - - ret = clk_set_rate(clk, rate_target); - if (ret) - return ret; - - i2s_tdm->clk_ppm = ppm; - - return 0; -} - -static int rockchip_i2s_tdm_calibrate_mclk(struct rk_i2s_tdm_dev *i2s_tdm, - struct snd_pcm_substream *substream, - unsigned int lrck_freq) -{ - struct clk *mclk_root; - struct clk *mclk_parent; - unsigned int mclk_root_freq; - unsigned int mclk_root_initial_freq; - unsigned int mclk_parent_freq; - unsigned int div, delta; - u64 ppm; - int ret; - - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) - mclk_parent = i2s_tdm->mclk_tx_src; - else - mclk_parent = i2s_tdm->mclk_rx_src; - - switch (lrck_freq) { - case 8000: - case 16000: - case 24000: - case 32000: - case 48000: - case 64000: - case 96000: - case 192000: - mclk_root = i2s_tdm->mclk_root0; - mclk_root_freq = i2s_tdm->mclk_root0_freq; - mclk_root_initial_freq = i2s_tdm->mclk_root0_initial_freq; - mclk_parent_freq = DEFAULT_MCLK_FS * 192000; - break; - case 11025: - case 22050: - case 44100: - case 88200: - case 176400: - mclk_root = i2s_tdm->mclk_root1; - mclk_root_freq = i2s_tdm->mclk_root1_freq; - mclk_root_initial_freq = i2s_tdm->mclk_root1_initial_freq; - mclk_parent_freq = DEFAULT_MCLK_FS * 176400; - break; - default: - dev_err(i2s_tdm->dev, "Invalid LRCK frequency: %u Hz\n", - lrck_freq); - return -EINVAL; - } - - ret = clk_set_parent(mclk_parent, mclk_root); - if (ret) - return ret; - - ret = rockchip_i2s_tdm_clk_set_rate(i2s_tdm, mclk_root, - mclk_root_freq, 0); - if (ret) - return ret; - - delta = abs(mclk_root_freq % mclk_parent_freq - mclk_parent_freq); - ppm = div64_u64((uint64_t)delta * 1000000, (uint64_t)mclk_root_freq); - - if (ppm) { - div = DIV_ROUND_CLOSEST(mclk_root_initial_freq, mclk_parent_freq); - if (!div) - return -EINVAL; - - mclk_root_freq = mclk_parent_freq * round_up(div, 2); - - ret = clk_set_rate(mclk_root, mclk_root_freq); - if (ret) - return ret; - - i2s_tdm->mclk_root0_freq = clk_get_rate(i2s_tdm->mclk_root0); - i2s_tdm->mclk_root1_freq = clk_get_rate(i2s_tdm->mclk_root1); - } - - return clk_set_rate(mclk_parent, mclk_parent_freq); -} - -static int rockchip_i2s_tdm_set_mclk(struct rk_i2s_tdm_dev *i2s_tdm, - struct snd_pcm_substream *substream, - struct clk **mclk) -{ - unsigned int mclk_freq; - int ret; - - if (i2s_tdm->clk_trcm) { - if (i2s_tdm->mclk_tx_freq != i2s_tdm->mclk_rx_freq) { - dev_err(i2s_tdm->dev, - "clk_trcm, tx: %d and rx: %d should be the same\n", - i2s_tdm->mclk_tx_freq, - i2s_tdm->mclk_rx_freq); - return -EINVAL; - } - - ret = clk_set_rate(i2s_tdm->mclk_tx, i2s_tdm->mclk_tx_freq); - if (ret) - return ret; - - ret = clk_set_rate(i2s_tdm->mclk_rx, i2s_tdm->mclk_rx_freq); - if (ret) - return ret; - - /* mclk_rx is also ok. */ - *mclk = i2s_tdm->mclk_tx; - } else { - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) { - *mclk = i2s_tdm->mclk_tx; - mclk_freq = i2s_tdm->mclk_tx_freq; - } else { - *mclk = i2s_tdm->mclk_rx; - mclk_freq = i2s_tdm->mclk_rx_freq; - } - - ret = clk_set_rate(*mclk, mclk_freq); - if (ret) - return ret; - } - - return 0; -} - static int rockchip_i2s_ch_to_io(unsigned int ch, bool substream_capture) { if (substream_capture) { @@ -853,19 +650,17 @@ static int rockchip_i2s_tdm_hw_params(struct snd_pcm_substream *substream, struct snd_soc_dai *dai) { struct rk_i2s_tdm_dev *i2s_tdm = to_info(dai); - struct clk *mclk; - int ret = 0; unsigned int val = 0; unsigned int mclk_rate, bclk_rate, div_bclk = 4, div_lrck = 64; + int err; if (i2s_tdm->is_master_mode) { - if (i2s_tdm->mclk_calibrate) - rockchip_i2s_tdm_calibrate_mclk(i2s_tdm, substream, - params_rate(params)); + struct clk *mclk = (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) ? + i2s_tdm->mclk_tx : i2s_tdm->mclk_rx; - ret = rockchip_i2s_tdm_set_mclk(i2s_tdm, substream, &mclk); - if (ret) - return ret; + err = clk_set_rate(mclk, DEFAULT_MCLK_FS * params_rate(params)); + if (err) + return err; mclk_rate = clk_get_rate(mclk); bclk_rate = i2s_tdm->frame_width * params_rate(params); @@ -973,96 +768,6 @@ static int rockchip_i2s_tdm_trigger(struct snd_pcm_substream *substream, return 0; } -static int rockchip_i2s_tdm_set_sysclk(struct snd_soc_dai *cpu_dai, int stream, - unsigned int freq, int dir) -{ - struct rk_i2s_tdm_dev *i2s_tdm = to_info(cpu_dai); - - /* Put set mclk rate into rockchip_i2s_tdm_set_mclk() */ - if (i2s_tdm->clk_trcm) { - i2s_tdm->mclk_tx_freq = freq; - i2s_tdm->mclk_rx_freq = freq; - } else { - if (stream == SNDRV_PCM_STREAM_PLAYBACK) - i2s_tdm->mclk_tx_freq = freq; - else - i2s_tdm->mclk_rx_freq = freq; - } - - dev_dbg(i2s_tdm->dev, "The target mclk_%s freq is: %d\n", - stream ? "rx" : "tx", freq); - - return 0; -} - -static int rockchip_i2s_tdm_clk_compensation_info(struct snd_kcontrol *kcontrol, - struct snd_ctl_elem_info *uinfo) -{ - uinfo->type = SNDRV_CTL_ELEM_TYPE_INTEGER; - uinfo->count = 1; - uinfo->value.integer.min = CLK_PPM_MIN; - uinfo->value.integer.max = CLK_PPM_MAX; - uinfo->value.integer.step = 1; - - return 0; -} - -static int rockchip_i2s_tdm_clk_compensation_get(struct snd_kcontrol *kcontrol, - struct snd_ctl_elem_value *ucontrol) -{ - struct snd_soc_dai *dai = snd_kcontrol_chip(kcontrol); - struct rk_i2s_tdm_dev *i2s_tdm = snd_soc_dai_get_drvdata(dai); - - ucontrol->value.integer.value[0] = i2s_tdm->clk_ppm; - - return 0; -} - -static int rockchip_i2s_tdm_clk_compensation_put(struct snd_kcontrol *kcontrol, - struct snd_ctl_elem_value *ucontrol) -{ - struct snd_soc_dai *dai = snd_kcontrol_chip(kcontrol); - struct rk_i2s_tdm_dev *i2s_tdm = snd_soc_dai_get_drvdata(dai); - int ret = 0, ppm = 0; - int changed = 0; - unsigned long old_rate; - - if (ucontrol->value.integer.value[0] < CLK_PPM_MIN || - ucontrol->value.integer.value[0] > CLK_PPM_MAX) - return -EINVAL; - - ppm = ucontrol->value.integer.value[0]; - - old_rate = clk_get_rate(i2s_tdm->mclk_root0); - ret = rockchip_i2s_tdm_clk_set_rate(i2s_tdm, i2s_tdm->mclk_root0, - i2s_tdm->mclk_root0_freq, ppm); - if (ret) - return ret; - if (old_rate != clk_get_rate(i2s_tdm->mclk_root0)) - changed = 1; - - if (clk_is_match(i2s_tdm->mclk_root0, i2s_tdm->mclk_root1)) - return changed; - - old_rate = clk_get_rate(i2s_tdm->mclk_root1); - ret = rockchip_i2s_tdm_clk_set_rate(i2s_tdm, i2s_tdm->mclk_root1, - i2s_tdm->mclk_root1_freq, ppm); - if (ret) - return ret; - if (old_rate != clk_get_rate(i2s_tdm->mclk_root1)) - changed = 1; - - return changed; -} - -static struct snd_kcontrol_new rockchip_i2s_tdm_compensation_control = { - .iface = SNDRV_CTL_ELEM_IFACE_PCM, - .name = "PCM Clock Compensation in PPM", - .info = rockchip_i2s_tdm_clk_compensation_info, - .get = rockchip_i2s_tdm_clk_compensation_get, - .put = rockchip_i2s_tdm_clk_compensation_put, -}; - static int rockchip_i2s_tdm_dai_probe(struct snd_soc_dai *dai) { struct rk_i2s_tdm_dev *i2s_tdm = snd_soc_dai_get_drvdata(dai); @@ -1072,9 +777,6 @@ static int rockchip_i2s_tdm_dai_probe(struct snd_soc_dai *dai) if (i2s_tdm->has_playback) snd_soc_dai_dma_data_set_playback(dai, &i2s_tdm->playback_dma_data); - if (i2s_tdm->mclk_calibrate) - snd_soc_add_dai_controls(dai, &rockchip_i2s_tdm_compensation_control, 1); - return 0; } @@ -1115,7 +817,6 @@ static const struct snd_soc_dai_ops rockchip_i2s_tdm_dai_ops = { .probe = rockchip_i2s_tdm_dai_probe, .hw_params = rockchip_i2s_tdm_hw_params, .set_bclk_ratio = rockchip_i2s_tdm_set_bclk_ratio, - .set_sysclk = rockchip_i2s_tdm_set_sysclk, .set_fmt = rockchip_i2s_tdm_set_fmt, .set_tdm_slot = rockchip_dai_tdm_slot, .trigger = rockchip_i2s_tdm_trigger, @@ -1444,35 +1145,6 @@ static void rockchip_i2s_tdm_path_config(struct rk_i2s_tdm_dev *i2s_tdm, rockchip_i2s_tdm_tx_path_config(i2s_tdm, num); } -static int rockchip_i2s_tdm_get_calibrate_mclks(struct rk_i2s_tdm_dev *i2s_tdm) -{ - int num_mclks = 0; - - i2s_tdm->mclk_tx_src = devm_clk_get(i2s_tdm->dev, "mclk_tx_src"); - if (!IS_ERR(i2s_tdm->mclk_tx_src)) - num_mclks++; - - i2s_tdm->mclk_rx_src = devm_clk_get(i2s_tdm->dev, "mclk_rx_src"); - if (!IS_ERR(i2s_tdm->mclk_rx_src)) - num_mclks++; - - i2s_tdm->mclk_root0 = devm_clk_get(i2s_tdm->dev, "mclk_root0"); - if (!IS_ERR(i2s_tdm->mclk_root0)) - num_mclks++; - - i2s_tdm->mclk_root1 = devm_clk_get(i2s_tdm->dev, "mclk_root1"); - if (!IS_ERR(i2s_tdm->mclk_root1)) - num_mclks++; - - if (num_mclks < 4 && num_mclks != 0) - return -ENOENT; - - if (num_mclks == 4) - i2s_tdm->mclk_calibrate = 1; - - return 0; -} - static int rockchip_i2s_tdm_path_prepare(struct rk_i2s_tdm_dev *i2s_tdm, struct device_node *np, bool is_rx_path) @@ -1610,11 +1282,6 @@ static int rockchip_i2s_tdm_probe(struct platform_device *pdev) i2s_tdm->io_multiplex = of_property_read_bool(node, "rockchip,io-multiplex"); - ret = rockchip_i2s_tdm_get_calibrate_mclks(i2s_tdm); - if (ret) - return dev_err_probe(i2s_tdm->dev, ret, - "mclk-calibrate clocks missing"); - regs = devm_platform_get_and_ioremap_resource(pdev, 0, &res); if (IS_ERR(regs)) { return dev_err_probe(i2s_tdm->dev, PTR_ERR(regs), @@ -1667,13 +1334,6 @@ static int rockchip_i2s_tdm_probe(struct platform_device *pdev) goto err_disable_hclk; } - if (i2s_tdm->mclk_calibrate) { - i2s_tdm->mclk_root0_initial_freq = clk_get_rate(i2s_tdm->mclk_root0); - i2s_tdm->mclk_root1_initial_freq = clk_get_rate(i2s_tdm->mclk_root1); - i2s_tdm->mclk_root0_freq = i2s_tdm->mclk_root0_initial_freq; - i2s_tdm->mclk_root1_freq = i2s_tdm->mclk_root1_initial_freq; - } - pm_runtime_enable(&pdev->dev); regmap_update_bits(i2s_tdm->regmap, I2S_DMACR, I2S_DMACR_TDL_MASK,
The sample rates set by the rockchip_i2s_tdm driver in master mode are inaccurate up to 5% in several cases, due to the driver logic to configure clocks and a nasty interaction with the Common Clock Framework. To understand what happens, here is the relevant section of the clock tree (slightly simplified), along with the names used in the driver: vpll0 _OR_ vpll1 "mclk_root" clk_i2s2_8ch_tx_src "mclk_parent" clk_i2s2_8ch_tx_mux clk_i2s2_8ch_tx "mclk" or "mclk_tx" This is what happens when playing back e.g. at 192 kHz using audio-graph-card (when recording the same applies, only s/tx/rx/): 0. at probe, rockchip_i2s_tdm_set_sysclk() stores the passed frequency in i2s_tdm->mclk_tx_freq (*) which is 50176000, and that is never modified afterwards 1. when playback is started, rockchip_i2s_tdm_hw_params() is called and does the following two calls 2. rockchip_i2s_tdm_calibrate_mclk(): 2a. selects mclk_root0 (vpll0) as a parent for mclk_parent (mclk_tx_src), which is OK because the vpll0 rate is a good for 192000 (and sumbultiple) rates 2b. sets the mclk_root frequency based on ppm calibration computations 2c. sets mclk_tx_src to 49152000 (= 256 * 192000), which is also OK as it is a multiple of the required bit clock 3. rockchip_i2s_tdm_set_mclk() 3a. calls clk_set_rate() to set the rate of mclk_tx (clk_i2s2_8ch_tx) to the value of i2s_tdm->mclk_tx_freq (*), i.e. 50176000 which is not a multiple of the sampling frequency -- this is not OK 3a1. clk_set_rate() reacts by reparenting clk_i2s2_8ch_tx_src to vpll1 -- this is not OK because the default vpll1 rate can be divided to get 44.1 kHz and related rates, not 192 kHz The result is that the driver does a lot of ad-hoc decisions about clocks and ends up in using the wrong parent at an unoptimal rate. Step 0 is one part of the problem: unless the card driver calls set_sysclk at each stream start, whatever rate is set in mclk_tx_freq during boot will be taken and used until reboot. Moreover the driver does not care if its value is not a multiple of any audio frequency. Another part of the problem is that the whole reparenting and clock rate setting logic is conflicting with the CCF algorithms to achieve largely the same goal: selecting the best parent and setting the closest clock rate. And it turns out that only calling once clk_set_rate() on clk_i2s2_8ch_tx picks the correct vpll and sets the correct rate. The fix is based on removing the custom logic in the driver to select the parent and set the various clocks, and just let the Clock Framework do it all. As a side effect, the set_sysclk() op becomes useless because we now let the CCF compute the appropriate value for the sampling rate. It also implies that the whole calibration logic is now dead code and so it is removed along with the "PCM Clock Compensation in PPM" kcontrol, which has always been broken anyway. The handling of the 4 optional clocks also becomes dead code and is removed. The actual rates have been tested playing 30 seconds of audio at various sampling rates before and after this change using sox: time play -r <sample_rate> -n synth 30 sine 950 gain -3 The time reported in the table below is the 'real' value reported by the 'time' command in the above command line. rate before after --------- ------ ------ 8000 Hz 30.60s 30.63s 11025 Hz 30.45s 30.51s 16000 Hz 30.47s 30.50s 22050 Hz 30.78s 30.41s 32000 Hz 31.02s 30.43s 44100 Hz 30.78s 30.41s 48000 Hz 29.81s 30.45s 88200 Hz 30.78s 30.41s 96000 Hz 29.79s 30.42s 176400 Hz 27.40s 30.41s 192000 Hz 29.79s 30.42s While the tests are running the clock tree confirms that: * without the patch, vpll1 is always used and clk_i2s2_8ch_tx always produces 50176000 Hz, which cannot be divided for most audio rates except the slowest ones, generating inaccurate rates * with the patch: - for 192000 Hz vpll0 is used - for 176400 Hz vpll1 is used - clk_i2s2_8ch_tx always produces (256 * <rate>) Hz Fixes: 081068fd6414 ("ASoC: rockchip: add support for i2s-tdm controller") Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> --- I don't know why such a complex yet incorrect logic is present, but it is possibly a remnant of the downstream 4.19 driver (https://github.com/rockchip-linux/kernel/blob/develop-4.19/sound/soc/rockchip/rockchip_i2s_tdm.c). It might be that the CCF was not yet capable of optimally reparenting back in 4.19, so it did make sense back then. Changed in v3: - this patch is new in v3, and it replaces and supersedes the patch "ASoC: rockchip: i2s-tdm: Fix clk_id usage in .set_sysclk()" --- sound/soc/rockchip/rockchip_i2s_tdm.c | 352 +--------------------------------- 1 file changed, 6 insertions(+), 346 deletions(-)