From: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
To: Arnd Bergmann <arnd@kernel.org>,
Liam Girdwood <lgirdwood@gmail.com>,
Peter Ujfalusi <peter.ujfalusi@linux.intel.com>,
Bard Liao <yung-chuan.liao@linux.intel.com>,
Ranjani Sridharan <ranjani.sridharan@linux.intel.com>,
Daniel Baluta <daniel.baluta@nxp.com>,
Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>,
Kai Vehmanen <kai.vehmanen@linux.intel.com>,
Mark Brown <broonie@kernel.org>, Jaroslav Kysela <perex@perex.cz>,
Takashi Iwai <tiwai@suse.com>,
Nathan Chancellor <nathan@kernel.org>,
Nick Desaulniers <ndesaulniers@google.com>,
Bill Wendling <morbo@google.com>,
Justin Stitt <justinstitt@google.com>,
Brent Lu <brent.lu@intel.com>,
sound-open-firmware@alsa-project.org,
linux-sound@vger.kernel.org, linux-kernel@vger.kernel.org,
llvm@lists.linux.dev
Subject: Re: [PATCH] sound: sof: ioc4-topology: avoid extra dai_params copy
Date: Wed, 7 Aug 2024 10:37:31 +0200 [thread overview]
Message-ID: <250c63d7-d81e-49ea-ac8f-2e3496075f20@linux.intel.com> (raw)
In-Reply-To: <20240807080302.2372297-1-arnd@kernel.org>
On 8/7/24 10:02, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The snd_pcm_hw_params structure is really too large to fit on the
> stack. Because of the way that clang inlines functions, it ends up
> twice in one function, which exceeds the 1024 byte limit for 32-bit
> architecutes:
>
> sound/soc/sof/ipc4-topology.c:1700:1: error: stack frame size (1304) exceeds limit (1024) in 'sof_ipc4_prepare_copier_module' [-Werror,-Wframe-larger-than]
> sof_ipc4_prepare_copier_module(struct snd_sof_widget *swidget,
>
>>From what I can tell, this was unintentional, as both
> sof_ipc4_prepare_dai_copier() and sof_ipc4_prepare_copier_module() make a
> copy for the same purpose, but copying it once has the exact same effect.
Humm, not sure. I think the copy was intentional so that if one of the
fixups fails, then the initial hw_params structure is not modified.
Also not sure why a compiler would think inlining such a large function
is a good idea?
Probably need Ranjani, Peter and Bard to comment on the format handling.
> Remove the extra copy and change the direct struct assignment to
> an explicit memcpy() call to make it clearer to the reader that this
> is what happens. Note that gcc treats struct assignment as a memcpy()
> that may be inlined anyway, so the resulting object code is the same.
>
> Fixes: f9209644ae76 ("ASoC: SOF: ipc4-topology: Correct DAI copier config and NHLT blob request")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> sound/soc/sof/ipc4-topology.c | 19 +++++++------------
> 1 file changed, 7 insertions(+), 12 deletions(-)
>
> diff --git a/sound/soc/sof/ipc4-topology.c b/sound/soc/sof/ipc4-topology.c
> index 87be7f16e8c2..77f6850010f6 100644
> --- a/sound/soc/sof/ipc4-topology.c
> +++ b/sound/soc/sof/ipc4-topology.c
> @@ -1651,7 +1651,6 @@ sof_ipc4_prepare_dai_copier(struct snd_sof_dev *sdev, struct snd_sof_dai *dai,
> struct snd_pcm_hw_params *params, int dir)
> {
> struct sof_ipc4_available_audio_format *available_fmt;
> - struct snd_pcm_hw_params dai_params = *params;
> struct sof_ipc4_copier_data *copier_data;
> struct sof_ipc4_pin_format *pin_fmts;
> struct sof_ipc4_copier *ipc4_copier;
> @@ -1676,7 +1675,7 @@ sof_ipc4_prepare_dai_copier(struct snd_sof_dev *sdev, struct snd_sof_dai *dai,
> num_pin_fmts = available_fmt->num_input_formats;
> }
>
> - ret = sof_ipc4_adjust_params_to_dai_format(sdev, &dai_params, pin_fmts,
> + ret = sof_ipc4_adjust_params_to_dai_format(sdev, params, pin_fmts,
> num_pin_fmts);
> if (ret)
> return ret;
> @@ -1684,15 +1683,11 @@ sof_ipc4_prepare_dai_copier(struct snd_sof_dev *sdev, struct snd_sof_dai *dai,
> single_bitdepth = sof_ipc4_copier_is_single_bitdepth(sdev, pin_fmts,
> num_pin_fmts);
> ret = snd_sof_get_nhlt_endpoint_data(sdev, dai, single_bitdepth,
> - &dai_params,
> + params,
> ipc4_copier->dai_index,
> ipc4_copier->dai_type, dir,
> &ipc4_copier->copier_config,
> &copier_data->gtw_cfg.config_length);
> - /* Update the params to reflect the changes made in this function */
> - if (!ret)
> - *params = dai_params;
> -
> return ret;
> }
>
> @@ -1784,9 +1779,9 @@ sof_ipc4_prepare_copier_module(struct snd_sof_widget *swidget,
> * for capture.
> */
> if (dir == SNDRV_PCM_STREAM_PLAYBACK)
> - ref_params = *fe_params;
> + memcpy(&ref_params, fe_params, sizeof(ref_params));
> else
> - ref_params = *pipeline_params;
> + memcpy(&ref_params, pipeline_params, sizeof(ref_params));
>
> copier_data->gtw_cfg.node_id &= ~SOF_IPC4_NODE_INDEX_MASK;
> copier_data->gtw_cfg.node_id |=
> @@ -1819,7 +1814,7 @@ sof_ipc4_prepare_copier_module(struct snd_sof_widget *swidget,
> * In case of capture the ref_params returned will be used to
> * find the input configuration of the copier.
> */
> - ref_params = *fe_params;
> + memcpy(&ref_params, fe_params, sizeof(ref_params));
> ret = sof_ipc4_prepare_dai_copier(sdev, dai, &ref_params, dir);
> if (ret < 0)
> return ret;
> @@ -1829,7 +1824,7 @@ sof_ipc4_prepare_copier_module(struct snd_sof_widget *swidget,
> * input configuration of the copier.
> */
> if (dir == SNDRV_PCM_STREAM_PLAYBACK)
> - ref_params = *pipeline_params;
> + memcpy(&ref_params, pipeline_params, sizeof(ref_params));
>
> break;
> }
> @@ -1838,7 +1833,7 @@ sof_ipc4_prepare_copier_module(struct snd_sof_widget *swidget,
> ipc4_copier = (struct sof_ipc4_copier *)swidget->private;
> copier_data = &ipc4_copier->data;
> available_fmt = &ipc4_copier->available_fmt;
> - ref_params = *pipeline_params;
> + memcpy(&ref_params, pipeline_params, sizeof(ref_params));
>
> break;
> }
next prev parent reply other threads:[~2024-08-07 8:37 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-07 8:02 [PATCH] sound: sof: ioc4-topology: avoid extra dai_params copy Arnd Bergmann
2024-08-07 8:37 ` Pierre-Louis Bossart [this message]
2024-08-07 8:47 ` Arnd Bergmann
2024-08-07 10:00 ` Mark Brown
2024-08-07 15:09 ` Mark Brown
2024-08-07 15:18 ` Arnd Bergmann
2024-08-07 16:21 ` Ranjani Sridharan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=250c63d7-d81e-49ea-ac8f-2e3496075f20@linux.intel.com \
--to=pierre-louis.bossart@linux.intel.com \
--cc=arnd@arndb.de \
--cc=arnd@kernel.org \
--cc=brent.lu@intel.com \
--cc=broonie@kernel.org \
--cc=daniel.baluta@nxp.com \
--cc=justinstitt@google.com \
--cc=kai.vehmanen@linux.intel.com \
--cc=lgirdwood@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-sound@vger.kernel.org \
--cc=llvm@lists.linux.dev \
--cc=morbo@google.com \
--cc=nathan@kernel.org \
--cc=ndesaulniers@google.com \
--cc=perex@perex.cz \
--cc=peter.ujfalusi@linux.intel.com \
--cc=ranjani.sridharan@linux.intel.com \
--cc=seppo.ingalsuo@linux.intel.com \
--cc=sound-open-firmware@alsa-project.org \
--cc=tiwai@suse.com \
--cc=yung-chuan.liao@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox