From: "Mukunda, Vijendar" <Vijendar.Mukunda@amd.com>
To: Yu Zhao <yuzhao@google.com>, Liam Girdwood <lgirdwood@gmail.com>,
Mark Brown <broonie@kernel.org>, aroslav Kysela <perex@perex.cz>,
Takashi Iwai <tiwai@suse.com>
Cc: Daniel Kurtz <djkurtz@chromium.org>,
"Agrawal, Akshu" <Akshu.Agrawal@amd.com>,
"Deucher, Alexander" <Alexander.Deucher@amd.com>,
"alsa-devel@alsa-project.org" <alsa-devel@alsa-project.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/2] ASoC: use DMA addr rather than CPU pa for acp_audio_dma
Date: Wed, 5 Dec 2018 06:42:43 +0000 [thread overview]
Message-ID: <0282a69b-51dc-b1e9-a017-2f18c08b9c50@amd.com> (raw)
In-Reply-To: <20181204224253.216075-1-yuzhao@google.com>
On 05/12/18 4:12 AM, Yu Zhao wrote:
> We shouldn't assume CPU physical address we get from page_to_phys()
> is same as DMA address we get from dma_alloc_coherent(). On x86_64,
> we won't run into any problem with the assumption when dma_ops is
> nommu_dma_ops. However, DMA address is IOVA when IOMMU is enabled.
> And it's most likely different from CPU physical address when AMD
> IOMMU is not in passthrough mode.
This is really a good fix. We will apply this patch changes for Raven
platform also.
Thanks,
Vijendar
>
> Signed-off-by: Yu Zhao <yuzhao@google.com>
> ---
> sound/soc/amd/acp-pcm-dma.c | 15 +++++----------
> sound/soc/amd/acp.h | 2 +-
> 2 files changed, 6 insertions(+), 11 deletions(-)
>
> diff --git a/sound/soc/amd/acp-pcm-dma.c b/sound/soc/amd/acp-pcm-dma.c
> index cdebab2f8ce5..fd3db4c37882 100644
> --- a/sound/soc/amd/acp-pcm-dma.c
> +++ b/sound/soc/amd/acp-pcm-dma.c
> @@ -303,11 +303,10 @@ static void set_acp_to_i2s_dma_descriptors(void __iomem *acp_mmio, u32 size,
> }
>
> /* Create page table entries in ACP SRAM for the allocated memory */
> -static void acp_pte_config(void __iomem *acp_mmio, struct page *pg,
> +static void acp_pte_config(void __iomem *acp_mmio, dma_addr_t addr,
> u16 num_of_pages, u32 pte_offset)
> {
> u16 page_idx;
> - u64 addr;
> u32 low;
> u32 high;
> u32 offset;
> @@ -317,7 +316,6 @@ static void acp_pte_config(void __iomem *acp_mmio, struct page *pg,
> /* Load the low address of page int ACP SRAM through SRBM */
> acp_reg_write((offset + (page_idx * 8)),
> acp_mmio, mmACP_SRBM_Targ_Idx_Addr);
> - addr = page_to_phys(pg);
>
> low = lower_32_bits(addr);
> high = upper_32_bits(addr);
> @@ -333,7 +331,7 @@ static void acp_pte_config(void __iomem *acp_mmio, struct page *pg,
> acp_reg_write(high, acp_mmio, mmACP_SRBM_Targ_Idx_Data);
>
> /* Move to next physically contiguos page */
> - pg++;
> + addr += PAGE_SIZE;
> }
> }
>
> @@ -343,7 +341,7 @@ static void config_acp_dma(void __iomem *acp_mmio,
> {
> u16 ch_acp_sysmem, ch_acp_i2s;
>
> - acp_pte_config(acp_mmio, rtd->pg, rtd->num_of_pages,
> + acp_pte_config(acp_mmio, rtd->dma_addr, rtd->num_of_pages,
> rtd->pte_offset);
>
> if (rtd->direction == SNDRV_PCM_STREAM_PLAYBACK) {
> @@ -850,7 +848,6 @@ static int acp_dma_hw_params(struct snd_pcm_substream *substream,
> int status;
> uint64_t size;
> u32 val = 0;
> - struct page *pg;
> struct snd_pcm_runtime *runtime;
> struct audio_substream_data *rtd;
> struct snd_soc_pcm_runtime *prtd = substream->private_data;
> @@ -986,16 +983,14 @@ static int acp_dma_hw_params(struct snd_pcm_substream *substream,
> return status;
>
> memset(substream->runtime->dma_area, 0, params_buffer_bytes(params));
> - pg = virt_to_page(substream->dma_buffer.area);
>
> - if (pg) {
> + if (substream->dma_buffer.area) {
> acp_set_sram_bank_state(rtd->acp_mmio, 0, true);
> /* Save for runtime private data */
> - rtd->pg = pg;
> + rtd->dma_addr = substream->dma_buffer.addr;
> rtd->order = get_order(size);
>
> /* Fill the page table entries in ACP SRAM */
> - rtd->pg = pg;
> rtd->size = size;
> rtd->num_of_pages = PAGE_ALIGN(size) >> PAGE_SHIFT;
> rtd->direction = substream->stream;
> diff --git a/sound/soc/amd/acp.h b/sound/soc/amd/acp.h
> index dbbb1a85638d..e5ab6c6040a6 100644
> --- a/sound/soc/amd/acp.h
> +++ b/sound/soc/amd/acp.h
> @@ -123,7 +123,7 @@ enum acp_dma_priority_level {
> };
>
> struct audio_substream_data {
> - struct page *pg;
> + dma_addr_t dma_addr;
> unsigned int order;
> u16 num_of_pages;
> u16 i2s_instance;
>
next prev parent reply other threads:[~2018-12-05 6:42 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-04 22:42 Yu Zhao
2018-12-04 22:42 ` [PATCH 2/2] ASoC: use dma_ops of parent device " Yu Zhao
2018-12-05 6:45 ` Mukunda, Vijendar
2018-12-06 20:26 ` Applied "ASoC: use dma_ops of parent device for acp_audio_dma" to the asoc tree Mark Brown
2018-12-05 6:42 ` Mukunda, Vijendar [this message]
2018-12-06 20:26 ` Applied "ASoC: use DMA addr rather than CPU pa " Mark Brown
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=0282a69b-51dc-b1e9-a017-2f18c08b9c50@amd.com \
--to=vijendar.mukunda@amd.com \
--cc=Akshu.Agrawal@amd.com \
--cc=Alexander.Deucher@amd.com \
--cc=alsa-devel@alsa-project.org \
--cc=broonie@kernel.org \
--cc=djkurtz@chromium.org \
--cc=lgirdwood@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=perex@perex.cz \
--cc=tiwai@suse.com \
--cc=yuzhao@google.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
all inboxes | Powered by JetHome®