ASoC: SOF: ipc4-topology: Add deep buffer size to the debug message #4544
Conversation
yongzhi1
requested review from
lgirdwood,
plbossart,
ranj063,
kv2019i,
dbaluta,
bardliao,
ujfalusi,
vijendarmukunda,
kuanhsuncheng and
yaochunhung
as code owners
sound/soc/sof/ipc4-topology.c
Outdated
| dev_dbg(sdev->dev, "copier %s, IPC size is %d, deep buffer size %u", swidget->widget->name, | ||
| ipc_size, copier_data->gtw_cfg.dma_buffer_size); | ||
| else | ||
| dev_dbg(sdev->dev, "copier %s, IPC size is %d", swidget->widget->name, ipc_size); |
There was a problem hiding this comment.
IMHO, we can always print the dma buffer size.
There was a problem hiding this comment.
Ack, I was inquired several times about how to tell the current deep buff size when experimenting different DEEPBUFFER_FW_DMA_MS.
There was a problem hiding this comment.
the deepbuffer is only affecting swidget->id == snd_soc_dapm_aif_in, in other copiers and w/o DB the buffer size is 2ms worth of bytes.
The question is: what is important? The buffer size in bytes or the length in time? A printing around 1729 if deep_buffer_dma_ms is set might be enough?
Or print only for the host copiers unconditionally?
There was a problem hiding this comment.
deep_buffer_dma_ms is more clear as an indicator, is the following acceptable?
unconditionally for snd_soc_dapm_aif_in:
[ 93.872063] sof-audio-pci-intel-mtl 0000:00:1f.3: copier host-copier.1.playback, deep_buffer_dma_ms 0, dma_buffer_size 384
When deepbuf is used:
[ 156.859504] sof-audio-pci-intel-mtl 0000:00:1f.3: copier host-copier.31.playback, deep_buffer_dma_ms 10, dma_buffer_size 1920
e57dd7c
to
280603f
Compare
|
V2:
|
sound/soc/sof/ipc4-topology.c
Outdated
| max((u32)SOF_IPC4_MIN_DMA_BUFFER_SIZE, deep_buffer_dma_ms) * | ||
| copier_data->base_config.ibs; | ||
| dev_dbg(sdev->dev, "copier %s, deep_buffer_dma_ms %u, dma_buffer_size %u", | ||
| swidget->widget->name, deep_buffer_dma_ms, |
There was a problem hiding this comment.
I would use the max((u32)SOF_IPC4_MIN_DMA_BUFFER_SIZE, deep_buffer_dma_ms) otherwise the log could be questionable with values that don't really make sense with regular ms to byte conversions.
There was a problem hiding this comment.
Ack, let me update in next version, thanks for the review!!
yongzhi1
force-pushed
the
deepbuf-dma-size
branch
from
280603f
to
ff51e02
Compare
|
V3: Print max((u32)SOF_IPC4_MIN_DMA_BUFFER_SIZE, deep_buffer_dma_ms) instead of deep_buffer_dma_ms: I am OK to print the buffer info conditionally for (deep_buffer_dma_ms > 0) case only if above looks bit verbose/confusing. |
yongzhi1
force-pushed
the
deepbuf-dma-size
branch
from
ff51e02
to
21de9be
Compare
Print deep_buffer_dma_ms and dma_buffer_size for debug purpose. Signed-off-by: Yong Zhi <[email protected]>
yongzhi1
force-pushed
the
deepbuf-dma-size
branch
from
21de9be
to
129f6df
Compare
When deep buffer is used, print the dma size for debug.