Repository navigation
Backport 1bfb1978d ('dma: alloc separate buffers for DMAs') - #3712
Merged
brentlu merged 32 commits intoOct 22, 2021
Merged
Conversation
Member
|
@mwasko fyi - can you assign to cml production branch owner for review |
Adds new NOTIFIER_ID_DMA_DOMAIN_CHANGE. This id will be used by DMA single channel scheduling domain. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds DMA channel period value to dma_chan_data. We can now easily retrieve the period value of DMA channel transfers. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Some ll scheduling domains require to have interrupt cleared right at the beginning of handler. Clear domain earlier to fulfill that requirement. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds implementation of DMA single channel scheduling domain. This domain allows to schedule all the tasks on all the cores on single DMA channel interrupt. This way we have only one interrupt source and we don't need to worry about handling many interrupts. We always select the channel with the lowest period. This domain cannot be used with the DMAs, which require to manually reload the next data to be transferred e.g. DW-DMA on BYT or HSW. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Switches from DMA multi channel domain to single channel as the scheduling source for one of the low latency schedulers for cAVS platforms. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Doesn't copy sink component, if it's not active. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
With new schedulers implementations we don't need to trigger pipeline in atomic context anymore. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adjust memory sizes for heaps to provide enough memory for alignment in buffer heap. Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Fix alloc definitions for host build. Function definitions were using defines in their body which blocked usage of const values in said defines. Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Add alignment option for memory requests. Fix alloc definitions for UT. Function definitions were using defines in their body which blocked usage of const values in said defines. Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Fixes tasks initialization for connected pipelines. If we play on pipeline, which is not the owner of the scheduling component, then the pipeline_prepare() is not called on pipe owning that component and tasks stay unitialized. Fixes: de7d4c9 ("pipeline: allocate pipe_task only if needed") Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds new start_position field based on which we assess whether we should throw an xrun. It fixes rare cases of xruns on release with timer scheduled pipelines. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Returns an error, when there is no DAI component connected to the DAI index specified by DAI config IPC. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
In case of unsuccessful scheduler task init, the pipeline task init should also fail. Signed-off-by: Janusz Jankowski <janusz.jankowski@linux.intel.com>
Renames ipc_get_comp function to ipc_get_comp_by_id. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
When new pipeline is created we need to also verify whether given pipeline id is already taken. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Merged hp buffer heap with normal buffer heap. Since current buffer heap was actually using hp memory and it served no purpose to have two heaps. Removed extern capabilities and merged other capabiliteis to buffer heap. Currently none of those buffer heaps supports extern. Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Add seeting of alignment for DMA buffer that is required by Hardware configuration. Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Unify function naming to fix DEBUG_HEAP build Signed-off-by: Jakub Dabek <jakub.dabek@intel.com>
Adds and initializes id, pipeline_id and caps fields in buffer struct. This is the only data needed after buffer creation. It allows us to remove the whole sof_ipc_buffer struct from the buffers. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Extracts allocation part of buffer_new to separate function. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds new DMA attribute DMA_ATTR_BUFFER_PERIOD_COUNT and implements it for every DMA. This attribute says how many periods should DMA buffer consist of. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Implements simple buffer copy functions for 16 and 32 bit data. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Implements copy functions from/to DMA buffer. They take care of cache operations and also can perform custom processing during copy. By default it will be usual buffer copy, but in the future it can be replaced with pcm conversions, so we don't need volume into the pipelines just for that operation. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
DW-DMA has no specific alignments regarding buffer and copy sizes. However setting it to 32 bytes breaks stream formats with period sizes lower than 32 bytes, as linked list is being configured incorrectly. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Currently DMAs are using buffers on pipelines connected to either host or dai components. This causes many problems in firmware logic e.g. we need to preload playback pipelines with data or we need to manually override buffer sizes, if the buffer isn't complaint with specific DMA requirements. Also it requires from topology creators to know some hardware specific things on each platform like that for timer based scheduling on cAVS platforms we need to have 3 periods in buffers connected to dai components. This patch allocates separate buffers for DMAs, which are made complaint by using different DMA attributes first. We no longer need to preload playback streams, because we can start DMA right away and just let it transfer zeroes before any valid data will come. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Removes buffer cache operations done based on the fact, whether buffer is used by DMA or not. From now on DMAs have their own dedicated buffers with dedicated copy functions, so this code is no longer needed. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Change flow of components to not resize sink buffers. We should just check the size of the allocated buffers and let the topology define the maximum sizes. Components can easily work with non aligned buffer sizes. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Removes code related to preload as with separate DMA buffers we no longer need it. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Changes flow of dai_copy to call dma_copy even if there is no data available. This way we make sure that all DMA needed procedures are executed, since buffers pointers won't be changed anyway. Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Contributor
Author
|
Following pull requests have been selected: dma: alloc separate buffers for DMAs #1984 |
Member
|
@mwasko @abonislawski @slawblauciak fyi - who owns the CML production branch ? |
Contributor
|
@bkokoszx can help with creating new CML release if necessary, but the backporting itself of PRs to older CML releases is not supported on our end. |
Contributor
Author
|
The target branch, cml-008-patch-backport2, is a fresh one on HEAD of cml-008-patch-backport. Adrian helped us to create this branch. Does it work for you? |
Contributor
Sorry @brentlu, currently we are on cml-011 so our recommendation is to switch to latest CML release branch. |
Contributor
Author
|
FW has been deployed to CML Chromebooks since March without issue reported. Merge the code to the unused branch. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Backport the commit 1bfb197 ('dma: alloc separate buffers for DMAs') and other related commits to fix the audio stop issue of CML Chromebooks.
commit 1bfb197
Author: Tomasz Lauda tomasz.lauda@linux.intel.com
Date: Fri Oct 18 13:08:44 2019 +0200
dma: alloc separate buffers for DMAs
Currently DMAs are using buffers on pipelines connected to either
host or dai components. This causes many problems in firmware logic
e.g. we need to preload playback pipelines with data or we need
to manually override buffer sizes, if the buffer isn't complaint
with specific DMA requirements. Also it requires from topology
creators to know some hardware specific things on each platform
like that for timer based scheduling on cAVS platforms we need
to have 3 periods in buffers connected to dai components.
This patch allocates separate buffers for DMAs, which are made
complaint by using different DMA attributes first. We no longer
need to preload playback streams, because we can start DMA
right away and just let it transfer zeroes before any valid
data will come.
Signed-off-by: Tomasz Lauda tomasz.lauda@linux.intel.com