Skip to content

mux: add missing buffer locks #2545

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
tlauda merged 1 commit into from
Mar 17, 2020
Merged

mux: add missing buffer locks #2545

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 18 additions & 1 deletion src/audio/mux/mux.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -387,6 +387,7 @@ static int mux_copy(struct comp_dev *dev)
uint32_t frames = -1;
uint32_t sources_bytes[MUX_MAX_STREAMS] = { 0 };
uint32_t sink_bytes;
uint32_t flags = 0;

comp_dbg(dev, "mux_copy()");

Expand All @@ -399,11 +400,14 @@ static int mux_copy(struct comp_dev *dev)
/* align source streams with their respective configurations */
list_for_item(clist, &dev->bsource_list) {
source = container_of(clist, struct comp_buffer, sink_list);
buffer_lock(source, &flags);
if (source->source->state == dev->state) {
Comment thread
dbaluta marked this conversation as resolved.
Outdated
num_sources++;
i = get_stream_index(cd, source->pipeline_id);
sources[i] = source;
sources_stream[i] = &source->stream;
} else {
buffer_unlock(source, flags);
}
}

Expand All @@ -414,18 +418,31 @@ static int mux_copy(struct comp_dev *dev)
sink = list_first_item(&dev->bsink_list, struct comp_buffer,
source_list);

buffer_lock(sink, &flags);

/* check if sink is active */
if (sink->sink->state != dev->state)
if (sink->sink->state != dev->state) {
for (i = 0; i < MUX_MAX_STREAMS; i++) {
if (!sources[i])
continue;
buffer_unlock(sources[i], flags);
}

@dbaluta dbaluta Mar 13, 2020

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

where is the corresponding buffer_lock for this? Code looks strange.

@paulstelian97 paulstelian97 Mar 13, 2020

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@dbaluta I think it's the one on line 403 (it's in the same function).

@dbaluta dbaluta Mar 13, 2020

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see. thanks.

@lgirdwood lgirdwood Mar 13, 2020

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, I'm not seeing this, I can see lock(source) and unlock(sources[i]) which seem unaligned as they are different objects. Some inline comments may help here.

@tlauda tlauda Mar 13, 2020

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@lgirdwood They are the same objects.

@dbaluta dbaluta Mar 16, 2020

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tlauda this looks error prone.

@tlauda tlauda Mar 16, 2020

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@dbaluta What do you mean?

@lgirdwood lgirdwood Mar 16, 2020

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if simplification would be better here. i.e.

	/* align source streams with their respective configurations */
	list_for_item(clist, &dev->bsource_list) {
		source = container_of(clist, struct comp_buffer, sink_list);
		buffer_lock(source, &flags);
		if (source->source->state == dev->state) {
			num_sources++;
			i = get_stream_index(cd, source->pipeline_id);
			sources[i] = source;
			sources_stream[i] = &source->stream;
		} else {
			buffer_unlock(source, flags);
		}
	}

Means we only have certain source buffers locked, i.e only the buffers that out configuration is telling us to copy from. Would it make more sense to just lock all sources and then unlock all sources ?

Another thing that we should do is to change the buffer locks to reader and writer locking.

buffer_rlock(buffer, flags);
buffer_wlock(buffer, flags);

This way we only lock a buffer for reading or writing (i.e.reader lock updates r_ptr and avail, writer lock updates w_ptr and free). This would allow cross core buffers that would not block a core that the current implementation might.

@tlauda tlauda Mar 16, 2020 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@lgirdwood It's definitely candidate for future improvement. This patch is about aligning with rest of the components. mux_copy function is the only one missing locks currently.

@lgirdwood lgirdwood Mar 16, 2020

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok, great for the r/w buffer locking update.


buffer_unlock(sink, flags);
return 0;
}

for (i = 0; i < MUX_MAX_STREAMS; i++) {
if (!sources[i])
continue;
frames = MIN(frames,
audio_stream_avail_frames(sources_stream[i],
&sink->stream));
buffer_unlock(sources[i], flags);
}

buffer_unlock(sink, flags);

for (i = 0; i < MUX_MAX_STREAMS; i++) {
if (!sources[i])
continue;
Expand Down