Skip to content
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.

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

aop: 13.5 bring up + mics + als + generalize afk/epic #381

Merged
merged 22 commits into from
Oct 30, 2024
Merged

Conversation

eiln
Copy link
Contributor

@eiln eiln commented Jan 15, 2024

aop_audio: stream hpai device (high power audio input) microphone samples from aop to admac. output seems to be 32-bit LE floats in groups of 3 (3 mic array)
aop_als: monitor/report red, green, blue, clear, lux, gain etc information from ambient light sensor (ALS)

Tested j293ap 13.5

Also includes afk/epic/rbep generalizations which hopefully don't break DCP

Move from __init__.py to base.py so it can be imported.

Signed-off-by: Eileen Yoon <[email protected]>
@eiln eiln marked this pull request as draft January 16, 2024 15:55
@eiln eiln marked this pull request as ready for review January 19, 2024 09:45
@eiln eiln marked this pull request as draft January 20, 2024 17:12
@eiln
Copy link
Contributor Author

eiln commented Jan 20, 2024

nvm this broke dcp. fixing it


class AFKError(Exception):
pass
# TODO no one else but AOP uses AFKRingBuf right?
Copy link
Member

Choose a reason for hiding this comment

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

DCP uses AFKRingBuf on some endpoints via EPICEndpoint. Used in experiments/dcpext_iboot.py

Copy link
Contributor Author

Choose a reason for hiding this comment

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

New commits don't break DCP experiments. DCP/AOP AFKRingbuf are too different to share (e.g. different EPIC header, different ring size, AOP also writes ptrs to DART and etc), everything else is fine though. I think we can move afk to soc/apple/ and just have fn pointers for the rbep r/w calls

@eiln eiln marked this pull request as ready for review January 20, 2024 20:52
@eiln eiln changed the title aop: 13.5 bring up + we got mail aop: 13.5 bring up + stream microphone samples Jan 20, 2024
@eiln eiln changed the title aop: 13.5 bring up + stream microphone samples aop: 13.5 bring up + mics (kinda) + als Jan 21, 2024
Consistently use an enum for EPICSubtype (used later) and support the v2
EPICService intializer protocol used by AOP.

Signed-off-by: Eileen Yoon <[email protected]>
DCP only gets two reply types: a u32 retcode or a command reply. It
previously checked for retcode by checking length(data) == 4 and
assuming command otherwise. AOP can get other reply types (e.g. STRING
after probing device). It actually sends us the reply subtype in the
EPICSubHeader, so use that information and handle replies accordingly.

Signed-off-by: Eileen Yoon <[email protected]>
Much simpler than the command messages. We just write directly into the
TX ringbuffer.

Signed-off-by: Eileen Yoon <[email protected]>
Because AOP gets a lot of reports ;)

Signed-off-by: Eileen Yoon <[email protected]>
Default to version=4 as to not break DCP

Signed-off-by: Eileen Yoon <[email protected]>
Make the ringbuffer class robust to various block sizes to generalize to
both DCP and AOP.

The first three blocks of the ringbuffer is reserved for exchanging size,
rptr, wptr:

```
           bufsize      unk
00000000  00007e80 00070006 00000000 00000000 00000000 00000000 00000000 00000000
00000020  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000040  *   rptr
00000080  00000600 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000000a0  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000000c0  *   wptr
00000100  00000680 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000120  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000140  *
```

Each block is spread out by some block_size multiple of 0x40 (step). The
0th block holds the size of the ringbuffer contents, the 1st block holds
the rptr, and the 2nd block holds the wptr. The actual contents of the
ringbuffer starts after the first three blocks, which will be
collectively called the "header".

However, this block_size isn't constant. DCP seems to consistently use
0x40, but AOP can use both 0x40/0x80. Since we're not given the block_size,
so wemust bootstrap it. Recall we are given the total size of the
rinbuffer in the mailbox message. Since we're always given the size of
the ringbuffer `bufsize` at offset +block_size * 0 (or simply 0), and we
can find the header size by subtracting `bufsize` from the total size.
Since we also know that the header is always 3 blocks wide, we can
divide the header size by 3 to obtain the block_size.

Signed-off-by: Eileen Yoon <[email protected]>
Otherwise the previous commit is way too confusing.

Signed-off-by: Eileen Yoon <[email protected]>
@eiln
Copy link
Contributor Author

eiln commented Jan 23, 2024

@jannau I generalized AFK/EPIC code to both DCP and AOP. I feel like I should send these in a separate series but it obviously depends on the above. I think marcan's on holiday. From my tests (dcp/dcp_iboot; I can't test dptx/dcpext) it doesn't break DCP but I'd like to get your ack

@eiln eiln changed the title aop: 13.5 bring up + mics (kinda) + als aop: 13.5 bring up + mics + als + generalize afk/epic Jan 24, 2024
Output audio format still unknown, not sure if it's garbage (see lpai
commit) or some weird packed float encoding I'm not figuring out.

Signed-off-by: Eileen Yoon <[email protected]>
```
[syslog] * [ALSCT720.cpp:967]updateDynamicIntegrationParameters allAbove=0, anyBelow=1, threshold hit=0, lev
> 02:0x50000000000005 (TYPE=0x5, INDEX=0x5)
< 24:0x85000000000000
[als.als] Report REPORT/0xc4 #0
[als.als] Container:
    unk0 = 0xEC
    sequence = 4
    timestamp = 0x0000000019352268
    red = 11
    green = 18
    blue = 11
    clear = 40
    lux = 7.277451038360596
    unk_zero = 0
    status = 3
    gain = 256
    unk3 = 68
    unk4 = 35
    unk5 = 17874
    integration_time = 378080
< 02:0x50000000000006
[syslog] * [ALSCT720.cpp:454]handleInterrupt: result: 0 (status=0x3)
> 02:0x50000000000006 (TYPE=0x5, INDEX=0x6)
```

Signed-off-by: Eileen Yoon <[email protected]>
I'm getting garbage from the low-power audio input (lpai) mic which
exists solely for voicetrigger. The garbage specifically is `61 7b`
repeated. Obviously no voicetrigger report because there's nothing
useful in lpai output. I'm suspecting its an mca/clock issue (maybe even
SEP/permissions) because there's nothing suspicious from the aop RX/TX
IPC side.

```
[audio.audio] Report REPORT/0xee #0
[audio.audio] unknown report: 0xee
00000000  c5 96 20 01 00 00 00 00  9c f5 10 00 00 00 00 00  |.. .............|
00000010  b8 07 00 00 20 30 6e 69  01 00 00 00 43 02 00 00  |.... 0ni....C...|
00000020  00 00 00 00 c5 96 20 01  00 00 00 00 00 00 00 00  |...... .........|
00000030  a4 07 00 00 9a 07 00 00  61 7b 61 7b 61 7b 61 7b  |........a{a{a{a{|
00000040  61 7b 61 7b 61 7b 61 7b  61 7b 61 7b 61 7b 61 7b  |a{a{a{a{a{a{a{a{|
00000050  *
000007b0  61 7b 61 7b 61 7b 61 7b                           |a{a{a{a{        |
```

Signed-off-by: Eileen Yoon <[email protected]>
@pbasista
Copy link

Hi, could someone familiar with Asahi Linux and Apple audio help to move this PR or more generally the microphone support forward?

@marcan marcan force-pushed the main branch 3 times, most recently from 4f95305 to 95d67cf Compare April 9, 2024 03:58
@marcan marcan merged commit d649c88 into AsahiLinux:main Oct 30, 2024
2 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants