No sound on HP Pavilion dv5-1164er SBx00 Azalia (Intel HDA)

[X512]

This is hrev42190.

There are no sound in embedded speakers and headphones. Haiku recognise audio device as "HD Audio".

Open sound goes system into KDL.

[X512]

listdev

[axeld]

Can you provide a syslog as well? The HDA driver puts out lots of useful information during boot.

[6foot3]

This seems to be another HDA ticket in addition to my recent #7680. The first reply to #7680 pointed to possible duplicates as #4492, #4635, #4755, #6213 and #7144. My last comment on that ticket mentioned another laptop which was the subject of ticket #5866 which still has an issue with the HDA driver.

Would it be usefull to try and amalgamate some of these HDA tickets and have one ticket where we can submit sylogs for all these HDA problems so that any developer that wants to look into it can have all the syslogs in one place?

I have already attached the syslog for my netbook (to #7680) and would like to submit the syslog for my laptop without creating another (duplicate) ticket.

[X512]

syslog

[X512]

Still present in hrev43243.

[X512]

Still present in hrev44242.

[HAL]

I have the same bug in hrev1alpha4-44597. I occasionally hear a low volume crackle sound. My audio chip is the same according to listdev. My motherboard is different but same audio chipset. I will upload output of listdev.

[HAL]

listdev from system with msi 870a fuzion mb.

[HAL]

I just did a complete shutdown not a restart and sound is back. I thought before that a restart would be enough, before commenting here but it was not. There looks like some corruption in working memory that a restart wont fix. I have over 12Gb of RAM so it is not like it would run out.

[HAL]

The reason is probably I rebooted first from Ubuntu to Haiku and the RAM memory was not cleared. It was not cleared even after another reboot from Haiku to Haiku. RAM can still retain info from other OS.

[anevilyak]

(In #9750) Duplicate of #7703.

[pulkomandy]

Removing HDA audio tickets from Beta2 milestone, no one has worked on it and opensound is available at least as a stopgap.

[mmlr]

If the relevant hardware is still around, it might be worth trying a recent nightly. If this was just a delayed audio start it might be fixed by hrev54464, if it was related to HDA interrupts it might also see changes. Other, less recent changes to HDA might also have fixed this.

[X512]

Still no sound in hrev54465. Both speakers and headphones are not working. Nothing is heard when restarting media_server.

[ttcoder]

Replying to HAL:

The reason is probably I rebooted first from Ubuntu to Haiku and the RAM memory was not cleared. It was not cleared even after another reboot from Haiku to Haiku. RAM can still retain info from other OS.

There are other tickets and cases like this (including on my own old T410), where rebooting to another OS intializes the HDA chipset correctly, and sometimes the state is memorized for several hours after a shutdown, so long as you keep the laptop's battery and the mains power supply plugged in. To my semi-layman mind, it seems more likely that the state is memorized in the 'electronics' (HDA chipset?) of the motherboard, rather than in the RAM modules. Anyway, X512 might be able to confirm that the same is true for the HP pavillion : reboot to a BSD or Linux, then back to Haiku, see if sound works then. Maybe that could narrow down the possible causes a bit.

[mmlr]

As I'm not able to reproduce the issue on my local hardware, I'll need some info extracted from the relevant systems.

Initially I'd like to inspect the HDA register state in the failed state, and ideally compare with one from a working state after a warm reboot. The issue may not show in the registers at all as there's a lot more going on with verbs, pins, widgets...

Please dump the registers in KDL and attach that:

1. Enter KDL
2. Find the HDA register area: areas | grep hda_hw_regs
3. Find the base address in the third column
4. Dump the content: dw <baseAddressFromStep3> 150
5. Exit KDL
6. Attach the KDL output of the command in 4 from the syslog

[mmlr]

As a next step: More debug output.

I'm attaching a patch that enables debug output in the AudioMixer and MultiAudioNode as well as one entry in HDA. This is mostly the set of debug output that I used to follow the issue fixed in hrev54464 and I'd like to validate that this side looks sane. Please apply the patch and rebuild hda hmulti_audio.media_addon mixer.media_addon libmedia.so.

Boot the patched system, play some media, then get a syslog. Note that there is a lot of debug output, so the syslog will fill up quickly and it is important to get the initial part. That means the whole process until grabbing the syslog needs to be reasonably quick and both syslog and syslog.old should be copied out.

[X512]

Sound is not working when rebooted from Linux. Sound in Linux is working. This PC have touch volume slider and mute button that is working on Linux.

[X512]

Only front headphones are displayed as audio output and there are no audio input, but microphone socket is present. There are more inputs and outputs in ALSA mixer in Linux.

[mmlr]

Replying to X512:

Only front headphones are displayed as audio output and there are no audio input, but microphone socket is present. There are more inputs and outputs in ALSA mixer in Linux.

What's interesting is that there are two controllers according to your syslog. The first one has one output and no inputs, so would fit your description of only one output. It does fail init however and should essentially just be left unused. The second controller then has 4 outputs and 4 inputs which would presumably match ALSA. It sounds like the two controllers get mixed up?

I have not yet studied the memory dump, so can't yet tell if this would align with such a mixup.

[mmlr]

The register dump looks fine and definitely belongs to a 4 input 4 output stream device. There are also two streams configured, i.e. both playback and recording, although the link position (LPIB) for the recording stream seems to indicate the stream isn't actually doing anything (which might be fine, I haven't checked). So there shouldn't be any mixups.

[X512]

Is it possible to test audio driver bypassing media kit?

[ttcoder]

It was more necessary before mmlr's TimeSource fix, now the media kit should be ok to test HDA, but yes, if you wish to, you can follow the test procedure here like modeenf and me did a while back : https://dev.haiku-os.org/ticket/12356#comment:16

[korli]

Please check with a current nightly.