In theory the fxpi kernel has an ALSA-capable driver, and we're using ALSA in Gonk, so audio should work. But it hasn't been tested.

This work isn't blocked on the graphics stack.

With a basic ogg theora file, I get 01-01 00:15:58.520 E/AudioFlinger(14009): createTrack_l() Bad parameter: sampleRate 44100 format 1, channelMask 0x00000001 "for output 0x1c453e8 with format 1 01-01 00:15:58.520 E/AudioTrack(15662): AudioFlinger could not create track, status: -22 01-01 00:15:58.520 E/libOpenSLES(15662): AudioTrack::initCheck status 4294967274 01-01 00:15:58.520 W/libOpenSLES(15662): Leaving Object::Realize (SL_RESULT_CONTENT_UNSUPPORTED) 01-01 00:15:58.990 W/Music (15662): [JavaScript Warning: "Media resource blob:app://music.gaiamobile.org/c67f27bc-8163-441f-bcb8-0e406d1459d1 could not be decoded." {file: "app://music.gaiamobile.org/index.html" line: 0}] So looks like we didn't get lucky with audio.

Hm, I may have messed up something but I can't get basic WAV or OGG audio to play in my B rev2 Raspbian image (over HDMI anyway). My little cheap TV may be at fault but I'm going to deprioritize this a bit for now.

This may be a simple matter of not packing the right media configuration file.

using the following script var ac = new AudioContext(); var o = ac.createOscillator(); o.connect(ac.destination); o.start(); I get : W/AudioFlinger( 66): Thread AudioOut_1 cannot connect to the power manager service E/SchedPolicy( 2029): add_tid_to_cgroup failed; background=0 E/SchedPolicy( 2029): add_tid_to_cgroup failed; background=0 W/AudioFlinger( 66): Thread AudioOut_1 cannot connect to the power manager service E/SchedPolicy( 2029): add_tid_to_cgroup failed; background=0 and trying to set an alert in the setting menu: W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00e78400, server=00e77c00 W/AudioFlinger( 66): Thread AudioOut_1 cannot connect to the power manager service W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00e8c000, server=00e8b800 W/AudioFlinger( 66): Thread AudioOut_1 cannot connect to the power manager service [...] W/AudioTrack( 2142): obtainBuffer timed out (is the CPU pegged?) 0xb2a7c380 user=00004400, server=00003c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 W/AudioTrack( 2072): obtainBuffer timed out (is the CPU pegged?) 0xb2a2c980 user=00fa3400, server=00fa2c00 [...]

That means the opensles audio callback is called, but we kind of fail to return from it. Interesting.

Firefox OS is not being worked on