On Wed, 2003-11-26 at 15:15, Takashi Iwai wrote: > At Wed, 26 Nov 2003 15:10:06 +0100, > Ronny V. Vindenes wrote: > > > > On Wed, 2003-11-26 at 13:32, Takashi Iwai wrote: > > > At Wed, 26 Nov 2003 13:10:11 +0100, > > > Ronny V. Vindenes wrote: > > > > > > > > On Wed, 2003-11-26 at 11:44, Takashi Iwai wrote: > > > > > At Wed, 26 Nov 2003 01:11:07 +0100, > > > > > Ronny V. Vindenes wrote: > > > > > > > > > > > > On Tue, 2003-11-25 at 20:04, Takashi Iwai wrote: > > > > > > > Hi, > > > > > > > > > > > > > > right now the ALSA 1.0.0pre3 packages are released. > > > > > > > > > > > > > > In this version, alsa-tools package was divided to two several ones, > > > > > > > alsa-tools and alsa-firmware. The latter contains only the firmware > > > > > > > binary files required by loader programs in alsa-tools package. > > > > > > > Please test them if you have a board requiring the loader (hdsp, > > > > > > > mixart, vx, usx2y). > > > > > > > > > > > > > > Especially, HDSP users, please test and report whether it works, since > > > > > > > the hdsploader was modified to load the files dynamically. > > > > > > > > > > > > > > > > > > > > > > > > > > mmap'ed sound with the cs46xx driver seems to be broken, every other > > > > > > half second the sound is muted. Non-mmap'ed sound plays normally. This > > > > > > is under 2.6.0-test9-mm5. The version included in test9-mm5 works as > > > > > > expected (there was some patches in mm5 against ALSA, but applying those > > > > > > to 1.0.0pre3 made no difference for this particular problem). > > > > > > > > > > there is no big change in cs46xx driver itself, so the change is > > > > > either in kernel core or in the config. > > > > > do you use CONFIG_SND_CS46XX_NEW_DSP ? > > > > > > > > Yes. > > > > > > > > I've tested it with 2.6.0-test10 too now, and I've narrowed it down a > > > > bit: native alsa mmap'ed or not works, but oss emulation only works > > > > correctly when not using mmap. > > > > > > > > Things like wine, quake3, Enemy Territory all sound awful. Plain > > > > 2.6.0-test10 works fine in all cases. > > > > > > to be sure, try to set direct mode for oss-emulation via writing to > > > the proc file. see alsa-kernel/Documentation/OSS-Emulation.txt. > > > > Setting direct mode made no difference. > > > > > > > > also, please check /proc/asound/card0/pcm0p/sub0/* files in both > > > cases. > > > > I've attached cat /proc/asound/card0/pcm0p/sub0/* for both cases > > sorry, i meant the ALSA 1.0.0pre3 and the 2.6.0-test10 (oss-emu with > mmap). it's known that the mmap support of this chip is not perfect. > but i still suspect that the base oss-emu code change affected this. > > > Takashi >
ok, here's from plain test10 -- Ronny V. Vindenes <[EMAIL PROTECTED]>
access: MMAP_INTERLEAVED format: S16_LE subformat: STD channels: 2 rate: 22050 (22050/1) period_size: 512 buffer_size: 16384 tick_time: 1000 OSS format: S16_LE OSS channels: 2 OSS rate: 22050 OSS period bytes: 2048 OSS periods: 32 card: 0 device: 0 subdevice: 0 stream: PLAYBACK id: CS46xx name: CS46xx subname: subdevice #0 class: 0 subclass: 0 subdevices_count: 31 subdevices_avail: 30 64 state: RUNNING trigger_time: 1069856885.000853207 tstamp : 1069856889.000718741 delay : 16144 avail : 240 avail_max : 33200 ----- hw_ptr : 85232 appl_ptr : 101376 tstamp_mode: NONE period_step: 1 sleep_min: 0 avail_min: 512 xfer_align: 1 start_threshold: 1 stop_threshold: 1073741824 silence_threshold: 528 silence_size: 528 boundary: 1073741824
