Various apps (e.g. ogg123, mplayer, xmame...) crash with the error message: -- pcm.c:5889: snd_pcm_mmap_commit: Assertion `frames <= snd_pcm_mmap_avail(pcm)' failed. -- since I upgraded from alsa-lib-1.0.2 to alsa-lib-1.0.3. ogg123 lasts more depending on the song apparently, the others crash immediately. The problem disappears if I downgrade to alsa-lib-1.0.2. At boot my kernel (development-sources-2.6.3) shows: -- Advanced Linux Sound Architecture Driver Version 1.0.2c (Thu Feb 05 15:41:49 2004 UTC). PCI: Found IRQ 9 for device 0000:00:02.7 intel8x0_measure_ac97_clock: measured 49338 usecs intel8x0: clocking to 48000 ALSA device list: #0: SiS SI7012 at 0xe000, irq 9 -- Here's the emerge info, BTW I tried with less aggressive CFLAGS too and the problem remains. -- Portage 2.0.50-r1 (default-x86-2004.0, gcc-3.3.3, glibc-2.3.3_pre20040207-r0, 2.6.3) ================================================================= System uname: 2.6.3 i686 Intel(R) Pentium(R) 4 CPU 2.40GHz Gentoo Base System version 1.4.3.13p1 Autoconf: sys-devel/autoconf-2.59-r3 Automake: sys-devel/automake-1.8.2 ACCEPT_KEYWORDS="x86 ~x86" AUTOCLEAN="yes" CFLAGS="-march=pentium4 -Os -mfpmath=sse -pipe" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3/share/config /usr/lib/mozilla/defaults/pref /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=pentium4 -Os -mfpmath=sse -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs ccache sandbox" GENTOO_MIRRORS="ftp://sunsite.cnlab-switch.ch/mirror/gentoo/ http://212.219.247.17/sites/www.ibiblio.org/gentoo/ rsync://ftp.snt.utwente.nl/gentoo http://212.219.247.11/sites/www.ibiblio.org/gentoo/ http://212.219.247.15/sites/www.ibiblio.org/gentoo/" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="X aalib acpi alsa apache2 atm avi cdr crypt cups dga dillo dvd encode flac gif gimpprint gpm gtk gtk2 imap imlib ipv6 java jpeg mad maildir memlimit mmx mozilla moznocompose moznoirc moznomail mpeg ncurses nls nptl oggvorbis opengl pam pdflib perl png pnp ppds python quicktime readline samba sasl sdl sse ssl tcpd tetex theora tiff truetype unicode usb wmf wxwindows x86 xface xml xml2 xmms xv zlib" --
did you recompile packages after upgrade ?
Sorry I did not understand. Which packages?
ogg123, etc... Also, it might be a driver/lib version thing... try using alsa-driver-1.0.3 with alsa-lib-1.0.3
same problem here. After upgrading to alsa-*-1.0.3, all soundapps crashing. Recompiling the apps doesn't work, they still crash. Downgrade to 1.0.2 -> all works fine again kernel: 2.6.4-rc2 (development-sources) but gentoo-dev-sources have the same problem. using alsa-drivers-1.0.3 is not an option for Kernel 2.6.x users, IMHO.
"but gentoo-dev-sources have the same problem. using alsa-drivers-1.0.3 is not an option for Kernel 2.6.x users, IMHO." Stefan, why do you say that? I actually recommend this to prevent the problem you are describing here, so this is what I'd suggest unless you can't compile it in which case I'd say stay with -lib-1.0.2
I said 'IMHO'. ;-) I thought, it's dangerous to mix the 2.4 world with 2.6. But if you say, it's ok and working, I will try it (as soon as the KDE 3.2.1 emerge is finished ;)).
alsa-driver-1.0.3 doesn't work. It does not compile properly, because of wrong kernel version. ok, lets wait for alsa-*-1.0.4 ;-) btw: it seems to be safe to emerge alsa-firmware-1.0.3 and alsa-utils-1.0.3. Just leave alsa-lib and alsa-oss with version 1.0.2. btw: there's a BUG in alsa-oss in then aoss script. But I think, this is a Bug for the ALSA-Team, not the Gentoo Team. but FYI: [/usr/bin/aoss] LD_PRELOAD=${exec_prefix}/lib/libaoss.so $* should be done this way: LD_PRELOAD=${exec_prefix}/lib/libaoss.so exec "$@" since $* passes the options with spaces/blanks not correctly. And the 'exec' saves some bytes of RAM ;)
Stefan: Did you file the bug on alsa's bug tracking system? Also, 1.0.3's -firmware and -utils don't depend on any changes made to alsa-lib between 1.0.2 and 1.0.3 to function... if you did a -U --deep, it'd get updated.
> Stefan: Did you file the bug on alsa's bug tracking system? not yet. also, it's really only alsa-lib-1.0.3. All other packages can be installed. But alsa-lib-1.0.3 is broken, at least with Kernel 2.6.4 (development-sources).
There's already a bug listed on ALSA's bugtracker. You can find it under the id #117. Something must have gone wrong with dmix in ALSA's 1.0.3 release. perex has offered a first diff to alsa-lib-1.0.3a, I hope it fixes that problem. I 'll post a ebuild including the patch later on, so everyone who's got problems using the dmix-plugin can try it out simple. BTW: it seems to be a intel8x0-only problem. Can anyone else confirm that behavior with another chip/driver?
Created attachment 27245 [details, diff] perex's first official fix As promised here's a updated ebuild. Extract the file's to /usr/local/portage/media-libs/alsa-lib/ (or wherever you store your own ebuilds). Sadly the fix didn't work for me, but maybe you've got more luck! PS: I'm not really familiar with creating ebuilds, so it could be my misstake, too ;)
yes, snd_cmipci and snd_via82xx doesn't work, too.
Comment on attachment 27245 [details, diff] perex's first official fix doesn't fix the problem
Created attachment 27269 [details] Fixed ebuild for alsa-lib-1.0.3a perex's second shot seems to fix the problem. Please try again with the new atached files and report your results.
please only post text... not tarballs...
I'm testing this out for inclusion now... If you have a working dmix configuration for i8x0, I'd be very grateful if you could send it my way as I was having a lot of trouble setting it up... although perhaps that trouble stemmed from this bug and I didn't realize it at the time... anyways... it'll probably make it into portage in the next day or so...
alsa-lib-1.0.3b has been released a few minutes ago. So you just have to rename media-libs/alsa-lib/alsa-lib-1.0.3.ebuild to media-libs/alsa-lib/alsa-lib-1.0.3b.ebuild News on http://www.alsa-project.org: 2004-03-13 alsa-lib 1.0.3b released - it fixes SIGSEGV problem for dmix plugin (when a specific GCC version is used) PS: You can mark this bug as fixed then
I've installed alsa 1.0.3b and have the following .asoundrc: pcm.dmixed { ipc_key 1025 type dmix slave.pcm "hw:0,0" } pcm.dsnooped { ipc_key 1027 type dsnoop slave.pcm "hw:0,0" } pcm.asymed { type asym playback.pcm "dmixed" capture.pcm "dsnooped" } pcm.!default { type plug slave.pcm "asymed" } using that configuration i get (still) the same error from artsd: artsd: pcm.c:5899: snd_pcm_mmap_commit: Assertion `frames <= snd_pcm_mmap_avail(pcm)' failed. I'm not sure if it's related to this bug report...
I comitted 1.0.3b Dick, can you please comment about your problem in the bug upstream: https://bugtrack.alsa-project.org/alsa-bug/bug_view_advanced_page.php?bug_id=0000117
Dick, how did you start your artsd? Maybe it's even just a little configuration problem. I didn't get artsd running in fullduplex mode, yet, but playback only mode works fine (artsd -a toss). I followed the steps mentioned in the WIKI: http://alsa.opensrc.org/index.php?page=DmixPlugin
first of all: alsa-lib-1.0.3b finally works!!! :-D and then: > I didn't get artsd running in fullduplex mode, yet, but > playback only mode works fine (artsd -a toss). I know only to ways to get aRts working with dmix. 1. aoss artsd -a toss -r 48000 2. artsd -a esd -r 48000 Since I start aRts with KDE, I use the 2nd way, because I don't need a wrapperscript with aoss. ESD works great with dmix/dsnoop/asym. And so the soundserver cascade aRts -> ESD -> ALSA/dmix or ALSA/jack finally works 100% OK, though it isn't realtime anymore... ;-) aRts is really broken. :-(
setting fixed...
