On Sun, 2006-05-28 at 22:32 +0300, Paul Mundt wrote: > Hi Adrian, > > It's getting better, but still a ways to go.. > The bad news is that what I thought was working code wasn't. The dma_wait_for_completion function is still broken in the G2 DMA, so when the correct code is inserted the kernel oops: / # insmod /lib/modules/2.6.16-sh/kernel/sound/drivers/aica.ko ALSA /home/adrian/aica/aica/aica.c:630: ALSA Driver for Yamaha AICA Super Intelligent Sound Processor / # cat aine-email.wav > /dev/dsp scheduling while atomic: cat/0x00000002/1216 Call trace: [<8c1d41f2>] schedule+0x512/0x840 [<8c0fad20>] __sdivsi3+0x0/0x9a [<8c1d3ce0>] schedule+0x0/0x840 [<8c02f6a0>] prepare_to_wait+0x0/0xc0 [<8c14c704>] dma_wait_for_completion+0x84/0x100 [<8c01bca0>] getnstimeofday+0x0/0x40 [<8c032c4a>] ktime_get+0xa/0x20 [<8c14c704>] dma_wait_for_completion+0x84/0x100 [<8c0fad20>] __sdivsi3+0x0/0x9a [<8c1d3ce0>] schedule+0x0/0x840 [<8c02f6a0>] prepare_to_wait+0x0/0xc0 [<8c02f8c0>] autoremove_wake_function+0x0/0x40 [<8c02f8c0>] autoremove_wake_function+0x0/0x40 [<8c05ed2a>] vfs_write+0x6a/0x180 [<8c05ef14>] sys_write+0x34/0xa0 [<8c0051bc>] syscall_call+0xc/0xe [<8c05eee0>] sys_write+0x0/0xa0 _______________________________________________ Alsa-devel mailing list Alsa-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/alsa-devel