On Fri May 26 2006 00:32, Bob Tracy wrote: > Well, crud... The "es18xx.c" patch didn't make any difference for me. > Audio output still loops, and the machine still gets hosed (requires > hitting the reset switch to recover after successfully interrupting > "aplay" from the keyboard). I see you downloaded the patch. Any chance you could try using the "linux-image-2.6.14-2-alpha-generic_2.6.14-4_alpha.deb" package (that's what I am running) just to 100% make sure we are on same page? I also assume you modprobed the "snd-es18xx" module with the "isapnp=0" option? Then you used aplay to drop a raw sound clip through the card? Assuming nothing went wrong in the patching, I'm not quite sure what is wrong. The looping definitely does suggest that it is an interupt problem. What I'm guessing is happending is that the driver sets up the first 0.5s of sound and then never gets (or realizes it has got) the card's interupt to tell it that it is ready for the next 0.5s of sound. In my case it was because the interupt detection code in the driver was not managing to figure out that the interupt was for it. Clearly something extra is happening in yours. It looks like (from the code and experimentation) my card (an ES1888) can configure itself to whatever interupt I choose (currently 5, which is the default for the driver). Could it be that yours cannot? I know that 7 was also a common pick in early ISA sound cards. If you do a "modinfo", you see that the driver also has a "port", a "mpu_port", a "fm_port", a "irq", a "dma1", and a "dma2" option. The fact that you get some sound suggest to me that port and possibly a dma option is correct. I take it that you can use "amixer" (or "alsamixer") to turn up the volume on the endlessly repeated sound bit? What does "/etc/interupts" say? While aplay in running on my machine I see the IRQ (which ever I choose) incrementing relatively quickly and everything else staying pretty much stable (except the clock of course)? Before I came up with my patch it used to stay fixed. -T PS: I see there is also an "#undef REG_DEBUG" in the code. It looks like if you change that to a "#define REG_DEBUG" you get some extra printks. -- Tyson Whitehead (-twhitehe at uwo.ca -- WSC-) Computer Engineer Dept. of Applied Mathematics, Graduate Student- Applied Mathematics University of Western Ontario, GnuPG Key ID# 0xF7666BFF London, Ontario, Canada
Attachment:
pgpnE1AMlvvsW.pgp
Description: PGP signature