On Tue, 03 Jan 2017 14:49:40 +0100, Enrico Mioso wrote: > > hello Tahaski, > hello everyone reading this message. > > Thank you for the patience you had with me. Now I finally came to set up this system for crash analysis, compiling the kernel with debug informations and frame pointers and so on. I set up kdump also. Unfortunately, I can get no useful output. > The last things the system wrote before dying have been reconstructed. I noted down decimal values with the help of my mum. She wrote them down on a piece of paper. > I am hitting the following messages: > azx_get_response timeout, switching to polling mode: last cmd=0x014ba000 > azx_get_response timeout, switching to single_cmd mode: last cmd=0x014ba000 It's reading the output amp on NID 0x14, the speaker pin. It's weird that such a verb triggers the problem. Usually the unexpected behavior is triggered at the widget power up/down. If the switching to single cmd mode causes an issue, you can disable by the patch like below. This may lead to the continuous errors, though. Takashi --- diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c index 500878556578..43bf151d8712 100644 --- a/sound/pci/hda/hda_controller.c +++ b/sound/pci/hda/hda_controller.c @@ -861,6 +861,7 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr, return -EIO; } +#if 0 /* a fatal communication error; need either to reset or to fallback * to the single_cmd mode */ @@ -875,6 +876,7 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr, chip->single_cmd = 1; hbus->response_reset = 0; snd_hdac_bus_stop_cmd_io(bus); +#endif return -EIO; } _______________________________________________ Alsa-devel mailing list Alsa-devel@xxxxxxxxxxxxxxxx http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
