Discussion:
[linux-audio-user] RME Hammerfall DSP Multiface firmware loading problem
Matt Kleffner
19 years ago
Permalink
Hello,

Previously I could load the firmware onto my RME Multiface, but I
haven't used it for a few months. I recently reinstalled debian sarge
on my machine, but I can no longer get the firmware to load. I found
previous emails on this problem but I found no clear solution. I am
using a 2.6.14.3 kernel with the alsa drivers compiled as modules, and
the drivers are newer than the buggy (for RME) 1.0.8 versions. I've
also put snd_hdsp in my initrd image as mentioned in
http://lalists.stanford.edu/lad/2005/11/0149.html

The version info and error messages are listed below. Any thoughts on
what is wrong? Does the firmware and hdsploader version (1.0.9rc3)
have to match the alsa driver version (1.0.10rc1)? Did I miss a
solution on this or another list? Any help is appreciated.

- Matt

Kernel messages at boot time:
Hammerfall-DSP: wait for FIFO status <= 0 failed after 30 iterations
Hammerfall-DSP: couldn't get firmware from userspace. try using hdsploader
Hammerfall-DSP: card initialization pending : waiting for firmware

Results from hdsploader:
hdsploader - firmware loader for RME Hammerfall DSP cards
Looking for HDSP + Multiface or Digiface cards :
Card 0 : VIA 82C686A/B rev50 with ICE1232 at 0xa800, irq 9
Card 1 : RME Hammerfall DSP at 0xdd000000, irq 9
Upload firmware for card hw:1
Hwdep ioctl error on card hw:1 : Input/output error.

lspci | grep Hammerfall:
0000:00:0a.0 Multimedia audio controller: Xilinx Corporation RME
HammerfallDSP(rev 32)

cat /proc/asound/version:
Advanced Linux Sound Architecture Driver Version 1.0.10rc1 (Mon Sep 12
08:13:092005 UTC).

hdsploader source:
alsa-firmware-loaders deb package, AGNULA, package version 1.0.9rc3-0

firmware (from alsa-firmware-1.0.9rc3.tar.bz2):
multiface_firmware.bin
multiface_firmware_rev11.bin
Paul Davis
19 years ago
Permalink
...
try power-cycling *everything*. the computer, (including unplugging the
power cord (and battery if its a laptop), and the multiface box.

i've run into this condition once or twice, and the only solution i have
found is a complete power cycle of everything.


--p
Matt Kleffner
19 years ago
Permalink
I powered down and unplugged everything. After reconnecting everythingand turning the (desktop) computer on, the "Host Error" light on themultiface went out and the kernel log contained:
Hammerfall-DSP: wait for FIFO status <= 0 failed after 30 iterationsHammerfall-DSP: loading firmwareHammerfall-DSP: finished firmware loading
Now, when I run hdspconf and select any option (e.g. change samplingrate), I get an error for each option selected. For example, if Ichange the sampling rate to 44.1 kHz the radio button selection"jumps" back to 48 kHz.
The stdout of hdspconf looks like:
HDSPConf 1.4 - Copyright (C) 2003 Thomas Charbonnel <***@undata.org>This program comes WITH ABSOLUTELY NO WARRANTYHDSPConf is free software, see the file copying for details
Looking for HDSP cards :Card 0 : VIA 82C686A/B rev50 with ICE1232 at 0xa800, irq 9Card 1 : RME Hammerfall DSP + Multiface at 0xdd000000, irq 9Multiface found !1 Hammerfall DSP card found.Error accessing ctl interface on card hw:1Error accessing ctl interface on card hw:1Error accessing ctl interface on card hw:1
The last message is repeated every time I attempt to change a setting,and it doesn't show up at all until I try to change a setting. If Iclose hdspconf and run
amixer -c 1 cset numid=11 2(syntax from http://www.linuxjournal.com/article/7024)
the correct (44.1 kHz) setting shows up when I reopen hdspconf.Changing the rate back to 48 kHz in hdspconf results in the sameerror, except the selection now remains fixed at 44.1 kHz.
Any thoughts on why this is happening? Is the version of HDSPConf too old?
- Matt
On 12/16/05, Paul Davis <***@linuxaudiosystems.com> wrote:> try power-cycling *everything*. the computer, (including unplugging the> power cord (and battery if its a laptop), and the multiface box.>> i've run into this condition once or twice, and the only solution i have> found is a complete power cycle of everything.>>> --p
Paul Davis
19 years ago
Permalink
...
i have a nasty feeling that ALSA 1.0.10 may have broken something here.
this is the second report of this in a few weeks. i know its ugly and
horrible and time consuming, but it would really help if you could back
up to 1.0.9 and test that.

--p
Matt Kleffner
19 years ago
Permalink
Sorry about the formatting of my last message - I'm not sure how I got gmail
to do that! Hopefully this one will be formatted correctly...

Summary: hdspconf works with 1.0.9b in 2.6.13.5; FIFO error remains

The easiest way for me to revert to 1.0.9 was to compile a slightly older
vanilla 2.6 kernel. (1.0.10 is in the vanilla 2.6.14.3 kernel.) For those who
are interested: I could not find any online kernel source browsers that were
functional, so I started downloading other versions of the kernel. I got lucky
and the next lower version, 2.6.13.5, has the following in
linux-2.6.13.5/include/sound/version.h :
/* include/version.h. Generated by configure. */
#define CONFIG_SND_VERSION "1.0.9b"
#define CONFIG_SND_DATE " (Thu Jul 28 12:20:13 2005 UTC)"

Therefore, I compiled this kernel. I powered down, removed all cables, plugged
all of them back in and powered on. Here is the kernel log:
Hammerfall-DSP: wait for FIFO status <= 0 failed after 30 iterations
Hammerfall-DSP: loading firmware
Hammerfall-DSP: finished firmware loading

hdspconf now works (with 1.0.9b and 2.6.13.5) and I can change any of its
settings. However, the kernel log messages have not changed from 2.6.14.3.
Is the FIFO error important? Using the amixer workaround to set the sampling
rate and compiling 2.6.14.3 with the PREEMPT flags set, I was able to
successfully record audio with ardour - even though the FIFO error was present
at boot time.

If I find I need 1.0.9b with 2.6.14.3, how challenging is it to downgrade alsa
in the kernel source?

- Matt
...
Lee Revell
19 years ago
Permalink
Post by Matt Kleffner
If I find I need 1.0.9b with 2.6.14.3, how challenging is it to
downgrade alsa in the kernel source?
It's pointless, you need to identify what change introduced the bug so
it can be fixed for 2.6.14.4.

Lee

Loading...