This patch adds support for the RME RPM devices. It's mostly based on
Florian Faber's previous patch against hdspmixer 1.0.23,
forwarded-ported to the current hdspmixer code.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
As an addition to the previous commit, let's also cover the 3rd case
when a preset file was written with hdspmixer v1.0.24 or v1.0.24.1.
In this case, no magic header will be present, but the file size would
differ from the pre 1.0.24 format.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When reading a preset file, v1.0.23 only used 26 channels instead of 64.
Reading 64 channels from a 26 channel file won't work, hence set it
depending on the file format version.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The old hdspmixer (prior to e24e56795e,
that is pre-1.0.24) has used the hard-coded value "14" for
reading/storing 14 ints from/to preset files, however, it's actually
HDSP_MAX_DEST that should be used.
HDSP_MAX_DEST was bumped from 14 to 32 to allow for MADI cards (32
stereo pairs equal 64 output channels on MADI cards)
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cards like the multiface/digiface have additional headphones out. Those
were not initialized in the presets due to wrong loop boundaries:
maxdest represents the amount of physical stereo pairs, and chnls is
either equal or less, so the output fader array needs more iterations
than the playback section.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
With the new "store current settings to the virtual 9th preset" before
switching cards code, one needs to make sure the actual mixer state is
loaded with sane values, either from the preset file or a generic
builtin preset.
Calling preset_change(1) is sufficient, setting all the required data.
However, in case of more than one RME card in the system, one needs to
call this function for each card, otherwise, some of the cards store
uninitialized data to the 9th preset, resulting in a weird mixer state
afterwards.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When running with more than one card, switching cards would lose any
changes made to the current card. To avoid this inconvenience, save the
current settings to the virtual 9th preset and restore them when
switching back.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When switching cards, all current settings are lost. To have a place
where to store them, let's add a virtual 9th preset which can be used as
a temporary scratch pad.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Each preset can have a different level meter setting (FS samples for
OVR, release rate, minimum level and so on).
These settings were not saved/restore to/from the on-disk file. This
patch adds the missing functionality.
Unfortunately, the current on-disk format is a 1:1 binary dump from
memory without any header information. In other words, this commit will
break backward compatibility with older hdspmixers, that is, new preset
files cannot be read by older versions of hdspmixer. However, we can
still read the old mix files and save them in the new format.
I hence bumped the version, so users know to re-create their files after
upgrading to 1.11.
Bug discovered by Raphaël Doursenaud from ematech.fr.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When running with more than one card, it isn't obvious which card is
shown.
Store the ALSA cardname in the corresponding class and show it in the
window title upon switching cards.
Also, don't show "(null)" but "(unsaved)" in case the user hasn't
selected a preset file.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Some cards (like multiface) have more physical output ports than
playback ports, mostly because of additional headphones out.
For those cards, the old abstraction model of channels_input and
channels_output doesn't fit, so let's introduce channels_output.
Of course, channels_output is always 2*max_dest at the given speed_mode
(SS/DS/QS), so one could extend this idea, store all destination
settings in channels_output[3] (one for each speed mode) and rip off the
massive code duplication for setting maxdest or max_dest respectively.
Note that dest_map_whatever_speed_mode's array size indirectly defines
the right value for channels_output (read: even more unwanted
redundancy)
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When selecting preset 6 or 7 on AES(32), hdspmixer has caused a segfault
due to indirect out of bound access on the destination label array.
The amount of destinations is the number of physical stereo
pairs, so it's usually half the channel count, in some cases one more if
there are additional headphone jacks.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, hdsp and hdspm use different ioctls. Consequently, the metering
is wrong. To avoid code duplication, use pointers to the corresponding
struct members.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
If three (or more) RME cards are installed in one box, hdspmixer will
try to open a non-existing 4th card, causing an error in snd_ctl_open
and finally terminates itself.
cards[] is a static array, and one must not read beyond the last
element. The solution is far from elegant, however, it's a rather
unintrusive change.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The copyright list got longer, so we need more vertical space.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Code provided by Fredrik Lingvall <fredrik.lingvall@gmail.com>
It seems the PCIe (AES) and PCI (AES32) versions behave the same, so we
can kill two birds with one stone.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Merged the work by Florian Faber that's distributed separately as
hdspmixer64.
Code taken from http://wiki.linuxproaudio.org/index.php/App:hdspmixer_64
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
From debian bug#450805:
We are using Hammerfall DSP cards. After booting, their audio output
remains silent until hdspmixer is started. No interaction in the GUI
of hdspmixer is necessary to unmute the first HDSP card; however,
further cards are only unmuted when activating the respective GUI
page ("2", "3"). Apparently, hdspmixer does some automatic
initialization of the card when activating the page.
Since we'd like to have a fully automatic startup, the following
patch activates the page for each existing card on startup, thereby
initializing them. There are surely more elegant solutions, but this
patch is tested and solves the problem for us.
This patch changes .iface to SNDRV_CTL_ELEM_IFACE_MIXER whre _PCM or
_HWDEP was used in controls that are not associated with a specific PCM
(sub)stream or hwdep device, and changes some controls that got
inconsitent .iface values due to copy+paste errors. Furthermore, it
makes sure that all control that do use _PCM or _HWDEP use the correct
number in the .device field.
