Age | Commit message (Collapse) | Author | Files | Lines |
|
Also add the missing codec registration.
|
|
|
|
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Drop the own true/false definitions as well
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Both are broken, just dummy funcs
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The recent code has moved the system PM ops conditionally with
CONFIG_PM_SLEEP. Also the more PM ops have been added in the upstream
and we need to follow that, too.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Forcibly include the system linux/stat.h.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
A part of it is actually used now.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
It's nowhere accessed, but needed for the recent drm audio component
changes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This fixes compile warnings in the recent ca0132 driver.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The LED audio trigger wrappers are improved to notify the changes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Needed for building 5.8 kernel code.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
snd_hda_queue_unsol_event() got dropped. Use the direct call of
snd_hdac_bus_queue_event() as a quick fix.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Add some essential typedefs, workqueue workaround, and the new
control access flag definition for 5.6 kernel.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
|
|
|
|
A simplified version.
|
|
Just to pass the compilation.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
hda_codec.h is moved to the common include/sound, so we need to add
the check in configure script and change the relevant codes accordingly.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Copied from the upstream code, mixed up both uapi and the main ones.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Just do nothing, only for avoiding the compile error.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Just a build workaround.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Added snd_ctl_apply_vmaster_slaves()
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The call for generic driver init was missing, so -m generic didn't work.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The pm_runtime_get_if_in_use() should return -EINVAL when no runtime
PM is used. But we didn't check it, thus it resulted as if the codec
power were down.
Fix it by referring to the actual power_save value in
pm_runtime_get_if_in_use().
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This eliminates the possibly inconsistencies in future.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Include <sound/hda_chmap.h> from the dist.
Since the new code uses hweight_long(), include linux/bitops.h, too.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Our WARN_ON() macro wrapper makes compiler unhappy when a pointer is
passed to the argument:
warning: passing argument 1 of ‘_WARN_ON’ makes integer from pointer without a cast [-Wint-conversion]
Fix it up.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The recent restructuring of HDMI chmap helper requires the workarounds
in hda-emu side, too.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The header linux/mod_devicetable.h is directly included from the
kernel tree, so that struct hda_device_id definition can be moved
around. For that, a few more type definitions, the additional uevent
wrapper and some fixes in linux/dmi.h are needed, too.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Align with the original behavior, return -EBUSY if cache_only is set
but no cache is available.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This allows building hda-emu even after recent i915 -> hda patches.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This resulted in the empty string, eventually aborts.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Each item is applied sequentially instead of only the last one.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
It's annoying to see errors at each run.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Just for letting the build go.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
It's still incomplete, the duplicated ctls need to be resolved later.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Both are needed to build the latest drivers.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Otherwise it oopses at error.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Now hda controller code was split, and bus_ops was dropped.
It requires more wrappers in linux/*.h and sound/*.h files, as well as
the adaption of builds of hda-emu.c.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
Now we support regmap in the upstream, so follows the hda-emu, too.
The regmap support codes are added without checks in configure.
If you need to build the pre 4.1-rc1 version, use the previous git
checkout.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Mention the new --with-kerneldir option, and update download URLs.
|
|
Now the HD-audio code is split and the core helper codes are put in
sound/hda/ directory. hda-emu also follows this change.
The --with-hdadir configure option is replaced with --with-kerneldir
option. Instead of passing the HD-audio directory, pass the root
directory of Linux kernel here.
configure script has been a bit modified to take care of two
directories, sound/hda and sound/pci/hda. The former is built in
lib subdirectory while the latter is still symlinked to hda
subdirectory and built in hda subdirectory.
|
|
|
|
- kstrtoint() (incomplete, no error check)
- kasprintf()
Also fix the malloc call in kvasprintf() to use the our own helper.
Otherwise it'll screw up at freeing this memory.
|
|
Don't call remove callback unless the device is really added
|
|
The HD-audio driver is now using the standard bus to bind with
devices, and there have been a lot of code changes along with it.
Adapt hda-emu to follow the changes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
input_report_key(), input_report_sync(), keybits[] and a few KEY_*
constants definitions for the recent patch to support HP mic-mute key
handling.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
It accepts variable arguments actually.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
We recently started using it on the kernel side.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
These verbs exist according to the driver, but we know nothing about
them, except they are used to calibrate some EQ.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This updates the previously submitted alsa-infos to include machines
certified by Canonical in recent years, in the hope is that this will
help upstream driver QA.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
|
|
Previously, if a machine had a modem codec, it was not included
in the total count of machines.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The driver sends these verbs on purpose. I can't verify that it
is correct to do so, but I hope the driver is doing the right thing here.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This is enabled on any modern distro and necessary for modern HDMI
stuff (with more than two outputs) to function correctly.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This way we get a more complete test over the codecs.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Also improve runner's error handling a little.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
In some cases (when codec index parameter was specified),
selecting a modem codec resulted in "ERROR -77", which was
confusing.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
When using the -i switch to specify an index, we need to deal with
two common cases:
1) All codecs belong to one card
2) There are as many cards as there are codecs
This patch makes sure we pick the correct pci ID and SSID for
the second case, too.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
|
|
Just use snprintf(). It's not correct, but works so far as it's used
only in hda_generic.c debug prints.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Some new HDMI commands were added recently to the HDA driver.
This patch makes hda-emu support these new commands (or at least,
not output errors).
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
When the base value is given via -b option, don't override assoc, seq
and misc numbers but keep the original bits.
|
|
... and add missing hda_beep.c in kernel/Makefile.am
|
|
3.15 kernel has lots of changes in HD-audio device management. Here
are some hacks to adapt them.
|
|
Just print via pr_xxx().
|
|
... in addition to the dummy symbol_put_addr() wrapper.
|
|
|
|
The option -P allows to set only the initial pin configs which may be
overridden by the driver's own setup. We want to provide sometimes
the pin configs corresponding to user_pin_configs sysfs. The new
option, -U, serves for that purpose.
|
|
|
|
|
|
|
|
Let's just ignore COEF accesses.
|
|
Now almost all VIA-specific verbs and workarounds should have been
covered.
|
|
Copied from kernel tree. Now verb definitions are split from
hda_codec.h in the latest code.
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
It's obvious...
|
|
When something wrong happens during PCM operation, properly close and
manage refcounts instead of just powering down and returning.
|
|
Just like -P option does, check the input file contents and skips to
[hint] section if it's a firmware patch file.
|
|
Like -H option, now hda-emu accepts the pin config value directly
specified via -P option like 0x0a=0x12345678.
|
|
If a string with '=' is passed to -H option, hda-emu parses the value
and handles as a hint string directly.
|
|
We often want to pass the pin configs used for snd-hda-intel patch
option. This patch allows user to pass such a file to -P option as
is.
|
|
This is needed for the new acpi functionality in patch_realtek.c.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
MacBookAir 6,1
MacBookAir 6,2
MacBookPro 11,2
|
|
Taken from alsa-info db
|
|
Source of information: http://www.x.org/docs/AMD/AMD_HDA_verbs_v2.pdf
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This allows for testing the feature as well as avoiding a warning
when testing the relevant codecs.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
... and allocate PCM runtime struct on the fly, too, instead of
keeping in static dummy structs in hda_test_pcm().
This will fix the invalid get from channel map controls.
|
|
Just simplistic ones, not portable at all.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
So far, unconditionally enabled but symbol_request() returns always NULL.
|
|
Taken from
https://bugzilla.novell.com/show_bug.cgi?id=841645
|
|
|
|
|
|
|
|
This is with a broken SSID, which caused a problem on 3.9/3.10 kernels.
|
|
Reference: https://bugzilla.novell.com/show_bug.cgi?id=833585
|
|
Reference: https://bugzilla.kernel.org/show_bug.cgi?id=59791
|
|
There is sometimes a secret node 0x51 on Realtek codecs, controlling
an equally secret DSP.
At least make sure the automated test suite doesn't fail when trying
to access this secret node.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The upstream commit 1ba65ae4bdbd43265c51ee4c30ff21a48124b6d8
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Taken from the upstream change (commit 1ca2f2ec).
|
|
For example, the jack polling code needs the self rescheduling.
Also, flush the pending work at the end of each command, also for
the wrong / help command. So you can execute the pending work by
pressing RETURN on the prompt.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Otherwise the power-up sequence is hidden during the proc output is
taken when power_save option is set.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Also support __GFP_ZERO flag.
|
|
Sending zero as stream tag is discouraged, because it means no
data, and the real code doesn't do that. Avoid it here too.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
This typo caused snd_hda_codec_prepare never to be executed by hda-emu.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The pinctl in alsa-info is not the initial
pinctl but the current, so sanify the current pinctl before
accepting it as the initial pinctl. (The initial pinctl is used
during pm test.)
Several false positives in the test tool is removed
with this improvement.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
When the codec is powered down via power_save option, we shouldn't
trigger the unsol event via jack command. Otherwise it kicks off the
power-up sequence again.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
The build got broken due to recent patch changes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
patch_hdmi.c _is_ compiled in the recent kernels, of course.
So the real fix is to compile patch_hdmi.c only when patch_atihdmi.c &
co don't exist.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
patch_hdmi.c is actually included by patch_atihdmi.c & co.
It's not directly compiled.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Also add -H option to give the initial hints.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Add options to initialize/randomize pinctls and amps before resume so
that we can check the resume procedure more intensively.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Currently it just returns an error.
|
|
The initialization is missing because it's done in hda_hwdep.c in the
upstream code, which isn't built in hda-emu.
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Copied from the kernel code.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Added in wrapper.h.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Instead of symlinkin to each of kernel/*.c file, create a master symlink
to $hdadir -> hda directory, and include each file from "../hda/xxx.c"
in kernel/*.c file. This makes the kernel/ subdirectory more consistent.
The configure script is modified for that, and also for defining the
conditional for each patch_*.c so that kernel/Makefile.am can compile
files selectively.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
When PCM stream operation is performed in PCM command and the given
channel count isn't supported, currently hda-emu shows an error but
continues the operation. It's often bad as it's easily overseen.
If this error is detected, abort the operation to make it clear.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
|
|
Since the recent driver doesn't overwrite the pin default cfg values
any longer, we should use snd_hda_codec_get_pincfg() for querying the
current pincfg values. Also give -r option to show the "original"
default cfg values.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
There is another variant of iMac 9,1, apparently, too...
|
|
|
|
|
|
|
|
|
|
|