The fix in commit 77119d83a1 assumes that both ucm1 / ucm2 directories
have the configurations for the long name in the same directory. For v2
we changed the location to card_name/card_long_name aka
driver_name/driver_long_name to make the directory layout more structured.
Fixes: 77119d83a1 ("ucm: Fix opening of master-configs by the card's longname")
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
There is no need to manually 0 terminate the buffer with snprintf,
only strncpy has the very unfortunate behavior of not guaranteeing 0
termination.
Likewise there is no need to substract one from the buffer size of
the buffer passed to snprintf.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Unless environment variables are set, then configuration_filename()
when initially called by load_master_config() will first try to find
the requested master-config under <prefix>/alsa/ucm2 and then under
<prefix>/alsa/ucm.
Once a master-config is found this way, we should set conf_format to
match, so that subsequent lookups only look under the same directory
as where the master-config was found.
This fixes 2 problems:
1. uc_mgr_config_load() looking under <prefix>/alsa/ucm for includes for
UCM2 profiles because it is called with uc_mgr->conf_format as format
and before this commit that would stay 0 when autodetecion is used.
2. parse_verb_file() possibly loading an UCM2 verb-file for an UCM1 profile,
the chance of this happening is small as this means that even though
there is no UCM2 master-config there is an UCM2 profile dir matching
uc_mgr->conf_file_name, which would be weird.
Fixes: aba2260ae7 ("ucm: switch to ucm2 directory and v2 format, keep backward compatibility")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
uc_mgr_import_master_config() first calls load_master_config()
with the card's longname and if that fails then calls it again with the
card_name.
Before this commit configuration_filename() would force conf_format to 2
when it the access(fn, R_OK) test failed for the ucm1 longname
master-config filename.
This would cause configuration_filename() to blindly return a filename
under <prefix>/ucm2 without seeing if that is actually there and without
trying to fallback to the old profiles under <prefix>/ucm, breaking the
loading of UCM1 profiles by card_name.
This commit fixes this by modifying configuration_filename() to not set
conf_format when checking for the UCM1 config filename fails.
Instead, to make sure that any errors about opening the file still report
the new path, configuration_filename() now resets the filename to the UCM2
path if the access() check has failed for both the UCM2 and UCM1 paths,
without touching conf_format.
Fixes: aba2260ae7 ("ucm: switch to ucm2 directory and v2 format, keep backward compatibility")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Prior to commit aba2260ae7 ("ucm: switch to ucm2 directory and v2 format,
keep backward compatibility").
The filename build in parse_verb_file() was build like this:
<prefix>/<uc_mgr->conf_file_name>/<file>
Where uc_mgr->conf_file_name would contain either the card_name or the
card's longname depending on the detection of a longname based config in
uc_mgr_import_master_config().
While the filename used in load_master_config() was build like this:
<prefix>/<card_name>/<card_name>.conf
And uc_mgr_import_master_config() first calls load_master_config()
with the card's longname and if that succeeds it overwrites
uc_mgr->conf_file_name with the longname so that the subsequent uses
of uc_mgr->conf_file_name in parse_verb_file() correctly use the longname.
But the new configuration_filename() helper added in commit aba2260ae7
_always_ builds the filename like this:
<prefix>/<uc_mgr->conf_file_name>/<file><suffix>
This breaks the loading of the master-config by its longname, as when
the longname is tried uc_mgr->conf_file_name still contains the card_name.
This commit fixes this by adding a dir parameter to configuration_filename()
and restoring the old behavior by passing card_name as dir in
load_master_config().
Fixes: aba2260ae7 ("ucm: switch to ucm2 directory and v2 format, keep backward compatibility")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
I omitted to use parent paths and the proper error code
when no paths are defined.
Fixes: f600310954 ("conf: improve the include paths code")
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Because of the nature of the structured parser, it is not guaranteed
that the order of the parsed tree is equal to the lines in the source.
Add possibility to insert the If results before or after the selected
identifier like:
SectionDevice."Speaker" {
...
}
If.1 {
Condition { ... }
Before.SectionDevice "Speaker"
True { ... }
}
If.2 {
Condition { ... }
After.SectionDevice "Speaker"
True { ... }
}
If the "After" or "Before" identifiers are not found, the result is added
after the last entry in the block.
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Allows ucm to override minimum buffer level for platforms that
incorrectly report their buffer level.
Signed-off-by: Curtis Malainey <cujomalainey@chromium.org>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Add a sys path to the EDID file for the corresponding HDMI device so
userspace can use if needed
Signed-off-by: Curtis Malainey <cujomalainey@chromium.org>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
The syntax is simple:
If./any-if-identificator/ {
Condition {
Type /type_here/
/optional defines/
}
True {
/block used when condition is evaluated as true/
}
False {
/block used when condition is evaluated as false/
}
}
The Type "ControlExists" is implemented:
Condition {
Type ControlExists
Device "hw:${CardId}"
Control "iface=CARD,name='Headphone Jack'"
}
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Those two variables are supported:
${ConfName} - configuration file name
${CardId} - card identification string (like PCH)
${CardName} - card name (driver)
${CardLongName} - card long name (driver)
${env:ENV_NAME} - returns the environment variable ENV_NAME
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Some clients like pulseaudio wants to access the multiple instances
of sound cards. This patch adds prefixes like "hw:" to the card_name
argument to handle this. The card index (value) or card identification
(string) can be used for this prefix.
Also the prefix "strict:" was added to avoid the driver name and
driver long name matching. It might be useable for use case
configurations which are not bound to the one sound card.
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
- various cleanups (more straight code and allocations)
- do not add the base config path /usr/share/alsa to the
explicit include list of directories
(it's not wanted for ucm configs)
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Most of PCM API functions have snd_pcm_lock()/unlock() wraps for the
actual calls of ops, and some plugins try to unlock/relock internally
for the given PCM object. This, unfortunately, causes a problem in
some configurations and leads to the unexpected behavior or deadlock.
The main problem is that we call snd_pcm_lock() with the given PCM
object, while calling the ops with pcm->op_arg or pcm->fast_op_arg as
the slave PCM object. Meanwhile the plugin code assumes that the
passed PCM object is already locked, and calls snd_pcm_unlock().
This bug doesn't hit always because in most cases pcm->op_arg and
fast_op_arg are identical with pcm itself. But in some configurations
they have different values, so the problem surfaces.
This patch is an attempt to resolve these inconsistencies. It
replaces most of snd_pcm_lock()/unlock() calls with either pcm->op_arg
or pcm->fast_op_arg, depending on the call pattern, so that the plugin
code can safely run snd_pcm_unlock() to the given PCM object.
Fixes: 54931e5a54 ("pcm: Add thread-safety to PCM API")
Reported-by: Ben Russell <thematrixeatsyou@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_pcm_set_chmap() leaks the memory returned from snd_pcm_get_chmap()
without releasing. Add the missing free() call as well as a slight
code refactoring.
Reported-by: Conrad Jones
BugLink: https://github.com/alsa-project/alsa-lib/pull/11
Fixes: d20e24e5d1 ("chmap: Always succeed setting the map to what it already is")
Signed-off-by: Takashi Iwai <tiwai@suse.de>
