On Fri, Oct 24, 2008 at 09:09:33PM +0200, Robert Jarzmik wrote:
Mark Brown broonie@sirena.org.uk writes:
All this patch does is change the limit, there is still a limit on the length of control names in ALSA but as I say it's purely a cosmetic issue.
I don't quite agree on the "cosmetic".
If I recall correctly, that name is used in dapm path setup. If the name is incorrect (truncated), some dapm elements are not switched on. I'm speaking from memory, I'll try to browse my git repository to track it back. Notice that I'm really busy ATM, so I don't have enough manpower to test immediatly.
It looks like you might've been right after all. I still can't see where this makes a difference to the speaker paths but the patch below seems to resolve the issues I'm seeing here.
Oliver, could you test the patch below please?
diff --git a/sound/soc/soc-dapm.c b/sound/soc/soc-dapm.c index 28da235..fdb4d2c 100644 --- a/sound/soc/soc-dapm.c +++ b/sound/soc/soc-dapm.c @@ -293,7 +293,7 @@ static int dapm_new_mixer(struct snd_soc_codec *codec, struct snd_soc_dapm_widget *w) { int i, ret = 0; - char name[32]; + size_t name_len; struct snd_soc_dapm_path *path;
/* add kcontrol */ @@ -307,11 +307,16 @@ static int dapm_new_mixer(struct snd_soc_codec *codec, continue;
/* add dapm control with long name */ - snprintf(name, 32, "%s %s", w->name, w->kcontrols[i].name); - path->long_name = kstrdup (name, GFP_KERNEL); + name_len = 2 + strlen(w->name) + + strlen(w->kcontrols[i].name); + path->long_name = kmalloc(name_len, GFP_KERNEL); if (path->long_name == NULL) return -ENOMEM;
+ snprintf(path->long_name, name_len, "%s %s", + w->name, w->kcontrols[i].name); + path->long_name[name_len - 1] = '\0'; + path->kcontrol = snd_soc_cnew(&w->kcontrols[i], w, path->long_name); ret = snd_ctl_add(codec->card, path->kcontrol);