| Message ID | 1402221536-26473-1-git-send-email-anssi.hannula@iki.fi (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
At Sun, 8 Jun 2014 12:58:56 +0300, Anssi Hannula wrote: > > On most HDA Intel HDMI devices, the channel map list is approx. 500 ints > in size, making the 256-sized buffer used by > snd_pcm_query_chmaps_from_hw() too small and causing the query to fail > (NULL result to caller, ENOMEM). > > Bump the buffer size to 2048 ints. > > Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi> > --- > > Found while writing XBMC support. > > I don't like just bumping the buffer size, so I hope you have some idea > of a better solution? > I guess one way would be just to retry with a dynamically allocated > bigger buffer on ENOMEM, if there is nothing better... For user-space, this stack size doesn't matter too much, I suppose. Unless we'll hit really too big size, I guess we can keep in this way. So, I applied it as is. thanks, Takashi > > src/pcm/pcm_hw.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/src/pcm/pcm_hw.c b/src/pcm/pcm_hw.c > index 9ff83a9..ed83197 100644 > --- a/src/pcm/pcm_hw.c > +++ b/src/pcm/pcm_hw.c > @@ -1051,7 +1051,7 @@ snd_pcm_query_chmaps_from_hw(int card, int dev, int subdev, > { > snd_ctl_t *ctl; > snd_ctl_elem_id_t *id; > - unsigned int tlv[256], *start; > + unsigned int tlv[2048], *start; > snd_pcm_chmap_query_t **map; > int i, ret, nums; > > -- > 1.8.4.5 >
08.06.2014 13:03, Takashi Iwai kirjoitti: > At Sun, 8 Jun 2014 12:58:56 +0300, > Anssi Hannula wrote: >> >> On most HDA Intel HDMI devices, the channel map list is approx. 500 ints >> in size, making the 256-sized buffer used by >> snd_pcm_query_chmaps_from_hw() too small and causing the query to fail >> (NULL result to caller, ENOMEM). >> >> Bump the buffer size to 2048 ints. >> >> Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi> >> --- >> >> Found while writing XBMC support. >> >> I don't like just bumping the buffer size, so I hope you have some idea >> of a better solution? >> I guess one way would be just to retry with a dynamically allocated >> bigger buffer on ENOMEM, if there is nothing better... > > For user-space, this stack size doesn't matter too much, I suppose. > Unless we'll hit really too big size, I guess we can keep in this > way. So, I applied it as is. OK. I'm a bit afraid about HDMI 2.x, though, with its reported 32 channel audio support... but I guess we can tackle that when we actually know how exactly it works and get driver support. > > thanks, > > Takashi > > >> >> src/pcm/pcm_hw.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/src/pcm/pcm_hw.c b/src/pcm/pcm_hw.c >> index 9ff83a9..ed83197 100644 >> --- a/src/pcm/pcm_hw.c >> +++ b/src/pcm/pcm_hw.c >> @@ -1051,7 +1051,7 @@ snd_pcm_query_chmaps_from_hw(int card, int dev, int subdev, >> { >> snd_ctl_t *ctl; >> snd_ctl_elem_id_t *id; >> - unsigned int tlv[256], *start; >> + unsigned int tlv[2048], *start; >> snd_pcm_chmap_query_t **map; >> int i, ret, nums; >> >> -- >> 1.8.4.5 >>
diff --git a/src/pcm/pcm_hw.c b/src/pcm/pcm_hw.c index 9ff83a9..ed83197 100644 --- a/src/pcm/pcm_hw.c +++ b/src/pcm/pcm_hw.c @@ -1051,7 +1051,7 @@ snd_pcm_query_chmaps_from_hw(int card, int dev, int subdev, { snd_ctl_t *ctl; snd_ctl_elem_id_t *id; - unsigned int tlv[256], *start; + unsigned int tlv[2048], *start; snd_pcm_chmap_query_t **map; int i, ret, nums;
On most HDA Intel HDMI devices, the channel map list is approx. 500 ints in size, making the 256-sized buffer used by snd_pcm_query_chmaps_from_hw() too small and causing the query to fail (NULL result to caller, ENOMEM). Bump the buffer size to 2048 ints. Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi> --- Found while writing XBMC support. I don't like just bumping the buffer size, so I hope you have some idea of a better solution? I guess one way would be just to retry with a dynamically allocated bigger buffer on ENOMEM, if there is nothing better... src/pcm/pcm_hw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)