Bug: LADSPA implementation - no up/downmix handling

Bug: LADSPA implementation - no up/downmix handling

[Andreas Hartmann]

[-- Attachment #1.1: Type: text/plain, Size: 340 bytes --]

Hi,

Using the surround_encoder of the swh-plugins I noticed, that the alsa 
implementation only assumes an equal number of in- and output channels for 
ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and leads to 
a Segfault shortly after.
There needs to be a check and correct handling for up/downmix scenarios.

Andreas

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: Bug: LADSPA implementation - no up/downmix handling

[Andreas Hartmann]

[-- Attachment #1.1: Type: text/plain, Size: 428 bytes --]

Am Samstag, 12. September 2015, 22:10:40 schrieb Andreas Hartmann:
> Hi,
> 
> Using the surround_encoder of the swh-plugins I noticed, that the alsa
> implementation only assumes an equal number of in- and output channels for
> ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and leads
> to a Segfault shortly after.
> There needs to be a check and correct handling for up/downmix scenarios.
> 
> Andreas

ping

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: Bug: LADSPA implementation - no up/downmix handling

[Takashi Iwai]

On Sat, 12 Sep 2015 22:10:40 +0200,
Andreas Hartmann wrote:
> 
> Hi,
> 
> Using the surround_encoder of the swh-plugins I noticed, that the alsa 
> implementation only assumes an equal number of in- and output channels for 
> ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and leads to 
> a Segfault shortly after.
> There needs to be a check and correct handling for up/downmix scenarios.

Could you give the back trace?  Or at best submit a fix patch?


thanks,

Takashi

Re: Bug: LADSPA implementation - no up/downmix handling

[Andreas Hartmann]

[-- Attachment #1.1: Type: text/plain, Size: 1415 bytes --]

Am Mittwoch, 23. September 2015, 16:26:58 schrieb Takashi Iwai:
> On Sat, 12 Sep 2015 22:10:40 +0200,
> 
> Andreas Hartmann wrote:
> > Hi,
> > 
> > Using the surround_encoder of the swh-plugins I noticed, that the alsa
> > implementation only assumes an equal number of in- and output channels for
> > ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and
> > leads to a Segfault shortly after.
> > There needs to be a check and correct handling for up/downmix scenarios.
> 
> Could you give the back trace?  Or at best submit a fix patch?
> 
> 
> thanks,
> 
> Takashi
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Sure here you go

#0  0x00007ffff78623ef in snd_pcm_ladspa_allocate_memory (pcm=0x626fa0, 
pcm=0x626fa0, pcm=0x626fa0, ladspa=0x621ad0) at pcm_ladspa.c:753
#1  snd_pcm_ladspa_init (pcm=0x626fa0) at pcm_ladspa.c:834
#2  0x00007ffff7842946 in snd_pcm_plugin_prepare (pcm=0x626fa0) at 
pcm_plugin.c:171
#3  0x00007ffff784290f in snd_pcm_plugin_prepare (pcm=0x62c760) at 
pcm_plugin.c:162
#4  0x000000000040256a in ?? ()
#5  0x00007ffff7222ec5 in __libc_start_main (main=0x401d80, argc=4, 
argv=0x7fffffffde28, init=<optimized out>, fini=<optimized out>, 
rtld_fini=<optimized out>,
    stack_end=0x7fffffffde18) at libc-start.c:287
#6  0x0000000000402fdd in ?? ()

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: Bug: LADSPA implementation - no up/downmix handling

[Takashi Iwai]

On Thu, 24 Sep 2015 12:44:53 +0200,
Andreas Hartmann wrote:
> 
> Am Mittwoch, 23. September 2015, 16:26:58 schrieb Takashi Iwai:
> > On Sat, 12 Sep 2015 22:10:40 +0200,
> > 
> > Andreas Hartmann wrote:
> > > Hi,
> > > 
> > > Using the surround_encoder of the swh-plugins I noticed, that the alsa
> > > implementation only assumes an equal number of in- and output channels for
> > > ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and
> > > leads to a Segfault shortly after.
> > > There needs to be a check and correct handling for up/downmix scenarios.
> > 
> > Could you give the back trace?  Or at best submit a fix patch?
> > 
> > 
> > thanks,
> > 
> > Takashi
> > _______________________________________________
> > Alsa-devel mailing list
> > Alsa-devel@alsa-project.org
> > http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
> 
> Sure here you go
> 
> #0  0x00007ffff78623ef in snd_pcm_ladspa_allocate_memory (pcm=0x626fa0, 
> pcm=0x626fa0, pcm=0x626fa0, ladspa=0x621ad0) at pcm_ladspa.c:753
> #1  snd_pcm_ladspa_init (pcm=0x626fa0) at pcm_ladspa.c:834
> #2  0x00007ffff7842946 in snd_pcm_plugin_prepare (pcm=0x626fa0) at 
> pcm_plugin.c:171
> #3  0x00007ffff784290f in snd_pcm_plugin_prepare (pcm=0x62c760) at 
> pcm_plugin.c:162
> #4  0x000000000040256a in ?? ()
> #5  0x00007ffff7222ec5 in __libc_start_main (main=0x401d80, argc=4, 
> argv=0x7fffffffde28, init=<optimized out>, fini=<optimized out>, 
> rtld_fini=<optimized out>,
>     stack_end=0x7fffffffde18) at libc-start.c:287
> #6  0x0000000000402fdd in ?? ()

It looks rather like a simply typo.  Does the patch below work?


Takashi

diff --git a/src/pcm/pcm_ladspa.c b/src/pcm/pcm_ladspa.c
index 631ee0f35b73..6552b4379ec0 100644
--- a/src/pcm/pcm_ladspa.c
+++ b/src/pcm/pcm_ladspa.c
@@ -749,7 +749,7 @@ static int snd_pcm_ladspa_allocate_memory(snd_pcm_t *pcm, snd_pcm_ladspa_t *lads
                                 return -ENOMEM;
                         }
 			for (idx = 0; idx < instance->input.channels.size; idx++) {
-			        chn = instance->output.channels.array[idx];
+			        chn = instance->input.channels.array[idx];
 			        if (pchannels[chn] == NULL && chn < ichannels) {
 			                instance->input.data[idx] = NULL;
 			                continue;

Re: Bug: LADSPA implementation - no up/downmix handling

[Takashi Iwai]

On Fri, 25 Sep 2015 20:51:38 +0200,
Andreas Hartmann wrote:
> 
> Am Donnerstag, 24. September 2015, 15:07:14 schrieben Sie:
> > On Thu, 24 Sep 2015 12:44:53 +0200,
> > 
> > Andreas Hartmann wrote:
> > > Am Mittwoch, 23. September 2015, 16:26:58 schrieb Takashi Iwai:
> > > > On Sat, 12 Sep 2015 22:10:40 +0200,
> > > > 
> > > > Andreas Hartmann wrote:
> > > > > Hi,
> > > > > 
> > > > > Using the surround_encoder of the swh-plugins I noticed, that the alsa
> > > > > implementation only assumes an equal number of in- and output channels
> > > > > for
> > > > > ladspa plugins. This occurs from Line 752 onwards in pcm_ladspa.c and
> > > > > leads to a Segfault shortly after.
> > > > > There needs to be a check and correct handling for up/downmix
> > > > > scenarios.
> > > > 
> > > > Could you give the back trace?  Or at best submit a fix patch?
> > > > 
> > > > 
> > > > thanks,
> > > > 
> > > > Takashi
> > > > _______________________________________________
> > > > Alsa-devel mailing list
> > > > Alsa-devel@alsa-project.org
> > > > http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
> > > 
> > > Sure here you go
> > > 
> > > #0  0x00007ffff78623ef in snd_pcm_ladspa_allocate_memory (pcm=0x626fa0,
> > > pcm=0x626fa0, pcm=0x626fa0, ladspa=0x621ad0) at pcm_ladspa.c:753
> > > #1  snd_pcm_ladspa_init (pcm=0x626fa0) at pcm_ladspa.c:834
> > > #2  0x00007ffff7842946 in snd_pcm_plugin_prepare (pcm=0x626fa0) at
> > > pcm_plugin.c:171
> > > #3  0x00007ffff784290f in snd_pcm_plugin_prepare (pcm=0x62c760) at
> > > pcm_plugin.c:162
> > > #4  0x000000000040256a in ?? ()
> > > #5  0x00007ffff7222ec5 in __libc_start_main (main=0x401d80, argc=4,
> > > argv=0x7fffffffde28, init=<optimized out>, fini=<optimized out>,
> > > rtld_fini=<optimized out>,
> > > 
> > >     stack_end=0x7fffffffde18) at libc-start.c:287
> > > 
> > > #6  0x0000000000402fdd in ?? ()
> > 
> > It looks rather like a simply typo.  Does the patch below work?
> > 
> > 
> > Takashi
> > 
> > diff --git a/src/pcm/pcm_ladspa.c b/src/pcm/pcm_ladspa.c
> > index 631ee0f35b73..6552b4379ec0 100644
> > --- a/src/pcm/pcm_ladspa.c
> > +++ b/src/pcm/pcm_ladspa.c
> > @@ -749,7 +749,7 @@ static int snd_pcm_ladspa_allocate_memory(snd_pcm_t
> > *pcm, snd_pcm_ladspa_t *lads return -ENOMEM;
> >                          }
> >  			for (idx = 0; idx < instance->input.channels.size; idx++) {
> > -			        chn = instance->output.channels.array[idx];
> > +			        chn = instance->input.channels.array[idx];
> >  			        if (pchannels[chn] == NULL && chn < ichannels) {
> >  			                instance->input.data[idx] = NULL;
> >  			                continue;
> 
> It seems to work

Sorry, I forgot to apply this to upstream.  Now pushed out.


thanks,

Takashi