| Message ID | 9c9a53a4905984a570ba5672cbab84f2027dedc1.1624560484.git.christophe.jaillet@wanadoo.fr (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 0cbbeaf370221fc469c95945dd3c1198865c5fe4 |
| Headers | show |
| Series | ALSA: firewire-lib: Fix 'amdtp_domain_start()' when no AMDTP_OUT_STREAM stream is found | expand |
Hi, On Thu, Jun 24, 2021 at 08:49:36PM +0200, Christophe JAILLET wrote: > The intent here is to return an error code if we don't find what we are > looking for in the 'list_for_each_entry()' loop. > > 's' is not NULL if the list is empty or if we scan the complete list. > Introduce a new 'found' variable to handle such cases. > > Fixes: 60dd49298ec5 ("ALSA: firewire-lib: handle several AMDTP streams in callback handler of IRQ target") > Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> > --- > We could test with" if (list_entry_is_head(s, &d->streams, list))" > instead, but I find it much less readable. > --- > sound/firewire/amdtp-stream.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/sound/firewire/amdtp-stream.c b/sound/firewire/amdtp-stream.c > index aad9778d1c4d..9be2260e4ca2 100644 > --- a/sound/firewire/amdtp-stream.c > +++ b/sound/firewire/amdtp-stream.c > @@ -1943,6 +1943,7 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, > unsigned int events_per_period = d->events_per_period; > unsigned int queue_size; > struct amdtp_stream *s; > + bool found = false; > int err; > > if (replay_seq) { > @@ -1955,10 +1956,12 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, > > // Select an IT context as IRQ target. > list_for_each_entry(s, &d->streams, list) { > - if (s->direction == AMDTP_OUT_STREAM) > + if (s->direction == AMDTP_OUT_STREAM) { > + found = true; > break; > + } > } > - if (!s) > + if (!found) > return -ENXIO; > d->irq_target = s; > > -- > 2.30.2 Indeed. Nice to catch it. The bug affects Linux kernel v5.5 or later. Acked-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> In kernel v5.10 or later, we can use 'list_entry_is_head()' macro added by a commit e130816164e2 ("include/linux/list.h: add a macro to test if entry is pointing to the head"). However the development of ALSA firewire stack is out-of-tree repository for my convenience to backport it to kernel v4.17 or later (I need testers...), so I prefer your change. Thanks Takashi Sakamoto
On Thu, 24 Jun 2021 20:49:36 +0200, Christophe JAILLET wrote: > > The intent here is to return an error code if we don't find what we are > looking for in the 'list_for_each_entry()' loop. > > 's' is not NULL if the list is empty or if we scan the complete list. > Introduce a new 'found' variable to handle such cases. > > Fixes: 60dd49298ec5 ("ALSA: firewire-lib: handle several AMDTP streams in callback handler of IRQ target") > Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Applied now. Thanks. Takashi > --- > We could test with" if (list_entry_is_head(s, &d->streams, list))" > instead, but I find it much less readable. > --- > sound/firewire/amdtp-stream.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/sound/firewire/amdtp-stream.c b/sound/firewire/amdtp-stream.c > index aad9778d1c4d..9be2260e4ca2 100644 > --- a/sound/firewire/amdtp-stream.c > +++ b/sound/firewire/amdtp-stream.c > @@ -1943,6 +1943,7 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, > unsigned int events_per_period = d->events_per_period; > unsigned int queue_size; > struct amdtp_stream *s; > + bool found = false; > int err; > > if (replay_seq) { > @@ -1955,10 +1956,12 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, > > // Select an IT context as IRQ target. > list_for_each_entry(s, &d->streams, list) { > - if (s->direction == AMDTP_OUT_STREAM) > + if (s->direction == AMDTP_OUT_STREAM) { > + found = true; > break; > + } > } > - if (!s) > + if (!found) > return -ENXIO; > d->irq_target = s; > > -- > 2.30.2 >
On Fri, Jun 25, 2021 at 06:05:50AM +0900, Takashi Sakamoto wrote: > > In kernel v5.10 or later, we can use 'list_entry_is_head()' macro added > by a commit e130816164e2 ("include/linux/list.h: add a macro to test if > entry is pointing to the head"). Someone needs to backport that all the way back... regards, dan carpenter
diff --git a/sound/firewire/amdtp-stream.c b/sound/firewire/amdtp-stream.c index aad9778d1c4d..9be2260e4ca2 100644 --- a/sound/firewire/amdtp-stream.c +++ b/sound/firewire/amdtp-stream.c @@ -1943,6 +1943,7 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, unsigned int events_per_period = d->events_per_period; unsigned int queue_size; struct amdtp_stream *s; + bool found = false; int err; if (replay_seq) { @@ -1955,10 +1956,12 @@ int amdtp_domain_start(struct amdtp_domain *d, unsigned int tx_init_skip_cycles, // Select an IT context as IRQ target. list_for_each_entry(s, &d->streams, list) { - if (s->direction == AMDTP_OUT_STREAM) + if (s->direction == AMDTP_OUT_STREAM) { + found = true; break; + } } - if (!s) + if (!found) return -ENXIO; d->irq_target = s;
The intent here is to return an error code if we don't find what we are looking for in the 'list_for_each_entry()' loop. 's' is not NULL if the list is empty or if we scan the complete list. Introduce a new 'found' variable to handle such cases. Fixes: 60dd49298ec5 ("ALSA: firewire-lib: handle several AMDTP streams in callback handler of IRQ target") Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> --- We could test with" if (list_entry_is_head(s, &d->streams, list))" instead, but I find it much less readable. --- sound/firewire/amdtp-stream.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)