Re: [PATCH v3 02/10] ALSA: hda: ext: Clean up streams if their initialization fails

[Cezary Rojewski <cezary.rojewski@intel.com>]

On 9/1/2026 1:42 PM, Takashi Iwai wrote:
> On Mon, 31 Aug 2026 18:42:25 +0200,
> Cezary Rojewski wrote:


>> @@ -94,6 +94,7 @@ int snd_hdac_ext_stream_init_all(struct hdac_bus *bus, int start_idx,
>>  	int (*setup_op)(struct hdac_stream *, bool);
>>  	int stream_tag = 0;
>>  	int i, tag, idx = start_idx;
>> +	struct hdac_stream *s, *_s;
>>  
>>  	if (pci->device == PCI_DEVICE_ID_INTEL_HDA_APL)
>>  		setup_op = snd_hdac_apl_host_stream_setup;
>> @@ -103,7 +104,7 @@ int snd_hdac_ext_stream_init_all(struct hdac_bus *bus, int start_idx,
>>  	for (i = 0; i < num_stream; i++) {
>>  		struct hdac_ext_stream *hext_stream = kzalloc_obj(*hext_stream);
>>  		if (!hext_stream)
>> -			return -ENOMEM;
>> +			goto err_nomem;
>>  		tag = ++stream_tag;
>>  		snd_hdac_ext_stream_init(bus, hext_stream, idx, dir, tag);
>>  		idx++;
>> @@ -112,6 +113,14 @@ int snd_hdac_ext_stream_init_all(struct hdac_bus *bus, int start_idx,
>>  
>>  	return 0;
>>  
>> +err_nomem:
>> +	list_for_each_entry_safe(s, _s, &bus->stream_list, list) {
>> +		struct hdac_ext_stream *hext_stream = stream_to_hdac_ext_stream(s);
>> +
>> +		list_del(&s->list);
>> +		kfree(hext_stream);
>> +	}
> 
> We may reuse snd_hdac_ext_stream_free_all() but without calling
> snd_hdac_ext_stream_decouple(), too?
Majority of code found within sound/hda/core/ext/ needs some lifting the
goal behind this patchset is not it.  Moving the decouple function away
from snd_hdac_ext_stream_free_all() is a refactor.  I'm against adding
an if-statement in snd_hdac_ext_stream_free_all().  However, I could
replace the error path with snd_hdac_ext_stream_free_all() for now and
live with the superfluous operation till the refactor : )