Abramo Bagnara wrote: >Tim Goetze wrote: >> >> Jaroslav Kysela wrote: >> >> >Hello all, >> > >> > I need your advice here. The current snd_pcm_mmap_commit() >> >function returns zero or an error. I've added a new function >> >snd_pcm_mmap_commit_partial() to detect partial transfers as well: >> > >> >int snd_pcm_mmap_commit(snd_pcm_t *pcm, >> > snd_pcm_uframes_t offset, >> > snd_pcm_uframes_t frames); >> >int snd_pcm_mmap_commit_partial(snd_pcm_t *pcm, >> > snd_pcm_uframes_t offset, >> > snd_pcm_uframes_t frames, >> > snd_pcm_uframes_t *commited); >> > >> >Abramo doesn't like new function so much, so I need your vote here, if: >> > >> >1) change snd_pcm_mmap_commit prototype to >> > >> >snd_pcm_sframes_t snd_pcm_mmap_commit(snd_pcm_t *pcm, >> > snd_pcm_uframes_t offset, >> > snd_pcm_uframes_t frames); >> > >> >2) leave the current API as is (two separate commit functions) >> > >> >3) remove snd_pcm_mmap_commit() and rename >> > snd_pcm_mmap_commit_partial() to snd_pcm_mmap_commit() although >> > it makes necessary to add an extra parameter to all existing >> > applications (NULL can be taken as "I don't care") >> > >> > >> >I see the benefit of the extra parameter, that we can return the error >> >code and transfer count in one pass. Otherwise the error code can be lost >> >or returned later using another function like snd_pcm_avail_update() etc.. >> >> as a programmer i'd say, yes i'd like to be informed about partial >> transfers. >> >> i'd make snd_pcm_mmap_commit return -errno in case of error and number of >> frames commited otherwise. >> that's approximately what read/write(2) do, and that's what i'd prefer. > >Jaroslav point is that in this way you can't know *both* error and >frames count. > >You lose the error info (or delay its return).
if the error case takes longer to complete, this is fine by me -- we'll have to stop the device in this case anyway. i think the API should be optimized for normal operations (and i haven't ever had an error return by mmap_commit afaicr). tim _______________________________________________ Alsa-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/alsa-devel
