Re: [REGRESSION] ALSA: firewire-lib: heavy digital distortion with Fireface 800

[Takashi Sakamoto <o-takashi@xxxxxxxxxxxxx>]

On Thu, Jul 25, 2024 at 08:08:14AM -0600, Gustavo A. R. Silva wrote:
> Hi!
> 
> On 25/07/24 07:07, Takashi Iwai wrote:
> > On Thu, 25 Jul 2024 00:24:29 +0200,
> > edmund.raile wrote:
> > > 
> > > Bisection revealed that the bitcrushing distortion with RME FireFace 800
> > > was caused by 1d717123bb1a7555
> > > ("ALSA: firewire-lib: Avoid -Wflex-array-member-not-at-end warning").
> > > 
> > > Reverting this commit yields restoration of clear audio output.
> > > I will send in a patch reverting this commit for now, soonTM.
> > > 
> > > #regzbot introduced: 1d717123bb1a7555
> > 
> > While it's OK to have a quick revert, it'd be worth to investigate
> > further what broke there; the change is rather trivial, so it might be
> > something in the macro expansion or a use of flex array stuff.
> > 
> 
> I wonder is there is any log that I can take a look at. That'd be really
> helpful.

The original designated initializer fills all of fields with 0.

The new designated initializer assigns CIP_HEADER_QUADLETS (=2) to
struct fw_iso_packet.header_length. It is wrong value in the case of
CIP_NO_HEADER. Additionally it is wrong value in another case since the
value of the field should be byte unit.

I'll post a patch soon.


Regards

Takashi Sakamoto