On Fri, Jan 08, 2010 at 11:36:59AM -0600, Miguel Aguilar wrote: > Mark Brown wrote: > > On Thu, Jan 07, 2010 at 04:17:21PM -0600, miguel.aguilar@xxxxxxxxxxxx wrote: > >> + if (device == 0) > >> + davinci_cfg_reg(DM365_EVT2_ASP_TX); > >> + else > >> + davinci_cfg_reg(DM365_EVT2_VC_TX); > > I'd be a bit more comfortable with this if it were using something more > > symbolic like a #define or enum rather than checking a bare number to > > work out which device it's talking to. > The idea of these function is check at runtime if the user space application is > requesting the AIC3x or the voice codec, then it will set the proper source for > the dma channels, since the ASP and the Voice Codec share the same dma channels, > so that's why use a #define doesn't make sense. I see what your code is doing but at the minute it's making this decision based on the device number that's being passed in by comparing it as a pure number. This seems fragile - something symbolic that joined things up a bit more wouldn't raise eyebrows in the same way. > Can you check the part of this patch related to registering both codecs AIC3x > and the voice codec? Like I say that all looks fine to me but I can't really check if the DaVinci code is idiomatic. _______________________________________________ Alsa-devel mailing list Alsa-devel@xxxxxxxxxxxxxxxx http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
