> >> /* DM_CM_INFO2 and DM_CM_INFO2_MASK */ > >> #define INFO2_DTRANERR1 BIT(17) > >> #define INFO2_DTRANERR0 BIT(16) > >> +#define INFO2_RESERVED_BITS GENMASK_ULL(32, 0) > > > > Same as above. Maybe we even need one define only? > > No, I think 2 separate regs deserve 2 separate masks. Whatever. I am not into bike-shedding. It works, too. > >> + /* > >> + * We don't use the DMA interrupts, but they might have been enabled > >> + * by a bootloader, so mask them to avoid an interrupt storm... > >> + */ > > > > Two spaces after ',' looks odd to me. Also, no need for "..." > > Both are my preferences. Yes, I probably should have worked in the > typesetting industry... :-) But where is the continuation of "..."? > > I'd even think with a name like CLEAR_MASK, the comment could even go. > > Disagree here. The register #defines don't provide enough info on what's > going on... Masking out all interrupts at the beginning is pretty standard behaviour in my book. Well, at least, please shorten it to sth like "Disable interrupts, we don't use them".
Attachment:
signature.asc
Description: PGP signature
