On 10/30/18 7:29 AM, Stefan Ring wrote: > On Sun, Oct 28, 2018 at 6:13 PM Eric Sandeen <sandeen@xxxxxxxxxxx> wrote: >> Check patch 2 as well, the "break;" is also causing corruption I think, >> and it should probably just be a no-op ";" > > Sorry about that. I think it's more than that. The single_fsb function > takes care to only write something out when there is something to > write and to only recalculate the crc when there has been a change. > I'd like to carry over this spirit, but I have the impression that in > the end the two functions would become so similar that they will just > merge into one. I think the "recalculate crc" flagging is a little inconsistent, tbh - or maybe I should say it consistently recalculates it more often than it needs to? It probably makes sense to follow the existing examples of when it's set, and if there's any cleanup/optimization to do, that can be done later. -Eric
