On Tue, Apr 30, 2019 at 04:02:23AM +0100, Al Viro wrote: > On Tue, Apr 30, 2019 at 04:22:38AM +0200, Nicholas Mc Guire wrote: > > On Mon, Apr 29, 2019 at 10:03:36AM -0400, Sven Van Asbroeck wrote: > > > On Mon, Apr 29, 2019 at 2:11 AM Nicholas Mc Guire <hofrat@xxxxxxxxx> wrote: > > > > > > > > V2: As requested by Sven Van Asbroeck <thesven73@xxxxxxxxx> make the > > > > impact of the patch clear in the commit message. > > > > > > Thank you, but did you miss my comment about creating a local variable > > > instead? See: > > > https://lkml.org/lkml/2019/4/28/97 > > > > Did not miss it - I just don't think that makes it any more > > understandable - the __force __be16 makes it clear I believe > > that this is correct, sparse does not like this though - so tell > > sparse. > > ... to STFU, 'cause you know better. The trouble is, how do we > (or yourself a year or two later) know *why* it is correct? > Worse, how do we (or yourself, etc.) know if a change about to be > done to the code won't invalidate the proof of yours? > > > The local variable would need to be explained as it is > > functionally not necessary - therefor I find it more confusing > > that using __force here. > > What's confusing is mixing host- and fixed-endian values in the > same variable at different times. Treat those as unrelated > types that happen to have the same sizeof. > > Quite a few of __force instances in the tree should be taken out > and shot. Don't add to their number. ok - my bad thn - I had assumed that using __force is reasonable if the handling is correct and its a localized conversoin only like var = be16_to_cpu(var) which evaded introducing additinal variables just to have different types but no different function. But the long-term issue of hiding bugs by __force makes sesne to me - will give it another shot at scripting this in coccinelle. thx! hofrat _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
