On 12/07/2012 12:01 PM, Christophe Gisquet wrote: > 2012/12/7 Justin Ruggles <[email protected]>: >>> +cglobal sbr_qmf_deint_neg, 2,3,4,v,src,vrev >> >> needs INIT_XMM sse > > Is one needed for each cglobal? Because there's already one prior to > that function, and no other one.
Not technically, but it makes maintaining the code easier since changing a function won't also require modifying the one below it. Plus it makes the code easier to read by not having to scan up the file to find the CPU flags for a particular function. >> I don't quite get the use of m2 here. It looks unnecessary. > > Indeed, but this patch is kind of "sent without proper checking", > because I couldn't find a sample for it, even in the fate suite. Sorry > about the ping, I could have just dropped it. > >>> + jl .loop >> >> Where is the return? > > And indeed it is missing, because this code is never run. > > Unless someone can point a sample, it is probably not worth working on > that patch. Ah, ok. At least we'll have it in the ML archive in case we need it. Thanks, Justin _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
