On Tue, Mar 26, 2019 at 1:07 PM Arnaldo Carvalho de Melo <arnaldo.melo@xxxxxxxxx> wrote: > > Em Tue, Mar 26, 2019 at 12:53:56PM -0700, Andrii Nakryiko escreveu: > > On Tue, Mar 26, 2019 at 12:18 PM Arnaldo Carvalho de Melo > > > So we have to show the sum bitfields, not only when there are bit holes, > > > that way we would see a: > > > > > > /* size: 1248, cachelines: 20, members: 85 */ > > > /* sum members: 1235, holes: 4, sum holes: 8 */ > > > /* sum bitfield members: 40 bits (5 bytes) */ > > > /* last cacheline: 32 bytes */ > > > > > > With the patch below in place on top of yours, please review, > > > > I like this idea! I think we should generalize it to all possible > > combinations of bit/byte holes. E.g., if there is bit hole, but no > > byte hole, we probably still want to emit total size of non-bitfield > > members. I went ahead and produced v2 of this patch, which > > incorporates idea from your patch below. Please check it out. > > Right, that covers all cases, I'll push out what we have so far, seems > much better than what was before this series. > > I'll go combing it all tomorrow, have to switch gears now to some other > stuff. > > Thanks a lot, I think we're getting close to a release, I'll just do > some tests with the --reorganize code to see how these changes affected > that codebase. Sounds great, thanks! > > - Arnaldo
