From: Mark Kettenis Subject: Re: lld/mips64: lazy bindings To: Kirill A. Korinsky Cc: rober@openbsd.org, tech@openbsd.org Date: Tue, 10 Nov 2026 06:28:44 +0100 > Date: Tue, 10 Nov 2026 01:51:54 +0100 > From: Kirill A. Korinsky > > On Tue, 10 Nov 2026 01:45:11 +0100, > Mark Kettenis wrote: > > > > > Date: Mon, 09 Nov 2026 11:42:05 +0100 > > > From: Kirill A. Korinsky > > > > > > robert@, > > > > > > here a diff which implemented a lazy bidning for mips64. > > > > > > I had tried to upstream it and upstream rejected it because, let me quote: > > > > > > > That is a lot of MIPS-specific machinery for a port with little active > > > > maintenance in lld, and it changes the output for every MIPS OS, not just > > > > OpenBSD (hence the updates to existing tests). > > > > > > Instead they suggested this hack: > > > https://gist.github.com/MaskRay/fd946027ac2134ad697c31128c92812c > > > > > > I think that lld fix is cleaner, so, shall we go it in and support as part > > > of our not that small diff? > > > > If upstream rejects this, I don't think we should do this. > > > > Upstream rejects this because they seems not really care about mips64. > > Do we need lazy bindings in mips64? Do we plan to switch mips64 to lld? If upstream doesn't care about mips64, maybe we shouldn't switch. Because they wouldn't care about other fixes.