Will Deacon <will@xxxxxxxxxx> writes: > Hi Michael, > > On Fri, Aug 16, 2019 at 02:52:40PM +1000, Michael Ellerman wrote: >> Will Deacon <will@xxxxxxxxxx> writes: >> > Although Alpha, Itanic and PowerPC all override NM, only PowerPC does it >> > conditionally so I agree with you that passing '--synthetic' unconditionally >> > would resolve the problem and is certainly my preferred approach if mpe is >> > ok with it. >> >> I'd rather we keep passing --synthetic, otherwise there's the potential >> that symbols go missing that were previously visible. > > Yup -- that was my suggestion above. > >> I think we can keep the new_nm check, but drop the dependency on >> CONFIG_PPC64, and that will fix it. Worst case is we start passing >> --synthetic on ppc32, but that's probably not a problem. >> >> This seems to fix it for me, and 32-bit builds fine. > > Brill, thanks for confirming! > >> Do you want me to send a proper patch for this, or do you want to squash >> it into the original series? > > I'd prefer not to rebase the arm64 queue, so if you send this as a proper > patch, please, then I can queue it on top before reverting the hack we > currently have. Cool, just sent a patch. cheers
