hi On 6/7/18 4:03 PM, Darren Tucker wrote: > On 8 June 2018 at 07:09, PGNet Dev <pgnet.dev@xxxxxxxxx> wrote: >> Verifying a report I just got pinged about, building vanilla openssh 7.7p1 on linux configures ok, but fails build around 'retpoline' > [...] >> Should the retpoline flag be getting added? If so, what's needed to make LD happy with it? > > configure checks to see if the linker accepts those flags, so my guess > is that something is added later in configure that causes it to fail. > I suggest looking at config.log to see what the actaul compiler > messages are from these parts: > > OSSH_CHECK_CFLAG_COMPILE([-mretpoline]) # clang > OSSH_CHECK_LDFLAG_LINK([-Wl,-z,retpolineplt]) > > and the actual compiler output show up in config.log. > > I wrote and tested this patch against the development versions of gcc > and clang which was all that was available at the time, it's possible > the released versions don't behave quite the same. here's the complete config.log: https://paste.fedoraproject.org/paste/WtZP~Oe1~es2ddeqCBtCqQ the 'fun' appears to start at line# 652 gcc-8: error: unrecognized command line option '-mretpoline'; did you mean '-Wtrampolines'? nooooo, no trampolines (is that a thing?) ;-) at 1st quick look-thru, I'm just seeing that the flag failed, but not _why_. yet. looking closer now. also, just fyi, during my testing I did try clang6 et al as well. failed as well, with different/additional errors. didn't spend any further time on it yet. _______________________________________________ openssh-unix-dev mailing list openssh-unix-dev@xxxxxxxxxxx https://lists.mindrot.org/mailman/listinfo/openssh-unix-dev
