[PATCH 3/3] elf/tst-dl_find_object: Disable subtests for non-contiguous maps (bug 28732)

[H.J. Lu]

On Fri, Jan 14, 2022 at 7:56 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> * H. J. Lu:
>
> > On Fri, Jan 14, 2022 at 7:51 AM Florian Weimer <fweimer@redhat.com> wrote:
> >>
> >> * H. J. Lu:
> >>
> >> > On Fri, Jan 14, 2022 at 7:39 AM Florian Weimer <fweimer@redhat.com> wrote:
> >> >>
> >> >> * H. J. Lu:
> >> >>
> >> >> >> We never set l_contiguous for the main executable, so it doesn't matter
> >> >> >> what the link editor does.  And none of the glibc fixes went in so far.
> >> >> >>
> >> >> >> Thanks,
> >> >> >> Florian
> >> >> >>
> >> >> >
> >> >> > Should l_contiguous be set on the main executable? If not, why?
> >> >>
> >> >> I think it should be set even if the kernel loads the main executable.
> >> >> See patch 2:
> >> >>
> >> >>   [PATCH 2/3] elf: Set l_contiguous to 1 for the main map in more cases
> >> >>   <https://sourceware.org/pipermail/libc-alpha/2022-January/134894.html>
> >> >
> >> > If we have the second patch, do we still need the 3rd one?
> >>
> >> I think gaps are expected on some architectures for main executable
> >> (ia64?).
> >>
> >
> > Please add a header to indicate that the gap is expected only on some
> > architectures.
>
> Sorry, do you mean there should be a .h file to indicate whether the
> architecture expects a gap?
>
> How do I tell whether the gap is expected or the result of binutils
> bugs?
>

You can add a linker test to check for the linker bug.

-- 
H.J.