Bug 1581224

I am not sure when brp-strip and brp-strip-shared are used, but for find-debuginfo.sh the fix is easy (although untested):

diff --git a/scripts/find-debuginfo.sh b/scripts/find-debuginfo.sh
index a7c2db0..90a4494 100755
--- a/scripts/find-debuginfo.sh
+++ b/scripts/find-debuginfo.sh
@@ -235,6 +235,7 @@ strip_to_debug()
   $strip_g && case "$(file -bi "$2")" in
   application/x-sharedlib*) g=-g ;;
   application/x-executable*) g=-g ;;
+  application/x-pie-executable*) g=-g ;;
   esac
   eu-strip --remove-comment $r $g ${keep_remove_args} -f "$1" "$2" || exit
   chmod 444 "$1" || exit

Note that instead of using find_debuginfo_args="-g" (or _find_debuginfo_opts), which runs [eu-]strip to only remove the .debug sections, but not other non-allocated sections, you could also use find_debuginfo_args="--keep-section .symtab" to explicitly keep the .symtab section (but still strip away everything else).

(In reply to Mark Wielaard from comment #1)
> I am not sure when brp-strip and brp-strip-shared are used, but for
> find-debuginfo.sh the fix is easy (although untested):
> 
> diff --git a/scripts/find-debuginfo.sh b/scripts/find-debuginfo.sh
> index a7c2db0..90a4494 100755
> --- a/scripts/find-debuginfo.sh
> +++ b/scripts/find-debuginfo.sh
> @@ -235,6 +235,7 @@ strip_to_debug()
>    $strip_g && case "$(file -bi "$2")" in
>    application/x-sharedlib*) g=-g ;;
>    application/x-executable*) g=-g ;;
> +  application/x-pie-executable*) g=-g ;;
>    esac
>    eu-strip --remove-comment $r $g ${keep_remove_args} -f "$1" "$2" || exit
>    chmod 444 "$1" || exit

I patched this manually into a buildroot, and can confirm that this change restores the .symtab section for ld.so, so valgrind works again.  Thanks!

(In reply to Florian Weimer from comment #2)
> I patched this manually into a buildroot, and can confirm that this change
> restores the .symtab section for ld.so, so valgrind works again.  Thanks!

Thanks for testing.
I posted the patch upstream:
http://lists.rpm.org/pipermail/rpm-maint/2018-May/007976.html
And added it to fedora rpm and did a build for rawhide (rpm-4.14.1-9.fc29)
http://koji.fedoraproject.org/koji/buildinfo?buildID=1084068

If it works out, it should be trivial to add to f28.

glibc-2.27.9000-19.fc29.x86_64 unbreaks valgrind.  However, that version was built against file-5.33-4.fc29 as well, so it is not clear which change actually fixed matters. 8-)

(In reply to Florian Weimer from comment #4)
> glibc-2.27.9000-19.fc29.x86_64 unbreaks valgrind.  However, that version was
> built against file-5.33-4.fc29 as well, so it is not clear which change
> actually fixed matters. 8-)

Aha, I see file also added a countermeasure in file-5.33-4:

- do not classify shared libraries as pie executables (#1581343)

Good, then either this or that going into f28 would resolve things.

We already have an update for file-5.33-4.fc28 at Bodhi:

https://bodhi.fedoraproject.org/updates/FEDORA-2018-d1425c85a7

Despite its karma is +3, it has not yet reached updates-testing, and I see no button to move it forward.

# rpm -q file
file-5.33-4.fc28.x86_64
# file --mime /sbin/ldconfig 
/sbin/ldconfig: application/x-pie-executable; charset=binary
# file --mime /bin/bash
/bin/bash: application/x-pie-executable; charset=binary

So the file update is definitely NOT sufficient to fix the find-debuginfo.sh issue.  We need the patch from comment 1.

I have submitted a new update to fix the MIME output, too:

https://bodhi.fedoraproject.org/updates/file-5.33-5.fc28

Sorry for the troubles!

rpm-4.14.1-9.fc28 has been submitted as an update to Fedora 28. https://bodhi.fedoraproject.org/updates/FEDORA-2018-1894f7892a

rpm-4.14.1-9.fc28 has been pushed to the Fedora 28 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-1894f7892a

rpm-4.14.1-9.fc28 has been pushed to the Fedora 28 stable repository. If problems still persist, please make note of it in this bug report.