Bug 2007412

Summary:	glibc: ldd segfaults when inspecting vdso/vdso64.so
Product:	Red Hat Enterprise Linux 8	Reporter:	Milos Malik <mmalik>
Component:	glibc	Assignee:	glibc team <glibc-bugzilla>
Status:	CLOSED WONTFIX	QA Contact:	qe-baseos-tools-bugs
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	8.5	CC:	aoliva, arjun.is, ashankar, codonell, dj, extras-qa, fweimer, law, mcermak, mfabian, mnewsome, pfrankli, rth, sipoyare
Target Milestone:	rc	Keywords:	Bugfix, Triaged
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	2002756	Environment:
Last Closed:	2021-10-01 15:06:56 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	2002756, 2007417
Bug Blocks:

Description Milos Malik 2021-09-23 19:18:14 UTC

+++ This bug was initially created as a clone of Bug #2002756 +++

Description of problem:

Version-Release number of selected component (if applicable):
glibc-common-2.28-164.el8.x86_64
glibc-langpack-en-2.28-164.el8.x86_64
glibc-2.28-164.el8.x86_64
glibc-headers-2.28-164.el8.x86_64
glibc-devel-2.28-164.el8.x86_64

How reproducible:
 * always

Steps to Reproduce:
# dmesg -c >& /dev/null
# find /usr -name vdso64.so
/usr/lib/modules/4.18.0-343.el8.x86_64/vdso/vdso64.so
# ldd `find /usr -name vdso64.so`
ldd: exited with unknown exit code (139)
# dmesg
[  190.771399] ld-linux-x86-64[5275]: segfault at 7fab7ddd9408 ip 00007fab7dbb5180 sp 00007ffd1cf21880 error 7 in ld-2.28.so[7fab7dbae000+2c000]
[  190.771408] Code: de ea 01 00 31 c0 e8 1f 96 00 00 e9 34 f4 ff ff 4c 8d 0d 9b e8 01 00 e9 c3 fb ff ff 4d 85 c0 74 71 48 8b 43 60 48 85 c0 74 04 <4c> 01 40 08 48 8b 43 58 48 85 c0 74 04 4c 01 40 08 48 8b 43 68 48
# 

Actual results:
 * segfault

Expected results:
 * no segfault

Comment 1 Milos Malik 2021-09-23 19:20:52 UTC

# coredumpctl info -1
           PID: 5275 (ld-linux-x86-64)
           UID: 0 (root)
           GID: 0 (root)
        Signal: 11 (SEGV)
     Timestamp: Thu 2021-09-23 15:14:19 EDT (5min ago)
  Command Line: /lib64/ld-linux-x86-64.so.2 --verify /usr/lib/modules/4.18.0-343.el8.x86_64/vdso/vdso64.so
    Executable: /usr/lib64/ld-2.28.so
 Control Group: /user.slice/user-0.slice/session-3.scope
          Unit: session-3.scope
         Slice: user-0.slice
       Session: 3
     Owner UID: 0 (root)
       Boot ID: b3cf94a4ba904efdbdf7f81400639733
    Machine ID: 4d0e3a63c207450cb7488bd9b38fda5b
      Hostname: rhel8-machine
       Storage: none
       Message: Process 5275 (ld-linux-x86-64) of user 0 dumped core.
#

Comment 2 Florian Weimer 2021-09-23 19:23:50 UTC

Note that the upstream patch is broken: https://sourceware.org/pipermail/libc-alpha/2021-September/131287.html

I had to back it out in rawhide.

Comment 3 Carlos O'Donell 2021-10-01 13:40:35 UTC

Milos,

I'm evaluting this for inclusion in 8.6, but I wanted to know a little more about your requirements. Does this also impact testing in RHEL8?

Comment 4 Milos Malik 2021-10-01 14:58:28 UTC

From my SELinux QE point of view, it does not impact our testing at all. I found it by mistake :-)

Comment 5 Carlos O'Donell 2021-10-01 15:06:56 UTC

(In reply to Milos Malik from comment #4)
> From my SELinux QE point of view, it does not impact our testing at all. I
> found it by mistake :-)

In that case we're going to look at fixing this only for RHEL9, but it will be CLOSED/WONTFIX for RHEL8. Thanks for the report, it is helpful, and sparked a bunch of conversation around object formats :-)