[Bug libc/28624] New: openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[Bug libc/28624] New: openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[arekm at maven dot pl]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

            Bug ID: 28624
           Summary: openjdk 8/9 assume uni processor and gets stuck due to
                    lack of cpu counting /proc fallback with glibc 2.34
           Product: glibc
           Version: 2.34
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: libc
          Assignee: unassigned at sourceware dot org
          Reporter: arekm at maven dot pl
                CC: drepper.fsp at gmail dot com
  Target Milestone: ---

commit f13fb81ad3159543741e9132685335002a6d5df2
Author: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Date:   Thu Mar 25 14:04:37 2021 -0300

    linux: Remove /proc/cpuinfo fallback on alpha and sparc

introduced in glibc 2.34 also removed generic /proc/cpuinfo fallback
(sysdeps/unix/sysv/linux/getsysstats.c change).

That causes breakage with openjdk 8 and 9 (and variants like icedtea) when no
sysfs is mounted (two my cases where this got visible was chroot()ed system and
linux vserver guest). openjdk assumes then that system is uniprocessor,
disables MP things in virtual machine and even simple "java --help" gets stuck
on futex operations.

openjdk sources:
hotspot/src/os/linux/vm/os_linux.cpp for number of cpus checking
hotspot/src/share/vm/runtime/globals.hpp for AssumeMP

Newer openjdk like 11 defaults to multi processor, so the bug doesn't happen
there.

Easy reproducer by Jan Palus:

systemd-run --wait -t -p InaccessiblePaths=/sys
/usr/lib64/jvm/openjdk8/bin/java -version

(needs 1-15 runs)


Bisecting confirms:

f13fb81ad3159543741e9132685335002a6d5df2 is the first bad commit
commit f13fb81ad3159543741e9132685335002a6d5df2
Author: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Date:   Thu Mar 25 14:04:37 2021 -0300

    linux: Remove /proc/cpuinfo fallback on alpha and sparc

    There is no much gain in fallback to cpuinfo if sysfs is no present,
    usually on restricted environment neither will be present.  It also
    simplifies the code and make all architecture use the sched_getaffinity
    as the sysfs fallback.

    Checked on sparc64-linux-gnu.

 sysdeps/unix/sysv/linux/alpha/getsysstats.c | 38 -----------------------------
 sysdeps/unix/sysv/linux/getsysstats.c       | 22 +----------------
 sysdeps/unix/sysv/linux/sparc/getsysstats.c | 38 -----------------------------
 3 files changed, 1 insertion(+), 97 deletions(-)
 delete mode 100644 sysdeps/unix/sysv/linux/alpha/getsysstats.c
 delete mode 100644 sysdeps/unix/sysv/linux/sparc/getsysstats.c

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[sam at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

Sam James <sam at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dilfridge at gentoo dot org,
                   |                            |sam at gentoo dot org

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[dilfridge at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

Andreas K. Huettel <dilfridge at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://bugs.gentoo.org/sho
                   |                            |w_bug.cgi?id=807832

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |adhemerval.zanella at linaro dot o
                   |                            |rg
   Last reconfirmed|                            |2021-11-24
           Assignee|unassigned at sourceware dot org   |adhemerval.zanella at linaro dot o
                   |                            |rg
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |ASSIGNED

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |2.35
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #1 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
Fixed on 2.35.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

--- Comment #2 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
I backported to 2.34 as well.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/28624] openjdk 8/9 assume uni processor and gets stuck due to lack of cpu counting /proc fallback with glibc 2.34

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28624

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fweimer at redhat dot com

--- Comment #3 from Florian Weimer <fweimer at redhat dot com> ---
Sorry, I didn't see this bug in time.

The bug here is that the /sys code returns a uniprocessor if /sys is not
available. Then then /proc fallback is no longer needed.

A mounted /sys is more or less required these days.

-- 
You are receiving this mail because:
You are on the CC list for the bug.