linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Robert Stupp <snazy@snazy.de>
To: linux-kernel@vger.kernel.org
Subject: mlockall(MCL_CURRENT) blocking infinitely
Date: Thu, 24 Oct 2019 09:36:37 +0200	[thread overview]
Message-ID: <4576b336-66e6-e2bb-cd6a-51300ed74ab8@snazy.de> (raw)

Hi guys,

I've got an issue with `mlockall(MCL_CURRENT)` after upgrading Ubuntu 
19.04 to 19.10 - i.e. kernel version change from 5.0.x to 5.3.x.

The following simple program hangs forever with one CPU running at 100% 
(kernel):

#include <stdio.h>
#include <sys/mman.h>
int main(char** argv) {
   printf("Before mlockall(MCL_CURRENT)\n");
   // works in 5.0
   // hangs forever w/ 5.1 and newer
   mlockall(MCL_CURRENT);
   printf("After mlockall(MCL_CURRENT)\n");
}

All kernel versions since 5.1 (tried 5.1.0, 5.1.21, 5.2.21, 5.3.0-19, 
5.3.7, 5.4-rc4) show the same symptom (hanging in mlockall(MCL_CURRENT) 
with 100% kernel-CPU). 5.0 kernel versions (5.0.21) are fine.

First, I thought, that it's something generic, so I tried the above 
program in a fresh install of Ubuntu eoan (5.3.x) in a VM in virtualbox, 
but it works fine there. So I suspect, that it has to do with something 
that's specific to my machine.

My first suspicion was that some library "hijacks" mlockall(), but 
calling the test program with `LD_DEBUG=all` shows that glibc gets 
called directly:
      12248:    symbol=mlockall;  lookup in file=./test [0]
      12248:    symbol=mlockall;  lookup in 
file=/lib/x86_64-linux-gnu/libc.so.6 [0]
      12248:    binding file ./test [0] to 
/lib/x86_64-linux-gnu/libc.so.6 [0]: normal symbol `mlockall' [GLIBC_2.2.5]
An `strace` doesn't show anything meaningful (beside that mlockall's 
been called but never returns). dmesg and syslog don't show anything 
obvious (to me) as well.

Some information about the machine:
- Intel(R) Core(TM) i7-6900K, Intel X99 chipset
- NVMe 1.1b
- 64GB RAM (4x 16GB)

I've also reverted all changes for sysctl and ld.conf and checked for 
other suspicious software without any luck.

I also tried a bunch of variations of the above program, but only 
`mlockall(MCL_CURRENT)` or `mlockall(MCL_FUTURE | MCL_CURRENT)` hang.

A `git diff v5.0..v5.1 mm/` doesn't show anything obvious (to me).

It seems, there's no debug/trace information that would help to find out 
what exactly it's doing.

I'm kinda lost at the moment.


PS: Variations of the above test program:

#include <stdio.h>
#include <sys/mman.h>
char foo[65536];
int main(char** argv) {
   printf("Before mlock()\n");
   int e = mlock(foo, 8192); // works in 5.0, 5.1, 5.2, 5.3, 5.4
   printf("After mlock()=%d\n", e);
}


#include <stdio.h>
#include <stdlib.h>
#include <sys/mman.h>
int main(char** argv) {
   printf("Before mlockall(MCL_FUTURE)\n");
   int e = mlockall(MCL_FUTURE); // works in 5.0, 5.1, 5.2, 5.3, 5.4
   printf("After mlockall(MCL_FUTURE) = %d\n", e);
   void* mem = malloc(1024 * 1024 * 1024);
   printf("After malloc()\n");
   mem = malloc(1024 * 1024 * 1024);
   printf("After malloc()\n");
   mem = malloc(1024 * 1024 * 1024);
   printf("After malloc()\n");
   // works in 5.0, 5.1, 5.2, 5.3, 5.4
}


#include <stdio.h>
#include <sys/mman.h>
int main(char** argv) {
   printf("Before munlockall()\n");
   int e = munlockall(); // works in 5.0, 5.1, 5.2, 5.3, 5.4
   printf("After munlockall() = %d\n", e);
}


#include <stdio.h>
#include <sys/mman.h>
int main(char** argv) {
   printf("Before mlockall(MCL_CURRENT|MCL_FUTURE)\n");
   // works in 5.0
   // hangs forever w/ 5.1 and newer
   int e = mlockall(MCL_CURRENT|MCL_FUTURE);
   printf("After mlockall(MCL_CURRENT|MCL_FUTURE) = %d\n", e);
}

PPS: Kernel version images installed from 
https://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D

-- 
Robert Stupp
@snazy


             reply	other threads:[~2019-10-24  7:59 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-24  7:36 Robert Stupp [this message]
2019-10-24 23:34 ` mlockall(MCL_CURRENT) blocking infinitely Randy Dunlap
2019-10-25  9:21   ` Michal Hocko
2019-10-25 11:02     ` Robert Stupp
2019-10-25 11:46       ` Michal Hocko
2019-10-25 11:50         ` Michal Hocko
2019-10-25 11:59           ` Robert Stupp
2019-10-25 13:19             ` Robert Stupp
2019-10-25 11:55         ` Robert Stupp
2019-10-25 12:05           ` Michal Hocko
2019-10-25 12:11             ` Michal Hocko
2019-10-25 13:10               ` Robert Stupp
2019-10-25 13:27                 ` Michal Hocko
2019-10-25 13:45                   ` Robert Stupp
2019-10-25 13:57                     ` Michal Hocko
2019-10-25 14:00                       ` Michal Hocko
2019-10-25 15:58                         ` Robert Stupp
2019-11-05 13:23                           ` Robert Stupp
2019-11-05 15:28                             ` Vlastimil Babka
2019-11-05 18:22                               ` Johannes Weiner
2019-11-05 20:05                                 ` Robert Stupp
2019-11-06 10:25                                   ` Robert Stupp
2019-11-06 11:26                                     ` Robert Stupp
2019-11-06 12:04                                       ` Jan Kara
2019-11-06 12:24                                       ` Robert Stupp
2019-11-06 12:03                                 ` Jan Kara
2019-11-06 13:45                                   ` Robert Stupp
2019-11-06 14:35                                     ` Jan Kara
2019-11-06 15:32                                       ` Robert Stupp
2019-11-06 14:38                                     ` Jan Kara
2019-11-06 14:56                                     ` Josef Bacik
2019-11-06 15:05                                       ` Jan Kara
2019-11-06 15:14                                         ` Josef Bacik
2019-11-06 15:25                                           ` Jan Kara
2019-11-06 16:39                                           ` Robert Stupp
2019-11-06 17:03                                             ` Robert Stupp
2019-11-06 17:25                                               ` Jan Kara
2019-11-07  8:08                                                 ` Robert Stupp
2019-11-20 12:42                                                   ` Robert Stupp
2019-10-25 13:55                   ` Robert Stupp

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4576b336-66e6-e2bb-cd6a-51300ed74ab8@snazy.de \
    --to=snazy@snazy.de \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).