linux-kernel - [BUG] mem leak (not: info leak) in perf (kernel 4.0.0-rc2)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <CAP145pij-nZcuKytP-HUkJx7RtZO3_99vS0Coi24nG8BQgAkNw@mail.gmail.com>
Date:	Thu, 12 Mar 2015 18:14:59 +0100
From:	Robert Święcki <robert@...ecki.net>
To:	linux-kernel@...r.kernel.org
Subject: [BUG] mem leak (not: info leak) in perf (kernel 4.0.0-rc2)

Hi,

I suppose I observe a substantial mem leak in the kernel while using
perf_event_open()

Reproducible under:
Kernel: $ uname -a
Linux test1 4.0.0-0.rc2.git0.1.fc22.x86_64 #1 SMP Tue Mar 3 21:24:17
UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Not reproducible under:
Linux test2 3.16.0-30-generic #40~14.04.1-Ubuntu SMP Thu Jan 15
17:43:14 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Repro - with: https://code.google.com/p/honggfuzz/source/checkout

Note: It might be related to the CPU, as Intel vs AMD code differs in
a few places here. I tested it under Intel only though.

$ ~/src/honggfuzz/honggfuzz -Di -f /bin/ls -s -- /bin/ls
...

$ while [ 1 ]; do cat /proc/meminfo | grep -E "Stack|Slab"; sleep 1; done
Slab:             437584 kB
KernelStack:     1037472 kB
Slab:             439160 kB
KernelStack:     1046528 kB
Slab:             443036 kB
KernelStack:     1055488 kB
Slab:             446360 kB
KernelStack:     1065888 kB
Slab:             448540 kB
KernelStack:     1073008 kB
Slab:             452428 kB
KernelStack:     1082912 kB
Slab:             455548 kB
KernelStack:     1094640 kB
Slab:             458620 kB
KernelStack:     1102448 kB
Slab:             462488 kB
KernelStack:     1114752 kB
Slab:             466632 kB
KernelStack:     1123936 kB
Slab:             468024 kB
KernelStack:     1129040 kB
Slab:             470780 kB
KernelStack:     1135712 kB
Slab:             473148 kB
KernelStack:     1145856 kB
Slab:             477684 kB
KernelStack:     1156528 kB
Slab:             480804 kB
KernelStack:     1165904 kB
Slab:             482524 kB
KernelStack:     1174032 kB
Slab:             485496 kB
KernelStack:     1183120 kB

# slabtop
  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
 90528  90528 100%    2.47K   7544 12    241408K task_struct

[a few sec sleep]

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
 97356  97356 100%    2.47K   8113 12    259616K task_struct

[a few sec sleep]

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
103668 103668 100%    2.47K   8639 12    276448K task_struct

It seems to be not released after all involved processes have terminated.

Will try to debug it more, but others might see some obvious problem here.

-- 
Robert Święcki
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/