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: <bed0e6da-55be-ff26-64e7-42e62163dba5@ashwireless.com>
Date:   Fri, 15 Mar 2019 10:07:53 +0000
From:   Michael Haben <mh@...wireless.com>
To:     linux-kernel@...r.kernel.org
Subject: Memory-compaction not working on Atmel SAMA5 ARM platform, kernel
 v4.14 and v4.20

Hello all,


Memory-compaction is not working for me on an Atmel SAMA5D2 SoC with 
128Mbyte of RAM, on kernel versions 4.14.73. 4.14.88 and 4.20.0

Using "cat /proc/buddyinfo" to observe the state of memory 
fragmentation, and "echo 1 > /proc/sys/vm/compact_memory" to (try to) 
trigger compaction.

My kernel's /proc/config.gz includes "CONFIG_COMPACTION=y" and 
"CONFIG_MIGRATION=y"

I have rebuilt v4.20 several++ times, enabling various "CONFIG_xxx" 
items trying to enable tracing, but without success 
(/sys/kernel/debug/tracing is not present), so I haven't been able to 
observe whether the compaction-trigger is having any effect at all.


Some circumstantial evidence: memory-compaction was working on another 
ARM platform (Raspberry Pi) with kernel v4.9 (Raspbian Jessie), but 
stopped working when I dist-upgraded it to v4.14.98 (Raspbian Stretch).  
The "
/proc/sys/vm/compact_memory" device-node is present, but writing to it 
has no effect on the output from "cat /proc/buddyinfo".


So my questions are:

1.  Are "CONFIG_COMPACTION=y" + "CONFIG_MIGRATION=y" the only settings 
necessary for memory-compaction to be present and working in a kernel?  
I have Done My Own Research, but Am Not An Expert on Linux kernel features!

2.  Must I persevere with trying to get tracing working, or is there 
another way I can debug this issue?


Thank you,

Mike Haben

ASH Wireless

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ