Gentoo Archives: gentoo-hardened

From: Brian Kroth <bpkroth@××××.edu>
To: gentoo-hardened@l.g.o
Subject: Re: [gentoo-hardened] kernel upgrade problems: bad page state
Date: Tue, 30 Oct 2007 23:20:50
Message-Id: 4727BBD9.3030509@wisc.edu
In Reply to: Re: [gentoo-hardened] kernel upgrade problems: bad page state by Remy Bosch
1 Remy Bosch wrote:
2 > Brian Kroth wrote:
3 >> I started getting these messages after an upgrade from
4 >> hardened-sources-2.6.20-r10 to 2.6.22.r8. Nothing changed from the
5 >> old config. I just did "make oldconfig" and answered N to everything
6 >> since they weren't relevant. In particular, nothing in PAX or GrSec
7 >> menus were changed, though whenever I get big nasty messages like this
8 >> I suspect one of them. Can anyone give me some direction as to how to
9 >> fix this or what might be causing it? I can post the full configs if
10 >> need be.
11 >>
12 >> Thanks,
13 >> Brian
14 >
15 > I had bad-page errors as well. What I did was wait to see if the adres
16 > stayed the same. In my case it was, and I can only conclude it was bad
17 > ram. And with 2GB for only desktop useage, I didnot get the massage
18 > often. Besides, in my case memtest did not get the faulty page. Even
19 > after 8+ runs. atm I add the badram-patch to my kernel, so i can pass
20 > the faulty page address to the kernel, in order to have it ignored.
21 >
22 > I got this page broken:
23 > page:c1b1fdc0 flags:0x80000000 mapping:00000001 mapcount:0 count:0
24 >
25 >
26 > Goodluck!
27 >
28 > Remy
29 >
30 >
31
32 Where's this badram-patch and who supports it? memtest also didn't
33 report any errors for me. I let it run for 9 passes and decided to try
34 again. This is a server with ECC ram, so I would think it would catch
35 it, but I can't say for certain since I don't know the details well
36 enough (at least at the moment). A quick grep/sed of kern.log shows the
37 output below, so I don't think it's the same page, but with the
38 randomization features I'd have to think about it some more to be sure.
39 It also seems strange to me that only nagios and cactid would crap out
40 on this. If I run the cacti poller just through php I get no errors.
41 The server also hosts ntp, portage over NFS, syslog-ng, mysql, dovecot,
42 sec, postfix, and a bunch of management scripts for other servers. It's
43 a busy thing, so if bad ram were the case, I'd expect to see them in
44 other places. Possibly just bad code on cactid/nagios' part? Should
45 PAX complain then?
46
47 For now, I'm going to roll back to my 2.6.20-r10 config to see if I can
48 reproduce the errors there. I'm also adding a rule to sec to warn me
49 about this and attempt to stop the service the next time it happens.
50 The first time it happened cactid drove the load up so high on the
51 machine that I had no other recourse than to power cycle it via IPMI.
52
53 Also including emerge --info for anyone who's interested. Should I be
54 opening a bug for this?
55
56 # pcregrep " page:" /var/log/kern.log | sed 's/.*\(page:\S\+\).*/\1/' |
57 perl -e 'while (<>) { $data{$_}++ }; for (keys %data) { print $data{$_}
58 . " " . $_ ; };'
59
60 1 page:c14320e0
61 1 page:c12afe80
62 1 page:c12cb860
63 1 page:c159a480
64 1 page:c1392da0
65 1 page:c1203f60
66 1 page:c112b560
67 1 page:c124b620
68 1 page:c154bcc0
69 1 page:c1012fe0
70 1 page:c1416200
71 1 page:c11c53c0
72 1 page:c1358a00
73 1 page:c16baec0
74 1 page:c10a20c0
75 1 page:c1278ac0
76 1 page:c14bf880
77 1 page:c105ef00
78 1 page:c1211b20
79 1 page:c10ab540
80 1 page:c11fd320
81 1 page:c10cb440
82 1 page:c1434c20
83 1 page:c1412fc0
84 1 page:c1060e80
85 1 page:c1278ae0
86 1 page:c1101ae0
87 1 page:c13071e0
88 1 page:c11fd400
89 1 page:c1467da0
90 1 page:c16beb20
91 1 page:c152d5e0
92 1 page:c11bc800
93 1 page:c138c720
94 1 page:c1132400
95 1 page:c17fdd40
96 1 page:c11c54a0
97 1 page:c11d0f20
98 1 page:c105cae0
99 1 page:c16ce580
100 1 page:c170fac0
101
102 emerge --info
103 Portage 2.1.3.16 (hardened/x86/2.6, gcc-3.4.6, glibc-2.6.1-r0,
104 2.6.22-hardened-r8 i686)
105 =================================================================
106 System uname: 2.6.22-hardened-r8 i686 Intel(R) Xeon(TM) CPU 2.80GHz
107 Timestamp of tree: Tue, 30 Oct 2007 06:00:01 +0000
108 distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632)
109 [disabled]
110 ccache version 2.4 [enabled]
111 app-shells/bash: 3.2_p17
112 dev-lang/python: 2.4.4-r6
113 dev-python/pycrypto: 2.0.1-r6
114 dev-util/ccache: 2.4-r7
115 sys-apps/baselayout: 1.12.9-r2
116 sys-apps/sandbox: 1.2.18.1-r2
117 sys-devel/autoconf: 2.13, 2.61-r1
118 sys-devel/automake: 1.7.9-r1, 1.10
119 sys-devel/binutils: 2.18-r1
120 sys-devel/gcc-config: 1.3.16
121 sys-devel/libtool: 1.5.24
122 virtual/os-headers: 2.6.22-r2
123 ACCEPT_KEYWORDS="x86"
124 CBUILD="i686-pc-linux-gnu"
125 CFLAGS="-march=pentium4 -O2 -pipe -fforce-addr"
126 CHOST="i686-pc-linux-gnu"
127 CONFIG_PROTECT="/etc"
128 CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/hotplug /etc/hotplug.d
129 /etc/init.d /etc/php/apache2-php5/ext-active/
130 /etc/php/cgi-php5/ext-active/ /etc/php/cli-php5/ext-active/
131 /etc/revdep-rebuild /etc/terminfo /etc/udev /etc/udev/rules.d"
132 CXXFLAGS="-march=pentium4 -O2 -pipe -fforce-addr"
133 DISTDIR="/mnt/build/distfiles"
134 EMERGE_DEFAULT_OPTS="--nospinner"
135 FEATURES="buildpkg ccache collision-protect metadata-transfer sandbox
136 sfperms strict unmerge-orphans userfetch"
137 GENTOO_MIRRORS="http://gentoo.chem.wisc.edu/gentoo/
138 ftp://gentoo.chem.wisc.edu/gentoo/ http://gentoo.mirrors.tds.net/gentoo
139 ftp://gentoo.mirrors.tds.net/gentoo http://gentoo.osuosl.org/
140 ftp://distro.ibiblio.org/pub/linux/distributions/gentoo/
141 http://distro.ibiblio.org/pub/linux/distributions/gentoo/
142 http://distfiles.gentoo.org"
143 MAKEOPTS="-j5"
144 PKGDIR="/mnt/build/packages"
145 PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times
146 --compress --force --whole-file --delete --delete-after --stats
147 --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages
148 --filter=H_**/files/digest-*"
149 PORTAGE_TMPDIR="/var/tmp"
150 PORTDIR="/mnt/build/portage"
151 PORTDIR_OVERLAY="/mnt/build/portage-local"
152 SYNC="rsync://rsync.gentoo.org/gentoo-portage"
153 USE="acl acpi apache2 bash-completion berkdb bzip2 caps chroot cracklib
154 crypt erandom fam gmp gpm hardened jpeg lm_sensors logrotate maildir mmx
155 ncurses nls nptl pam pcre perl pic png python readline smp snmp sse sse2
156 ssl syslog tcpd threads vhosts x86 xattr xml xpm"
157 ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug
158 file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null
159 plug rate route share shm softvol" ELIBC="glibc" INPUT_DEVICES="mouse
160 keyboard" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk
161 hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU"
162 Unset: CPPFLAGS, CTARGET, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS,
163 PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS

Attachments

File name MIME type
smime.p7s application/x-pkcs7-signature

Replies

Subject Author
Re: [gentoo-hardened] kernel upgrade problems: bad page state pageexec@××××××××.hu
Re: [gentoo-hardened] kernel upgrade problems: bad page state Remy Bosch <remybosch@××××××.nl>