1.1.9-1 Release Notes
06/24/2014 13:36
Release 1.1.9-1 includes the following changes:
Overview
This release includes a kernel panic bug fix that caused issues for some customers and additional fixes for CloakFS and iothrottled.
CloakFS
- Fix for removing cloakFS rules. CloakFS defined rules were not being undefined when changes were made to the configuration file to remove them.
- Performance enhancement for the cloak_file /etc/passwd feature.
- The cloak_file feature does not currently support jailshell environments, we are adding this support in the next release.
Kernel
- Critical kernel panic fix. The problem is caused by a race condition, tsk->thread.xstate can be NULL during a elf_core_dump(). This patch returns -ENODEV in xfpregs_get() when this condition in met. The result is a core dump with an inaccurate state for the floating point registers. The fix will prevent the kernel from crashing.
- Fixed buffer overflow bug. At startup, iothrottled was not checking the maximum amount of mount points per device and giving a buffer overflow / SIGSEGV error.
MySQL
No changes
CPUD
No changes
cPanel Configs
No changes
blstat
No changes
Upgrading BetterLinux:
Run this:
yum --disableexcludes=all upgrade betterlinux\*
Then reboot your system.
Changelog for BetterLinux Release 1.1.9-1
CPUD and CLOAKFS:
commit 14ba123affe28ddd2afc3eeec62d89c98140eba3
Author: David
Date: Wed May 14 15:29:53 2014 -0600
Rework cloak_file caching.
IOTHROTTLED:
commit cedad5ee03370e073213ede9287b5cb047e5c0bc
Author: Andrea Righi <andrea@betterlinux.com>
Date: Wed May 28 19:17:47 2014 +0200
fix a buffer overflow bug at initialization
Correctly check if the maximum amount of mount points per device is
exceeded in init_disk_list() to avoid a buffer overflow / SIGSEGV bug
when iothrottled is started.
Signed-off-by: Andrea Righi
CPANEL CONFIGS:
No changes
KERNEL:
commit 95b2ab0a46e1200a8e2ce32118217390882fe615
Author: Andrea Righi
Date: Wed May 14 13:58:11 2014 +0200
centos6-2.6.32-431.17.1.el6: i387: fix NULL pointer dereference
A customer reported the following kernel panic bug:
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [] memcpy+0xb/0x120
PGD 6c667e067 PUD 177215067 PMD 0
Oops: 0000 [#1] SMP
last sysfs file: /sys/devices/system/cpu/online
CPU 0
Modules linked in: tcp_diag inet_diag nfs_throttle(P) nfs lockd fscache auth_rpcgss nfs_acl sunrpc blockio_throttle(P) sch_htb cls_uid(P) connlimit(P) fuse cloakfs(P) group_lib(P) pid_group klibgrp(P) hlist kinterval bl_license xt_limit xt_DSCP xt_TCPMSS iptable_mangle ipt_REJECT ipt_LOG ipv6 nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack iptable_filter ip_tables ext3 jbd microcode iTCO_wdt iTCO_vendor_support power_meter dcdbas sb_edac edac_core lpc_ich mfd_core shpchp sg tg3 ptp pps_core ext4 jbd2 mbcache sr_mod cdrom sd_mod crc_t10dif ahci wmi megaraid_sas dm_mirror dm_region_hash dm_log dm_mod [last unloaded: scsi_wait_scan]
Pid: 25297, comm: php Tainted: P --------------- 2.6.32-431.11.2.el6.bl1.1.7_1.x86_64 #1 Dell Inc. PowerEdge R720/0C4Y3R
RIP: 0010:[] [] memcpy+0xb/0x120
RSP: 0018:ffff88074edcfa80 EFLAGS: 00010246
RAX: ffff88010941c200 RBX: ffff88005e220240 RCX: 0000000000000040
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff88010941c200
RBP: ffff88074edcfac8 R08: ffff88010941c200 R09: 0000000000000000
R10: 0000000000000200 R11: 0000000000000000 R12: ffff88010941c200
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000200
FS: 0000000000000000(0000) GS:ffff880028200000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 00000009ecf21000 CR4: 00000000000407f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process php (pid: 25297, threadinfo ffff88074edce000, task ffff88005e220240)
Stack:
ffffffff810176b6 ffffffff81019d75 0000000000000000 ffffffff81019ddc
ffffffff81c17078 0000000000000001 ffff88017c2b8800 0000000000000001
ffffffff81601920 ffff88074edcfc08 ffffffff811ec30c 0000000000000001
Call Trace:
[] ? xfpregs_get+0x66/0xc0
[] ? getreg+0x105/0x130
[] ? genregs_get+0x3c/0xa0
[] elf_core_dump+0x51c/0xfa0
[] ? __wake_up+0x53/0x70
[] ? call_usermodehelper_exec+0xab/0x120
[] do_coredump+0x963/0x1000
[] ? fixup_exception+0x1d/0x80
[] ? __sigqueue_free+0x3d/0x50
[] get_signal_to_deliver+0x1ed/0x460
[] ? wake_up_state+0x10/0x20
[] do_signal+0x75/0x800
[] ? force_sig_info+0xa2/0x110
[] do_notify_resume+0x90/0xc0
[] retint_signal+0x48/0x8c
Code: 49 89 70 50 19 c0 49 89 70 58 41 c6 40 4c 04 83 e0 fc 83 c0 08 41 88 40 4d c9 c3 90 90 90 90 90 48 89 f8 89 d1 c1 e9 03 83 e2 07 48 a5 89 d1 f3 a4 c3 20 48 83 ea 20 4c 8b 06 4c 8b 4e 08 4c
RIP [] memcpy+0xb/0x120
RSP
CR2: 0000000000000000
This seems to be exactly this problem:
https://access.redhat.com/site/solutions/640843
It is not a BetterLinux related bug, but it is good to have this fixed.
The problem is that, due to a race condition, tsk->thread.xstate can be
NULL during a elf_core_dump().
This patch returns -ENODEV in xfpregs_get() if this particular buggy
condition in met. The result is that we may have a core dump with an
inaccurate state of the floating point registers, but it saves us from
crashing the kernel.
Signed-off-by: Andrea Righi
