2008-07-16 14:23:34 |
Angel |
bug |
|
|
added bug |
2008-07-22 14:32:22 |
Angel |
description |
Hi, we use to boot our machines the diskless method through pxe + initramfs + squashfs +unionfs, and it worked great with kernel 2.6.20-16, but we need to migrate to kernel 2.6.24 and something strange is happening.
We use motherboard intel dg35ec which has a "Ethernet controller: Intel Corporation 82566DC Gigabit Network Connection" (e1000 module).
When it boot for first time,
can't open /tmp/net-eth0.conf
kernel panic
That mean that the kernel couldn't bring up the eth0 interface, but a few lines before that appear:
ipconfig:segfault at 400bf4 rip 400bf4 rsp "ex_number" error 14
So, I managed to put "strace" into the initramfs and add to the pxelinux.conf "break=mount" in order to get a shell in the initram and debug the problem.
the module e1000 is load successfully and without problem
$>modprobe e1000
Intel(R) PRO/1000 Network Driver - version 7.3.20-k2-NAPI
Copyright (c) 1999-2007 Intel Corporation.
ACPI: PCI Interrupt 0000:00:19.0[A] -> GSI 20 (level, low) -> IRQ 20
e1000: 0000:00:19.0: e1000_probe: (PCI Express:2.5Gb/s:Width x1) MAC
e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
but ipconfig still fail.
$>ipconfig
ipconfig:segfault at 400bf4 rip 400bf4 rsp numero_en_hex_largo error 14
using strace:
$>strace -Ff -tt ipconfig
execve ("/bin/ipconfig","ipconfig","/*21 vars*/")=0
--- sigsegv (seg fault) @ 0 (0) ---
+++ killed by sigsegv +++
the packages are from official hardy :
linux-ubuntu-modules-2.6.24-18-generic (2.6.24-18.26)
linux-image-2.6.24-18-generic (2.6.24-18.32)
linux-restricted-modules2.6.24-18-generic (2.6.24.13-18.41)
Thanks |
Hi, we use to boot our machines the diskless method through pxe + initramfs + squashfs +unionfs, and it worked great with kernel 2.6.20-16, but we need to migrate to kernel 2.6.24 and something strange is happening.
We use motherboard intel dg35ec which has a "Ethernet controller: Intel Corporation 82566DC Gigabit Network Connection" (e1000 module).
When it boot for first time,
can't open /tmp/net-eth0.conf
kernel panic
That mean that the kernel couldn't bring up the eth0 interface, but a few lines before that appear:
ipconfig:segfault at 400bf4 rip 400bf4 rsp "ex_number" error 14
So, I managed to put "strace" into the initramfs and add to the pxelinux.conf "break=mount" in order to get a shell in the initram and debug the problem.
the module e1000 is load successfully and without problem
$>modprobe e1000
Intel(R) PRO/1000 Network Driver - version 7.3.20-k2-NAPI
Copyright (c) 1999-2007 Intel Corporation.
ACPI: PCI Interrupt 0000:00:19.0[A] -> GSI 20 (level, low) -> IRQ 20
e1000: 0000:00:19.0: e1000_probe: (PCI Express:2.5Gb/s:Width x1) MAC
e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
but ipconfig still fail.
$>ipconfig
ipconfig:segfault at 400bf4 rip 400bf4 rsp numero_en_hex_largo error 14
using strace:
$>strace -Ff -tt ipconfig
execve ("/bin/ipconfig","ipconfig","/*21 vars*/")=0
--- sigsegv (seg fault) @ 0 (0) ---
+++ killed by sigsegv +++
the packages are from official hardy :
linux-ubuntu-modules-2.6.24-18-generic (2.6.24-18.26)
linux-image-2.6.24-18-generic (2.6.24-18.32)
linux-restricted-modules2.6.24-18-generic (2.6.24.13-18.41)
I noticed that the e1000 driver is old, so I get the new one (8.0.3.1) from the intel home page, and the forked e1000e (0.4.1.7) which now include the PCI-expres's cards (my case)
e1000-8.0.3.1 http://downloadmirror.intel.com/9180/eng/e1000-8.0.3.1.tar.gz
e1000e-0.4.1.7 http://downloadmirror.intel.com/15817/eng/e1000e-0.4.1.7.tar.gz
But still SEG FAULT in ipconfig.
Thanks |
|
2008-07-29 15:18:20 |
Angel |
description |
Hi, we use to boot our machines the diskless method through pxe + initramfs + squashfs +unionfs, and it worked great with kernel 2.6.20-16, but we need to migrate to kernel 2.6.24 and something strange is happening.
We use motherboard intel dg35ec which has a "Ethernet controller: Intel Corporation 82566DC Gigabit Network Connection" (e1000 module).
When it boot for first time,
can't open /tmp/net-eth0.conf
kernel panic
That mean that the kernel couldn't bring up the eth0 interface, but a few lines before that appear:
ipconfig:segfault at 400bf4 rip 400bf4 rsp "ex_number" error 14
So, I managed to put "strace" into the initramfs and add to the pxelinux.conf "break=mount" in order to get a shell in the initram and debug the problem.
the module e1000 is load successfully and without problem
$>modprobe e1000
Intel(R) PRO/1000 Network Driver - version 7.3.20-k2-NAPI
Copyright (c) 1999-2007 Intel Corporation.
ACPI: PCI Interrupt 0000:00:19.0[A] -> GSI 20 (level, low) -> IRQ 20
e1000: 0000:00:19.0: e1000_probe: (PCI Express:2.5Gb/s:Width x1) MAC
e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
but ipconfig still fail.
$>ipconfig
ipconfig:segfault at 400bf4 rip 400bf4 rsp numero_en_hex_largo error 14
using strace:
$>strace -Ff -tt ipconfig
execve ("/bin/ipconfig","ipconfig","/*21 vars*/")=0
--- sigsegv (seg fault) @ 0 (0) ---
+++ killed by sigsegv +++
the packages are from official hardy :
linux-ubuntu-modules-2.6.24-18-generic (2.6.24-18.26)
linux-image-2.6.24-18-generic (2.6.24-18.32)
linux-restricted-modules2.6.24-18-generic (2.6.24.13-18.41)
I noticed that the e1000 driver is old, so I get the new one (8.0.3.1) from the intel home page, and the forked e1000e (0.4.1.7) which now include the PCI-expres's cards (my case)
e1000-8.0.3.1 http://downloadmirror.intel.com/9180/eng/e1000-8.0.3.1.tar.gz
e1000e-0.4.1.7 http://downloadmirror.intel.com/15817/eng/e1000e-0.4.1.7.tar.gz
But still SEG FAULT in ipconfig.
Thanks |
In order to fix this issue, upgrade klibc-utils and libklibc to at least version 1.5.7-4ubuntu3
Here was the original repot with the problem,
#####
Hi, we use to boot our machines the diskless method through pxe + initramfs + squashfs +unionfs, and it worked great with kernel 2.6.20-16, but we need to migrate to kernel 2.6.24 and something strange is happening.
We use motherboard intel dg35ec which has a "Ethernet controller: Intel Corporation 82566DC Gigabit Network Connection" (e1000 module).
When it boot for first time,
can't open /tmp/net-eth0.conf
kernel panic
That mean that the kernel couldn't bring up the eth0 interface, but a few lines before that appear:
ipconfig:segfault at 400bf4 rip 400bf4 rsp "ex_number" error 14
So, I managed to put "strace" into the initramfs and add to the pxelinux.conf "break=mount" in order to get a shell in the initram and debug the problem.
the module e1000 is load successfully and without problem
$>modprobe e1000
Intel(R) PRO/1000 Network Driver - version 7.3.20-k2-NAPI
Copyright (c) 1999-2007 Intel Corporation.
ACPI: PCI Interrupt 0000:00:19.0[A] -> GSI 20 (level, low) -> IRQ 20
e1000: 0000:00:19.0: e1000_probe: (PCI Express:2.5Gb/s:Width x1) MAC
e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
but ipconfig still fail.
$>ipconfig
ipconfig:segfault at 400bf4 rip 400bf4 rsp numero_en_hex_largo error 14
using strace:
$>strace -Ff -tt ipconfig
execve ("/bin/ipconfig","ipconfig","/*21 vars*/")=0
--- sigsegv (seg fault) @ 0 (0) ---
+++ killed by sigsegv +++
the packages are from official hardy :
linux-ubuntu-modules-2.6.24-18-generic (2.6.24-18.26)
linux-image-2.6.24-18-generic (2.6.24-18.32)
linux-restricted-modules2.6.24-18-generic (2.6.24.13-18.41)
I noticed that the e1000 driver is old, so I get the new one (8.0.3.1) from the intel home page, and the forked e1000e (0.4.1.7) which now include the PCI-expres's cards (my case)
e1000-8.0.3.1 http://downloadmirror.intel.com/9180/eng/e1000-8.0.3.1.tar.gz
e1000e-0.4.1.7 http://downloadmirror.intel.com/15817/eng/e1000e-0.4.1.7.tar.gz
But still SEG FAULT in ipconfig.
Thanks
#####
thanks |
|
2008-07-29 15:18:20 |
Angel |
title |
ipconfig segmentation fault in kernel 2.6.24-18 (module e1000) |
ipconfig segmentation fault in kernel 2.6.24-18 with klibc-utils 1.4.30-3ubuntu2 |
|
2008-07-29 15:19:48 |
Angel |
linux: status |
New |
Fix Released |
|