You are not logged in.
This is the information about my Netbook:
Lenovo Ideapad S10
Memory = 995,8 MiB
Processor = Intel® Atom™ CPU N270 @ 1.60GHz × 2
Graphics = Intel® 945GME x86/MMX/SSE2
System base = Arch Linux 32-bit
This is my journalctl:
[root@arch julio]# journalctl
-- Logs begin at jue 2013-09-26 14:43:49 COT, end at mié 2013-10-02 15:36:17 COT. --
sep 26 14:43:49 arch systemd-journal[103]: Runtime journal is using 184.0K (max 49.7M, leaving 74.6M of free 497.7M, current l
sep 26 14:43:49 arch systemd-journal[103]: Runtime journal is using 188.0K (max 49.7M, leaving 74.6M of free 497.7M, current l
sep 26 14:43:49 arch kernel: Initializing cgroup subsys cpuset
sep 26 14:43:49 arch kernel: Initializing cgroup subsys cpu
sep 26 14:43:49 arch kernel: Initializing cgroup subsys cpuacct
sep 26 14:43:49 arch kernel: Linux version 3.11.1-2-ARCH (tobias@testing-i686) (gcc version 4.8.1 20130725 (prerelease) (GCC)
sep 26 14:43:49 arch kernel: Atom PSE erratum detected, BIOS microcode update recommended
sep 26 14:43:49 arch kernel: e820: BIOS-provided physical RAM map:
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x0000000000000000-0x000000000009f7ff] usable
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x000000000009f800-0x000000000009ffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000000d2000-0x00000000000d3fff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000000dc000-0x00000000000fffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x0000000000100000-0x000000003f6cffff] usable
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x000000003f6d0000-0x000000003f6e1fff] ACPI data
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x000000003f6e2000-0x000000003f6e2fff] ACPI NVS
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x000000003f6e3000-0x000000003fffffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000e0000000-0x00000000efffffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000fec00000-0x00000000fec0ffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000fed00000-0x00000000fed003ff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000fed14000-0x00000000fed19fff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000fed1c000-0x00000000fed8ffff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000fee00000-0x00000000fee00fff] reserved
sep 26 14:43:49 arch kernel: BIOS-e820: [mem 0x00000000ff000000-0x00000000ffffffff] reserved
sep 26 14:43:49 arch kernel: Notice: NX (Execute Disable) protection cannot be enabled: non-PAE kernel!
sep 26 14:43:49 arch kernel: SMBIOS 2.5 present.
sep 26 14:43:49 arch kernel: DMI: LENOVO 433336U /Kuril , BIOS 14C
sep 26 14:43:49 arch kernel: e820: update [mem 0x00000000-0x00000fff] usable ==> reserved
sep 26 14:43:49 arch kernel: e820: remove [mem 0x000a0000-0x000fffff] usable
sep 26 14:43:49 arch kernel: e820: last_pfn = 0x3f6d0 max_arch_pfn = 0x100000
sep 26 14:43:49 arch kernel: MTRR default type: uncachable
lines 1-31
I did what you recommended and this product is the report
https://wiki.archlinux.org/index.php/Microcode
Any idea?
Offline
Please use code tags: https://bbs.archlinux.org/help.php#bbcode
I don't understand your post; if you have that module loaded and the intel-ucode package installed, that's all you need. Look further down your dmesg output to see if it is applied. Example on my Atom 300:
% dmesg
...
[ 0.000000] Atom PSE erratum detected, BIOS microcode update recommended
...
[ 2.581238] microcode: CPU0 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.583464] microcode: CPU0 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.586684] microcode: CPU0 updated to revision 0x219, date = 2009-04-10
[ 2.588567] microcode: CPU1 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.588775] microcode: CPU1 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.590016] microcode: CPU1 updated to revision 0x219, date = 2009-04-10
[ 2.594813] microcode: CPU2 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.594988] microcode: CPU2 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.597279] microcode: CPU2 updated to revision 0x219, date = 2009-04-10
[ 2.600178] microcode: CPU3 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.600381] microcode: CPU3 sig=0x106c2, pf=0x8, revision=0x20d
[ 2.603348] microcode: CPU3 updated to revision 0x219, date = 2009-04-10
[ 2.605672] microcode: Microcode Update Driver: v2.00 <tigran@aivazian.fsnet.co.uk>, Peter OrubaOffline
That meaning in your demesg:
% dmesg
...
[ 0.000000] Atom PSE erratum detected, BIOS microcode update recommended
...
Is neccesary an update of microcode?
Offline
I believe so, but since that occurs before the microcode module is loaded, I believe that you can ignore it. In other words, the microcode is updated AFTER that message is displayed.
Offline
What part of "BIOS microcode update recommended" you don't understand?
This particular hardware bug affects the kernel before microcode updater has any chance to run and installing intel-ucode won't fix it. It can only be fixed by updating BIOS (provided that the update actually contains microcode with fixes for this bug, what isn't guaranteed) or booting some bootloader or temporary kernel with microcode updater and then switching to the main kernel.
This bug forces Linux to disable support for "huge pages". They are supposed to slightly improve performance in some applications, but Linux didn't use them by default until one or two years ago and stuff worked so it's not a big deal if you won't do anything about this problem.
Last edited by mich41 (2013-10-03 16:01:36)
Offline
Post the output of this command:
dmesg | grep microcodeThat will give you an informative message about whether or not your microcode has been updated. The microcode updates are in linux-firmware which you probably have installed and if they are able they will be applied automatically.
Offline
Post the output of this command:
dmesg | grep microcodeThat will give you an informative message about whether or not your microcode has been updated. The microcode updates are in linux-firmware which you probably have installed and if they are able they will be applied automatically.
That is my understanding as well.
I believe so, but since that occurs before the microcode module is loaded, I believe that you can ignore it. In other words, the microcode is updated AFTER that message is displayed.
Offline
A quick google search resulted in this:
The error message "Atom PSE erratum detected" is not only cosmetic in nature. As soon as you see this, the kernel has disabled some functionality in order to keep your system stable and you can't fix it with a microcode update in userspace. (source)
You'll have to update the bios or use the early microcode update as described here: https://www.kernel.org/doc/Documentatio … rocode.txt
Edit: You can download the microcode data here: https://downloadcenter.intel.com/Search … crocode%22
Last edited by progandy (2013-10-03 17:07:05)
| alias CUTF='LANG=en_XX.UTF-8@POSIX ' | alias ENGLISH='LANG=C.UTF-8 ' |
Offline
You'll have to update the bios or use the early microcode update as described here: https://www.kernel.org/doc/Documentatio … rocode.txt
Interesting... perhaps this is handled some how by mkinitcpio? I didn't see anything obvious in the hooks section.
Edit: You can download the microcode data here: https://downloadcenter.intel.com/Search … crocode%22
This gives the same file that [extra]/intel-ucode uses, but neither the source nor the package provides "GenuineIntel.bin" mentioned in the document.
EDIT: If the info contained in the blog post you referenced is indeed accurate, I fail to see what value the intel-ucore packages provides at all. In other words, it loads the updates after the kernel has already disabled the features the updates are fixing. Seems as though either:
1) The package provides superfluous files.
2) It needs to generate a modified initrc image as described in the early-microcode.txt file you referenced.
3) We are wrong and the package does indeed provide a solution.
EDIT2: http://lists.debian.org/debian-user/201 … 00395.html
Last edited by graysky (2013-10-03 18:27:03)
Offline
This gives the same file that [extra]/intel-ucode uses, but neither the source nor the package provides "GenuineIntel.bin" mentioned in the document.
The archive provides a microcode.dat. The document shows a copy from "microocode.bin" to .../GenuineIntel.bin, so I guess that this is acceptable and the kernel can load intel microcode files as long as they are stored with that particular "GenuineIntel.bin" as name.
intel-ucode uses this file as the source, but it splits it in ucode files and the early kernel image needs the microcode in one single file.
3) We are wrong and the package does indeed provide a solution.
The package provides a solution for most updates since they don't result in changes applied in the kernel. In the case of the Atom PSE erratum PSE will be disabled on boot, though. I'm not totally sure it won't be reenabled after the microcode update, but this is probably not done.
http://git.kernel.org/cgit/linux/kernel … 3cfbaef730
http://git.kernel.org/cgit/linux/kernel … 24fec1fa3a
Last edited by progandy (2013-10-03 18:48:49)
| alias CUTF='LANG=en_XX.UTF-8@POSIX ' | alias ENGLISH='LANG=C.UTF-8 ' |
Offline