Hi
I've installed CentOS 5.2 (i386) in a computer with a DG33BU motherboard ... When I installed it it crashed because of ACPI issues. Searching Google I found that I need to turn ACPI off in order to successfully install CentOS, which I did. After a while I noticed that because of ACPI is off the Linux kernel finds, initializes and uses only 1 CPU core in the Q6600 quad core CPU that is installed in this computer :-( How can I force it to find, initialize and use all the 4 cores in the computer even if ACPI is off?
-- TIA Paolo
On Tuesday 03 February 2009, Paolo Supino wrote:
Hi
I've installed CentOS 5.2 (i386) in a computer with a DG33BU motherboard ... When I installed it it crashed because of ACPI issues. Searching Google I found that I need to turn ACPI off in order to successfully install CentOS, which I did. After a while I noticed that because of ACPI is off the Linux kernel finds, initializes and uses only 1 CPU core in the Q6600 quad core CPU that is installed in this computer :-( How can I force it to find, initialize and use all the 4 cores in the computer even if ACPI is off?
Try to leave ACPI on and boot with "pci=nommconf" instead. This works on my DG33 (mine is a TL not a BU though).
/Peter
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
-- TIA Paolo
On Tue, Feb 3, 2009 at 12:44 PM, Peter Kjellstrom cap@nsc.liu.se wrote:
On Tuesday 03 February 2009, Paolo Supino wrote:
Hi
I've installed CentOS 5.2 (i386) in a computer with a DG33BU
motherboard
... When I installed it it crashed because of ACPI issues. Searching
I found that I need to turn ACPI off in order to successfully install CentOS, which I did. After a while I noticed that because of ACPI is off the Linux kernel finds, initializes and uses only 1 CPU core in the Q6600 quad core CPU
that
is installed in this computer :-( How can I force it to find, initialize and use all the 4 cores in the computer even if ACPI is off?
Try to leave ACPI on and boot with "pci=nommconf" instead. This works on my DG33 (mine is a TL not a BU though).
/Peter
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Paolo Supino wrote:
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
You're probably booting the single CPU kernel.
Try booting the kernel ending with "-ELsmp".
Agile Aspect wrote:
Paolo Supino wrote:
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
You're probably booting the single CPU kernel.
Try booting the kernel ending with "-ELsmp".
All CentOS 5 kernels are SMP kernels, there is no ELsmp kernel anymore. Problem really seems to be missing ACPI functionality which leads to the problem that only one core is found.
Ralph
Hi Ralph
Which brings me back to my original post: how do I enable multicore on a DG33BU motherboard, booting with pci=nommconf didn't change the situation :-(
-- TIA Paolo
On Tue, Feb 3, 2009 at 1:40 PM, Ralph Angenendt <ra+centos@br-online.dera%2Bcentos@br-online.de
wrote:
Agile Aspect wrote:
Paolo Supino wrote:
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
You're probably booting the single CPU kernel.
Try booting the kernel ending with "-ELsmp".
All CentOS 5 kernels are SMP kernels, there is no ELsmp kernel anymore. Problem really seems to be missing ACPI functionality which leads to the problem that only one core is found.
Ralph
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Hi Agile
There's no non SMP kernel in CentOS 5.x (check out: http://wiki.centos.org/FAQ/CentOS5#head-d70935212ce3b7b072b0075c1807a4bd3ea1...).
dmesg boot output (taken from /var/log/dmesg) shows the following lines: Linux version 2.6.18-92.1.22.el5 (mockbuild@builder16.centos.org) (gcc version 4.1.2 20071124 (Red Hat 4.1.2-42)) #1 SMP Tue Dec 16 12:03:43 EST
and ...
found SMP MP-table at 000fe200 Memory for crash kernel (0x0 to 0x0) notwithin permissible range disabling kdump Using x86 segment limits to approximate NX protection On node 0 totalpages: 1048576 DMA zone: 4096 pages, LIFO batch:0 Normal zone: 225280 pages, LIFO batch:31 HighMem zone: 819200 pages, LIFO batch:31 DMI 2.4 present. Using APIC driver default Intel MultiProcessor Specification v1.4 Virtual Wire compatibility mode. OEM ID: Product ID: APIC at: 0xFEE00000 Processor #0 6:15 APIC version 20 Enabling APIC mode: Flat. Using 0 I/O APICs BIOS bug, no explicit IRQ entries, using default mptable. (tell your hw vendor) Processors: 1 Allocating PCI resources starting at d2000000 (gap: d0000000:20000000) Detected 2400.268 MHz processor. Built 1 zonelists. Total pages: 1048576 Kernel command line: ro root=LABEL=/ acpi=off mapped APIC to ffffd000 (fee00000) Enabling fast FPU save and restore... done. Enabling unmasked SIMD FPU exception support... done. Initializing CPU#0 CPU 0 irqstacks, hard=c074c000 soft=c072c000 PID hash table entries: 4096 (order: 12, 16384 bytes) Console: colour VGA+ 80x25 Dentry cache hash table entries: 131072 (order: 7, 524288 bytes) Inode-cache hash table entries: 65536 (order: 6, 262144 bytes) Memory: 3353184k/4194304k available (2097k kernel code, 40176k reserved, 877k data, 228k init, 2477052k highmem) Checking if this processor honours the WP bit even in supervisor mode... Ok. Calibrating delay using timer specific routine.. 4802.07 BogoMIPS (lpj=2401035) Security Framework v1.0.0 initialized SELinux: Initializing. SELinux: Starting in permissive mode selinux_register_security: Registering secondary module capability Capability LSM initialized as secondary Mount-cache hash table entries: 512 CPU: After generic identify, caps: bfebfbff 20000000 00000000 00000000 0000e3bd 00000000 00000001 CPU: After vendor identify, caps: bfebfbff 20000000 00000000 00000000 0000e3bd 00000000 00000001 monitor/mwait feature present. using mwait in idle threads. CPU: L1 I cache: 32K, L1 D cache: 32K CPU: L2 cache: 4096K CPU: Physical Processor ID: 0 CPU: Processor Core ID: 0 CPU: After all inits, caps: bfebf3ff 20000000 00000000 00000940 0000e3bd 00000000 00000001 Intel machine check architecture supported. Intel machine check reporting enabled on CPU#0. Checking 'hlt' instruction... OK. SMP alternatives: switching to UP code Freeing SMP alternatives: 14k freed CPU0: Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz stepping 0b Total of 1 processors activated (4802.07 BogoMIPS). Brought up 1 CPUs . . . . .
Unfortunately I still have only 1 core used ...
-- TIA Paolo
On Tue, Feb 3, 2009 at 1:12 PM, Agile Aspect agile.aspect@gmail.com wrote:
Paolo Supino wrote:
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
You're probably booting the single CPU kernel.
Try booting the kernel ending with "-ELsmp".
-- Article. VI. Clause 3 of the constitution of the United States states:
"The Senators and Representatives before mentioned, and the Members of the several State Legislatures, and all executive and judicial Officers, both of the United States and of the several States, shall be bound by Oath or Affirmation, to support this Constitution; but no religious Test shall ever be required as a Qualification to any Office or public Trust under the United States."
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Hi Ralph
You're right and I'm partially wrong: I changed /boot/grub/menu.lst from acpi=off to pci=nommconf and rebooted the system expecting it to come up with the new command line parameters. As you've shown me in the paste of the dmesg output I posted this isn't the case. Which makes me wonder, and ask, where does it take the boot parameters from if not from /boot/grub/menu.lst?
-- TIA Paolo
On Tue, Feb 3, 2009 at 1:46 PM, Ralph Angenendt <ra+centos@br-online.dera%2Bcentos@br-online.de
wrote:
Paolo Supino wrote:
Kernel command line: ro root=LABEL=/ acpi=off
^^^^^^^^Unfortunately I still have only 1 core used ...
As said: Try with "pci=nommconf" only. And please trim your mails >:)
Cheers,
Ralph
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Tuesday 03 February 2009, Paolo Supino wrote:
Hi Ralph
You're right and I'm partially wrong: I changed /boot/grub/menu.lst from acpi=off to pci=nommconf and rebooted the system expecting it to come up with the new command line parameters. As you've shown me in the paste of the dmesg output I posted this isn't the case. Which makes me wonder, and ask, where does it take the boot parameters from if not from /boot/grub/menu.lst?
On CentOS the grub config file is /boot/grub/grub.conf. There are normally two symlinks pointing to this, /etc/grub.conf and /boot/grub/menu.lst.
If your /boot/grub/menu.lst is broken (now a file not a symlink) then this is expected behaviour. Check it out with the file command.
/Peter
Hi Peter
The symlinks aren't broken: the grub.conf file is located in /boot/grub/. /etc/grub/menu.lst points to it and so does /etc/grub.conf ... [root@server grub]# ls -l /etc/grub.conf /boot/grub/menu.lst /boot/grub/grub.conf -rw------- 1 root root 974 Feb 3 13:59 /boot/grub/grub.conf lrwxrwxrwx 1 root root 11 Dec 15 10:04 /boot/grub/menu.lst -> ./grub.conf lrwxrwxrwx 1 root root 22 Dec 15 10:04 /etc/grub.conf -> ../boot/grub/grub.conf
-- TIA Paolo
On Tue, Feb 3, 2009 at 3:07 PM, Peter Kjellstrom cap@nsc.liu.se wrote:
On Tuesday 03 February 2009, Paolo Supino wrote:
Hi Ralph
You're right and I'm partially wrong: I changed /boot/grub/menu.lst from acpi=off to pci=nommconf and rebooted the system expecting it to come up with the new command line parameters. As you've shown me in the paste of the dmesg output I posted this isn't the case. Which makes me wonder, and ask, where does it take the boot parameters from if not from /boot/grub/menu.lst?
On CentOS the grub config file is /boot/grub/grub.conf. There are normally two symlinks pointing to this, /etc/grub.conf and /boot/grub/menu.lst.
If your /boot/grub/menu.lst is broken (now a file not a symlink) then this is expected behaviour. Check it out with the file command.
/Peter
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Paolo Supino wrote:
Hi Peter
The symlinks aren't broken: the grub.conf file is located in /boot/grub/. /etc/grub/menu.lst points to it and so does /etc/grub.conf ... [root@server grub]# ls -l /etc/grub.conf /boot/grub/menu.lst /boot/grub/grub.conf -rw------- 1 root root 974 Feb 3 13:59 /boot/grub/grub.conf lrwxrwxrwx 1 root root 11 Dec 15 10:04 /boot/grub/menu.lst -> ./grub.conf lrwxrwxrwx 1 root root 22 Dec 15 10:04 /etc/grub.conf -> ../boot/grub/grub.conf
Typically there's more than one kernel version installed with corresponding lines in grub.conf. Perhaps you're not booting the kernel entry you think you are? Check your "default=" line.
Hi Here's my /boot/grub/grub.conf content: grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You do not have a /boot partition. This means that # all kernel and initrd paths are relative to /, eg. # root (hd0,0) # kernel /boot/vmlinuz-version ro root=/dev/mapper/sil_aibhcbccdhagp1 # initrd /boot/initrd-version.img #boot=/dev/mapper/sil_aibhcbccdhag default=0 timeout=10 splashimage=(hd0,0)/boot/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-92.1.22.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.1.22.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.1.22.el5.img title CentOS (2.6.18-92.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.el5.img
As you can see there's no mentioning of "acpi=off" anywhere .... and unless I stop it and change the command line parameter from acpi=off to pci=nommconf it still boots with acpi=off ... :-(
-- TIA Paolo
On Tue, Feb 3, 2009 at 3:23 PM, Toby Bluhm tkb@alltechmedusa.com wrote:
Paolo Supino wrote:
Hi Peter
The symlinks aren't broken: the grub.conf file is located in /boot/grub/. /etc/grub/menu.lst points to it and so does /etc/grub.conf
...
[root@server grub]# ls -l /etc/grub.conf /boot/grub/menu.lst /boot/grub/grub.conf -rw------- 1 root root 974 Feb 3 13:59 /boot/grub/grub.conf lrwxrwxrwx 1 root root 11 Dec 15 10:04 /boot/grub/menu.lst ->
./grub.conf
lrwxrwxrwx 1 root root 22 Dec 15 10:04 /etc/grub.conf -> ../boot/grub/grub.conf
Typically there's more than one kernel version installed with corresponding lines in grub.conf. Perhaps you're not booting the kernel entry you think you are? Check your "default=" line.
-- tkb _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Tuesday 03 February 2009, Paolo Supino wrote:
Hi Here's my /boot/grub/grub.conf content: grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You do not have a /boot partition. This means that # all kernel and initrd paths are relative to /, eg. # root (hd0,0) # kernel /boot/vmlinuz-version ro root=/dev/mapper/sil_aibhcbccdhagp1 # initrd /boot/initrd-version.img #boot=/dev/mapper/sil_aibhcbccdhag default=0 timeout=10 splashimage=(hd0,0)/boot/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-92.1.22.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.1.22.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.1.22.el5.img title CentOS (2.6.18-92.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.el5.img
A machine with this /boot/grub/grub.conf should for sure not boot with "acpi=off". The only idea I have left is that your install is messed up with regards to /boot as a partition or as part of /. In any way grub does not read the file above when it boots. So what does it read on your machine... Maybe you could try a:
find / -name grub.conf
/Peter
As you can see there's no mentioning of "acpi=off" anywhere .... and unless I stop it and change the command line parameter from acpi=off to pci=nommconf it still boots with acpi=off ... :-(
On Tue, 2009-02-03 at 15:39 +0200, Paolo Supino wrote:
Hi Here's my /boot/grub/grub.conf content: grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You do not have a /boot partition. This means that # all kernel and initrd paths are relative to /, eg. # root (hd0,0) # kernel /boot/vmlinuz-version ro root=/dev/mapper/sil_aibhcbccdhagp1 # initrd /boot/initrd-version.img #boot=/dev/mapper/sil_aibhcbccdhag default=0 timeout=10 splashimage=(hd0,0)/boot/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.18-92.1.22.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.1.22.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.1.22.el5.img title CentOS (2.6.18-92.el5) root (hd0,0) kernel /boot/vmlinuz-2.6.18-92.el5 ro root=LABEL=/ pci=nommconf initrd /boot/initrd-2.6.18-92.el5.img
As you can see there's no mentioning of "acpi=off" anywhere .... and unless I stop it and change the command line parameter from acpi=off to pci=nommconf it still boots with acpi=off ... :-(
Sounds like your boot/root is not what you think it is. I had an LFS system I built long ago and added CentOS to it. It could be booted from either HD. I forgot to change my root= stuff in fstab when I finally converted fully. Took me awhile to figure out why changes I made had no effect on booting. I changed my root in fstab and all worked well after that.
<snip>
HTH
Hi William
You're right ... I do have 2 HDs and my root is not what I thought it is! Here is what is happening with this system: It has a SIL SATA RAID0/1 card installed with 2x 160GB SATA HDs connected to it. The card is configured to mirror between the HDs (RAID1). But as I saw now Linux actually sees both HDs twice as dmesg shows: SCSI subsystem initialized libata version 3.00 loaded. sata_sil 0000:06:00.0: version 2.3 ACPI: PCI Interrupt 0000:06:00.0[A] -> GSI 21 (level, low) -> IRQ 66 scsi0 : sata_sil scsi1 : sata_sil ata1: SATA max UDMA/100 mmio m512@0xe0004800 tf 0xe0004880 irq 66 ata2: SATA max UDMA/100 mmio m512@0xe0004800 tf 0xe00048c0 irq 66
ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) ata1.00: ATA-7: WDC WD1600AAJS-00PSA0, 05.06H05, max UDMA/133 ata1.00: 312581808 sectors, multi 16: LBA48 NCQ (depth 0/32) ata1.00: configured for UDMA/100
ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310) ata2.00: ATA-7: WDC WD1600AAJS-00PSA0, 05.06H05, max UDMA/133 ata2.00: 312581808 sectors, multi 16: LBA48 NCQ (depth 0/32) ata2.00: configured for UDMA/100 Vendor: ATA Model: WDC WD1600AAJS-0 Rev: 05.0 Type: Direct-Access ANSI SCSI revision: 05 SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) sda: Write Protect is off sda: Mode Sense: 00 3a 00 00 SCSI device sda: drive cache: write back SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) sda: Write Protect is off sda: Mode Sense: 00 3a 00 00 SCSI device sda: drive cache: write back sda: sda1 sda2 sda3 sda4 sd 0:0:0:0: Attached scsi disk sda Vendor: ATA Model: WDC WD1600AAJS-0 Rev: 05.0 Type: Direct-Access ANSI SCSI revision: 05 SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) sdb: Write Protect is off sdb: Mode Sense: 00 3a 00 00 SCSI device sdb: drive cache: write back SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) sdb: Write Protect is off sdb: Mode Sense: 00 3a 00 00 SCSI device sdb: drive cache: write back sdb: sdb1 sdb2 sdb3 sdb4 sd 1:0:0:0: Attached scsi disk sdb
/etc/fstab looks for root on /dev/sdb1, while grub looks (and finds) it's configuration (hd0,0) which corresponds to sda1. The thing that strikes me is that the mirror works for grub (as it only sees 1 HD), but doesn't work for Linux. This defeats the purpose of having a RAID1 to begin with :-( Unlike your scenario this system was installed from scratch (CentOS is the only install) and it was done after the mirror was setup in SIL card bios. BTW: the SIL card boot messages claims that the mirror set is valid and active. I will hace to check into this.
-- TIA Paolo
the sil is fraid or fakeraid..it's actually software raid done by a binary often proprietary driver,..if you want to do raid one buy a real hardware raid card or use the linux mdraid.
On 2/3/2009 6:06 AM, Paolo Supino wrote:
Hi Peter
I tried that and it the result is the same: Linux only finds, initializes and uses only 1 core ... :-( Is there anything else I can try to have Linux initialize all 4 cores?
-- TIA Paolo
On Tue, Feb 3, 2009 at 12:44 PM, Peter Kjellstrom <cap@nsc.liu.se mailto:cap@nsc.liu.se> wrote:
On Tuesday 03 February 2009, Paolo Supino wrote: > Hi > > I've installed CentOS 5.2 (i386) in a computer with a DG33BU motherboard > ... When I installed it it crashed because of ACPI issues. Searching Google > I found that I need to turn ACPI off in order to successfully install > CentOS, which I did. > After a while I noticed that because of ACPI is off the Linux kernel > finds, initializes and uses only 1 CPU core in the Q6600 quad core CPU that > is installed in this computer :-( > How can I force it to find, initialize and use all the 4 cores in the > computer even if ACPI is off? Try to leave ACPI on and boot with "pci=nommconf" instead. This works on my DG33 (mine is a TL not a BU though). /Peter _______________________________________________ CentOS mailing list CentOS@centos.org <mailto:CentOS@centos.org> http://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Did you turn acpi back on AND try that boot command?
Hi William
I didn't trun acpi back on (it's a sure way to freeze the computer). I did however find out that Linux ignores the RAID1 that is set by the SIL SATA card I have installed :-( See my last post (just a few minutes ago) with more details.
-- TIA Paolo
On Tue, 2009-02-03 at 17:17 +0200, Paolo Supino wrote:
Hi William
I didn't trun acpi back on (it's a sure way to freeze the computer). I did however find out that Linux ignores the RAID1 that is set by the SIL SATA card I have installed :-( See my last post (just a few minutes ago) with more details.
Glad you discovered the _symptom_ of the raid problem. However, I'm _totally_ ignorant_ about raid. I suggest you 1) search the archives for stuff about raid and 2) start another thread describing your raid issue. There's a _lot_ of folks on the list who are expert with lots of types of raid.
What I think I recall from the threads I've seen (I don't know if this is generally applicable) is that the HD be partitioned for 2 approximately 100GB partitions to be used as boot and the rest defined for raid. Some suggest that all of the "rest" be an LVM partition for ease of management.
But, as I said, I'm not knowledgeable, so do start another thread about your raid concerns after searching the archives. You should find a lot in there.
<snip>
Paolo Supino wrote:
Hi
I've installed CentOS 5.2 (i386) in a computer with a DG33BU motherboard ... When I installed it it crashed because of ACPI issues. Searching Google I found that I need to turn ACPI off in order to successfully install CentOS, which I did. After a while I noticed that because of ACPI is off the Linux kernel finds, initializes and uses only 1 CPU core in the Q6600 quad core CPU that is installed in this computer :-( How can I force it to find, initialize and use all the 4 cores in the computer even if ACPI is off?
you might check to see if there's a BIOS update on http://support.intel.com and flash that, as ACPI uses tables in the flash bios to describe the hardware. Intel is pretty good about fixing stuff like this. http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&Inst=Yes&Prod...
indeed, the latest BIOS for that board is 0517 (DPP3510J.86A.0517.2009.0107.2203), and there is a fairly extensive list of bugfixes in the release notes http://downloadmirror.intel.com/17260/eng/DP_0517_ReleaseNotes.pdf
do read the readme on the flash processes, there are 5 files, you only need the right one for the method you want to use to flash it. http://downloadmirror.intel.com/17260/eng/BIOS%20Update%20Readme.pdf
-
Agile Aspect -
John R Pierce -
Paolo Supino -
Peter Kjellstrom -
Ralph Angenendt -
Toby Bluhm -
William L. Maltby -
William Warren