Hide Forgot
Description of problem: dmidecode reports bad Socket Designation where there is more than 9 CPU, for example on an KVM guest with 10 CPUs, here what Socket Designations fields shows: -CPU 1 -CPU 2 ... -CPU 9 -CPU : The bug is on the numbering that display ':' instead of '10'. Version-Release number of selected component (if applicable): dmidecode-2.1-3.el5 How reproducible: dmidecode on a system with more than 9 CPU (KVM guest in my case) Steps to Reproduce: 1. dmidecode|grep "Socket Designation:" 2. 3. Actual results: Socket Designation: CPU 1 Socket Designation: CPU 2 Socket Designation: CPU 3 Socket Designation: CPU 4 Socket Designation: CPU 5 Socket Designation: CPU 6 Socket Designation: CPU 7 Socket Designation: CPU 8 Socket Designation: CPU 9 Socket Designation: CPU : Expected results: Socket Designation: CPU 1 Socket Designation: CPU 2 Socket Designation: CPU 3 Socket Designation: CPU 4 Socket Designation: CPU 5 Socket Designation: CPU 6 Socket Designation: CPU 7 Socket Designation: CPU 8 Socket Designation: CPU 9 Socket Designation: CPU 10 Additional info:
Thanks for the report, Herve! 1. try to reproduce the issue with the latest version of dmidecode*. 2. please attach the binary dump from this machine. --dump-bin option. *dmidecode v2.11-1: http://people.redhat.com/aarapov/pkgs/
Hi Anton, I've installed your dmidecode package (it has the same version number !) and tested it: $ rpm -Uvh /tmp/dmidecode-2.11-1.el5.x86_64.rpm Preparing... ########################################### [100%] 1:dmidecode ########################################### [100%] [root@dona ~] $ dmidecode |grep Socket Socket Designation: CPU 1 Socket Designation: CPU 2 Socket Designation: CPU 3 Socket Designation: CPU 4 Socket Designation: CPU 5 Socket Designation: CPU 6 Socket Designation: CPU 7 Socket Designation: CPU 8 Socket Designation: CPU 9 Socket Designation: CPU : Still the same problem. $ dmidecode --dump-bin dona.dmidecode # dmidecode 2.11 SMBIOS 2.4 present. # Writing 884 bytes to dona.dmidecode. # Writing 31 bytes to dona.dmidecode. Please find the file in attachments.
Created attachment 517420 [details] dmidecode binary dump
In the default display for CPU 9 the displayed handle is 0x0409. For the CPU 10, the handle is 0x040A: Handle 0x0409, DMI type 4, 32 bytes Processor Information Socket Designation: CPU 9 Type: Central Processor Family: Other Manufacturer: QEMU ID: 63 06 00 00 FD FB 8B 07 Version: Not Specified Voltage: Unknown ... Handle 0x040A, DMI type 4, 32 bytes Processor Information Socket Designation: CPU : Type: Central Processor Family: Other Manufacturer: QEMU ID: 63 06 00 00 FD FB 8B 07 Version: Not Specified Voltage: Unknown ... It seems that the CPU number can be simply found by "anding 0xFF" the handle. Maybe the dmidecode code get this value and add it to the zero character ascii code to get CPU number. That works for one to ten CPUs but fail beyond. Sorry but, I have not enough time to dive in dmidecode code ;)
Herve, that's the problem corrupted data that the BIOS of your HV exposes to the guests. I've moved the bug to seabios maintainer, perhaps he aware of it already. dmidecode's codebase is stable wrt decode the cpu data, and works just fine at my local machine. I will dig into your binary though... thanks,
Tried with rhel6.3 guest over rhel63 host, this problem does not exist. cmd: # /usr/libexec/qemu-kvm -M rhel6.3.0 -enable-kvm -m 4G -smp 11 -nodefconfig -rtc base=utc,driftfix=slew -no-shutdown -device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 -drive file=rhel63_64_virtio.qcow2,if=none,id=virtio0,format=qcow2,cache=none,aio=native -device virtio-blk-pci,drive=virtio0,id=virtio0,scsi=off,bootindex=1 -netdev tap,id=hostnet0,vhost=on -device virtio-net-pci,netdev=hostnet0,id=net0,mac=20:20:20:11:22:19 -vnc :10 -monitor stdio -boot menu=on In guest: #dmidecode|grep "Socket Designation:" Socket Designation: CPU 1 Socket Designation: CPU 2 Socket Designation: CPU 3 Socket Designation: CPU 4 Socket Designation: CPU 5 Socket Designation: CPU 6 Socket Designation: CPU 7 Socket Designation: CPU 8 Socket Designation: CPU 9 Socket Designation: CPU a Socket Designation: CPU b
Closing this RHEL-6 bug, as RHEL-6 hosts don't have the bug. A bug for RHEL-5 needs to be proposed if a fix is reqeuired on RHEL-5.