Hello,
I've notice on both Fedora 24 4.5.5-300 and 4.6.5-300 systems, that Fedora 24 starts 5 out of 8 cpus. Is there a limit being set somewhere or is this a bug?
U-Boot 2016.05 (May 30 2016 - 20:03:50 +0000) for ODROID-XU3
CPU: Exynos5422 @ 800 MHz Model: Odroid XU3 based on EXYNOS5422 Board: Odroid XU3 based on EXYNOS5422 Type: xu4 DRAM: 2 GiB MMC: EXYNOS DWMMC: 0, EXYNOS DWMMC: 1 *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: No ethernet found. Hit any key to stop autoboot: 0 switch to partitions #0, OK mmc1(part 0) is current device Scanning mmc 1:1... switch to partitions #0, OK mmc0 is current device Scanning mmc 0:1... Found /extlinux/extlinux.conf Retrieving file: /extlinux/extlinux.conf 936 bytes read in 18 ms (50.8 KiB/s) Ignoring unknown command: ui Ignoring malformed menu command: autoboot Ignoring malformed menu command: hidden Ignoring unknown command: totaltimeout Ignoring unknown command: default=Fedora Fedora-Mate-armhfp-24-1.2 Boot Options. 1: Fedora (4.6.5-300.fc24.armv7hl) 24 (Twenty Four) 2: Fedora-Mate-armhfp-24-1.2 (4.5.5-300.fc24.armv7hl) Enter choice: 1: Fedora (4.6.5-300.fc24.armv7hl) 24 (Twenty Four) Retrieving file: /initramfs-4.6.5-300.fc24.armv7hl.img 17514609 bytes read in 1641 ms (10.2 MiB/s) Retrieving file: /vmlinuz-4.6.5-300.fc24.armv7hl 6102880 bytes read in 587 ms (9.9 MiB/s) append: rd.driver.pre=dw_mmc-exynos,exynosdrm,ehci-exynos,ohci-exynos ro root=UUID=dfac3a15-7d31-46ed-842f-2f7599871202 LANG=en_US .UTF-8 Retrieving file: /dtb-4.6.5-300.fc24.armv7hl/exynos5422-odroidxu4.dtb 47808 bytes read in 124 ms (376 KiB/s) Kernel image @ 0x42000000 [ 0x000000 - 0x5d1f60 ] ## Flattened Device Tree blob at 43000000 Booting using the fdt blob at 0x43000000 Loading Ramdisk to 4ef4b000, end 4ffff071 ... OK Loading Device Tree to 4ef3c000, end 4ef4aabf ... OK
Starting kernel ...
[ 0.000000] Booting Linux on physical CPU 0x100 [ 0.000000] Linux version 4.6.5-300.fc24.armv7hl (mockbuild@arm04-builder03.arm.fedoraproject.org) (gcc version 6.1.1 20160621 (Red Hat 6.1.1-3) (GCC) ) #1 SMP Thu Jul 28 03:09:19 UTC 2016 [ 0.000000] CPU: ARMv7 Processor [410fc073] revision 3 (ARMv7), cr=10c5387d [ 0.000000] CPU: div instructions available: patching division code [ 0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT aliasing instruction cache [ 0.000000] Machine model: Hardkernel Odroid XU4 [ 0.000000] efi: Getting EFI parameters from FDT: [ 0.000000] efi: UEFI not found. [ 0.000000] cma: Reserved 16 MiB at 0xbd800000 [ 0.000000] Memory policy: Data cache writealloc [ 0.000000] Samsung CPU ID: 0xe5422001 [ 0.000000] Running under secure firmware. [ 0.000000] percpu: Embedded 13 pages/cpu @eed8a000 s24332 r8192 d20724 u53248 [ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 512832 [ 0.000000] Kernel command line: rd.driver.pre=dw_mmc-exynos,exynosdrm,ehci-exynos,ohci-exynos ro root=UUID=dfac3a15-7d31-46ed- 842f-2f7599871202 LANG=en_US.UTF-8 [ 0.000000] PID hash table entries: 4096 (order: 2, 16384 bytes) [ 0.000000] Dentry cache hash table entries: 131072 (order: 7, 524288 bytes) [ 0.000000] Inode-cache hash table entries: 65536 (order: 6, 262144 bytes) [ 0.000000] Memory: 1991052K/2058240K available (7609K kernel code, 1140K rwdata, 3376K rodata, 1108K init, 854K bss, 50804K re served, 16384K cma-reserved, 1271808K highmem) [ 0.000000] Virtual kernel memory layout: [ 0.000000] vector : 0xffff0000 - 0xffff1000 ( 4 kB) [ 0.000000] fixmap : 0xffc00000 - 0xfff00000 (3072 kB) [ 0.000000] vmalloc : 0xf0800000 - 0xff800000 ( 240 MB) [ 0.000000] lowmem : 0xc0000000 - 0xf0000000 ( 768 MB) [ 0.000000] pkmap : 0xbfe00000 - 0xc0000000 ( 2 MB) [ 0.000000] modules : 0xbf000000 - 0xbfe00000 ( 14 MB) [ 0.000000] .text : 0xc0208000 - 0xc0cc28f8 (10987 kB) [ 0.000000] .init : 0xc0cc3000 - 0xc0dd8000 (1108 kB) [ 0.000000] .data : 0xc0dd8000 - 0xc0ef52b4 (1141 kB) [ 0.000000] .bss : 0xc0ef52b4 - 0xc0fcabb8 ( 855 kB) [ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=8, Nodes=1 [ 0.000000] Hierarchical RCU implementation. [ 0.000000] Build-time adjustment of leaf fanout to 32. [ 0.000000] NR_IRQS:16 nr_irqs:16 16 [ 0.000000] Switching to timer-based delay loop, resolution 41ns [ 0.000000] clocksource: mct-frc: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 79635851949 ns [ 0.000007] sched_clock: 32 bits at 24MHz, resolution 41ns, wraps every 89478484971ns [ 0.005176] Console: colour dummy device 80x30 [ 0.006058] console [tty0] enabled [ 0.006118] Calibrating delay loop (skipped), value calculated using timer frequency.. 48.00 BogoMIPS (lpj=120000) [ 0.006177] pid_max: default: 32768 minimum: 301 [ 0.006426] Security Framework initialized [ 0.006461] Yama: becoming mindful. [ 0.006505] SELinux: Initializing. [ 0.006849] Mount-cache hash table entries: 2048 (order: 1, 8192 bytes) [ 0.006893] Mountpoint-cache hash table entries: 2048 (order: 1, 8192 bytes) [ 0.008093] CPU: Testing write buffer coherency: ok [ 0.008169] ftrace: allocating 31745 entries in 63 pages [ 0.071546] CPU0: update cpu_capacity 448 [ 0.071596] CPU0: thread -1, cpu 0, socket 1, mpidr 80000100 [ 0.071832] Setting up static identity map for 0x40208280 - 0x40208318 [ 0.079277] EFI services will not be available. [ 0.081758] CPU1: update cpu_capacity 448 [ 0.081766] CPU1: thread -1, cpu 1, socket 0, mpidr 80000001 [ 0.083684] CPU2: update cpu_capacity 448 [ 0.083691] CPU2: thread -1, cpu 2, socket 0, mpidr 80000002 [ 0.085564] CPU3: update cpu_capacity 448 [ 0.085571] CPU3: thread -1, cpu 3, socket 0, mpidr 80000003 [ 0.087421] CPU4: update cpu_capacity 1535 [ 0.087428] CPU4: thread -1, cpu 0, socket 0, mpidr 80000000 [ 1.084983] CPU5: failed to boot: -110 [ 2.084997] CPU6: failed to boot: -110 [ 3.085014] CPU7: failed to boot: -110 [ 3.085076] Brought up 5 CPUs [ 3.085104] SMP: Total of 5 processors activated (240.00 BogoMIPS). [ 3.085131] CPU: All CPU(s) started in HYP mode.
Hello,
Is anyone else having this problem with the Odroid XU4 and Fedora 24? I've downloaded the 25 alpha Mate Spin and it too seems to have the same problem. Is anyone from the kernel team addressing this?
Thanks. Stewart
On Fri, Sep 2, 2016 at 11:56 AM, Stewart Samuels searider74@gmail.com wrote:
Hello,
Is anyone else having this problem with the Odroid XU4 and Fedora 24? I've downloaded the 25 alpha Mate Spin and it too seems to have the same problem. Is anyone from the kernel team addressing this?
It's a big.LITTLE archtecture some will be online some offline. What does lscpu report?
Peter
Hi Peter,
Here is the result of lscpu.
--------------------------------------------------------------------------------------- [root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000 ----------------------------------------------------------------------------------------
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
Stewart
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
Hi Peter,
I am not doing anything with the system other than booting up and logging in. This is true for the Ubuntu build as well.
Where are these policies set and can you provide any me any direction for documentation on them? seeming these are distro specific, I would expect something relative to Fedora.
Thanks.
Stewart
On 09/08/2016 05:29 AM, Peter Robinson wrote:
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
On Thu, Sep 8, 2016 at 1:19 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
I am not doing anything with the system other than booting up and logging in. This is true for the Ubuntu build as well.
Where are these policies set and can you provide any me any direction for documentation on them? seeming these are distro specific, I would expect something relative to Fedora.
Nope, they are upstream kernel (and possibly even upstream u-boot) specific. The only default we set in this regard that may, or may not, be Fedora specific is we use the On Demand governor as the default. This is architecture in dependent default across Fedora.
I doubt the Ubuntu build ships an upstream mainline kernel but then I don't follow any of what they do so TBH not sure there, I also have no idea what they set their default policy to.
So doing a quick google for "cpufreq" I get some of the following links that look remotely relevant, no idea how much they are, sorry.
https://docs.fedoraproject.org/en-US/Fedora/15/html/Power_Management_Guide/c... https://wiki.archlinux.org/index.php/CPU_frequency_scaling https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt https://www.pantz.org/software/cpufreq/usingcpufreqonlinux.html http://www.ibm.com/developerworks/library/l-cpufreq-1/index.html https://wiki.debian.org/HowTo/CpuFrequencyScaling
On 09/08/2016 05:29 AM, Peter Robinson wrote:
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
Thanks Peter,
This will be very disappointing if we cannot enable all the cpus.
BTW, when you refer to upstream here, is it the Redhat team or the Kernel team beyond?
Stewart
On 09/08/2016 08:40 AM, Peter Robinson wrote:
On Thu, Sep 8, 2016 at 1:19 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
I am not doing anything with the system other than booting up and logging in. This is true for the Ubuntu build as well.
Where are these policies set and can you provide any me any direction for documentation on them? seeming these are distro specific, I would expect something relative to Fedora.
Nope, they are upstream kernel (and possibly even upstream u-boot) specific. The only default we set in this regard that may, or may not, be Fedora specific is we use the On Demand governor as the default. This is architecture in dependent default across Fedora.
I doubt the Ubuntu build ships an upstream mainline kernel but then I don't follow any of what they do so TBH not sure there, I also have no idea what they set their default policy to.
So doing a quick google for "cpufreq" I get some of the following links that look remotely relevant, no idea how much they are, sorry.
https://docs.fedoraproject.org/en-US/Fedora/15/html/Power_Management_Guide/c... https://wiki.archlinux.org/index.php/CPU_frequency_scaling https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt https://www.pantz.org/software/cpufreq/usingcpufreqonlinux.html http://www.ibm.com/developerworks/library/l-cpufreq-1/index.html https://wiki.debian.org/HowTo/CpuFrequencyScaling
On 09/08/2016 05:29 AM, Peter Robinson wrote:
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
On Thu, Sep 8, 2016 at 1:57 PM, Stewart Samuels searider74@gmail.com wrote:
Thanks Peter,
This will be very disappointing if we cannot enable all the cpus.
BTW, when you refer to upstream here, is it the Redhat team or the Kernel team beyond?
Never Red Hat. Fedora is upstream to Red Hat, so upstream means the linux kernel upstream at kernel.org and the kernel at large.
On 09/08/2016 08:40 AM, Peter Robinson wrote:
On Thu, Sep 8, 2016 at 1:19 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
I am not doing anything with the system other than booting up and logging in. This is true for the Ubuntu build as well.
Where are these policies set and can you provide any me any direction for documentation on them? seeming these are distro specific, I would expect something relative to Fedora.
Nope, they are upstream kernel (and possibly even upstream u-boot) specific. The only default we set in this regard that may, or may not, be Fedora specific is we use the On Demand governor as the default. This is architecture in dependent default across Fedora.
I doubt the Ubuntu build ships an upstream mainline kernel but then I don't follow any of what they do so TBH not sure there, I also have no idea what they set their default policy to.
So doing a quick google for "cpufreq" I get some of the following links that look remotely relevant, no idea how much they are, sorry.
https://docs.fedoraproject.org/en-US/Fedora/15/html/Power_Management_Guide/c... https://wiki.archlinux.org/index.php/CPU_frequency_scaling https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt https://www.pantz.org/software/cpufreq/usingcpufreqonlinux.html http://www.ibm.com/developerworks/library/l-cpufreq-1/index.html https://wiki.debian.org/HowTo/CpuFrequencyScaling
On 09/08/2016 05:29 AM, Peter Robinson wrote:
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
Based on what I've read in the Fedora Docs, I should be able to change the cpu governor by using the following command:
cpupower frequency-set --governor [governor]
However, if I replace the [governor] string with "userspace" (quotes not include) nothing seems to happen. I do not get any errors from the command and "ondemand" seems to remain the governor, as shown by the following command:
[root@myodroid ~]# cpupower frequency-info --policy analyzing CPU 0: current policy: frequency should be within 200 MHz and 1.30 GHz. The governor "ondemand" may decide which speed to use within this range.
"userspace" is an available governor on my system, as shown by the following command:
[root@myodroid ~]# cpupower frequency-info --governors analyzing CPU 0: available cpufreq governors: conservative userspace powersave ondemand performance
In fact, it doesn't seem to matter to which governor I set it. Nothing changes.
Is there something I am missing? What could be happening here?
Stewart
On 09/08/2016 09:01 AM, Peter Robinson wrote:
On Thu, Sep 8, 2016 at 1:57 PM, Stewart Samuels searider74@gmail.com wrote:
Thanks Peter,
This will be very disappointing if we cannot enable all the cpus.
BTW, when you refer to upstream here, is it the Redhat team or the Kernel team beyond?
Never Red Hat. Fedora is upstream to Red Hat, so upstream means the linux kernel upstream at kernel.org and the kernel at large.
On 09/08/2016 08:40 AM, Peter Robinson wrote:
On Thu, Sep 8, 2016 at 1:19 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
I am not doing anything with the system other than booting up and logging in. This is true for the Ubuntu build as well.
Where are these policies set and can you provide any me any direction for documentation on them? seeming these are distro specific, I would expect something relative to Fedora.
Nope, they are upstream kernel (and possibly even upstream u-boot) specific. The only default we set in this regard that may, or may not, be Fedora specific is we use the On Demand governor as the default. This is architecture in dependent default across Fedora.
I doubt the Ubuntu build ships an upstream mainline kernel but then I don't follow any of what they do so TBH not sure there, I also have no idea what they set their default policy to.
So doing a quick google for "cpufreq" I get some of the following links that look remotely relevant, no idea how much they are, sorry.
https://docs.fedoraproject.org/en-US/Fedora/15/html/Power_Management_Guide/c... https://wiki.archlinux.org/index.php/CPU_frequency_scaling https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt https://www.pantz.org/software/cpufreq/usingcpufreqonlinux.html http://www.ibm.com/developerworks/library/l-cpufreq-1/index.html https://wiki.debian.org/HowTo/CpuFrequencyScaling
On 09/08/2016 05:29 AM, Peter Robinson wrote:
On Wed, Sep 7, 2016 at 8:44 PM, Stewart Samuels searider74@gmail.com wrote:
Hi Peter,
Here is the result of lscpu.
[root@myodroid ~]# lscpu Architecture: armv7l Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-4 Off-line CPU(s) list: 5-7 Thread(s) per core: 1 Core(s) per socket: 2 Socket(s): 2 Model name: ARMv7 Processor rev 3 (v7l) CPU max MHz: 1300.0000 CPU min MHz: 200.0000
Is there any way to enable these other cpus? My Ubuntu 16.04 installation has them all enabled and the Ubuntu responsiveness is much quicker. I suspect this has something to do with it.
So it's shut them off, it's something with the way the big.LITTLE stuff works, so it's basically as expected. I believe it's handled as part of the cpufreq policies from user space but I've done little with the b.L stuff so I'm not sure. I'd try with the performance policy first.
In terms of speed vs other distros, it would likely depend on a lot more than just the cores that are running but I have no idea what you're doing with it (remote server/desktop/what ever) so there's likely a lot that will come into play.
-
Peter Robinson -
Stewart Samuels