Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Only one core using kernel 4.9
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4  Next  
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Tue Jan 10, 2017 7:09 pm    Post subject: Reply with quote

I've copied the .config from my running 4.8.10er kernel and then make oldconfig.
I used the default answer mostly, except for some amazon stuff i don't need.

I've pasted the dmesg-kernel-4.9.2.txt

http://pastebin.com/9x0nf1CQ

and the config-kernel-4.9.2.txt

http://pastebin.com/PEhPM0hQ

The error is gone, as you can see.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Wed Jan 11, 2017 5:49 pm    Post subject: Reply with quote

The interesting part is now this:

4.9.2
[ 0.117526] smpboot: CPU0: AMD FX(tm)-4100 Quad-Core Processor (family: 0x15, model: 0x1, stepping: 0x2)
[ 0.117529] Performance Events: Fam15h core perfctr, AMD PMU driver.
[ 0.117533] ... version: 0
[ 0.117534] ... bit width: 48
[ 0.117534] ... generic registers: 6
[ 0.117535] ... value mask: 0000ffffffffffff
[ 0.117536] ... max period: 00007fffffffffff
[ 0.117536] ... fixed-purpose events: 0
[ 0.117537] ... event mask: 000000000000003f
[ 0.125577] MCE: In-kernel MCE decoding enabled.
[ 0.127538] x86: Booted up 1 node, 1 CPUs
[ 0.127540] smpboot: Total of 1 processors activated (7184.53 BogoMIPS)

4.8.10
[ 0.117520] smpboot: CPU0: AMD FX(tm)-4100 Quad-Core Processor (family: 0x15, model: 0x1, stepping: 0x2)
[ 0.117523] Performance Events: Fam15h core perfctr, AMD PMU driver.
[ 0.117527] ... version: 0
[ 0.117527] ... bit width: 48
[ 0.117528] ... generic registers: 6
[ 0.117529] ... value mask: 0000ffffffffffff
[ 0.117529] ... max period: 00007fffffffffff
[ 0.117530] ... fixed-purpose events: 0
[ 0.117530] ... event mask: 000000000000003f
[ 0.123570] MCE: In-kernel MCE decoding enabled.
[ 0.128552] x86: Booting SMP configuration:
[ 0.128553] .... node #0, CPUs: #1 #2 #3
[ 0.322534] x86: Booted up 1 node, 4 CPUs
[ 0.322537] smpboot: Total of 4 processors activated (28734.89 BogoMIPS)
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3235
Location: Gainesville, Florida

PostPosted: Mon Jan 16, 2017 11:19 pm    Post subject: Reply with quote

Something appears wrong in your 4.9.2 kernel config file. ([4.9.2 is missing x86: Booting SMP configuration:) Try this:

1. Remove the /usr/src/linux-4.9.2-gentoo directory, and then get a fresh new source directory:
Code:
 emerge =sys-kernel/gentoo-sources-4.9.2

Remove (if any) .config file in the new source directory. NOT the .cocciconfig file

2. Then copy the working kernel .config file which does enable all 4 cores from your 4.8.10 directory into the your new linux-4.9.2-gentoo directory. Make sure it's just named .config (nothing else) Forget menuconfig, or oldconfig, etc.

3. Then cd into the new 4.9.2 and run:
Code:
 make bzImage && make modules && make modules_install

4. then
Code:
cp arch/x86/boot/bzImage /boot/kernel-4.9.2-gentoo


5, Update grub
Code:
 grub-mkconfig -o /boot/grub/grub.cfg

Reboot into the new 4.9.2, and all four cpu cores should be recognized, and active. If it works, you can basically use it (the .config file) on this hardware for future kernels.

You should also weed out your .config for your hardware, only enabling what you actually need. Then you can just use a nice gui like make gconfig or make xconfig, and easily check if need be for any new features/drivers. The kernel config GUI's really lay out everything so it's quick and easy to load and save your custom file into the new kernel, and deeply explore/edit all the config options if you wish.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.29-r2, gcc-9.1.0 kernel-5.1.16-gentoo USE=experimental


Last edited by wrc1944 on Tue Jan 17, 2017 2:14 pm; edited 1 time in total
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Tue Jan 17, 2017 1:18 am    Post subject: Reply with quote

Quote:
Then copy the working kernel .config file from your 4.8.10 directory into the your new linux-4.9.2-gentoo. Make sure it's just named .config (nothing else) Forget menuconfig, or oldconfig, etc.

3. Then cd into the new 4.9.2 and run:
Code:
make bzImage && make modules && make modules_install
NO! He MUST run make oldconfig first or new flags for 4.9 will not be set!

He should run make oldconfig first taking all the default choices.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3235
Location: Gainesville, Florida

PostPosted: Tue Jan 17, 2017 5:45 am    Post subject: Reply with quote

Tony0945,
I've done this many hundreds of times over the last 15-20 years, and it never fails. and yes the options selected in any .config file you place in a fresh source directory and then manually run
Code:
 make bzImage && make modules && make modules_install
are recognized and take effect on the new kernel.

I assume by "new flags for 4.9 will not be set! " you mean the kernel options, correct?

Of course I don't usually do kernels this way- this was just be sure donmartio quickly got a kernel recognizing all 4 cores because he seems to be having problems with running make oldconfig/menu config. We know his 4.8.10 config file is getting all cores recognized, and since he's compiling the 4.9.2 kernel on the same hardware that 4.8.10 config fle will work fine. I think that while going through the oldconfig options he probably left out a needed option. If he uses that 4.8.10 .config, I'll be very surprised if all 4 cores AREN'T then being utilized.

I always use the GUI make xconfig or make gconfig on desktop systems. Compared to their ease for loading custom .config files, and complete control and immediate search/help access for any actual new features, menuconfig or oldconfig (while eventually acomplishing the same thing) are to me awkward, more prone to making a mistake selecting options, not to mention being far more time consuming. Unless I needed to add new hardware or change file systems, enable new support or modules, I can just use the same .config for many, many subsequent kernel updates- no big deal.

All such kernel options like drivers, kernel updates, etc., etc. are automatically picked up from the new versions in the new kernel sources, even if you are still using your custom .config file.

Or, donmartio could try building a full 4.9.2 with "genkernel all" and worry about weeding out all unecessary hardware and drivers, etc. options in the .config file later.

Of course If you think anything I've mentioned is bad practice, I'm all ears, and welcome any feedback. :wink:
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.29-r2, gcc-9.1.0 kernel-5.1.16-gentoo USE=experimental
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Tue Jan 17, 2017 8:52 pm    Post subject: Reply with quote

Thanks for the replies.

I tried as you suggested with kernel 4.9.4 with no success.

May i try 'genkernel all' just to verify if i get the kernel recognize my cores.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Tue Jan 17, 2017 9:12 pm    Post subject: Reply with quote

Donmartio, what is the output of "lscpu"?
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Tue Jan 17, 2017 9:18 pm    Post subject: Reply with quote

Hi Tony0945,

with kernel 4.8.10:

Code:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    2
Core(s) per socket:    2
Socket(s):             1
Vendor ID:             AuthenticAMD
CPU family:            21
Model:                 1
Model name:            AMD FX(tm)-4100 Quad-Core Processor
Stepping:              2
CPU MHz:               3600.000
CPU max MHz:           3600,0000
CPU min MHz:           1400,0000
BogoMIPS:              7184.37
Virtualization:        AMD-V
L1d cache:             16K
L1i cache:             64K
L2 cache:              2048K
L3 cache:              8192K
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core perfctr_nb cpb hw_pstate vmmcall arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold

_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Tue Jan 17, 2017 9:38 pm    Post subject: Reply with quote

With kernel-4.9.4

Code:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                1
On-line CPU(s) list:   0
Thread(s) per core:    1
Core(s) per socket:    1
Socket(s):             1
Vendor ID:             AuthenticAMD
CPU family:            21
Model:                 1
Model name:            AMD FX(tm)-4100 Quad-Core Processor
Stepping:              2
CPU MHz:               3600.000
CPU max MHz:           3600,0000
CPU min MHz:           1400,0000
BogoMIPS:              7184.73
Virtualization:        AMD-V
L1d cache:             16K
L1i cache:             64K
L2 cache:              2048K
L3 cache:              8192K
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core perfctr_nb cpb hw_pstate vmmcall arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold

_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Tue Jan 17, 2017 11:10 pm    Post subject: Reply with quote

with 4.9.3 on one machine
Code:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                3
On-line CPU(s) list:   0-2
Thread(s) per core:    1
Core(s) per socket:    3
Socket(s):             1
Vendor ID:             AuthenticAMD
CPU family:            16
Model:                 5
Model name:            AMD Athlon(tm) II X3 440 Processor
Stepping:              2
CPU MHz:               800.000
CPU max MHz:           3000.0000
CPU min MHz:           800.0000
BogoMIPS:              6000.14
Virtualization:        AMD-V
L1d cache:             64K
L1i cache:             64K
L2 cache:              512K
and on another
Code:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    2
Core(s) per socket:    2
Socket(s):             1
Vendor ID:             AuthenticAMD
CPU family:            21
Model:                 48
Model name:            AMD A8-7600 Radeon R7, 10 Compute Cores 4C+6G
Stepping:              1
CPU MHz:               1400.000
CPU max MHz:           3100.0000
CPU min MHz:           1400.0000
BogoMIPS:              6188.74
Virtualization:        AMD-V
L1d cache:             16K
L1i cache:             96K
L2 cache:              2048K
It seems like there is a bug involving only the FX series, I would file a kernel bug.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Wed Jan 18, 2017 3:37 am    Post subject: Reply with quote

OK, I've taken your pastebin 4.9.2 config and only appended local version "donmartio" so that it does not overlay my kernel. I built it without "make oldconfig"

I used " make -j4 && make -j4 modules_install && make -j4 install" from my script on /usr/src/linux pointed to 4.9.3 Tomorrow I will go to the local machine and boot that kernel, since I probably can't ssh in due to hardware differences. I'll boot non-X and see what lscpu says. If it shows all three CPU's then something is definitely different for FX-4100 than for Athlon II X3.

Watch this space!

UPDATE1: Build failed with "*** No rule to make target '/lib/firmware/rtl_nic//rtl8168e-3.fw', needed by 'firmware/rtl8168e-3.fw.gen.o"
I had to remove CONFIG_EXTRA_FIRMWARE && CONFIG_EXTRA_FIRMWARE_DIR to progress. I don't have this hardware anyway.

That was the only problem building. Added the lines
Code:
title=test
root (hd0,0)
kernel /boot/vmlinuz-4.9.3donmartio-gentoo  root=/dev/sda3 vga=0x375
to /boot/grub/grub.conf and will test it tomorrow.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3235
Location: Gainesville, Florida

PostPosted: Wed Jan 18, 2017 4:54 am    Post subject: Reply with quote

donmartio,
Hmmm... one other idea: Have you tried booting the Gentoo live cd (or any other live cd) and seeing what
Code:
cat /proc/cpuinfo

reports?

A live cd would supposedly detect and properly configure its kernel for the AMD FX-4100 Quad-Core Processor . You could then look at the .config file it produced, and use it, or edit yours with what you found in the live cd kernel.

Then again, Tony0945 might very well be right about a weird undiscovered FX series kernel bug, somehow only affecting the FX-4100 Quad on certain motherboards, or who knows?

My AMD -8320 Eight-Core Processor is also family 21, and never had this problem over countless kernel versions, on different Gentoo installations and several other distros all on on my multi-boot testing machine. That fact still makes me think we're missing something in the .config file. :?
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.29-r2, gcc-9.1.0 kernel-5.1.16-gentoo USE=experimental
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3235
Location: Gainesville, Florida

PostPosted: Wed Jan 18, 2017 4:09 pm    Post subject: Reply with quote

donmartio,
I noticed your .config file has CONFIG_MK8=y set, which is enabling only the old AMD 64 3000+ flags (Venice core type cpus, circa 2006).
The amd FX 4100 quad is the vishera core (i.e. piledriver, as my FX 8320 is). EDIT: Actually, the FX-4100 Quad is zambezi/bulldozer, NOT piledriver- Sorry. :oops:

My thinking is that while mk8=y seems to still work with kernel 4.8.10, it might not be adequate for the 4.9 series?

You only will get the expanded list of possible gcc cpu optimizations by using the "expermental" USE flag for gentoo-sources, enabled with placing
Quote:
sys-kernel/gentoo-sources experimental
in /etc/portage/package.use. It also adds a few more patches like BFQ, but they aren't enabled by default, so no worries there.

Or, you could get the latest gcc opts patch and apply it yourself from https://github.com/graysky2/kernel_gcc_patch

FWIW, here's my relevant .config file section. When I first got this cpu, I set piledriver, but then went with CONFIG_MNATIVE=y which lets gcc itself detect and set a few more flags the specific cpu is designed for than setting any of the actual core names. I confirmed this when I made the change. Notice I also set CONFIG_NR_CPUS=8, the exact number of my cpu cores, while you have CONFIG_NR_CPUS=64. Not sure, but maybe in kernel 4.9.x setting way more than the actual cores might cause it to default to one core because it fails to find the number of cores the .config setting allows? Who knows?

Have you ever run across this?
Code:
#error "CONFIG_NR_CPUS is too large, please lower it."
Trying to figure this out, I ran across it several times. I would think CONFIG_NR_CPUS=64 would only apply for server cpus, which the amd FX4100 is not. At this point, I'm running out of ideas. :roll:

Quote:
# Processor type and features
#
CONFIG_ZONE_DMA=y
CONFIG_SMP=y
CONFIG_X86_FEATURE_NAMES=y
CONFIG_X86_FAST_FEATURE_TESTS=y
# CONFIG_X86_MPPARSE is not set
# CONFIG_GOLDFISH is not set
# CONFIG_X86_EXTENDED_PLATFORM is not set
# CONFIG_X86_INTEL_LPSS is not set
# CONFIG_X86_AMD_PLATFORM_DEVICE is not set
# CONFIG_IOSF_MBI is not set
CONFIG_X86_SUPPORTS_MEMORY_FAILURE=y
CONFIG_SCHED_OMIT_FRAME_POINTER=y
# CONFIG_HYPERVISOR_GUEST is not set
CONFIG_NO_BOOTMEM=y
# CONFIG_MK8 is not set
# CONFIG_MK8SSE3 is not set
# CONFIG_MK10 is not set
# CONFIG_MBARCELONA is not set
# CONFIG_MBOBCAT is not set
# CONFIG_MBULLDOZER is not set
# CONFIG_MPILEDRIVER is not set
# CONFIG_MSTEAMROLLER is not set
# CONFIG_MJAGUAR is not set
# CONFIG_MPSC is not set
# CONFIG_MATOM is not set
# CONFIG_MCORE2 is not set
# CONFIG_MNEHALEM is not set
# CONFIG_MWESTMERE is not set
# CONFIG_MSILVERMONT is not set
# CONFIG_MSANDYBRIDGE is not set
# CONFIG_MIVYBRIDGE is not set
# CONFIG_MHASWELL is not set
# CONFIG_MBROADWELL is not set
# CONFIG_MSKYLAKE is not set
# CONFIG_GENERIC_CPU is not set
CONFIG_MNATIVE=y
CONFIG_X86_INTERNODE_CACHE_SHIFT=6
CONFIG_X86_L1_CACHE_SHIFT=6
CONFIG_X86_INTEL_USERCOPY=y
CONFIG_X86_USE_PPRO_CHECKSUM=y
CONFIG_X86_TSC=y
CONFIG_X86_CMPXCHG64=y
CONFIG_X86_CMOV=y
CONFIG_X86_MINIMUM_CPU_FAMILY=64
CONFIG_X86_DEBUGCTLMSR=y
CONFIG_PROCESSOR_SELECT=y
# CONFIG_CPU_SUP_INTEL is not set
CONFIG_CPU_SUP_AMD=y
# CONFIG_CPU_SUP_CENTAUR is not set
CONFIG_HPET_TIMER=y
CONFIG_HPET_EMULATE_RTC=y
CONFIG_DMI=y
CONFIG_GART_IOMMU=y
# CONFIG_CALGARY_IOMMU is not set
CONFIG_SWIOTLB=y
CONFIG_IOMMU_HELPER=y
# CONFIG_MAXSMP is not set
CONFIG_NR_CPUS=8
# CONFIG_SCHED_SMT is not set
CONFIG_SCHED_MC=y
# CONFIG_PREEMPT_NONE is not set
# CONFIG_PREEMPT_VOLUNTARY is not set
CONFIG_PREEMPT=y
CONFIG_PREEMPT_COUNT=y
CONFIG_X86_LOCAL_APIC=y
CONFIG_X86_IO_APIC=y
CONFIG_X86_REROUTE_FOR_BROKEN_BOOT_IRQS=y
CONFIG_X86_MCE=y
# CONFIG_X86_MCE_INTEL is not set
CONFIG_X86_MCE_AMD=y
CONFIG_X86_MCE_THRESHOLD=y
# CONFIG_X86_MCE_INJECT is not set

_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.29-r2, gcc-9.1.0 kernel-5.1.16-gentoo USE=experimental


Last edited by wrc1944 on Wed Jan 18, 2017 11:14 pm; edited 6 times in total
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Wed Jan 18, 2017 4:26 pm    Post subject: Reply with quote

I booted your kernel. First I had a kernel panic because your kernel renamed /dev/sr0 to /dev/sda and /dev/sda to /dev/sdb. After fixing the kernel parameter "root=/dev/sdb3", it booted.
Of course, because of the hardware differences many things failed like net.eth0 and X, but I was able to run lscpu
Code:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                3
On-line CPU(s) list:   0-2
Thread(s) per core:    1
Core(s) per socket:    3
Socket(s):             1
Vendor ID:             AuthenticAMD
CPU family:            16
Model:                 5
Model name:            AMD Athlon(tm) II X3 440 Processor
Stepping:              2
CPU MHz:               3000.000
CPU max MHz:           3000.0000
CPU min MHz:           800.0000
BogoMIPS:              6000.22
Virtualization:        AMD-V
L1d cache:             64K
L1i cache:             64K
L2 cache:              512K


Along with wrc1944, I had noticed that you had selected the processor type as basic K8. It's a shot in the dark, but try selecting BULLDOZER or NATIVE.
I always build NATIVE.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Wed Jan 18, 2017 6:28 pm    Post subject: Reply with quote

Wow, thanks for your suggestions.

I'll try that immediately.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Wed Jan 18, 2017 6:48 pm    Post subject: Reply with quote

No success... :(

I tried

USE="experimental" emerge =sys-kernel/gentoo-sources-4.9.4

and selected NATIVE.

Ill try Bulldozer now and remerge the kernel sources to compile the kernel from 'clean' sources.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 7089
Location: Saint Amant, Acadiana

PostPosted: Wed Jan 18, 2017 7:22 pm    Post subject: Reply with quote

If everything fails I'd bisect and file a bug. Although it is possible you got "lucky" and your CPU has some rare hardware quirk.
_________________
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Wed Jan 18, 2017 7:30 pm    Post subject: Reply with quote

Well no... no success...

since the kernel config runs on your maschine Tony0945, it would be strange if this caused the problem.

It seems i'm running out of options now.

How and where do i post a kernel bug?
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 7089
Location: Saint Amant, Acadiana

PostPosted: Wed Jan 18, 2017 7:38 pm    Post subject: Reply with quote

You should bisect first, this gives your bug filing more weight.
How to file a Linux kernel bug.
_________________
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Wed Jan 18, 2017 8:14 pm    Post subject: Reply with quote

Hi Jaglover,

thanks... i'll bisect. Never heard about that before. Sounds reasonable.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Thu Jan 19, 2017 7:04 pm    Post subject: Reply with quote

Thanks again for this tip. I would have done this earlier if i had known this cool git feature.
I started git bisect with 4.8.17 and 4.9.0.

This was still a way to go, but here is the result:

Code:

8f54969dc8d6704632b42cbb5e47730cd75cc713 is the first bad commit
commit 8f54969dc8d6704632b42cbb5e47730cd75cc713
Author: Gu Zheng <guz.fnst@cn.fujitsu.com>
Date:   Thu Aug 25 16:35:16 2016 +0800

    x86/acpi: Introduce persistent storage for cpuid <-> apicid mapping
   
    The whole patch-set aims at making cpuid <-> nodeid mapping persistent. So that,
    when node online/offline happens, cache based on cpuid <-> nodeid mapping such as
    wq_numa_possible_cpumask will not cause any problem.
    It contains 4 steps:
    1. Enable apic registeration flow to handle both enabled and disabled cpus.
    2. Introduce a new array storing all possible cpuid <-> apicid mapping.
    3. Enable _MAT and MADT relative apis to return non-present or disabled cpus' apicid.
    4. Establish all possible cpuid <-> nodeid mapping.
   
    This patch finishes step 2.
   
    In this patch, we introduce a new static array named cpuid_to_apicid[],
    which is large enough to store info for all possible cpus.
   
    And then, we modify the cpuid calculation. In generic_processor_info(),
    it simply finds the next unused cpuid. And it is also why the cpuid <-> nodeid
    mapping changes with node hotplug.
   
    After this patch, we find the next unused cpuid, map it to an apicid,
    and store the mapping in cpuid_to_apicid[], so that cpuid <-> apicid
    mapping will be persistent.
   
    And finally we will use this array to make cpuid <-> nodeid persistent.
   
    cpuid <-> apicid mapping is established at local apic registeration time.
    But non-present or disabled cpus are ignored.
   
    In this patch, we establish all possible cpuid <-> apicid mapping when
    registering local apic.
   
    Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
    Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
    Signed-off-by: Zhu Guihua <zhugh.fnst@cn.fujitsu.com>
    Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
    Acked-by: Ingo Molnar <mingo@kernel.org>
    Cc: mika.j.penttila@gmail.com
    Cc: len.brown@intel.com
    Cc: rafael@kernel.org
    Cc: rjw@rjwysocki.net
    Cc: yasu.isimatu@gmail.com
    Cc: linux-mm@kvack.org
    Cc: linux-acpi@vger.kernel.org
    Cc: isimatu.yasuaki@jp.fujitsu.com
    Cc: gongzhaogang@inspur.com
    Cc: tj@kernel.org
    Cc: izumi.taku@jp.fujitsu.com
    Cc: cl@linux.com
    Cc: chen.tang@easystack.cn
    Cc: akpm@linux-foundation.org
    Cc: kamezawa.hiroyu@jp.fujitsu.com
    Cc: lenb@kernel.org
    Link: http://lkml.kernel.org/r/1472114120-3281-4-git-send-email-douly.fnst@cn.fujitsu.com
    Signed-off-by: Thomas Gleixner <tglx@linutronix.de>


I'll try to look into this later but it looks absolutely that they miss a mapping here.
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
donmartio
Apprentice
Apprentice


Joined: 11 Dec 2004
Posts: 233

PostPosted: Thu Jan 19, 2017 7:10 pm    Post subject: Reply with quote

Just for the record, i followed this howto:

https://wiki.gentoo.org/wiki/Kernel_git-bisect
_________________
Always code as if the person who ends up maintaining your code will be a violent psychopath who knows where you live.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Thu Jan 19, 2017 8:29 pm    Post subject: Reply with quote

Quote:
which is large enough to store info for all possible cpus.
Which explains why there is no problem with my old Athlon II nor my newer but very common A8. I would have thought that FX-4100 would be also, but it was very silly to assume that a table held ALL possible CPU's since new versions and steppings are released all the time. In fact, FX-4100 may even be in the table but your stepping is slightly different.


Thanks very much for the bisect link.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3235
Location: Gainesville, Florida

PostPosted: Thu Jan 19, 2017 11:37 pm    Post subject: Reply with quote

Guess I've been lucky to never have run into something like this, or else it's just very rare.
I've literally used every amd cpu family and all the cpus within a family released since the old amd k2 & k3 days, including most of the steppings of each cpu, if available.

Plus, all while testing every linux kernel version from 2.6.x -> 4.9.4 as they were released over the years. Only one I ever skipped was the first Bulldozer FX release, after looking at all the benches and reviews about its questionable single thread performance. Almost went with Intel. Finally got a Piledriver/AM3+ rig to hold me over, anticipating the promised upgrade the Steamroller or Excavator, which never were released. :roll: Looks like Zen is going to pan out, so if the benches and reviews look good that's the next one for me.

Many thanks for the great tip & info about bisecting!
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.29-r2, gcc-9.1.0 kernel-5.1.16-gentoo USE=experimental


Last edited by wrc1944 on Sat Jan 21, 2017 4:14 pm; edited 1 time in total
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 3076
Location: Illinois, USA

PostPosted: Fri Jan 20, 2017 9:02 pm    Post subject: Reply with quote

Just learned that 4.9.5 has this fix in it.
Quote:
Thomas Gleixner (1):
x86/bugs: Separate AMD E400 erratum and C1E bug


Might be worth a try.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Goto page Previous  1, 2, 3, 4  Next
Page 2 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum