| View previous topic :: View next topic |
| Author |
Message |
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
 Posted: Thu Dec 27, 2012 8:45 pm Post subject: IRQ seems to be using only CPU0. |
 |
|
For some unknown to me reason the IRQ seems to use pretty much only CPU0, smp_affinity the '00000f' suppose to use ALL cpus but it does use only the first one, I have googled the http://docs.oracle.com/cd/E24290_01/coh.371/e22838/tune_perftune.htm which let me understand a little about how can I force a specified irq to use specified cpu, I have written a hacky script to balance it between 4 threads (I own core i5-2430M, two cores with ht) and it seems to be working, kind of...
| Code: | #!/bin/sh
# irq-rebalance.sh
for i in /proc/irq/*/smp_affinity; do
case "${smp_affinity}" in
''|'8')
smp_affinity=1
;;
'1')
smp_affinity=2
;;
'2')
smp_affinity=4
;;
'4')
smp_affinity=8
;;
esac
echo "$i: 00000${smp_affinity}"
echo "00000${smp_affinity}" >"$i"
done |
Without it, the /proc/interrupts looks like:
| Code: | CPU0 CPU1 CPU2 CPU3
0: 24 0 0 0 IO-APIC-edge timer
1: 62188 0 0 0 IO-APIC-edge i8042
8: 267 0 0 0 IO-APIC-edge rtc0
9: 264195 0 0 0 IO-APIC-fasteoi acpi
12: 136278 0 0 0 IO-APIC-edge i8042
16: 18718 0 0 0 IO-APIC-fasteoi i915, ehci_hcd:usb1
17: 7697397 0 0 0 IO-APIC-fasteoi ehci_hcd:usb2, iwlwifi
18: 1083624 0 0 0 IO-APIC-fasteoi ahci, mmc0, eth0
22: 860678 0 0 0 IO-APIC-fasteoi snd_hda_intel
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 3362075 3341920 3416035 3924691 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 3 0 0 0 APIC ICR read retries
RES: 275352 310805 137941 66781 Rescheduling interrupts
CAL: 4294963368 4294964404 4294964540 4294964464 Function call interrupts
TLB: 261929 332582 224693 200127 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 147 147 147 147 Machine check polls
ERR: 0
MIS: 0 |
And yes, I have no idea what I am doing, if anyone have any idea about the IRQ and can tell me why its not balanced I would be thankful. |
|
| Back to top |
|
 |
aCOSwt
Bodhisattva
Joined: 19 Oct 2007
Posts: 2537
Location: Hilbert space
|
| Posted: Thu Dec 27, 2012 9:02 pm Post subject: |
|
|
Could be not related at all, just curious to know the status of CONFIG_HOTPLUG_CPU on the kernel your system is running.
_________________
|
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Thu Dec 27, 2012 9:08 pm Post subject: |
|
|
| It is enabled, I think I need it for suspend (s2disk) to work. |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Thu Dec 27, 2012 9:19 pm Post subject: |
|
|
| There's little point in balancing IRQs on a non-NUMA system, since most of their work is done in kernel threads already. |
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Thu Dec 27, 2012 9:21 pm Post subject: |
|
|
| Well, the thing is, when I have really busy usb ports (like copy a lot of data with avg 28 MBps from one usb device to another) my touchpad is... laggy. I also notice other small issues that could be IRQ fault. Appreantly manual rerouting is the only possible solution. Thats just sad. |
|
| Back to top |
|
|
PaulBredbury
Watchman
Joined: 14 Jul 2005
Posts: 7310
|
| Posted: Thu Dec 27, 2012 10:30 pm Post subject: |
|
|
| SlashBeast wrote: | | my touchpad is... laggy |
From my messing with IRQs, trying to optimize USB mouse input, soundcard and Nvidia IRQs (I just created contention and more latency), I would say just let the scheduler do it. I.e.:
| Code: | echo f > /proc/irq/default_smp_affinity
for i in /proc/irq/*/smp_affinity; do
echo f > $i
done |
(Is it supposed to be "f" or "0f" or 15?)
I don't know why this isn't automatic for you. Maybe:
| Quote: | | So if on your system CPU0 handles all interrupts by default, this probably means that APIC configured ambiguously. |
|
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Thu Dec 27, 2012 10:36 pm Post subject: |
|
|
| PaulBredbury wrote: | | I don't know why this isn't automatic for you. |
It is, but because of HOTPLUG_CPU (as I googled) it does force them on CPU0, the great read http://www.alexonlinux.com/smp-affinity-and-proper-interrupt-handling-in-linux
Long story short, when you have f (00000f), instead of using all cpus, it uses CPU0.
For now I adjusted the routes manualy, to put sata, wifi and sound card on different cores/threads.
| Code: | #!/bin/sh
set_irq() {
irq="$1"
case "$2" in
cpu0)
cpu=1
;;
cpu1)
cpu=2
;;
cpu2)
cpu=4
;;
cpu3)
cpu=8
;;
esac
echo "${cpu}" >"/proc/irq/${irq}/smp_affinity"
}
# 17: ehci_hcd:usb2, iwlwifi
set_irq 17 cpu3
# 18: ahci, mmc0, eth0
set_irq 18 cpu2
# 22: snd_hda_intel
set_irq 22 cpu1 |
And after few hours I can see that it does make things better, at least on paper.
| Code: | sabre ~ # ./summarize_irq_stats.awk old_interrupts
cpu0: 14022875
cpu1: 3985454
cpu2: 3778816
cpu3: 4191746
sabre ~ # ./summarize_irq_stats.awk /proc/interrupts
cpu0: 605714
cpu1: 547066
cpu2: 625991
cpu3: 825810 |
|
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Fri Dec 28, 2012 1:30 am Post subject: |
|
|
| If you want another (slightly crazy) solution you could try booting with "threadirqs=1", which forces all IRQ handlers possible into separate threads. IIRC the realtime kernel project uses this, so it should be safe. |
|
| Back to top |
|
|
dmpogo
Advocate
Joined: 02 Sep 2004
Posts: 3488
Location: Canada
|
| Posted: Fri Dec 28, 2012 1:38 am Post subject: |
|
|
That cannot be general situation. I have CONFIG_HOTPLUG_CPU=y on all four different machines I manage, and they all show proper IRQ balancing.
Note also that your reference refers to quite old kernels <= 2.6.24 |
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Fri Dec 28, 2012 10:31 am Post subject: |
|
|
| dmpogo wrote: | That cannot be general situation. I have CONFIG_HOTPLUG_CPU=y on all four different machines I manage, and they all show proper IRQ balancing.
Note also that your reference refers to quite old kernels <= 2.6.24 |
Then I am absolutly without idea, because it does keep all devices, acpi and rtc on one cpu. What is interesing I just tested on another sandybridge cpu, i7-2600, and I also see that devices' irq occupate single core.
| Code: | # cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7
0: 25 0 0 0 0 0 0 0 IO-APIC-edge timer
1: 249 0 0 0 0 0 0 0 IO-APIC-edge i8042
8: 1 0 0 0 0 0 0 0 IO-APIC-edge rtc0
9: 0 0 0 0 0 0 0 0 IO-APIC-fasteoi acpi
23: 60 0 0 0 0 0 0 0 IO-APIC-fasteoi ehci_hcd:usb3, ehci_hcd:usb4
40: 0 0 0 0 0 0 0 0 PCI-MSI-edge PCIe PME
41: 0 0 0 0 0 0 0 0 PCI-MSI-edge PCIe PME
42: 0 0 0 0 0 0 0 0 PCI-MSI-edge PCIe PME
43: 0 0 0 0 0 0 0 0 PCI-MSI-edge PCIe PME
44: 98076556 0 0 0 0 0 0 0 PCI-MSI-edge ahci
45: 90452375 0 0 0 0 0 0 0 PCI-MSI-edge eth0
46: 63 0 0 0 0 0 0 0 PCI-MSI-edge i915
47: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
48: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
49: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
50: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
51: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
52: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
53: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
54: 0 0 0 0 0 0 0 0 PCI-MSI-edge xhci_hcd
NMI: 218863 198034 192376 194477 42378 50328 52237 54105 Non-maskable interrupts
LOC: 1401606099 1421191629 1426138094 1432049557 479262595 490584201 490445234 478404349 Local timer interrupts
SPU: 0 0 0 0 0 0 0 0 Spurious interrupts
PMI: 218863 198034 192376 194477 42378 50328 52237 54105 Performance monitoring interrupts
IWI: 0 0 0 0 0 0 0 0 IRQ work interrupts
RTR: 7 0 0 0 0 0 0 0 APIC ICR read retries
RES: 26000157 9166845 2035661 1303693 788125 845999 904785 973690 Rescheduling interrupts
CAL: 7014953 7022613 7070873 7139248 670430 726545 757997 763109 Function call interrupts
TLB: 7950685 7524917 7394655 7322978 5088431 5147306 5228957 5433623 TLB shootdowns
TRM: 0 0 0 0 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 0 0 0 0 Machine check exceptions
MCP: 8683 8683 8683 8683 8683 8683 8683 8683 Machine check polls
ERR: 0 |
And thats on 3.4.x kernel with debian's config.
| Ant P. wrote: | | If you want another (slightly crazy) solution you could try booting with "threadirqs=1", which forces all IRQ handlers possible into separate threads. IIRC the realtime kernel project uses this, so it should be safe. |
There is so little info about threadirqs, I wouild like to know about some cons and pros of it. |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Fri Dec 28, 2012 1:24 pm Post subject: |
|
|
that's so weird, it should be automagically spreaded out over irqs.
I suspect bad kernel config, only explanation. And why are u using a Debian .config??
There shouldn't be done any manual intervention to spread out interrupts on multiple cores.
Also why don't you put i915, iwlwifi, eth0, snd_hda_intel on MSI irq? Those are doable, prolly ahci and mmc0 as well. |
|
| Back to top |
|
|
BillWho
Veteran
Joined: 03 Mar 2012
Posts: 1600
Location: US
|
| Posted: Fri Dec 28, 2012 7:40 pm Post subject: |
|
|
SlashBeast,
I've never checked this until I read through this thread. Interestingly enough on a Intel(R) Core(TM) i3-2350M CPU @ 2.30GHz laptop:
| Code: | laptop linux # cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 122 0 0 0 IO-APIC-edge timer
1: 4959 0 0 0 IO-APIC-edge i8042
8: 1 0 0 0 IO-APIC-edge rtc0
9: 0 0 0 0 IO-APIC-fasteoi acpi
12: 214633 0 0 0 IO-APIC-edge i8042
16: 38356 0 0 0 IO-APIC-fasteoi ehci_hcd:usb1
19: 242556 0 0 0 IO-APIC-fasteoi ath9k
23: 82206 0 0 0 IO-APIC-fasteoi ehci_hcd:usb2
41: 78453 0 0 0 PCI-MSI-edge ahci
42: 0 0 0 0 PCI-MSI-edge eth0
43: 552 0 0 0 PCI-MSI-edge snd_hda_intel
44: 88 0 0 0 PCI-MSI-edge i915
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 1073823 938123 908729 929452 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 2 0 0 0 APIC ICR read retries
RES: 259554 39511 30216 21274 Rescheduling interrupts
CAL: 257645 252898 233423 233077 Function call interrupts
TLB: 0 0 0 0 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 35 35 35 35 Machine check polls
ERR: 0
MIS: 0
|
On a AMD Phenom(tm) 9150e Quad-Core Processor desktop:
| Code: | stable ~ # cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 118 0 0 0 IO-APIC-edge timer
1: 0 0 0 2 IO-APIC-edge i8042
7: 1 0 0 0 IO-APIC-edge
8: 0 0 0 1 IO-APIC-edge rtc0
9: 0 0 0 0 IO-APIC-fasteoi acpi
12: 0 0 0 4 IO-APIC-edge i8042
16: 0 0 6 1898 IO-APIC-fasteoi ohci_hcd:usb3, ohci_hcd:usb4, snd_hda_intel
17: 0 0 0 3 IO-APIC-fasteoi ehci_hcd:usb1
18: 12 86 1512 1968847 IO-APIC-fasteoi ohci_hcd:usb5, ohci_hcd:usb6, ohci_hcd:usb7, radeon
19: 2 8 390 62737 IO-APIC-fasteoi ehci_hcd:usb2, snd_hda_intel
22: 0 3 427 48467 IO-APIC-fasteoi ahci
42: 0 2 134 23787 PCI-MSI-edge eth0
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 1578605 2145196 1964563 2957637 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 0 0 0 0 APIC ICR read retries
RES: 2728849 898339 755100 712052 Rescheduling interrupts
CAL: 52242 85795 65919 59867 Function call interrupts
TLB: 0 0 0 0 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 27 27 27 27 Machine check polls
ERR: 13
MIS: 0
|
Both are running version 3.6.11-gentoo kernel with similar config settings except of course for specific hardware differences. I have absolutely no clue 
_________________
Good luck 
Since installing gentoo, my life has become one long emerge  |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Sat Dec 29, 2012 1:07 pm Post subject: |
|
|
all my irqs are evenly spreaded over all the cores and the same number. Let's say the sound card is around 200 accross all the cores. And NOT 2 in one, 400 in another and 502342 in another and 0 in another.
This can't be that hard, double check these sections in menuconfig:
| Code: | General setup -->
Processor type and features -->
|
Boot with systemrescuecd, the latest release and check its behaviour. |
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Sun Dec 30, 2012 12:09 am Post subject: |
|
|
| _______0 wrote: | | I suspect bad kernel config, only explanation. And why are u using a Debian .config?? |
Because its a Debian.
| _______0 wrote: | all my irqs are evenly spreaded over all the cores and the same number. Let's say the sound card is around 200 accross all the cores. And NOT 2 in one, 400 in another and 502342 in another and 0 in another.
This can't be that hard, double check these sections in menuconfig:
| Code: | General setup -->
Processor type and features -->
|
|
Check for? |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Sun Dec 30, 2012 12:46 pm Post subject: |
|
|
| SlashBeast wrote: |
Because its a Debian.
|
I fail to see how's that's a gentoo problem then.
All the /proc/interrupts pasted here look plain wrong.
This is how ALL devices should appear like:
| Code: | | 16: 257 257 258 IO-APIC-fasteoi snd_hda_intel |
|
|
| Back to top |
|
|
BillWho
Veteran
Joined: 03 Mar 2012
Posts: 1600
Location: US
|
| Posted: Sun Dec 30, 2012 9:06 pm Post subject: |
|
|
SlashBeast.
Here's a dual Intel(R) Xeon(R) CPU 5160 @ 3.00GHz
| Code: | ws490-gentoo ~ # cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 163 0 0 0 IO-APIC-edge timer
1: 1 1 0 0 IO-APIC-edge i8042
8: 5 5 3 2 IO-APIC-edge rtc0
9: 0 0 0 0 IO-APIC-fasteoi acpi
12: 1 0 1 2 IO-APIC-edge i8042
14: 509 520 523 515 IO-APIC-edge ata_piix
15: 0 0 0 0 IO-APIC-edge ata_piix
16: 78 69 81 76 IO-APIC-fasteoi nouveau
18: 0 0 0 0 IO-APIC-fasteoi uhci_hcd:usb4
20: 1480 1480 1475 1476 IO-APIC-fasteoi ahci
21: 7954 7958 7949 7960 IO-APIC-fasteoi ehci_hcd:usb1, uhci_hcd:usb2
22: 0 0 0 0 IO-APIC-fasteoi uhci_hcd:usb3
23: 0 0 0 0 IO-APIC-fasteoi uhci_hcd:usb5
65: 37 35 36 35 PCI-MSI-edge snd_hda_intel
66: 1 0 0 2 PCI-MSI-edge eth0
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 15559 16094 16099 15744 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 0 0 0 0 APIC ICR read retries
RES: 3554 2081 2352 1626 Rescheduling interrupts
CAL: 1455 1445 2263 2244 Function call interrupts
TLB: 0 0 0 0 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 1 1 1 1 Machine check polls
ERR: 0
MIS: 0
|
It's a mystery why the I3 and I5 differ so drastically
_________________
Good luck
Since installing gentoo, my life has become one long emerge |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Sun Dec 30, 2012 11:28 pm Post subject: |
|
|
odd. Can you paste the configs of all systems to narrow down the problem?
Mm.. according to some intarwebz comments could be two things:
Turn these two off:
| Code: | CONFIG_HOTPLUG_CPU=y
CONFIG_ACPI_HOTPLUG_CPU=y
|
I doubt this cuz my IRQs are all distributed with these options on.
A comment on this post http://www.alexonlinux.com/smp-affinity-and-proper-interrupt-handling-in-linux mentions,
which seems more correct. Try the later one.
And maybe this as well?
Last edited by _______0 on Mon Dec 31, 2012 12:05 am; edited 1 time in total |
|
| Back to top |
|
|
BillWho
Veteran
Joined: 03 Mar 2012
Posts: 1600
Location: US
|
| Posted: Sun Dec 30, 2012 11:52 pm Post subject: |
|
|
| _______0 wrote: | | odd. Can you paste the configs of all systems to narrow down the problem? |
_______0,
For all it's worth:
Core(TM) i3-2350M CPU @ 2.30GHz laptop http://bpaste.net/show/67350/
AMD Phenom(tm) 9150e Quad-Core Processor desktop http://bpaste.net/show/67351/
Dual Xeon(R) CPU 5160 @ 3.00GHz http://bpaste.net/show/67352/
The xeon machine is old and neglected. I refreshed it with a clean install several months ago.
It was just intended to be used to backup files from the AMD and laptop, but over time, it morphed into a crippled desktop - maybe neglect is the secret 
_________________
Good luck
Since installing gentoo, my life has become one long emerge |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Mon Dec 31, 2012 12:32 am Post subject: |
|
|
I think I found it:
try that one along with:
and possibly try this in kernel boot params:
| Quote: | x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
default x2apic cluster mode on platforms
supporting x2apic.
|
|
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Mon Dec 31, 2012 12:36 am Post subject: |
|
|
try to set this one on:
| Code: | | # CONFIG_IRQ_REMAP is not set |
|
|
| Back to top |
|
|
SlashBeast
Retired Dev
Joined: 23 May 2006
Posts: 2922
|
| Posted: Mon Dec 31, 2012 12:50 am Post subject: |
|
|
| _______0 wrote: | | SlashBeast wrote: |
Because its a Debian.
|
I fail to see how's that's a gentoo problem then.
|
.
Uh huh... I said I saw it too on i7 sandy bridge debian. I do run into this issue regardless of distro on newer intel chips.
| _______0 wrote: | This is how ALL devices should appear like:
| Code: | | 16: 257 257 258 IO-APIC-fasteoi snd_hda_intel |
|
You don't say!
And of course every and each of the kernels here have SMP support enabled, without it you can't even see more than CPU0 ... |
|
| Back to top |
|
|
BillWho
Veteran
Joined: 03 Mar 2012
Posts: 1600
Location: US
|
| Posted: Mon Dec 31, 2012 1:24 am Post subject: |
|
|
_______0,
I had high hopes, but:
| Code: | laptop linux # diff .config .config.save
314d313
< CONFIG_X86_X2APIC=y
2442c2441
< CONFIG_IRQ_REMAP=y
---
> # CONFIG_IRQ_REMAP is not set
|
Config changes:
| Code: | Device drivers
IOMMU Hardware Support
Support for Interrupt Remapping (EXPERIMENTAL)
Processor type and features
Support x2apic |
And the results:
| Code: | bill@laptop ~ $ cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 121 0 0 0 IO-APIC-edge timer
1: 82 0 0 0 IO-APIC-edge i8042
8: 1 0 0 0 IO-APIC-edge rtc0
9: 0 0 0 0 IO-APIC-fasteoi acpi
12: 1363 0 0 0 IO-APIC-edge i8042
16: 540 0 0 0 IO-APIC-fasteoi ehci_hcd:usb1
19: 827 0 0 0 IO-APIC-fasteoi ath9k
23: 389 0 0 0 IO-APIC-fasteoi ehci_hcd:usb2
41: 6889 0 0 0 PCI-MSI-edge ahci
42: 0 0 0 0 PCI-MSI-edge eth0
43: 551 0 0 0 PCI-MSI-edge snd_hda_intel
44: 13 0 0 0 PCI-MSI-edge i915
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 19254 13568 11128 11235 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 2 0 0 0 APIC ICR read retries
RES: 954 527 607 284 Rescheduling interrupts
CAL: 906 891 864 787 Function call interrupts
TLB: 0 0 0 0 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 2 2 2 2 Machine check polls
ERR: 0
MIS: 0
|
Well, I'm still  Good try though
_________________
Good luck
Since installing gentoo, my life has become one long emerge |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Mon Dec 31, 2012 2:02 am Post subject: |
|
|
woah! This is getting more interesting by the minute!!
Although the ultimate option might be change to amd proc for proper /proc/interrupt sake, until then we should find out WHY late intels have this odd behaviour.
lol@this:
| Code: | | IWI: 0 0 0 0 IRQ work interrupts |
On mine there's activity, and evenly spread out (as proper) over cores.
I wonder whether has to do with timer used:
| Code: | | cat /sys/bus/clocksource/devices/clocksource0/current_clocksource |
Could you try booting with these kernel parameters??
| Code: | | hpet=force x2apic_phys |
ps: how did you put ahci on an MSI interrupt??
ps2: what about some un-obvious BIOS/UEFI settings?
This gentoo fella https://forums.gentoo.org/viewtopic-t-937042-start-0.html has it right on his i5, according to his uname but his sig says i7:
| eccerr0r wrote: | Is it like this on fresh boot? Single user mode?
Do you have any special hardware? Can the drivers be removed one at a time to see if any of the drivers require frequent interrupts by the clock tick timer?
Do you have HPET enabled on the newer machines? APIC?
Hmm... thought I had a 3.4.9 machine, and doesn't seem to be exhibiting interrupt storm in hw INT0... But I get a lot of local interrupts...
| Code: | mikuru $ uname -a
Linux mikuru 3.4.9-gentoo #2 SMP Sun Aug 26 09:37:03 MDT 2012 x86_64 Intel(R) Core(TM) i5-3317U CPU @ 1.70GHz GenuineIntel GNU/Linux
mikuru $ cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 127 0 0 0 IO-APIC-edge timer
1: 292 9536 146 311 IO-APIC-edge i8042
8: 2 11 0 0 IO-APIC-edge rtc0
9: 250 205 5 85 IO-APIC-fasteoi acpi
12: 7240 260815 3375 7179 IO-APIC-edge i8042
16: 84 132 30 38 IO-APIC-fasteoi ehci_hcd:usb1
23: 0 23 5 4 IO-APIC-fasteoi ehci_hcd:usb2
40: 5886 9660 3690 2082 PCI-MSI-edge i915
41: 19673 16300 4217 5245 PCI-MSI-edge ahci
42: 12821 30427 9406 4772 PCI-MSI-edge xhci_hcd
43: 6457 24409 3702 3569 PCI-MSI-edge snd_hda_intel
44: 0 0 0 0 PCI-MSI-edge eth0
45: 70034 267897 31547 40622 PCI-MSI-edge iwlwifi
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 688301 412667 713302 374754 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 3 0 0 0 APIC ICR read retries
RES: 389680 266397 32705 15704 Rescheduling interrupts
CAL: 67 82 73 77 Function call interrupts
TLB: 11227 13067 8305 12849 TLB shootdowns
TRM: 8 8 8 8 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 19 19 19 19 Machine check polls
ERR: 0
MIS: 0
|
|
|
|
| Back to top |
|
|
BillWho
Veteran
Joined: 03 Mar 2012
Posts: 1600
Location: US
|
| Posted: Mon Dec 31, 2012 2:35 am Post subject: |
|
|
_______0,
| Code: | bill@laptop /proc $ cat cmdline
BOOT_IMAGE=/boot/kernel root=/dev/sda8 rootfstype=ext3 hpet=force x2apic_phys ro
bill@laptop ~ $ cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 122 0 0 0 IO-APIC-edge timer
1: 54 0 0 0 IO-APIC-edge i8042
8: 1 0 0 0 IO-APIC-edge rtc0
9: 0 0 0 0 IO-APIC-fasteoi acpi
12: 487 0 0 0 IO-APIC-edge i8042
16: 436 0 0 0 IO-APIC-fasteoi ehci_hcd:usb1
19: 1643 0 0 0 IO-APIC-fasteoi ath9k
23: 270 0 0 0 IO-APIC-fasteoi ehci_hcd:usb2
41: 6538 0 0 0 PCI-MSI-edge ahci
42: 0 0 0 0 PCI-MSI-edge eth0
43: 12 0 0 0 PCI-MSI-edge i915
44: 550 0 0 0 PCI-MSI-edge snd_hda_intel
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 17806 10814 11136 10214 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 IRQ work interrupts
RTR: 2 0 0 0 APIC ICR read retries
RES: 1159 518 432 222 Rescheduling interrupts
CAL: 572 554 567 492 Function call interrupts
TLB: 0 0 0 0 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
MCE: 0 0 0 0 Machine check exceptions
MCP: 2 2 2 2 Machine check polls
ERR: 0
MIS: 0
|
No cigar here
UPDATE: Almost forgot
| Code: | bill@laptop /proc $ cat /sys/bus/clocksource/devices/clocksource0/current_clocksource
tsc
|
_________________
Good luck
Since installing gentoo, my life has become one long emerge |
|
| Back to top |
|
|
_______0
Guru
Joined: 15 Oct 2012
Posts: 521
|
| Posted: Mon Dec 31, 2012 3:08 am Post subject: |
|
|
| BillWho wrote: |
UPDATE: Almost forgot
| Code: | bill@laptop /proc $ cat /sys/bus/clocksource/devices/clocksource0/current_clocksource
tsc
|
|
oops I forgot this to add in kernel parameters:
In dmesg you should see this:
| Code: | | Switching to clocksource hpet |
Mmm... what does dmesg say about x2apic?? I am curious.
Also i8042 can be disabled in BIOS/UEFI (serial/parallel port) and in kernel config uncheck keyboard and ps/m mouse, no modern machine use none of those. Safe to get rid of them. |
|
| Back to top |
|
|
|
Kernel & Hardware
