[Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Anders Blomdell]

Hi

I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)

This is what msi.c says at bootup (after attached patch is applied):

  kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)

and this is what lspci says:

  00:00.0 Host bridge: Intel Corporation 4th Gen Core Processor DRAM Controller (rev 06)
  00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller (rev 06)
  00:03.0 Audio device: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor HD Audio Controller (rev 06)
  00:14.0 USB controller: Intel Corporation 8 Series/C220 Series Chipset Family USB xHCI (rev 05)
  00:16.0 Communication controller: Intel Corporation 8 Series/C220 Series Chipset Family MEI Controller #1 (rev 04)
  00:1a.0 USB controller: Intel Corporation 8 Series/C220 Series Chipset Family USB EHCI #2 (rev 05)
  00:1b.0 Audio device: Intel Corporation 8 Series/C220 Series Chipset High Definition Audio Controller (rev 05)
  00:1c.0 PCI bridge: Intel Corporation 8 Series/C220 Series Chipset Family PCI Express Root Port #1 (rev d5)
  00:1c.1 PCI bridge: Intel Corporation 8 Series/C220 Series Chipset Family PCI Express Root Port #2 (rev d5)
  00:1c.3 PCI bridge: Intel Corporation 82801 PCI Bridge (rev d5)
  00:1c.4 PCI bridge: Intel Corporation 8 Series/C220 Series Chipset Family PCI Express Root Port #5 (rev d5)
  00:1d.0 USB controller: Intel Corporation 8 Series/C220 Series Chipset Family USB EHCI #1 (rev 05)
  00:1f.0 ISA bridge: Intel Corporation H87 Express LPC Controller (rev 05)
  00:1f.2 SATA controller: Intel Corporation 8 Series/C220 Series Chipset Family 6-port SATA Controller 1 [AHCI mode] (rev 05)
  00:1f.3 SMBus: Intel Corporation 8 Series/C220 Series Chipset Family SMBus Controller (rev 05)
  02:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 06)
  03:00.0 PCI bridge: ASMedia Technology Inc. ASM1083/1085 PCIe to PCI Bridge (rev 04)
  05:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)
  05:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)

Anybody who has a good idea how to fix this?

Best regards

Anders Blomdell



-- 
Anders Blomdell                  Email: anders.blomdell@control.lth.se
Department of Automatic Control
Lund University                  Phone:    +46 46 222 4625
P.O. Box 118                     Fax:      +46 46 138118
SE-221 00 Lund, Sweden

-------------- next part --------------
A non-text attachment was scrubbed...
Name: xenomai-3.0-rc7-smi.patch
Type: text/x-patch
Size: 597 bytes
Desc: not available
URL: <http://xenomai.org/pipermail/xenomai/attachments/20151006/06ea1114/attachment.bin>

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Gilles Chanteperdrix]

On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
> Hi
> 
> I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)
> 
> This is what msi.c says at bootup (after attached patch is applied):
> 
>   kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
> 
> and this is what lspci says:
> 
> 
> Anybody who has a good idea how to fix this?

You should first check that the way smi.c retrieves the address of
the SMI_EN register is still valid for your chipset. If it is, then
try fiddling with the BIOS settings to see if some setting can
disable generation of SMIs. If unsuccessful, you can try contacting
the motherboard support to see if they have any advice.

-- 
					    Gilles.
https://click-hack.org

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Johann Obermayr]

Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
>> Hi
>>
>> I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)
>>
>> This is what msi.c says at bootup (after attached patch is applied):
>>
>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
>>
>> and this is what lspci says:
>>
>>
>> Anybody who has a good idea how to fix this?
> You should first check that the way smi.c retrieves the address of
> the SMI_EN register is still valid for your chipset. If it is, then
> try fiddling with the BIOS settings to see if some setting can
> disable generation of SMIs. If unsuccessful, you can try contacting
> the motherboard support to see if they have any advice.
>
On new Mainboards, you can't disable SMI.
Because this is a security problem, so SMI is locked by BIOS.
There are some registery on the mainboard chipset, hat lock the SMI 
register.
So you can't disable the SMI.
We also had this trouble, and we get a own BIOS with disabled SMI.

Regards
   Johann

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Gilles Chanteperdrix]

On Wed, Oct 07, 2015 at 09:25:23AM +0200, Johann Obermayr wrote:
> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
> > On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
> >> Hi
> >>
> >> I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)
> >>
> >> This is what msi.c says at bootup (after attached patch is applied):
> >>
> >>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
> >>
> >> and this is what lspci says:
> >>
> >>
> >> Anybody who has a good idea how to fix this?
> > You should first check that the way smi.c retrieves the address of
> > the SMI_EN register is still valid for your chipset. If it is, then
> > try fiddling with the BIOS settings to see if some setting can
> > disable generation of SMIs. If unsuccessful, you can try contacting
> > the motherboard support to see if they have any advice.
> >
> On new Mainboards, you can't disable SMI.
> Because this is a security problem, so SMI is locked by BIOS.

Yes, there is a bit in a register to lock the SMI_EN register, and
when the SMI_EN register is locked, there is no way to unlock it.

However, the BIOS may offer you some setting to disable the
problematic SMI source. This happened to me recently. It was not
possible to modify the SMI_EN register, but with the right BIOS
setting, no SMI was generated.

-- 
					    Gilles.
https://click-hack.org

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Johann Obermayr]

Am 07.10.2015 um 09:25 schrieb Johann Obermayr:
> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
>> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
>>> Hi
>>>
>>> I have problems disabling SMI-interrupts on "MSI H87-G43 
>>> motherboard" (http://www.msi.com/product/mb/H87-G43.html)
>>>
>>> This is what msi.c says at bootup (after attached patch is applied):
>>>
>>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
>>>
>>> and this is what lspci says:
>>>
>>>
>>> Anybody who has a good idea how to fix this?
>> You should first check that the way smi.c retrieves the address of
>> the SMI_EN register is still valid for your chipset. If it is, then
>> try fiddling with the BIOS settings to see if some setting can
>> disable generation of SMIs. If unsuccessful, you can try contacting
>> the motherboard support to see if they have any advice.
>>
> On new Mainboards, you can't disable SMI.
> Because this is a security problem, so SMI is locked by BIOS.
> There are some registery on the mainboard chipset, hat lock the SMI 
> register.
> So you can't disable the SMI.
> We also had this trouble, and we get a own BIOS with disabled SMI.
>
> Regards
>   Johann
>
> _______________________________________________
> Xenomai mailing list
> Xenomai@xenomai.org
> http://xenomai.org/mailman/listinfo/xenomai
>
Found this:

GEN_PMCON_1—General PM Configuration 1 Register (PM—D31:F0)
Offset Address: A0h Attribute: R/W, RO, R/WO
Default Value: 0000h Size: 16-bit
Lockable: No Usage: ACPI, Legacy
Power Well: Core

Bit 4: SMI_LOCK — R/WO. When this bit is set, writes to the GLB_SMI_EN 
bit (PMBASE
+ 30h, bit 0) will have no effect. Once the SMI_LOCK bit is set, writes 
of 0 to
SMI_LOCK bit will have no effect (that is, once set, this bit can only 
be cleared by
PLTRST#).

Regards
   Johann

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Anders Blomdell]

On 2015-10-07 10:04, Johann Obermayr wrote:
> Am 07.10.2015 um 09:25 schrieb Johann Obermayr:
>> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
>>> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
>>>> Hi
>>>>
>>>> I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)
>>>>
>>>> This is what msi.c says at bootup (after attached patch is applied):
>>>>
>>>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
>>>>
>>>> and this is what lspci says:
>>>>
>>>>
>>>> Anybody who has a good idea how to fix this?
>>> You should first check that the way smi.c retrieves the address of
>>> the SMI_EN register is still valid for your chipset. If it is, then
>>> try fiddling with the BIOS settings to see if some setting can
>>> disable generation of SMIs. If unsuccessful, you can try contacting
>>> the motherboard support to see if they have any advice.
>>>
>> On new Mainboards, you can't disable SMI.
>> Because this is a security problem, so SMI is locked by BIOS.
>> There are some registery on the mainboard chipset, hat lock the SMI register.
>> So you can't disable the SMI.
>> We also had this trouble, and we get a own BIOS with disabled SMI.
>>
>> Regards
>>   Johann
>>
>> _______________________________________________
>> Xenomai mailing list
>> Xenomai@xenomai.org
>> http://xenomai.org/mailman/listinfo/xenomai
>>
> Found this:
> 
> GEN_PMCON_1—General PM Configuration 1 Register (PM—D31:F0)
> Offset Address: A0h Attribute: R/W, RO, R/WO
> Default Value: 0000h Size: 16-bit
> Lockable: No Usage: ACPI, Legacy
> Power Well: Core
> 
> Bit 4: SMI_LOCK — R/WO. When this bit is set, writes to the GLB_SMI_EN bit (PMBASE
> + 30h, bit 0) will have no effect. Once the SMI_LOCK bit is set, writes of 0 to
> SMI_LOCK bit will have no effect (that is, once set, this bit can only be cleared by
> PLTRST#).
> 

Thanks everybody, seems like I have to give up on that motherboard:

1. Nothing found to disable in the BIOS
2. Disabling all sourcecs with smictrl (git://git.kiszka.org/smictrl.git)
   still gives high latencies with RTNet (+1.5 milliseconds). 
3. latency test does not show anything suspicious, so it might
   be some other issue.

Anybody that has a suggestion on a recent motherboard that has been used with
RTNet and showing good results?

Thanks everyone!

/Anders


-- 
Anders Blomdell                  Email: anders.blomdell@control.lth.se
Department of Automatic Control
Lund University                  Phone:    +46 46 222 4625
P.O. Box 118                     Fax:      +46 46 138118
SE-221 00 Lund, Sweden

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Gilles Chanteperdrix]

On Wed, Oct 07, 2015 at 02:31:06PM +0200, Anders Blomdell wrote:
> On 2015-10-07 10:04, Johann Obermayr wrote:
> > Am 07.10.2015 um 09:25 schrieb Johann Obermayr:
> >> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
> >>> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
> >>>> Hi
> >>>>
> >>>> I have problems disabling SMI-interrupts on "MSI H87-G43 motherboard" (http://www.msi.com/product/mb/H87-G43.html)
> >>>>
> >>>> This is what msi.c says at bootup (after attached patch is applied):
> >>>>
> >>>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
> >>>>
> >>>> and this is what lspci says:
> >>>>
> >>>>
> >>>> Anybody who has a good idea how to fix this?
> >>> You should first check that the way smi.c retrieves the address of
> >>> the SMI_EN register is still valid for your chipset. If it is, then
> >>> try fiddling with the BIOS settings to see if some setting can
> >>> disable generation of SMIs. If unsuccessful, you can try contacting
> >>> the motherboard support to see if they have any advice.
> >>>
> >> On new Mainboards, you can't disable SMI.
> >> Because this is a security problem, so SMI is locked by BIOS.
> >> There are some registery on the mainboard chipset, hat lock the SMI register.
> >> So you can't disable the SMI.
> >> We also had this trouble, and we get a own BIOS with disabled SMI.
> >>
> >> Regards
> >>   Johann
> >>
> >> _______________________________________________
> >> Xenomai mailing list
> >> Xenomai@xenomai.org
> >> http://xenomai.org/mailman/listinfo/xenomai
> >>
> > Found this:
> > 
> > GEN_PMCON_1—General PM Configuration 1 Register (PM—D31:F0)
> > Offset Address: A0h Attribute: R/W, RO, R/WO
> > Default Value: 0000h Size: 16-bit
> > Lockable: No Usage: ACPI, Legacy
> > Power Well: Core
> > 
> > Bit 4: SMI_LOCK — R/WO. When this bit is set, writes to the GLB_SMI_EN bit (PMBASE
> > + 30h, bit 0) will have no effect. Once the SMI_LOCK bit is set, writes of 0 to
> > SMI_LOCK bit will have no effect (that is, once set, this bit can only be cleared by
> > PLTRST#).
> > 
> 
> Thanks everybody, seems like I have to give up on that motherboard:
> 
> 1. Nothing found to disable in the BIOS
> 2. Disabling all sourcecs with smictrl (git://git.kiszka.org/smictrl.git)
>    still gives high latencies with RTNet (+1.5 milliseconds). 
> 3. latency test does not show anything suspicious, so it might
>    be some other issue.
> 
> Anybody that has a suggestion on a recent motherboard that has been used with
> RTNet and showing good results?

If the issue is with RTNet, does not it make more sense to find the
RTNet issue than to change motherboard? If you are able to detect a
high latency, can you not trigger a trace from the I-pipe tracer?

-- 
					    Gilles.
https://click-hack.org

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Anders Blomdell]

On October 7, 2015 2:41:07 PM GMT+02:00, Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org> wrote:
>On Wed, Oct 07, 2015 at 02:31:06PM +0200, Anders Blomdell wrote:
>> On 2015-10-07 10:04, Johann Obermayr wrote:
>> > Am 07.10.2015 um 09:25 schrieb Johann Obermayr:
>> >> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
>> >>> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
>> >>>> Hi
>> >>>>
>> >>>> I have problems disabling SMI-interrupts on "MSI H87-G43
>motherboard" (http://www.msi.com/product/mb/H87-G43.html)
>> >>>>
>> >>>> This is what msi.c says at bootup (after attached patch is
>applied):
>> >>>>
>> >>>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
>> >>>>
>> >>>> and this is what lspci says:
>> >>>>
>> >>>>
>> >>>> Anybody who has a good idea how to fix this?
>> >>> You should first check that the way smi.c retrieves the address
>of
>> >>> the SMI_EN register is still valid for your chipset. If it is,
>then
>> >>> try fiddling with the BIOS settings to see if some setting can
>> >>> disable generation of SMIs. If unsuccessful, you can try
>contacting
>> >>> the motherboard support to see if they have any advice.
>> >>>
>> >> On new Mainboards, you can't disable SMI.
>> >> Because this is a security problem, so SMI is locked by BIOS.
>> >> There are some registery on the mainboard chipset, hat lock the
>SMI register.
>> >> So you can't disable the SMI.
>> >> We also had this trouble, and we get a own BIOS with disabled SMI.
>> >>
>> >> Regards
>> >>   Johann
>> >>
>> >> _______________________________________________
>> >> Xenomai mailing list
>> >> Xenomai@xenomai.org
>> >> http://xenomai.org/mailman/listinfo/xenomai
>> >>
>> > Found this:
>> > 
>> > GEN_PMCON_1—General PM Configuration 1 Register (PM—D31:F0)
>> > Offset Address: A0h Attribute: R/W, RO, R/WO
>> > Default Value: 0000h Size: 16-bit
>> > Lockable: No Usage: ACPI, Legacy
>> > Power Well: Core
>> > 
>> > Bit 4: SMI_LOCK — R/WO. When this bit is set, writes to the
>GLB_SMI_EN bit (PMBASE
>> > + 30h, bit 0) will have no effect. Once the SMI_LOCK bit is set,
>writes of 0 to
>> > SMI_LOCK bit will have no effect (that is, once set, this bit can
>only be cleared by
>> > PLTRST#).
>> > 
>> 
>> Thanks everybody, seems like I have to give up on that motherboard:
>> 
>> 1. Nothing found to disable in the BIOS
>> 2. Disabling all sourcecs with smictrl
>(git://git.kiszka.org/smictrl.git)
>>    still gives high latencies with RTNet (+1.5 milliseconds). 
>> 3. latency test does not show anything suspicious, so it might
>>    be some other issue.
>> 
>> Anybody that has a suggestion on a recent motherboard that has been
>used with
>> RTNet and showing good results?
>
>If the issue is with RTNet, does not it make more sense to find the
>RTNet issue than to change motherboard? If you are able to detect a
>high latency, can you not trigger a trace from the I-pipe tracer?

Well, same kernel identical ethernetcard on motherboard without SMI, works flawlessly (network jitter in 30 ms range).

BTW: only the RTNet drivers, raw ethernet.

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Gilles Chanteperdrix]

On Wed, Oct 07, 2015 at 03:03:25PM +0200, Anders Blomdell wrote:
> 
> 
> On October 7, 2015 2:41:07 PM GMT+02:00, Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org> wrote:
> >On Wed, Oct 07, 2015 at 02:31:06PM +0200, Anders Blomdell wrote:
> >> On 2015-10-07 10:04, Johann Obermayr wrote:
> >> > Am 07.10.2015 um 09:25 schrieb Johann Obermayr:
> >> >> Am 06.10.2015 um 21:23 schrieb Gilles Chanteperdrix:
> >> >>> On Tue, Oct 06, 2015 at 12:03:55PM +0200, Anders Blomdell wrote:
> >> >>>> Hi
> >> >>>>
> >> >>>> I have problems disabling SMI-interrupts on "MSI H87-G43
> >motherboard" (http://www.msi.com/product/mb/H87-G43.html)
> >> >>>>
> >> >>>> This is what msi.c says at bootup (after attached patch is
> >applied):
> >> >>>>
> >> >>>>    kernel: [Xenomai] SMI workaround failed! (1830: 1 & ~1 -> 1)
> >> >>>>
> >> >>>> and this is what lspci says:
> >> >>>>
> >> >>>>
> >> >>>> Anybody who has a good idea how to fix this?
> >> >>> You should first check that the way smi.c retrieves the address
> >of
> >> >>> the SMI_EN register is still valid for your chipset. If it is,
> >then
> >> >>> try fiddling with the BIOS settings to see if some setting can
> >> >>> disable generation of SMIs. If unsuccessful, you can try
> >contacting
> >> >>> the motherboard support to see if they have any advice.
> >> >>>
> >> >> On new Mainboards, you can't disable SMI.
> >> >> Because this is a security problem, so SMI is locked by BIOS.
> >> >> There are some registery on the mainboard chipset, hat lock the
> >SMI register.
> >> >> So you can't disable the SMI.
> >> >> We also had this trouble, and we get a own BIOS with disabled SMI.
> >> >>
> >> >> Regards
> >> >>   Johann
> >> >>
> >> >> _______________________________________________
> >> >> Xenomai mailing list
> >> >> Xenomai@xenomai.org
> >> >> http://xenomai.org/mailman/listinfo/xenomai
> >> >>
> >> > Found this:
> >> > 
> >> > GEN_PMCON_1—General PM Configuration 1 Register (PM—D31:F0)
> >> > Offset Address: A0h Attribute: R/W, RO, R/WO
> >> > Default Value: 0000h Size: 16-bit
> >> > Lockable: No Usage: ACPI, Legacy
> >> > Power Well: Core
> >> > 
> >> > Bit 4: SMI_LOCK — R/WO. When this bit is set, writes to the
> >GLB_SMI_EN bit (PMBASE
> >> > + 30h, bit 0) will have no effect. Once the SMI_LOCK bit is set,
> >writes of 0 to
> >> > SMI_LOCK bit will have no effect (that is, once set, this bit can
> >only be cleared by
> >> > PLTRST#).
> >> > 
> >> 
> >> Thanks everybody, seems like I have to give up on that motherboard:
> >> 
> >> 1. Nothing found to disable in the BIOS
> >> 2. Disabling all sourcecs with smictrl
> >(git://git.kiszka.org/smictrl.git)
> >>    still gives high latencies with RTNet (+1.5 milliseconds). 
> >> 3. latency test does not show anything suspicious, so it might
> >>    be some other issue.
> >> 
> >> Anybody that has a suggestion on a recent motherboard that has been
> >used with
> >> RTNet and showing good results?
> >
> >If the issue is with RTNet, does not it make more sense to find the
> >RTNet issue than to change motherboard? If you are able to detect a
> >high latency, can you not trigger a trace from the I-pipe tracer?
> 
> Well, same kernel identical ethernetcard on motherboard without SMI, works flawlessly (network jitter in 30 ms range).
> 
> BTW: only the RTNet drivers, raw ethernet.

Using the I-pipe tracer and user-space msr-tools, you should be able
to confirm whether the issue is due to SMIs.

-- 
					    Gilles.
https://click-hack.org

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Jeroen Van den Keybus]

>> >> Thanks everybody, seems like I have to give up on that motherboard:
>> >>
>> >> 1. Nothing found to disable in the BIOS
>> >> 2. Disabling all sourcecs with smictrl
>> >(git://git.kiszka.org/smictrl.git)
>> >>    still gives high latencies with RTNet (+1.5 milliseconds).
>> >> 3. latency test does not show anything suspicious, so it might
>> >>    be some other issue.
>> >>
>> >> Anybody that has a suggestion on a recent motherboard that has been
>> >used with
>> >> RTNet and showing good results?

As you found out, most MB BIOSes set the SMI lock bit. No way around it.

How frequently do the high latencies occur => we had a MB that issued
an SMI to speed up the fans when the CPU temperature was rising. We
set the fans to max in BIOS to avoid this, successfully. (We think
that the SMI is issued as a last resort when the kernel is not
handling ACPI thermal events. So disabling every CPU ACPI feature in
an effort to avoid latencies may actually cause SMIs.)

Do you use the onboard LAN => some Intel (i217) uses a low-performance
version of PCIe, called Kumeran, see
http://sourceforge.net/p/e1000/mailman/message/32283346. In this post,
I also tested a patch that seemed to help a bit. Maybe it also affects
the 1.5ms latency ? Do you have the same issue with an external PCI(e)
NIC ? Ans in that case, also test different PCIe slots; some are
behind bridges, others not.

>> Well, same kernel identical ethernetcard on motherboard without SMI, works flawlessly (network jitter in 30 ms range).

30ms ? That's also quite significant, I'd say.

Which MB are you using that does not have SMI (or at least lets you
disable it completely) ?


Jeroen.

Re: [Xenomai] Problems disabling SMI interrupts on MSI H87-G43 motherboard

[Gilles Chanteperdrix]

On Thu, Oct 08, 2015 at 11:08:46AM +0200, Jeroen Van den Keybus wrote:
> >> >> Thanks everybody, seems like I have to give up on that motherboard:
> >> >>
> >> >> 1. Nothing found to disable in the BIOS
> >> >> 2. Disabling all sourcecs with smictrl
> >> >(git://git.kiszka.org/smictrl.git)
> >> >>    still gives high latencies with RTNet (+1.5 milliseconds).
> >> >> 3. latency test does not show anything suspicious, so it might
> >> >>    be some other issue.
> >> >>
> >> >> Anybody that has a suggestion on a recent motherboard that has been
> >> >used with
> >> >> RTNet and showing good results?
> 
> As you found out, most MB BIOSes set the SMI lock bit. No way around it.

According to the documentation the SMI lock bit should only lock the
SMI_EN register bit 0, that is the "global disable" bit. However, on
the last machines where I had SMI issues, the SMI_EN register was
entirely read-only, so there must be other ways than the SMI lock
bit to lock the SMI_EN register, which is why smi.c write then reads
the register, instead of checking the SMI lock bit.

-- 
					    Gilles.
https://click-hack.org