Community
cancel
Showing results for 
Search instead for 
Did you mean: 
shonk
Beginner
2,567 Views

DQ77MK PCIE BIOS BUG

Jump to solution

So i have two machines with Intel DQ77MK Boards

And have found two bios bugs

On S3 Suspend to ram (UEFI OS) the PCH PCIE ports on wakeup only initialize at PCIE 1.0 x 4 Speed

and are stuck that way a straight reboot fixes this so the bios manages to initialize fine on boot

and just has issues on wakeup

I350-T2=Locked at PCIE 1.0 x 4 after S3 UEFI

Toshiba XG3=Locked at PCIE 1.0 x 4 after S3 UEFI

PCIE 3.0 x 16 Slot doesnt have this problem

PCIE 2.0 x 1 Slots i have no idea as my Intel AC-7260's are 1.1

Windows Event Log moans about a HAL problem after S3 in UEFI Mode

and points at a faulty bios if the OS isnt in UEFI Mode this doesnt happen after S3 in the eventlog

No idea if the PCH PCIE 1.0 x 4 Bug happens in non UEFI OS as i only noticed this bug the other day

but noticed the Event Log UEFI Mode S3 Problem quite some time ago

this isnt just a case of PCIE power management before anyone mentions it

but the device is locked at PCIE 1.0 after S3 as can be confirmed by NVME SSD benchmarks of

700 MB/S instead of 1400 MB/S after S3

I have 100% confirmed this is a bios bug im not looking for help on my specific devices

BUT Requesting a Firmware Update that fixes this Problem

Machine 1

Board=Intel Desktop Board DQ77MK

CPU=Intel Xeon E3 1240 V2

Ram=32GB 1600 Mhz

PCIE 3.0 16x Slot=MSI R9 280 3GB OC

PCIE 2.0 4x Slot=Intel I350-T2 (PCIE 2.0 x 4 Card)

PCIE 2.0 1x Mini Slot=Intel AC-7260 (PCIE 1.0 x 1 Card)

OS = Windows 10 Fall Creators Update Uefi Mode

Boot Drive =KINGSTON SV300S37A240G 240GB SSD

PSU=Corsair HX750

Machine 2

Board=Intel Desktop Board DQ77MK

CPU=Intel i7 3770

Ram=32GB 1600 Mhz

PCIE 3.0 16x Slot=Sapphire R9 290 4GB Tri-X OC

PCIE 2.0 4x Slot=Toshiba THNSN5256GPU7 XG3 256GB M.2 2280 NVMe (PCIE 3.0 x 4 Card) (non bootable)

PCIE 2.0 1x Mini Slot=Intel AC-7260 (PCIE 1.0 x 1 Card)

OS = Windows 10 Fall Creators Update Uefi Mode

Boot Drive=Crucial M550 512GB SSD

PSU=Corsair HX750

Maximum Link Width: 4xCurrent Link Width: 4xMaximum Link Speed: 8.0 GT/sCurrent Link Speed: 2.5 GT/sDevice/Port Type: PCI Express Endpoint

Maximum Link Width: 4xCurrent Link Width: 4xMaximum Link Speed: 5.0 GT/sCurrent Link Speed: 2.5 GT/sDevice/Port Type: PCI Express Endpoint
0 Kudos
1 Solution
Shonk1
New Contributor I
794 Views

I have just fitted an Intel I350 T2 in the PCIe 2.0 1x Slot that also does it after standby

 

So to reiterate

PCIe3.0 16x = Fine

PCIe 2.0 4x = Bug

PCIe 2.0 1x = Bug

mPCIe 2.0 1x = Fine (with an Intel AC 9260 atleast)

 

Fix.bat

@ECHO OFF SET V=1.0 TITLE DQ77MK STR PCIE 2.5GT/s Fix v%V% cd \Program Files\pciutils @ECHO ON setpci -s 0:1c.4 CAP_EXP+0x30.w setpci -s 0:1c.4 CAP_EXP+0x30.w=0x0002:0x000f setpci -s 0:1c.4 CAP_EXP+0x10.w=0x0020:0x0020 setpci -s 0:1c.4 CAP_EXP+0x30.w setpci -s 0:1c.0 CAP_EXP+0x30.w setpci -s 0:1c.0 CAP_EXP+0x30.w=0x0002:0x000f setpci -s 0:1c.0 CAP_EXP+0x10.w=0x0020:0x0020 setpci -s 0:1c.0 CAP_EXP+0x30.w

Addresses are system-dependent. list devices by typing lspci in an admin command prompt

in my case i want to fix the

4x slot which is 0:1c.0 (PCI Express Root Port 1)

1x slot which is 0:1c.4 (PCI Express Root Port 5)

 

 

 

View solution in original post

11 Replies
shonk
Beginner
794 Views

This was posted in the support forums a few times when support was still active for desktop boards

Intel support came back that it couldnt be reproduced so it must be an error at the users end

but the reality was consumers where using the board in uefi mode

and support where using legacy mode and nobody put two and two together

In the past i would have fixed the bios myself but intel sign the bios

and as such this can only be fixed by intel

large bugs are still being fixed by intel for desktop boards

e.g. ME bug was patched a few months ago

I request a fix for broken UEFI mode as this has been broken since day1 of release

Log Name: SystemSource: Microsoft-Windows-HALDate: 24/02/2190 20:50:32Event ID: 12

Task Category: None

Level: ErrorKeywords: (1)User: N/AComputer: DQ77MK2

Description:

The platform firmware has corrupted memory across the previous system power transition. Please check for updated firmware for your system.

Event Xml:

http://schemas.microsoft.com/win/2004/08/events/event http://schemas.microsoft.com/win/2004/08/events/event">

1202000x80000000000000015047SystemDQ77MK2

1159159

n_scott_pearson
Super User Retired Employee
794 Views

Intel exited from the Desktop Boards business a number of years ago. All Desktop Boards are subsequently long past their End-Of-Life dates and long past their End-Of-Interactive-Support dates as well. The sustaining team was disbanded quite some time ago. There is simply no one left to work on any bugs. Yes, it is totally true that Intel delivered updates last year that delivered fixes for a set of security flaws in the Management Engine (ME) firmware, but what you need to understand is that the ME firmware team is part of the Chipset organization and thus active sustaining for this portion of the firmware is (still) ongoing. This is not the case here. There is no one to work on any BIOS bugs. Putting together a team at this point to deliver a fix for this issue would cost many thousands of dollars. It simply isn't going to happen.

Sorry (and remember that I am just the messenger),

...S

shonk
Beginner
794 Views

Im sure you know alot more about the teams than me

but they clearly still have bios engineers on site as the nuc's also suffered with this problem

which they promptly replaced a bios module and fixed it

being such a large bug that has been present since release

the least they could do is get one of the nuc team to package up a fix

they clearly know what fixes it as they fixed the nuc bios's

i could fix this myself if i had the capability to sign the firmware

in what 10 mins

Its a simple case of saying fred fix that uefi bug that the nuc's had

on the dq77mk firmware ok jim will have it fixed in 10 mins

n_scott_pearson
Super User Retired Employee
794 Views

No, it is not that simple. First of all, these BIOS were developed on completely different code bases, which means that a fix would need to be independently developed. As I said, doing so and then going through the required regression, validation and release cycles (nothing leaves Intel without it) would take a team of engineers and cost many thousands of dollars. I am sorry, but this is the reality of the situation; it simply isn't going to happen.

...S

shonk
Beginner
794 Views

This is the problem with signed firmware nobody can fix it apart from the oem

My only hope tbh is if it doesnt happen in legacy mode

I know the os doesnt moan after coming out of standby in legacy

as mine only started doing it after a fresh install in uefi mode

will wipe one of them in the next few days and check

the event log didnt bother me i just stopped windows logging it

but didnt realise until yesterday that the 4x slot was being downgraded after standby

shonk
Beginner
794 Views

I did a fresh legacy install

it fixed the event log error after standby

but the 4x slot still gets dropped down to gen 1 after standby

so its two different bug's

n_scott_pearson
Super User Retired Employee
794 Views

Thanks for the into. Unfortunately, at this point, all we can do is document the issues in case someone else runs into it.

...S

shonk
Beginner
794 Views

Im looking into forcing a link speed now in the registry with RMPcieLinkSpeed

i just need to find the area for the i350-t2 and the nvme ssd

Here's the reg fix to stop windows moaning about s3

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Control\WMI\Autologger\EventLog-System\{63d1e632-95cc-4443-9312-af927761d52a}]

"Enabled"=dword:00000000

Shonk1
New Contributor I
794 Views

I know this is an old post

but i have recently worked out how to manually fix this after every Suspend to Ram

and think its good to document incase anyone else wants to fix this

 

as i have recently put an nvme drive in this slot i spent a good while looking up how to fix this

as i didnt want the nvme drive degraded down to gen1 4x after the first sleep as this pc is on 24/7

DQ77MK isnt usually bootable via nvme due to a lack of nvme option rom

but i inject the nvme module into the uefi on every boot with rEFInd Boot Manager

i wont go into how thats done here but suffice to say its doable if you know what you are doing

 

I have created a scheduled task that is triggerd on Event ID 107 (e.g. The system has resumed from sleep.)

to run a batch file (C:\Program Files\pciutils\Fix.bat)

 

for this you need a copy of pciutils that has been built for windows (its a linux utility)

I used pciutils-3.5.5-win64 (has to be run with administrator priv's ofc)

the batch file consists of below you can test it works by manually running it as administrator then checking link speed with hwinfo64

 

The first line with setpci shows the current max link rate which will be Gen1 (so this doesnt have to be used its just for info)

The second line with setpci sets the max link rate to Gen2

The third line with setpci triggers a retrain of PCI Express Root Port 1

The forth line with setpci shows the current max link rate which will be Gen2 (so this doesnt have to be used its just for info)

 

Fix.bat

@ECHO OFF SET V=1.0 TITLE DQ77MK STR PCIE 2.5GT/s Fix v%V% cd \Program Files\pciutils @ECHO ON setpci -s 0:1c.0 CAP_EXP+0x30.w setpci -s 0:1c.0 CAP_EXP+0x30.w=0x0002:0x000f setpci -s 0:1c.0 CAP_EXP+0x10.w=0x0020:0x0020 setpci -s 0:1c.0 CAP_EXP+0x30.w

 

Addresses are system-dependent. list devices by typing lspci in an admin command prompt

in my case i want to fix the 4x slot which is 0:1c.1 (PCI Express Root Port 1)

the mini pcie slot doesnt exhibit this problem as my Intel AC 9260 runs just fine in Gen2 after STR)

the 16x slot is also fine it sits on gen1 at idle then goes to gen3 when the gpu is loaded

 

Shonk1
New Contributor I
794 Views

I suspect there may be a much cleaner fix via the registry with a flag forcing gen2

 

e.g. on my device

Intel Panther Point PCH - PCI Express Port 1 HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\PCI\VEN_8086&DEV_1E10&SUBSYS_20358086&REV_C4   Toshiba AIS NVMe SSD Controller HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\PCI\VEN_1179&DEV_010F&SUBSYS_00011179&REV_01

 

but after alot of looking up said flags i couldnt find info on it so will have to make do with the fix above

Shonk1
New Contributor I
795 Views

I have just fitted an Intel I350 T2 in the PCIe 2.0 1x Slot that also does it after standby

 

So to reiterate

PCIe3.0 16x = Fine

PCIe 2.0 4x = Bug

PCIe 2.0 1x = Bug

mPCIe 2.0 1x = Fine (with an Intel AC 9260 atleast)

 

Fix.bat

@ECHO OFF SET V=1.0 TITLE DQ77MK STR PCIE 2.5GT/s Fix v%V% cd \Program Files\pciutils @ECHO ON setpci -s 0:1c.4 CAP_EXP+0x30.w setpci -s 0:1c.4 CAP_EXP+0x30.w=0x0002:0x000f setpci -s 0:1c.4 CAP_EXP+0x10.w=0x0020:0x0020 setpci -s 0:1c.4 CAP_EXP+0x30.w setpci -s 0:1c.0 CAP_EXP+0x30.w setpci -s 0:1c.0 CAP_EXP+0x30.w=0x0002:0x000f setpci -s 0:1c.0 CAP_EXP+0x10.w=0x0020:0x0020 setpci -s 0:1c.0 CAP_EXP+0x30.w

Addresses are system-dependent. list devices by typing lspci in an admin command prompt

in my case i want to fix the

4x slot which is 0:1c.0 (PCI Express Root Port 1)

1x slot which is 0:1c.4 (PCI Express Root Port 5)

 

 

 

View solution in original post

Reply