Really bizarre LAN chip/port issue

[Aquinus]

It really has nothing to do with operating systems today. That setting is all about what happens before the boot drive is even touched so the hardware (including legacy hardware) can communicate at a really basic level. Once the OS boots - whether Linux or Windows, their own protocols take over. At least that's my understanding.

It does not help that there are different meanings for PnP.

How PnP Works (simplified) - though I think simplified is a relative term here.

You're confusing plug and play devices with a plug and play OS. Turning on Plug and Play OS makes it so the BIOS only initializes the core components, letting the OS configure (and manage,) everything else. To my knoledge, this should *always* be set to off so all components are initialized by the BIOS and then handed over to the OS as opposed to letting the OS initialize them in the first place. This page seems to describe it pretty well:

For x86-based systems, the way the BIOS interacts with PnP devices can vary, depending on whether the BIOS or the OS is responsible for configuring the hardware.

• For a fully compliant Advanced Configuration and Power Interface (ACPI) OS, such as Windows XP, the OS assigns device resources and ignores BIOS settings. This includes reassigning IRQ, DMA, and IO resources as well as arbitrating conflicts for all PCI devices. Because the PnP BIOS setting is ignored and ACPI is used instead, the BIOS setting can be left at either Yes/Enabled or No/Disabled, although Microsoft recommends that you set the option to No/Disabled.
• For a non-compliant ACPI OS, the BIOS assigns device resources prior to the loading of the operating systems, and the PnP BIOS setting must be No/Disabled.
• For a non-ACPI OS, the BIOS assigns device resources prior to the loading of the operating system, and the PnP BIOS setting must be No/Disabled.
• For a multi-boot system, this is especially important. In this case, the PnP BIOS settings must be No/Disabled. Otherwise, the system check for PnP may pass on one OS but fail on another OS.

In summary, the PnP BIOS setting should be set to No/Disabled. Because this setting only controls what system component is responsible for resource allocation, it will have little impact on the system's behavior and performance.

http://help.cognex.com/Default.htm#KB_Topics/VisionPro/Drivers/2905.htm

So my comment to the "OS/2 thing" was the fact that OS/2 was a Plug and Play OS that initializes devices from the get go instead of relying on the BIOS to do it.

 

[Bill_Bright]

You're confusing plug and play devices with a plug and play OS.

Actually, that was my point about PnP having different meanings.

To my knoledge, this should *always* be set to off so all components are initialized by the BIOS and then handed over to the OS as opposed to letting the OS initialize them in the first place.

Which is why I also said I never change that setting - I leave it at the default, which is No or Off (or Disabled) - depending on the BIOS maker and how it is worded.

as opposed to letting the OS initialize them in the first place.

That's the problem because the BIOS must do this on a hardware level so the various components can communicate BEFORE the OS is booted. Otherwise, you would see nothing on the monitor, you would have no keyboard or mouse input to enter or make changes in the BIOS Setup Menu - you could not boot to an external drive.

The bottom line here is, we are in agreement and the setting should remain at the default setting, which is No, Off or Disabled.

 

[RejZoR]

It says it's only for non boot related devices. You don't need LAN (unless boot ROM is enabled) or soundcard for booting of system...

 

[tabascosauz]

@RejZoR do you notice anything going on with the LEDs in the LAN port? IIRC if everything is working as it is supposed to then the Gigabit LED is supposed to light up. Is 8111C supposed to be similar to / have the same driver as the more common 8111G?

 

[RejZoR]

You mean when I don't see the LAN device in Windows, does the LAN LED glow then? Thx for the idea, didn't realyl consider checking it up. Though LED's seem to work as soon as you plug in the cable, even if PC is turned off. Hm, I'll observe that more.

 

[tabascosauz]

You mean when I don't see the LAN device in Windows, does the LAN LED glow then? Thx for the idea, didn't realyl consider checking it up. Though LED's seem to work as soon as you plug in the cable, even if PC is turned off. Hm, I'll observe that more.

Yeah, the 2 tiny LEDs on either side of the port. Since the only difference is a letter, I'm going to try and pull up the manual for my H81I, which uses a 8111G to figure out which LED is which, whether it should be flashing/continuous, etc.

Here is what my manual says for the 8111G:

I guess blinking/solid yellow and orange is what we're after.​

 

[remixedcat]

You might also try reinstalling the LAN drivers. I know the Nvidia drivers SHOULDN'T have made a difference or screwed with it, but it's always possible.

Advice for when you do upgrade, that I always do, since it could just be coincidence that the LAN port went bad at that exact time, is next motherboard get one with two LAN ports. This has saved me before when one went bad.

Nforce LAN devices fail quickly and in my hubby's system it fails and doesn't even work after countless reinstalls. They are crappy ETH PHYs and nvidia should feel bad. Best getting Intel ones. Hell even realtek is better.

 

[RejZoR]

Ok, I've been observing the LAN lights when it died again (btw, that PnP thing isn't the reason apparently) and the lights are completelly dead. Which means the problem is electrical. I just have no fucking clue what. I've inspected the PCB and components around the LAN circuitry and nothing seems to be out of the ordinary. Might have to open up the actual LAN connector maybe and check it out in there...

 

[Mussels]

Ok, I've been observing the LAN lights when it died again (btw, that PnP thing isn't the reason apparently) and the lights are completelly dead. Which means the problem is electrical. I just have no fucking clue what. I've inspected the PCB and components around the LAN circuitry and nothing seems to be out of the ordinary. Might have to open up the actual LAN connector maybe and check it out in there...

possibly just a dead chip. it can happen.

 

[RejZoR]

Well, it's only dead when it randomly decides to be dead. When it's not, it works absolutely perfectly. No lag or bandwidth stalling. How on Earth can that be with a "dead" chip!?

 

[Mussels]

Well, it's only dead when it randomly decides to be dead. When it's not, it works absolutely perfectly. No lag or bandwidth stalling. How on Earth can that be with a "dead" chip!?

because it has a fault that under certain conditions causes it to fail.

 

[tabascosauz]

@RejZoR maybe it's just time for a new board. It's out of warranty, and if you buy a third party NIC it's probably not going to be used when you inevitably decide to upgrade.

 

[Mussels]

@RejZoR maybe it's just time for a new board. It's out of warranty, and if you buy a third party NIC it's probably not going to be used when you inevitably decide to upgrade.

he can use a third party NIC for dirty cheap, depending on his needs even USB wifi could work depending on his setup and needs.

 

[RejZoR]

@RejZoR maybe it's just time for a new board. It's out of warranty, and if you buy a third party NIC it's probably not going to be used when you inevitably decide to upgrade.

I've already bought a PCI LAN card, but having a microATX board with huge graphic card makes it a bit inconvenient to use. That's why I was trying to get onboard LAN working again. I can use the LAN card, but it complicates stuff inside the case and restricts airflow. I'd prefer to wait till Skylake-E and AMD Zen are out, but taht's so far away and this crap is frustrating me. Oh well.

 

[tabascosauz]

I've already bought a PCI LAN card, but having a microATX board with huge graphic card makes it a bit inconvenient to use. That's why I was trying to get onboard LAN working again. I can use the LAN card, but it complicates stuff inside the case and restricts airflow. I'd prefer to wait till Skylake-E and AMD Zen are out, but taht's so far away and this crap is frustrating me. Oh well.

Well, perhaps the 980 could be relegated to the lower x16 slot? I mean, x16 => x8 shouldn't be a big difference, should it? Then again, it's PCIe 2.0...

If it's in the lower slot then a PCIe x1 style 7260AC could fit.

 

[RejZoR]

Actually, X58 can run in full x16 mode on either slot (yeah so old chipset and it can do that). The thing is, if I place it in the second one it actually covers the PCI slot so I can't stick in the LAN card anymore. I have it solved with PCIe raiser cable and soundcard connected to it outside the motherboard in a lower expansion slot, but the cable is now actualyl wrapping LAN card which is nto exactly ideal lol. But it works for now.

It would work better if I'd fit soundcard and PCIe LAn card above it, I think I could stick both above the graphic card. Anyway, I've migrated to a mid tower now and the next board will be normal ATX again. microATX is cute but unnecessarily restricts you so I'll avoid that from now on...

 

[tabascosauz]

Actually, X58 can run in full x16 mode on either slot (yeah so old chipset and it can do that). The thing is, if I place it in the second one it actually covers the PCI slot so I can't stick in the LAN card anymore. I have it solved with PCIe raiser cable and soundcard connected to it outside the motherboard in a lower expansion slot, but the cable is now actualyl wrapping LAN card which is nto exactly ideal lol. But it works for now.

It would work better if I'd fit soundcard and PCIe LAn card above it, I think I could stick both above the graphic card. Anyway, I've migrated to a mid tower now and the next board will be normal ATX again. microATX is cute but unnecessarily restricts you so I'll avoid that from now on...

Well, such a thing could happen to any board, but I see what you mean. It's infinitely harder having to deal with this on mITX like me (that is, unless you're like me and aren't a gamer, and can make do with a simple USB-N13). mITX is nice because often times they now come with a vertical mSATA or M.2 for the Wifi card, which is usually AC, and a long M.2 parallel to the board for an SSD. It's hard to find an ATX board that comes with a place for a wifi card and cutouts for the antennas in the I/O shield.

 

[RejZoR]

I loved microATX board and my old Lian Li miniATX case. But then I bought this gigantic GTX 980 and then I had a problem XD

 

[Pill Monster]

That's true. But there are millions of motherboards that use that same chipset and support the same CPUs so while each motherboard will have a unique BIOS, the differences, even among motherboards brands are slight. But to your point, if an update for W8.1 broke something with that particular motherboard, ASUS is very good at putting out updates. But still, Microsoft does not code to specific motherboards. It codes to standards so if something broke, then it would be because the board did not comply with standards. And if that were the case, there would be a whole lot of Rampage II Gene owners complaining about this problem. And I don't see it in the ASUS forums.

Even if in the same suite of tools, they are still different tests.

It really has nothing to do with operating systems today. That setting is all about what happens before the boot drive is even touched so the hardware (including legacy hardware) can communicate at a really basic level. Once the OS boots - whether Linux or Windows, their own protocols take over. At least that's my understanding.

Absolutely correct.

You're confusing plug and play devices with a plug and play OS. Turning on Plug and Play OS makes it so the BIOS only initializes the core components, letting the OS configure (and manage,) everything else. To my knoledge, this should *always* be set to off so all components are initialized by the BIOS and then handed over to the OS as opposed to letting the OS initialize them in the first place. This page seems to describe it pretty well:

http://help.cognex.com/Default.htm#KB_Topics/VisionPro/Drivers/2905.htm

So my comment to the "OS/2 thing" was the fact that OS/2 was a Plug and Play OS that initializes devices from the get go instead of relying on the BIOS to do it.

Sorry but you're way off......kinda surprising that someone who claims to be a developer with a Computer Science Degree doesn't don't know what PnP is?

PnP option is so bios can hand out IRQ assignments to legacy operating systems which don't support ACPI 2.0, aka everything up to XP, including OS/2.

This affects all devices with an IRQ, and isn't just for "initializing core components" at boot...



Sorry to quote myself but....as an example of why the option is there, some vendors used to force an IRQ address on PCI devices... Nvidia cards used IRQ 16 iirc, same as Creative:

Your system must be very old because IRQ assignment since XP is virtualized and handled by the Operating System, not the BIOS.
Even if IRQ assignment ordering is a bios option it's translated by ACPI and Windows will assign whatever IRQ address it wants.

You're talking about a specific problem which affected nF2/3 boards with Soundblaster installed. So you have a very old motherboard or are running Windows 98/XP.........


The driver reset causes the black screens, trick is to find out why.
Driver reset is called by the scheduler after IRQ timeout happens 7 times in a row.
Windows will BSOD if the reset fails, but sometimes the reset gets stuck and hangs causing a soundloop. TDR basically means the GPU is hung..

 

[RejZoR]

And yet I never claimed I'm a developer or have a degree of anything. Stop making shit up just because you don't like me for whatever reason...

 

[Aquinus]

Sorry but you're way off......kinda surprising that someone who claims to be a developer with a Computer Science Degree doesn't don't know what PnP is?

PnP option is so bios can hand out IRQ assignments to legacy operating systems which don't support ACPI 2.0, aka everything up to XP, including OS/2.

This affects all devices with an IRQ, and isn't just for "initializing core components" at boot...

Initializing core components was me dumbing down interrupt assignments. Stop nit picking and trying to start a fight. You just said what I said in more detail.

 

[Mussels]

ladies, no fighting. if you don't deal well with each other, stick with facts and sources and leave personal comments out of it.

 

[eidairaman1]

Anyone ever had this weird problem where LAN device integrated on motherboard (in my case on ASUS Rampage II Gene) almost 99% of time doesn't work at all with red cross in Windows 8.1 (it's not even detected at all in Device manager) when you boot system after having it turned off for few hours. Rebooting system doesn't resolve the issue. However, if I shut down the system, turn PSU switch to OFF, wait for all lights to go out on motherboard, power it back on and in same 99% of time, LAN magically appears.

This crap started happening all of a sudden for no real reason. I've rechecked everything and everything is in order. I've also tried stock clocks and everything and it's still happening.

Only thing that seems to somewhat corelate is installation of GTX 980. With HD7950 I nevr had such issues. Might be just weird coincidence though.

Any ideas? I'm really not in the mood for changing the whole system right now (was hoping to do that decision when AMD Zen will be released) so I'm looking for solutions. Any ideas?

Might be time to grab a lan card for testing

PS i use realtek's direct driver (not off of Asus)

 

[RejZoR]

I have a Realtek PCI card now, but still wanted to solve the problem since it's restricting the airflow a bit...

 

[Mussels]

I have a Realtek PCI card now, but still wanted to solve the problem since it's restricting the airflow a bit...

how much bandwidth do you need? is USB 2.0/3.0 networking an option?