5640 drivers and Windows 8.1 ... some hardware issues

there is no such thing as a generic network adaptor driver in windows

every network card require a driver of their own

Oh. Then what's "Broadcom 802.11ac Network Adapter (Microsoft)"? I always thought the "(Microsoft)" suffix meant it was a Microsoft/generic driver, as opposed to the one ending with "(Broadcom)" being a non-Microsoft driver.

Anyways, Microsoft support got back to me and told me just what I was afraid of: it's a hardware thing, and I need to contact Apple. Granted, I have a bit more technical information to work with than just stating, "it's an Apple thing," so we'll see if Apple's gonna throw us the typical "No, it's a Microsoft thing" that always seems to happen.

Just an update: Apple support was quite helpful, although we still couldn't find a resolution. The support advisor I spoke to recommended I try to see if this happens under Windows 7 and a reinstall of Windows 8.1, so I'm going to try it and see what happens. Since I can't reinstall my Windows 8.1 due to too much stuff being on it, I'm going to see if I can quad-boot Windows 7, 8.1, another 8.1, and OS X. Wooo!

EDIT: Or I could just use external drives, I guess.

At this point, as this Wi-Fi issue is so bizarre, the fix could come from anywhere it seems. Fingers crossed.

Another update: So my crazy plan of quad-booting didn't work out (Windows 7 doesn't actually support full UEFI booting, and Disk Utility refused to modify my partition table, and my trick of using the Hybrid MBR to install multiple Windows didn't work this time since my system is using pure EFI/GPT, etc.). What I DID do was Refresh my Windows 8.1 install back to its initial state, so that no Boot Camp software got loaded.

And, guess what? Even without a driver, the network adapter switched its entry in the device manager after waking from sleep. I didn't have any Apple drivers loaded AT ALL, period, and it still happened.

Now, I also took the liberty of checking how memory addresses are supposed to be handled in sleep via another Windows 8.1 PC I have, and they don't change through any sleep cycles. I even checked it on the Mac half of this same computer, and the memory address of the network adapter does NOT change. See in the following image; the top command is before sleep, and the bottom command is after sleep.

In Windows on this Mac, however, the memory address range most certainly DOES change upon waking from sleep, where it shouldn't be. In other words, this means that the resources the network adapter needs are being reallocated, which makes no sense. In effect, this causes Windows to install a second driver because the first one is not where Windows thought it was supposed to be when it woke up form sleep mode. Windows doesn't recycle the same driver because, as far as Windows knows, that first device was there before sleep in a specific state, but some small thing changed, forced it into another state, and now it's not the exact same device Windows thought it put to sleep in the first place.

What happens when Windows duplicates a driver that's already been installed once? You get that magic #2 appended to the new one's name. This means one of a couple of things: either another device is physically causing this address change, which would be Apple's problem, or there's a fundamental flaw with how Windows handles PCI-Express devices.

From what it seems, as far as Windows knows, it's waking up to find a slightly different device than it put to sleep. So that would mean Windows is behaving normally, right? It's just reassigning a driver because the equivalent of someone messing with the contents of the RAM in sleep mode is happening. Sleep mode is supposed to be static--i.e. not one thing should change during sleep mode, or else potentially big problems could occur. So the question now becomes, "What is causing the driver to change while the system is asleep?"

Granted, that's the million dollar question right there, the answer to which will be the source of the fix. Unfortunately, I don't know how to examine or otherwise snoop this. If there is a way to watch what happens to the RAM contents while the system sleeps and wakes up, we'd be able to tell if this happens when the machine goes into OR comes out of sleep. That would be a first step. the next step is to somehow figure out exactly why the device (not the driver!!!) is getting messed with, and then we'd know if this is Apple's, Microsoft's, or even Intel's or Broadcom's fault.

We might have two clues, though: The first one is that sleeping in safe mode also causes the issue to occur. Since we know that the issue is not caused by the device driver itself, though, this is to be expected.

The second one, which I don't know if it's a red herring or not anymore, is the Windows Event Log entry: The system firmware has changed the processor's memory type range registers (MTRRs) across a sleep state transition (S4). ...What is going on here?? Why is the firmware doing anything other than putting things into a low power state??

So that's where I am right now.

Any suggestions on where to go next?

EDIT: Found another, possibly pertinent error:

"Miniport Broadcom 802.11ac Network Adapter, {[long hex string]}, had event Fatal error: The miniport has failed a power transition to operational power"

This seems to suggest the issue happens during wakeup.

Update again: So I just checked and the error "The system firmware has changed the processor's memory type range registers (MTRRs) across a sleep state transition (S4)." happens on every sleep, which is reminiscent of an old error I get on my 2010 MacBook Pro under Windows about some ACPI error. My 2010 MBP never had any issues even remotely similar to this under Windows 8.1, so that message may not, in fact, have anything to do with this problem. The error in the first edit above ONLY happens on the first sleep, however, which is when the adapter changes. That's another question right there, then: "Why does the adapter only change on the first sleep?"

I can verify that the memory addresses and everything stay the same on further sleeps; it's just that first one that makes the adapter change.

heehe threads like this is that I advice anybody who are not going to be using windows for gaming to install windows in a virtual machine

save them heartache and frustrations and having to be tech heads to solve their issue

😀

Some of us actually need Boot Camp Windows for the hardware access not possible with virtualization. For example, hardware acceleration is a nice feature when you need it for things like content creation and encoding/decoding purposes. On that note, it's just too bad that Intel Quick Sync can't be used under Windows without an ugly hack that disables the Nvidia. 😟 But that's for another thread another day.

@sakman74

You know, I'm surprised no one else has said anything about this issue. Shouldn't it be affecting all Late 2013 rMBP users with at least Windows 8/8.1?? Moreover, it seems those with 2013 MacBook Airs don't have this problem (or haven't said anything about it) either, and they have the same BRCM4360 wifi chip. The most I can turn up are people with Dells and Sonys who have this problem with their network adapters. I wonder if we got bad cards? I can try to go to an Apple Store and see if they'll swap it out.

Ok, I just went to my nearest Apple store, and after literally one of the worst customer service experiences I've ever had there, we determined it wasn't a hardware issue and I was redirected to Micro Center since they have both Windows and Apple people under the same roof. Within about 10 seconds of describing the issue, the associate at Micro Center told me to stop using Windows 8's sleep mode because it's horribly, horribly buggy, citing examples that would make even a veteran tech guy cringe. Apparently there was a case where someone's laptop basically died--literally almost bricked itself--because it went to sleep running Windows 8, and the only fix was to tear the laptop apart to remove and reinsert the non-user-accessible battery. I actually can personally attest to an experience where Windows 8 caused me to return 3 power supplies because going into standby produced some kind of major incompatibility that would suddenly cut power to my desktop. It seems we've all been unlucky enough to stumble upon another one of Windows 8's sleep bugs.

The fix? Set the laptop to never sleep, ever. Disable Sleep mode everywhere possible and, to quote the Micro Center associate, "avoid Windows 8 sleep mode like the plague." I guess it takes just the right combination of hardware in order to use Windows 8's sleep mode, which, unfortunately, the late 2013 Retina MacBook Pro does not have (at least when EFI Booting Windows 8; maybe BIOS booting is different since it works fine on my mid-2010 MacBook Pro that way).

And for us Mac users, or even anyone with an SSD, this also means don't use hibernation mode, either. Hibernation writes a huge 4-32GB file (depends on how much RAM you have) to the SSD every time it is used, and since SSD life is measured in writes and not years, hibernation will effectively destroy the drive. This also means disable "Fast Boot," since Windows 8's Fast Boot feature uses hibernation to work.

I don't know about anyone else, but after so much back-and-forth, I'm with the "It's Windows's fault" camp on this one.

So... Hopefully Windows 9 is the answer?

As for why the Thunderbolt to Gigabit Ethernet adapter changes the driver, I know from experience that Mac Thunderbolt Products + Windows = buggy mess, so, upon further thought, that may have not even been related.

Arrrrrrrrrrrrrrrrrrrrrrrrrrgggggggggggggggggghhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

I just thought of something else that might work: Use a third party Wi-Fi connection manager that can save the network profiles you need to use independently of the driver. That way, no matter what the Wi-Fi driver does, you can still connect to your networks before/after sleep. Unfortunately, I don't know of any such program, nor do I have the programming skills to make one. 😟

So I couldn't sleep well knowing that I couldn't use sleep mode ('scuse the terrible pun), and today a workaround hit me.

It's not pretty, but it does the job. Basically it involves using a batch script to export the wireless network profiles in bulk from the first "Wi-Fi" interface, then sleeping, then importing all of the networks into "Wi-Fi 2." It's the manual equivalent of using a network profile manager that works independently of the driver, I guess.

Anyways, download this:

http://fbe.am/tGZ

Extract the "NetworkProfiles" folder somewhere you like, and read the included README for all details.

Hope this helps!

-KNNSpeed

I'm guessing you installed Windows in BIOS mode (i.e. You picked "Windows" instead of "EFI Boot" when setting up Bootcamp). That used to happen on my mid-2010 MacBook Pro (still does, rarely). it seems to happen most frequently when you try to speed through the alt-boot menu and pick Windows really quickly. I've been able to avoid that by waiting a bit on the alt-boot menu before picking Windows. My explanation is that by picking the boot menu too fast screws up the EFI-->emulated BIOS handoff.

I haven't had it happen once since I installed Windows 8.1 via native EFI.

That seems like a graphics driver problem. The only recommendation I can make is to update the graphics driver from either NVidia or Intel depending on your model. The Apple-supplied drivers tend to be outdated before they're even released.

Also, you could also try and reset your PRAM, since that seems to fix things like that, too (though the driver update is more important).