5640 drivers and Windows 8.1 ... some hardware issues

maybe you are using the wrong drivers

http://appleinsider.com/articles/14/02/11/apple-issues-new-boot-camp-support-sof tware-updates-compressor-with-bug-fixes

I had 1/3 boot up into win8.1 result in a bad page header fatal bluescreen and installing 5.1.5621 fixed the issue in my imac

I did not run the setup though I installed each driver manually

I got an error when I tried to run the exe

so I updated the drivers manually the only one that really installed was my AMD videop card driver which it seemed was the only one newer then what I already had

I have also been having this issue on my newly bought Late 2013 Retina MacBook Pro (rMBP), running Windows 8.1 Pro (using native EFI) with Boot Camp's latest drivers (version 5.1.5640). After waking up from sleep, the "Broadcom 802.11ac Network Adapter" becomes the "Broadcom 802.11ac Network Adapter #2" and subsequently "Wi-Fi" becomes "Wi-Fi 2," requiring me to reconnect to all of my networks since Windows seems to tie saved wireless networks to the driver.

Here is what I've found so far (note that when I say "Wi-Fi 2," I mean the driver is using the Network Adapter #2 instead of the original, numberless one):

• This only happens to Windows--the Mac side is fine.
• This only happens when Windows is put to sleep; hibernation and Fast Boot shutdown do not change the driver. I'm actually using hibernation right now as an alternative to sleep mode because it works fine.
• Sleeping a second time within the same Windows session keeps "Wi-Fi 2" as "Wi-Fi 2."
• I need to restart to get the original driver to work again. Shutting down *may* work, though, but since hibernation keeps "Wi-Fi 2" as "Wi-Fi 2" and doesn't change the driver back, I doubt it. Maybe non-Fast Boot shutdown would also fix it, but restart works well enough.

Essentially, sleep mode is unusable in Windows, since this network shuffle makes it so all networks need to be re-added.

Now, I have also noticed that booting into Windows with the Thunderbolt to Gigabit Ethernet adapter causes the Network Adapter to become "Broadcom 802.11ac Network Adapter #3" (and thus creating a "Wi-Fi 3" entry on the connection manager). I'm assuming that the Ethernet Adapter would also cause the Network Adapter #2 to appear if it were uninstalled previously.

AND, in the Windows Event Viewer, I see this, and I wonder if it is related to the above because the Wi-Fi adapter uses the PCI-express bus on the Late 2013 rMBP:

"The system firmware has changed the processor's memory type range registers (MTRRs) across a sleep state transition (S4). This can result in reduced resume performance."

What the heck is going on here??

I have tried messing around with the Device Manager's various Power Management tabs, as well as the Link State Power Management stuff in the Windows Control Panel. Nothing worked, and I have since set everything back to normal and changed every "Sleep" option I could find to "Hibernate." This isn't an ideal workaround due to the writes this makes to the SSD, so if anyone has a fix or can help figure this out that would be most appreciated!

I can answer the keyboard backlighting issue, though. The keyboard backlight only works when the Boot Camp drivers are loaded. In other words, the backlight does not work until you see the little Boot Camp icon in the taskbar. Whether or not you see the backlight at the login screen just depends on how fast the SSD loaded Windows that time.

I have the exact same behavior on my MacBook Pro Late 2013 with Windows 8.1. Unfortunately re-adding wi-fi networks to the "Wi-Fi 2" adapter is not an option for me. I'm a developer and I have virtual switch networking setup with Hyper-V for the Windows Phone emulator and with Oracle Virtual Box for the Genymotion emulator. I need to make sure the standard Wi-Fi adapter "survives" the sleep/resume cycle like a normal computer.

I have been going deep into Microsoft's support levels to get this resolved--it's definitely a Windows issue, seeing how there are zero problems on the Mac side. Unfortunately, for someone like me who has a lot of networks that need to be saved, this is a pretty big deal--especially since it becomes "Wi-Fi 3" when using the Thunderbolt to Gigabit Ethernet adapter. I'll post back once I get some kind of result, though at the rate this has been going it might not be soon (but don't worry; I'm not the kind of person who'll just forget to follow up).

I personally have a feeling that this has something to do with Windows' hardware compatibility (maybe it doesn't support Haswell's new sleep states right? I'm just guessing here, honestly), although the fact that one of the memory addresses of the network adapter CHANGES after sleep is a bit weird...

@Sakman74,

Have you tried putting your finger/a dark piece of paper over the light sensor (right next to the Facetime camera)? That turns on my backlight.

yeah windows driver issue

the windows core just speak with the driver interface

Well, think of it this way: how can it be a driver issue when it happens with the Microsoft default/basic/generic network adapter driver, the Apple-supplied one, AND the Asus PCE-AC68 driver? "Driver" is just the individual bit of software that allows a device to work, and the adapter does work. The issue crops up when Windows deactivates/turns it off/etc. to go into Sleep mode and then re-activates it on wake up, which is not so much a driver problem as it is a Standby/S3 state resource issue since the driver itself is still working properly (it's just not being loaded in the right place).

If anything, this seems way more telling of a problem with how Windows works with the PCI-e bus, again with reference to how the memory address changes (!!!) during sleep. Of course, this could also be something to do with Apple's EFI firmware, as cited in the event log error "The system firmware has changed the processor's memory type range registers (MTRRs) across a sleep state transition (S4). This can result in reduced resume performance."

If it's the latter, I'll take this case and its mountain of ridiculousness and unacceptability up with Apple. I want to thoroughly check Microsoft's side of the issue first, though, since that would potentially be a faster fix.