Blank screen when using iPXE

[ArneLovius]

I've come across an odd problem.

I have a Dell E4200 laptop, which works perfectly doing a straight WDS install, it also works perfectly if I do a PXE boot to WDSLinux and then start WDS from there.

If however I do a PXE boot to undionly.kpxe and then chain load to WDS, it loads as far as boot.wim and then the screen goes blank.

I thought it might just be slow, but it was still blank after an hour.

I thought it might have been a memory issue, but running memtest86+ for two passes threw no errors.

I thought it might be a memory "mapping" issue, so I undef'd the protocols I don't need from src/config/general.h such as AoE etc and rebuilt to get a (slightly) smaller undionly.kpxe (63,115 B) but that didn't appear to have any improvement. It shows "591kB free base memory after PXE unload"

Just in case somebody knows of a specific device that might be causing the problem, the output of lspci is

Code:

00:00.0 Host bridge: Intel Corporation Mobile 4 Series Chipset Memory Controller Hub (rev 07)
00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)
00:02.1 Display controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)
00:19.0 Ethernet controller: Intel Corporation 82567LM Gigabit Network Connection (rev 03)
00:1a.0 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #4 (rev 03)
00:1a.1 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #5 (rev 03)
00:1a.2 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #6 (rev 03)
00:1a.7 USB controller: Intel Corporation 82801I (ICH9 Family) USB2 EHCI Controller #2 (rev 03)
00:1b.0 Audio device: Intel Corporation 82801I (ICH9 Family) HD Audio Controller (rev 03)
00:1c.0 PCI bridge: Intel Corporation 82801I (ICH9 Family) PCI Express Port 1 (rev 03)
00:1c.1 PCI bridge: Intel Corporation 82801I (ICH9 Family) PCI Express Port 2 (rev 03)
00:1c.3 PCI bridge: Intel Corporation 82801I (ICH9 Family) PCI Express Port 4 (rev 03)
00:1d.0 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #1 (rev 03)
00:1d.1 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #2 (rev 03)
00:1d.2 USB controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #3 (rev 03)
00:1d.7 USB controller: Intel Corporation 82801I (ICH9 Family) USB2 EHCI Controller #1 (rev 03)
00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev 93)
00:1f.0 ISA bridge: Intel Corporation ICH9M-E LPC Interface Controller (rev 03)
00:1f.2 IDE interface: Intel Corporation 82801IBM/IEM (ICH9M/ICH9M-E) 2 port SATA Controller [IDE mode] (rev 03)
00:1f.3 SMBus: Intel Corporation 82801I (ICH9 Family) SMBus Controller (rev 03)
00:1f.5 IDE interface: Intel Corporation 82801IBM/IEM (ICH9M/ICH9M-E) 2 port SATA Controller [IDE mode] (rev 03)
02:01.0 FireWire (IEEE 1394): Ricoh Co Ltd R5C832 IEEE 1394 Controller (rev 05)
02:01.1 SD Host controller: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 22)
0c:00.0 Network controller: Intel Corporation WiFi Link 5100

I also experience a blank screen when trying to do a wimboot using a winpe generated from the Windows 8 ADK using the howto here and when doing a wimboot to Windows 7 using the howto here

Although I have syslog enabled, once it boots wdsnbp.bin, there are no further syslog entries

Using other laptops, desktop and hyper-v guests does not show this problem, which should prove the rest of the PXE > iPXE > WDS process.

The laptop boots clonezilla, Parted Magic and Ubuntu installers quite happily from iPXE.

I do appreciate that this is not a direct problem with iPXE per se, as iPXE is working :-) but it is something within iPXE that would appear to be causing the problem :-(

I've noticed two other reports of blank screens on Dell hardware here and here but neither of them had any follow up.

It should go without saying, but I'd be happy to try any troubleshooting steps to resolve this.

Cheers

Arne

[robinsmidsrod]

I believe this might also be related to the Broadcom NetExtreme (aka tg3) in the Dell machines. There are some patches from Kevin Tran on the mailing-list that haven't been merged to mainline yet. You might have luck with those, but considering you're using undionly.kpxe that doesn't seem to be the issue, because undionly doesn't use the tg3 driver. You might have that issue relating to interrupts being advertised, but never sent. I believe there was another "hack" posted either to the forum or the mailing-list not so long ago (less than a month, I think) that just forcibly said that a certain piece of hardware used polling, regardless of what it advertised. Try to find that post, and report back.

[ArneLovius]

Thanks for replying, but the Intel 82567LM (as per the lspci output I posted) uses the e1000e kernel module, not the tg3 kernel module...

[robinsmidsrod]

Indeed, my fault. But if you're using undionly.kpxe you're not using an iPXE-provided driver, you're using UNDI provided by the vendor PXE implementation (in this case either Intel and/or Dell) that is most likely the problem. Hence my pointer to that interrupts/polling "hack" patch.

[ArneLovius]

And having replaced my KVM switch so it works with the IP KVM that I had on my main server, so I could connect the laptop to it and reboot it remotely while trying to work out why it won't boot WDS after iPXE.

Only to discover that the VGA output is not active until Windows has booted...