SOLVED PXE Boot HP X2 210 (Hybrid tablet Windows 10 Pro)

  • Hi,

    We’ve just bought 32 HP tablet X2 210 (not really a tablet because it runs under Win 10 Pro…), which has USB 3.0 but no network port.
    Like other people, I tried to boot on PXE with USB to RJ45 adapter I found on a store near my office (ASIX AX88772C)
    It’s an UEFI bios so I had first to disable secure boot and it didn’t work with undionly.kpxe.
    If I specified an efi boot file (ipxe.efi or snp.efi), I’m able to boot on network, retrieve IP address and display Fog menu.
    BUT… That’s all…Any option I choose, screen stay black after loading file bzimage and initzx (for quick image I can enter credentials and choose image, but same problem after that)

    What can I change to complete task as quick registration or else ?

    Thanks for help, and my apologize because there are several topics on that subject, but each computer/adapter seems specific…


  • @sebastian-roth @Wayne-Workman No news for the moment but I’ll keep you in touch 😉

  • Senior Developer

    @Wayne-Workman No, they didn’t get back to me. But Matthieu just had another try and sent a new message two days ago. Let’s see what happens this time.

    We have an old firmware binary that might be useful to flash those devices with a fix. Though this has never been officially released and is missing other fixes that have been released since then. Let me know if you need help on this.

  • @sebastian-roth said in PXE Boot HP X2 210 (Hybrid tablet Windows 10 Pro):

    I’ll finish it and send it off then.

    Did you ever get a response?

  • @sebastian-roth Ok thanks, I’ll email you as soon as possible.

  • Senior Developer

    @Matthieu-Jacquart I am sorry but disabling these setup_efi_pci routine is a really dirty hack to make it work on those HP devices. We’re not going to include this into FOG not even with a “Host Kernel Parameter” thing.

    Sure I can send a message to those same guys but I kind of doubt that the fix will ever make it into the official firmware - whatever reason they have?!?! Would you mind preparing a piece of text so I don’t have to do this alone? Just read through the mail I sent you and put together a couple of lines. I’ll finish it and send it off then.

  • @sebastian-roth You’re just a master chief Seb ! It works perfectly !
    Do you think there’s any way to contact HP people to modify bios, or maybe (but I’m dreaming) just add an argument in fog Web UI “Host Kernel Arguments” ?
    Thanks a lot, once again 😉

  • Senior Developer

    @Matthieu-Jacquart I really hope I remembered correctly the change I made more than a year ago. Just compiled a new kernel which does not call the hanging function setup_efi_pci(). This is very ugly but seems to be a way to workaround this dang HP firmware bug! Find bzImage_disabled_setup_efi_pci here.

    Turns out HP released at least two new firmware versions but didn’t include the (probably very tiny) fix we worked out together with them. As they even provided fixed firmware blobs to try out I have no idea why this didn’t make it into their official code base?!?

  • Senior Developer

    @Matthieu-Jacquart Oh yeah, I definitely remember this one. It’s been a long endeavor till we got this fixed. As far as I remember HP provided a fixed UEFI firmware. Just checked the mails in my archive and sent you the message via E-Mail:

    On Tue, 03 May 2016 16:11:34 +0200 JACQUART Matthieu <email deleted> wrote:
    Ok ! With Shell_Full.efi I was able to boot on EFI shell and to downgrade/upgrade bios, and everyhting works fine with Yu Max bios ! I was able to register hots (using ipxe.efi in DHCP options) without error, good !

    So I am a bit confused that you say things don’t work. I had the impression that HP actually fixed this and possibly also officially released a new firmware including that fix. But maybe I am wrong. Please read through the old message I send you via mail and let me know what’s up.

  • Moderator

    @matthieu-jacquart Ah ok, now I understand.

    Just be aware the developers are taking a short holiday for the rest of the month. So their response will be a bit slower. Just be aware so you can adjust your schedule.

  • @george1421 It’s a long story, but it’s an incompatibility between this HP model and fog kernel (HP never made modification for this model and last bios don’t change the problem), so the only solution in February 2016 was to modified fog kernel (big thanks to Seb for his work), and after that everyhting worked fine.
    But now I want to deploy these hybrid and with modified kernel (based on 4.3.3) I’ve got fog (1.5.0 RC8) message : kernel too old.

  • Moderator

    @matthieu-jacquart Has there been any movement on HP’s part to update the uefi firmware on this hardware?

    Is there a compatibility issue with the kernel your have vs the current release of FOG? Or is it just old and you want it updated? I’m just trying to understand the context.

  • @sebastian-roth Hi Seb and @tom-elliott, I re-open this topic because I need to deploy once again our hybrid HP X2 210, but modified bzimage that Seb provided to me (bzImage_rb) is now too old for fog (base on kernel 4.3.3).
    Without this I can’t deploy our 32 hybrid, is there any way to update the bzImage_rb ?

  • Senior Developer

    Marking this thread solved for now as the OP is able to capture and deploy his tablets with the modified kernel. Although this is not a proper solution (more a hackish kind of workaround) we cannot do anything else than waiting for the people from HP we got in contact with to possibly fix the UEFI firmware!

    As of now the HP devs confirmed our findings (screwed memory addresses) but we haven’t heard back from them for two weeks now. No solution yet…

  • @Sebastian-Roth oh I see, I thought that might be related. But you’re right, I’m booting via Grub. Thank’s for the info though

  • Senior Developer

    @schnecki We are more than happy to help people solving their linux problems. But please understand that this forum is about FOG and we focus on getting FOG or FOS (the buildroot linux system at its base to do the imaging) to run on the devices. We only have basic knowledge on kernel stuff and we won’t be able to help you with this one I suppose.

    The OP was about the linux kernel freezing when trying to enumerate PCI devices booted in UEFI mode. I guess you don’t see that issue as you probably use GRUB bootloader (which actually bypasses the UEFI STUB and boots the linux kernel as old school bzImage). What you ask is totally unrelated. Feel free to open a new discussion and see if anyone from the FOG community is able to help.

  • Hi,
    somehow this thread stopped. Was the discussion moved? I’m also interested in the topic. I have Linux running on the tablet. First I bootet with Ubuntu 16.04 beta release to be able to install ArchLinux from there. Ubuntu runs stable but does not show the battery status. ArchLinux runs instable, but does show the battery status. The problem is that the i2c designware platform module causes a kernel panic, and then freezes the system completely. I guess that’s also why it is not possible to boot any other Live system.

    Here you can find the dmes output:

    It looks loke there is a problem with a semaphore. Unfortunately, I’m not too familiar with the i2c module nor the kerenel itself. Could someone give me a hint what to try out?

    Disabling the kernel module, as given under the link above fixes the issue, but causes other problems.

    P.S.: Written from my x2 210 running ArchLinux 🙂

  • Senior Developer

    @Matthieu-Jacquart The firmware people got back to me asking about the BIOS version. Can you please find out and tell me?

  • Senior Developer

    @Matthieu-Jacquart Good question. I’d better leave it as unsolved as this really is just a quick dirty hack and I am very afraid of pushing this into our main kernel. Talked to Tom Elliott about this already and I think we should better wait. Do your cloning as this is working for you now and I will get back to you for some more testing when we hear back from the firmware guys I’d say.

  • @Sebastian-Roth ok, and thanks again so much for your help.
    Do you think we can mark this topic as resolved, or we are waiting for a fix from HP / Insyde ?