Experiences with the Dell M3800, The Good, The Bad (no Ugly, it’s a beautiful machine)

Remaining UNSOLVED

Suspend does not function; you must power off before you go mobile.

  • Early in the boot lifetime
    • it does work
    • after a few hours of uptime, it does not
  • Later in the boot cycle
    • The initial suspend works on powerbutton or lid close.
    • Yet five+ seconds later, it comes back alive at full power & heat.
    • Then the cpu is running hard, getting hot, very hot
      it will overheat if left in a suitcase.
    • The screen is dark & the network is unresponsive; it is hard hung.
    • A powercycle is required.

Following

The Good

  • The native 4k2k display is worth the buy-in.
  • The HDMI2-MHL runs 4K2K externally over the HDMI port.  Neat!

The Bad

  • Uses Broadcom componentry
    which is known Linux-unfriendly
    but maybe it’s different here, different now

    • Wireless → SOLVED (a binary blobs in rpms: broadcom-wl, kmod-wl)
    • Bluetooth → SOLVABLE (a binary blob & extraordinary non-standard reencoding & manual installation)
  • No wireline ethernet; use the USB-to-RJ45 dongle DP/N 0J1GH5.

Yet

remember the good

  • 4k2k display
  • super-thin
  • Intel i7
  • 1T SSD
  • Does run Fedora 21 … except for the exceptions.

The Ugly

No ugly, it’s a beautiful machine.  See?

Complaints

Yet, we came to complain. There is stuff that “just does not work.”

  • The bluetooth does not “just work,” whereas it does “just work” elsewhere (on Fedora 20)
    Potentially Solvable
  • The graphics hangs & crashes … alot. Especially when coming out of suspend. this is the nouveau driver.
    UNSOLVED
  • The unit cannot be reliably suspended and put in a closed space. It will go into overheat mode.  The CPU runs hard.
    UNSOLVED.

Evidences

  • Fedora 21
  • Kernel 4.0.7
    Linux devbox.example.com 4.0.7-200.fc21.x86_64 #1 SMP Mon Jun 29 22:11:52 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Actualities

Graphics Crashes

These manifest as the graphics becomes inoperable; the CPU is running hard, the fan is running hard. Even virtual consoles will not work. You must powercycle.

Jul 24 19:52:10 devbox.example.ccom kernel: nouveau E[    PBUS][0000:02:00.0] MMIO write of 0x00000002 FAULT at 0x4188ac [ IBUS ]
Jul 24 19:52:24 devbox.example.com kernel: nouveau E[    PBUS][0000:02:00.0] MMIO write of 0x00000002 FAULT at 0x4188ac [ IBUS ]
Jul 24 19:52:55 devbox.example.com kernel: nouveau E[    PBUS][0000:02:00.0] MMIO write of 0x00000002 FAULT at 0x4188ac [ IBUS ]

ACPI Scary Warnings

Jul 24 19:52:30 devbox.example.com kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150204/nsarguments-95)
Jul 24 19:52:30 devbox.example.com kernel: ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
Jul 24 19:52:30 devbox.example.com kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150204/nsarguments-95)
Jul 24 19:53:01 devbox.example.com kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150204/nsarguments-95)
Jul 24 19:53:01 devbox.example.com kernel: ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
Jul 24 19:53:01 devbox.example.com kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150204/nsarguments-95)

 

Comments are closed.