devsetup1-4  Exercise 3: A gpucheck Function
=============================================

THE FUNCTION (add to ~/.bashrc)
-------------------------------
    gpucheck() {
      local ok=0

      echo "Running kernel: $(uname -r)"

      echo "DKMS (nvidia):"
      if dkms status 2>/dev/null | grep -i nvidia; then
        :
      else
        echo "  (no nvidia entry)"
        ok=1
      fi

      if lsmod | grep -q '^nvidia'; then
        echo "Module loaded:  yes"
      else
        echo "Module loaded:  NO"
        ok=1
      fi

      if nvidia-smi >/dev/null 2>&1; then
        echo "nvidia-smi:     works"
      else
        echo "nvidia-smi:     FAILED"
        ok=1
      fi

      return "$ok"
    }

USING IT
--------
    source ~/.bashrc
    gpucheck
    echo "status: $?"

A healthy machine prints the kernel, an nvidia line from dkms status that
includes the word "installed", "Module loaded: yes", "nvidia-smi: works", and
"status: 0". Any failure sets the return status to 1, so it also works in
scripts:  gpucheck || echo "fix the GPU before continuing".

HOW IT WORKS
------------
- dkms status prints one line per module and kernel. grep -i nvidia both
  shows the lines and tells the if whether there were any.
- lsmod lists loaded kernel modules; the nvidia module's name starts with
  "nvidia".
- nvidia-smi exits non-zero if it cannot talk to the driver, so redirecting
  its output away and testing the status is enough.
- The function keeps going after a failure and reports everything, rather
  than stopping at the first problem, so one run shows the whole picture.

WHEN TO RUN IT
--------------
1. Right after installing the driver and rebooting (proves it worked).
2. After every apt full-upgrade that installed a new kernel, BEFORE
   rebooting. At that point dkms status should show an nvidia entry marked
   installed for the NEW kernel. The "Module loaded" and nvidia-smi lines will
   describe the kernel you are still running, which is fine; the dkms line is
   the one that matters at this stage.
3. Again after the reboot, to confirm the new kernel really works.

If step 2 shows the new kernel has no installed nvidia entry, do not reboot
into it. Your old kernel is still installed and still works.

NOTE
----
This is a health check, not a repair tool. It tells you that something is
wrong and roughly where; the troubleshooting table in the chapter tells you
what to do about it.

WHY THIS WORKS AS AN ANSWER
---------------------------
It turns the chapter's advice ("check dkms status before rebooting") into a
single command you will actually remember to run, with a return status so
other scripts (such as the Chapter 10 provisioning script) can use it.