panic EARLY in boot (XPsp2 Atom N270); can't scroll back

[Jjanel]

I'm using latest vb, on an OLD eeePC 1000HA 2G ram XP (sp2; won't sp3).
MANY [latest] Linux'es run ok (Ubuntu, Tinycore, ...)
BUT both slacko (puppy) and Android x86 (latest) panic **very early** in booting.
Mostly all I can see on the console screen is the common/GENERIC:
Kernel Panic - not syncing: Attempted to kill init!
I tried Shift-PgUp etc but no effect (panics EARLY in the boot process)

1) Is there any way to capture the messages leading to the panic?
(I couldn't figure out a note about using serial port to a net port capture)
2) IF someone has such a system (I think Qemu has a N270), could they try these two distros.

TIA! p.s. I did try: pfix=ram and acpi=off (didn't help)
me at live (hotmail/outlook alt. com)

[mpack]

Please post the log file for a problem VM. With the VM fully shut down, right click it in GUI. Select "Show Log" and save "VBox.log" (ONLY) to a zip file. Attach the zip here.

[Jjanel]

attached.

[mpack]

You've enabled 3D acceleration but only provided 12MB of graphics RAM. Those two things probably don't go together. I'd turn off 3D acceleration and increase graphics RAM to 32MB.

Also, I'd say that 996MB is too much for this host (which only has 2GB to start with) to allocate to a VM. I'd stick to 512MB or less for a VM.

I see you chose the "Linux 2.6" template for the guest. Surely not correct, though I have no idea what is correct for Puppy Linux. Wikipedia says that 5.7 borrow code from Ubuntu, so I guess I'd try the Ubuntu 32bit template first.

I assume you meant you disabled IO-APIC, not ACPI. You should turn it back on in the VM settings, as I believe most modern Linux's require it.

You should also do an md5 checksum test on the ISO you downloaded, to make sure that isn't corrupted.

Final note: XP is not an officially supported host for VirtualBox 5.x. The devs have said that they make a small effort not to break it - and they don't just refuse to install like many vendors do - but nor do they consider it a big priority to maintain compatibility. That said, I've had no problem myself though my XP host is much beefier.

[michaln]

There's nothing obviously wrong in the log but XP SP2? That's a bit too exotic.

[mpack]

Hmm. Well spotted. The "sp2" part hadn't registered on me, it was just noise.

[Jjanel]

Before I ramble about intel_idle.max_cstate=0 fixing [both! distro] panic
(android came up 'single' [yea!] BUT puppy *now hangs* a bit later in booting; new log of this hang attached),
let me re-focus my original question on **VBox** (vs. Linux issues!):
1) is there any way to scroll back the guest console window in VBox? (e.g. with guest 'dead'=halted)

I want to say a BIG/*HUGE thanks* to the [Oracle] people for VBox & this forum/replys!

(I did try the suggestions; VBox [guest XP process per TaskManager] doesn't seem to *take* 'much' [<<100meg] memory)
Well, I focused on the panic EIP: intel_idle [still visible without scrolling back guest console]
Googles suggested intel_idle.max_cstate=0 which WORKED [meaning just fixed panic]! BUT now: Googles of:
puppy hangs at "Making the filesystem usable" (just before where it SHOULD say depmod...)
return just a few results, most mentioning 'recent years' Puppy hangs here on VBox, some mentioning w/out VT-x.
My VBox system->acceleration tab is greyed-out (cuz my Atom N270 doesn't have VT-x).

Update: I dug out my barely-viable 1G ram/16G ssd now-old Toshiba WT7-C (Atom Z3735G 1.33GHz) Win8.1 tablet,
installed VBox, and same Puppy booted all the way to GUI. I tried turning off VBox VT-x and it hung as before. (all defaults).
(Idk if any of the zillions of kernel boot parameters will allow Puppy to boot withOUT VT-x)

NOW, I have a new 'puzzle': the [Xwin] cursor doesn't [viably] work on the tablet VBox. But that's another topic...

I'm guessing there may be NO way to scroll back my original boot panic messages (cuz the 'screen' is video, not text [duh!]).