After the death of t43, an IBM Thinkpad T43, I acquired this notebook computer.
In the store, it was $399 raw price, with $60 discount.
I started the install at about 9:30pm, had to twiddle a bit to get BIOS into legacy mode so I could do the install; it was avidly UEFI, initially. Apparently it comes with Windows 8, 64 bit, as the default, which won't work unless it's tightly integrated into UEFI, so that there's the full chain of vendor controls over video and sound. They wouldn't want anyone to be able to use Windows 8 to "pirate" movies; priority #1!.
Got it to the point of thinking about loading packages across the network at about 10:30-11. It then spent into the wee hours loading Debian Squeeze packages (some 1300 of them).
In the morning, I rebooted, threw in a couple packages to start bootstrapping configuration (I want git, zsh, cfengine for such), and got it to the point of having *some* of what I want on it.
Wireless wasn't up yet, needs a blob that looks like it's in Debian, likely some of firmware-linux-nonfree, firmware-ralink, firmware-realtek. TBD.
It then ran an "apt-upgrade" from 6.0.5 (what I had a net install disk handy for) to "testing", which, operating on 1300-ish packages, takes a good portion of time. Next steps were:
Upgrading to a 3.2 kernel
Installing the nonfree firmware package, and turning on wireless
A bit more DHCP cleanup, including setting up other things on network to know about host "hpaq"
Pull a modern emacs onto it
Set up Aurora/Firefox (I downloaded it; didn't have enough libraries yet last night to be able to run it)
apt-get install etckeeper
It has 4GB of memory, expandable to 8GB, which cost ~$60. It was very slightly tricky figuring out how to get at the memory slot; it hides behind the panel immediately "below" the battery. Removing the panel required:
Unscrewing the one screw visible in the battery compartment, along the "top" of the panel. Note that the screw will not come all the way out; it stays attached to the panel.
Once the screw has been loosened, fingernail pressure will pull the panel off; it is held in by small tabs.
There are two memory slots; only the inner one was occupied.
No change to BIOS or the like was needed to recognize the 8GB of RAM; it was automatically recognized upon reboot.
The one thing I'm noticing as a hardware-ish irritation is that the touchpad is more sensitive than I'd like. Just a matter of taste.
Memory Type: DDR3 PC3-12800, DDR3 PC3-10600, DDR3 (non-ECC)
Maximum Memory: 8GB
A trouble I faced when trying to upgrade from Linux 2.6.32 to 3.2 or 3.8 was that the latter kernels would lead to the screen heading to black shortly after boot time, and the screen thereafter declining to show anything. The machine was running and otherwise functioning, as I was able to log in remotely.
Web searches looking for similar troubles suggested ACPI configuration troubles that could be addressed by modifying /etc/default/grub, assortedly:
Which seemed to have no effect, hence I removed it.
Which seemed to have no effect.
Which did have the desired effect, in conjunction with the previous.
This seems to relate to control over screen brightness; apparently if ACPI is not "told" to do/ignore this, upon loading Linux, the screen will get shut off.
Debugging Backlight on the Ubuntu wiki appears to best describe what is going on.
Once the backlight configuration is set as described (with acpi parms passed to Grub), brightness configuration can be found in /sys/class/backlight/intel_backlight; the following are interesting settings:
Gets me 4882, the maximum permissible brightness
echo 0 > /sys/class/backlight/intel_backlight/brightness
Setting brightness to 0 turns out rather unhappily
echo 4882 > /sys/class/backlight/intel_backlight/brightness
Setting brightness to 4882 is the starting default
echo 50 > /sys/class/backlight/intel_backlight/brightness
Setting brightness to 50 leaves the screen quite dim. The lowest value that seems reasonably readable is around 10.
Sadly, the built-in wireless network adapter declines to function, and for a particularly sad reason.
Between having modern enough kernel and the Ralink firmware package installed, it is all properly recognized.
root@hpaq:~# lspci -v | grep -10 01:00.0 | tail -11 01:00.0 Network controller: Ralink corp. Device 539b Subsystem: Hewlett-Packard Company Device 18ed Flags: bus master, fast devsel, latency 0, IRQ 16 Memory at c2500000 (32-bit, non-prefetchable) [size=64K] Capabilities:  Power Management version 3 Capabilities:  MSI: Enable- Count=1/32 Maskable- 64bit+ Capabilities:  Express Endpoint, MSI 00 Capabilities:  Advanced Error Reporting Capabilities:  Device Serial Number 00-00-7b-66-63-31-17-a4 Kernel driver in use: rt2800pci
(By the way, see this OpenSuSE link on Ralink wireless driver help ; it indicates someone else surmounted related problems on the very same hardware.)
Unfortunately, the drivers are not enough. There is a "soft key" control (that calls itself a "hardware switch" control, which is a bit of a lie) for activating/deactivating the network adaptor. On many systems, such controls come in the form of a physical hardware switch which you would turn on or off in order to activate/deactivate the network adaptor.
On this system, that control comes in the form of a Windows 8 driver which is activated upon pressing a key combination Fn-F12.
One may see the consequences, on Linux, via an application called rfkill:
oot@hpaq:~# rfkill list 1 1: hp-wifi: Wireless LAN Soft blocked: no Hard blocked: yes root@hpaq:~# rfkill unblock 1 root@hpaq:~# rfkill list 1 1: hp-wifi: Wireless LAN Soft blocked: no Hard blocked: yes
Indications are that this "hard blocked" button is controlled by a Wireless Button Driver that would be in the ACPI tree, and that it might not be too desperately difficult to disassemble the driver to figure out how to activate the button. Note that it's fairly visible; there is a little orange light on the "F12" key that indicates that it is blocked; it is expected that this light would turn green (or possibly blue) if the block were shut off.
Instead of fighting with that further at this time, I acquired a USB wireless network adaptor, specifically a TP-Link TL-WN727N, which happens to use the very same driver, and which worked with little further ado, just via adding the basic relevant additions to /etc/network/interfaces
Suggested possible solution: Fixing Wireless on the HP Folio 13 when using Linux .
The idea: Deactivate ACPI briefly during bootup in order to allow rfkill to work.
Unblock wifi and bluetooth and remove hp-wmi (the ACPI module?) early after boot
Reload then unload hp-wmi as part of thaw/resume processes