linux kernel 4.9.6 startup on skylake "Blocked a compatibility format interrupt request"

linux kernel 4.9.6 startup on skylake "Blocked a compatibility format interrupt request"

[Luke Kenneth Casson Leighton]

hi just a heads-up on trying out debian's 4.9 kernel on an awesomely
powerful recent laptop, it's unable to run /sbin/init.  full report is
here:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=855558

screenshot is here:
http://hands.com/~lkcl/IMG_20170220_1127225_rewind.jpg

this error appears to be quite common, including this:
https://lkml.org/lkml/2016/8/26/196

but that was during startup of xorg - this is /sbin/init (from inside
debian's initramdisk)

just an fyi although advice which helps to track this down much appreciated,

thanks.

---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

Re: linux kernel 4.9.6 startup on skylake "Blocked a compatibility format interrupt request"

[Luke Kenneth Casson Leighton]

https://forum.manjaro.org/t/acpi-error-on-boot-after-updating-to-4-9/12894/3

weell, this is the oddest regression "fixed" with a workaround
recommended from the archlinux link on ACPI DSDT:
https://wiki.archlinux.org/index.php/DSDT

by setting acpi_os_name="Windows 2009" which, after disassembling the
ACPI DSDT code i noted that it was slightly different handling from
windows > 2012 and windows < 2009, i got a successful boot.

andrew thank you for noting that /init appeared not to exist: ben also
noted the same.  it was (is) definitely there: lsinitramfs showed that
it definitely, definitely exists.  which means that the above ACPI
weirdness is causing the linux kernel to *BELIEVE* that it doesn't
exist.

i appreciate that things are moving on at a fast pace in the linux
kernel, and it's great to see support for shit-hot hardware like the
stuff i managed to justify getting (8-core i7 overclockable to 4ghz,
16gb of 2400mhz DDR4 RAM, a 2500 *megabyte* per second NVMe SSD...
3200x1800 LCD and this is in a *laptop*...)....

... but i must apologise for saying this: because of its price-tag
there is absolutely no fucking way in seven hells i am doing *ANY*
kind of BIOS update, ACPI decompile-then-hand-edit-then-recompile
strategy.  i'm not even going to let the *manufacturer* do a BIOS
update even if it has to ever go for a warranty repair.

point is: whilst i have a workaround, it would be nice to see some of
these ACPI hiccups go away.  anything i can do (which does NOT involve
BIOS or ACPI firmware updates) to help make that happen i'm happy to
help.  it's not hugely stable hardware so is rebooting every 36-72
hours anyway (s2ram hanging either on powerdown or powerup is the
usual culprit), so i am fine with compiling kernels (under 20 minutes!
yippee!) and trying out random stuff.

l.