[9fans] USB boot fails

[Florian Limberger]

Hello list,

I'm currently trying to boot a Lenovo ThinkPad T410 from a usb pen drive,
using the image from 9legacy.org, but the boot sequence stops at

sdE0: LLBA 625,142,448 sectors: WDC WD3200BEVT-08A23T1 02.01A02
WD-WXDD-WXD1A1154275 [mediachange]

with every option I try.

When booting 9pcf I come sometimes as far as

init: starting /bin/rc

but after then the first message appears and everything seems stuck.
With 9pccpuf I get

usb/hub... usb/disk... bad authentication password
bad authentication id
bad authentication domain
[... I input some values ...]
mpintrenable: multible botch irq 11, tbdf 5000000, lo 00000049, n
00000049
mpintrenable: out of choices eisa -1 isa 14 tbdf 0xc00fa00 irq 11
intrenable: couldn't enable irq 11, tbdf 0xC00FA00 for sdE (iahci)

awaitspinup: sdE0 didn't spin up after 20 seconds
sdE1: LBA 0 sectors: Optiarc DVD RW AD-7930H 1.D1 [mediachange]
time...
fossil(/def/sdXX/fossil)...version...sdE0: LLBA 625,142,448 sectors: WDC
WD3200BEVT-08A23T1 02.01A02 WD-WXDD-WXD1A1154275 [mediachange]
and then the same thing happens.

Using another USB port on the laptop, I experienced a similar error like in

http://9fans.net/archive/2012/07/123

but unlike arisawa I don't get a rc shell at the end.
On this other port, the 9pcf kernel fails with

boot: can't connect to file server: '/boot/kfs' does not exist
panic: boot process died: unknown
[...]

I did then try to add

readparts=

to the plan9.ini, but the result only is

can't update /dev/sdXX/data's devsd partition table for plan9: partition
already in use: "part plan9 32 2097152"
can't update /dev/sdXX/data's devsd partition table for 9fat: partition
already in use: "part 9fat 32 204832"
can't update /dev/sdXX/data's devsd partition table for nvram: partition
already in use: "part nvram 204832 204833"
can't update /dev/sdXX/data's devsd partition table for fscfg: partition
already in use: "part fscfg 204833 204849"
partition already in use: "part fossil 204849 2097120"
sdE0: LLBA 625,142,448 sectors: WDC WD3200BEVT-08A23T1 02.01A02
WD-WXDD-WXD1A1154275 [mediachange]

The paecpu kernel suggested by erik fails nearly instantly with

panic: e820 scan failed
mpintrenable: multible botch irq 11, tbdf 5000000, lo 00000049, n
00000049
mpintrenable: out of choices eisa -1 isa 14 tbdf 0xc00c000 irq 11
intrenable: couldn't enable irq 11, tbdf 0xC00D000 for usbehci
mpintrenable: multible botch irq 11, tbdf 5000000, lo 00000049, n
00000049
mpintrenable: out of choices eisa -1 isa 14 tbdf 0xc00e800 irq 11
intrenable: couldn't enable irq 11, tbdf 0xC00E800 for usbehci
mpintrenable: multible botch irq 11, tbdf 5000000, lo 00000049, n
00000049
mpintrenable: out of choices eisa -1 isa 14 tbdf 0xc00fa00 irq 11
intrenable: couldn't enable irq 11, tbdf 0xC00FA00 for sdE (iahci)

Is there anything more I can try?

Thanks,

flo

[Florian Limberger]

i would give the 9atom kernel a chance at this. this looks like
an issue with PUIS, and hopefully that's solved.

How would I do this? I have played with Plan 9 in qemu, but do not have a
real Plan 9 system otherwise to prepare the pen drive (which is the reason
why I use the 9legacy image in the first place).

if you still get "out of choices", you might try *acpi=1 in your
plan9.ini.
the original problem is my fault, since the ahci driver was written
before the low-power stuff was put into the ahci standard.
- erik

Thanks for the fast answer,

flo

[erik quanstrom]

Post by Florian Limberger
How would I do this? I have played with Plan 9 in qemu, but do not have a
real Plan 9 system otherwise to prepare the pen drive (which is the reason
why I use the 9legacy image in the first place).

you could copy a 9atom kernel onto the pen drive.
http://ftp.9atom.org/other/9pcf
you would then need to modify the plan9.ini to point to 9pcf
instead of the current kernel. if they are the same name, then
there is no need to change anything else.

- erik

[Florian Limberger]

Post by erik quanstrom
you could copy a 9atom kernel onto the pen drive.
http://ftp.9atom.org/other/9pcf
you would then need to modify the plan9.ini to point to 9pcf
instead of the current kernel. if they are the same name, then
there is no need to change anything else.

With this kernel I get

Post by erik quanstrom
boot: can't connect to file server: '/boot/kfs' does not exist
panic: boot process died: unknown

(on all usb ports) but the boot process seems much faster.

The *acpi=1 flag does change nothing, neither with the 9legacy nor with the
9atom kernel, but I got the 9legacy 9pcf kernel working with *nomp=1.
That is at least a start.

flo

[erik quanstrom]

Post by Florian Limberger
Hello list,
I'm currently trying to boot a Lenovo ThinkPad T410 from a usb pen drive,
using the image from 9legacy.org, but the boot sequence stops at

sdE0: LLBA 625,142,448 sectors: WDC WD3200BEVT-08A23T1 02.01A02
WD-WXDD-WXD1A1154275 [mediachange]

i would give the 9atom kernel a chance at this. this looks like
an issue with PUIS, and hopefully that's solved.

if you still get "out of choices", you might try *acpi=1 in your plan9.ini.

the original problem is my fault, since the ahci driver was written
before the low-power stuff was put into the ahci standard.

- erik

[Florian Limberger]

i would give the 9atom kernel a chance at this. this looks like
an issue with PUIS, and hopefully that's solved.

I got hold of a cd burner and burned a image of 9atom. But while copying
the
data, I get lots of

cpu0: spurious interrupt 39, last 11

all over the screen (sometimes last 12). There does not seem to be any
progress. Is there a possibility to manipulate the boot options when
booting
from cd?

if you still get "out of choices", you might try *acpi=1 in your plan9.ini.
the original problem is my fault, since the ahci driver was written
before the low-power stuff was put into the ahci standard.
- erik

Greetings,

flo