kernel compiling in woof-ce

Under development: PCMCIA, wireless, etc.
Message
Author
stemsee

#481 Post by stemsee »

Thanks DF

Do you prefer the layout?

I tried to fix the links for build and source, actually took the original code from build.sh, but not quite right.

Maybe the solution to the date stamp would be to use +year only, that way your error from compiling over midnight could only be a problem once a year ... lol!

Aufs-utils problem for me was that internet disconnected during compile, after which aufs-utils should get compiled, probably should get the git locally at the start.

I wrapped the script so that it continues regardless of errors, knowing that a few manual post compile manoeuvres results in a usable package.

cheers
stemsee

stemsee

#482 Post by stemsee »

Here is huge-kernel package 4.15.8 x64
Intel AMD
https://drive.google.com/file/d/1S4YUzU ... pru9oeUCmE

Also a fixed ubuild.sh script that built the above package
Attachments
ubuild.sh.gz
fake gz
(65.28 KiB) Downloaded 210 times

User avatar
Revolverve
Posts: 255
Joined: Sat 08 Nov 2008, 21:01
Location: 45°17'28.8"N 72°16'08.8"W_avatar/ mira.ca

#483 Post by Revolverve »

Thanks stemsee,

Kernel 4.15.8 is the second enabling my Motion j3500 bezel buttons to work without acpi=off boot argument,funky acpi / close-source hardware...
Nice fat work!
Why that size?

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

#484 Post by Dry Falls »

Why that size?
Firmware in modules.sfs is too big. Accounts for most of the size. Replacing the firmware directory in lib with edit-sfs, the only problem I encountered was that LINUXPWRBTTN is not enabled by the kernel, so I couldn't shutdown.

df

stemsee

#485 Post by stemsee »

Dry Falls wrote:LINUXPWRBTTN is not enabled by the kernel, so I couldn't shutdown.
df
Oops!

User avatar
Revolverve
Posts: 255
Joined: Sat 08 Nov 2008, 21:01
Location: 45°17'28.8"N 72°16'08.8"W_avatar/ mira.ca

#486 Post by Revolverve »

Bump...it does shutdown here, in an unusual way...yep
Could it explain AT Translated Set 2 keyboard(isa0060/serio0/input0--bezel button) being recognized as first input device..and working..
Never saw before,on reboot,
wacom devices (embeded usb)
and PC Speaker(isa0061/input1) are given different input order/rank.

Anybody know if there's a way to specify the order at which kernel load and/or could give priority to an input device? edit:udev maybe.

stemsee

#487 Post by stemsee »

kernel 4.15.14 x64 is running very well. Here are the links for sources and kernel and modules
https://drive.google.com/drive/folders/ ... B_fsweQuxB

User avatar
Revolverve
Posts: 255
Joined: Sat 08 Nov 2008, 21:01
Location: 45°17'28.8"N 72°16'08.8"W_avatar/ mira.ca

#488 Post by Revolverve »

4.15.14 got
LINUXPWRBTTN not enabled by the kernel...again

stemsee,on what do you test kernel?
else then unusual shutdown with hardware button it does look fine and make bezel buttons recognised ,here...

Could we see another rt build test one day?
or ./configure --with-realtime=uspace?
Meanwhile ,if in audio x64 need,
grab jamesbond 4.4.52-rt63 or ask lowt3ch if he would share is labour...

Thanks Mr.C,up to date tested all your build .

stemsee

#489 Post by stemsee »

On my lenovo core i7 5500U laptop I get instant off when pressing powerbutton ... yes too quick ... no final save ... no spin down (fine if completely in ram)!

I did look for the LINUXPWRBTTN option, briefly, in drivers .. I grepped for it but nowhere to be found. Where in the config sections is it?

User avatar
Revolverve
Posts: 255
Joined: Sat 08 Nov 2008, 21:01
Location: 45°17'28.8"N 72°16'08.8"W_avatar/ mira.ca

#490 Post by Revolverve »

Same here for shutdown,sorry no kernel compiling knowledge .
from my reading,
/etc/acpi/actions/acpi_poweroff.sh manage pwr button.
Could LINUXPWRBTTN be puppy specific?
more fun,google it! only result found are here,in this forum thread...
hummm..neither inside any system file script..
Barry or shinobar must have a clue,in japanese,auslan maybe...

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

#491 Post by Dry Falls »

I did look for the LINUXPWRBTTN option
I thought it was an old kernel parameter but more likely part of an older xorg (there was an evdevpwrbttn). The LNXPWBTTN error message is itself an error. The problem, I think, is in the acpi configuration in the kernel. I'll try and track it down.

df

ps., BarryK's compilation of 4.14.* behaved the same (bad shutdown)

User avatar
Revolverve
Posts: 255
Joined: Sat 08 Nov 2008, 21:01
Location: 45°17'28.8"N 72°16'08.8"W_avatar/ mira.ca

#492 Post by Revolverve »

yes..lnxpwrbtn give more sense already..

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

acpi config

#493 Post by Dry Falls »

Here are, I think, the relevant sections from your dotconfig (k4.15.14) and mine (k4.14.27)
(# Power management and ACPI options):
Attachments
4.15.14.gz
fakegz
(1.85 KiB) Downloaded 166 times
4.14.27.gz
(1.78 KiB) Downloaded 182 times

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

#494 Post by Dry Falls »

Same section from Peebee's latest (k4.16.1) which I think is a woofCE/fatdog config:
Attachments
4.16.1.gz
fakegz
(1.86 KiB) Downloaded 162 times

stemsee

#495 Post by stemsee »

k-4.16.3 x86_64 vmlinuz kernel-modules.sfs kernel_sources*.sfs

https://drive.google.com/drive/folders/ ... WibtFghQF5

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

#496 Post by Dry Falls »

vis k4.16.3, everyhing works except...

shutdown. This may be the culprit: (line 536 or so)

Code: Select all

# CONFIG_NUMA is not set
df

stemsee

#497 Post by stemsee »

Thanks DF

I respin one now!

stemsee

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

#498 Post by Dry Falls »

np. However, this kernel itself is no good for me. I' running from your configuration now, but I've tried three different compilations and each one throws this error in syslog:

Code: Select all

dbus-daemon[21432]: Cannot setup inotify for '/usr/share/dbus-1/system.d'; error 'Permission denied'
Probably fine for a normal pup, but JL64 has polkit and consolekit2 which are required not only for pulse, but running other DE than openbox or JWM. About 20 fewer processses running.

To make sure it's not just my build environment, I tried under another lighthouse I know to be good. No diff.

Have to put in the aufs tools manually. Acording to the readme in the aufs sources package, this probably is because there isn't a package available as of yet. He says the previous one should be ok.

My problem is the kernel itself. It's like my old bank. They were so paranoid about security, the upgrades prevented even regular users from gaining access to their own accounts. Needless to say, everyone (welll, enough anyway) switched over to the local credit union and that franchise closed down.

Personally, I'm sticking with the less-new-and-improved kernels. Funny, but k4.16.1 didn't have this issue. Of the new batch of kernels, I've had the most luck wirh k4.15.8.

thanks, This is a real learning experience.

df

stemsee

#499 Post by stemsee »

Thanks for input DF, I will try compiling next one on your jlh.

But for now here is huge k4.16.5-x86_64
https://drive.google.com/open?id=1czFwV ... gYoSbB7wXs

sources
https://drive.google.com/open?id=1rfuJf ... WbJPcq3Hf7

Dry Falls
Posts: 616
Joined: Tue 16 Dec 2014, 23:37
Location: Upper Columbia

run levels

#500 Post by Dry Falls »

As near as I can tell, this kernel (> k4.16.2) is, independent of the DOTconfig used, designed for the big boys who don't use busybox (post X) and do use runlevels. Ie, root autologin is taboo. I've seen no documentation, but this is what's happening for me. I think it must be for further security. As well, this kernel doesn't load my usb keyboard but leaves it for xorg/sysinit to do and at shutdown, the late-loaded keyboard is dropped so there is no response available to console dialogs (this is also independent of firmware). A hunch anyway. I think my previous bank anecdote is a spot-on analogy.

df

Post Reply