unofficial linux-ck repo with cpu optimized packages is ONLINE

graysky · 2013-02-15 23:17:25

lano1106 wrote:

Hi Graysky,
I am using your linux-ck-atom build and I just noticed that CONFIG_X86_GENERIC was enabled which has the following effect to the compilation:
Makefile_32.cpu:cflags-$(CONFIG_X86_GENERIC) += $(call tune,generic,$(call tune,i686))
IMO, this might slightly defeat the purpose of having very targeted binaries. All 32 bits targets are possibly affected by that....

Interesting observation. If one manually selects any CPU optimization when building under i686, the resulting .config does indeed have CONFIG_X86_GENERIC=y as you pointed out. I am not knowledgeable enough to know what overrides what... perhaps you should considering posting your observation to lkml for feedback? If you do, please paste the link to your post in here so I can follow it.

blackout23 · 2013-02-17 10:14:56

I just pulled nvidia-304xx-ck-kx 304.64-10. I wonder what changes with these upgrades that go from -9 to -10? It is still the same codebase I would assume, because nvidia has not released anything newer for 304 branch.

graysky · 2013-02-17 10:32:31

blackout23 wrote:

I just pulled nvidia-304xx-ck-kx 304.64-10. I wonder what changes with these upgrades that go from -9 to -10?

http://pkgbuild.com/git/aur-mirror.git/ … d7a7eb0604

Jristz · 2013-02-20 05:30:04

uname -a

Linux Samsung 3.7.9-1-ck #1 SMP PREEMPT Sun Feb 17 18:22:16 EST 2013 x86_64 GNU/Linux

ck-atom

Problem

feb 19 17:58:14 Samsung kernel: ------------[ cut here ]------------
feb 19 17:58:14 Samsung kernel: WARNING: at drivers/gpu/drm/gma500/gtt.c:383 psb_gtt_free_range+0x41/0x80 [gma500_gfx]()
feb 19 17:58:14 Samsung kernel: Hardware name: N100SP
feb 19 17:58:14 Samsung kernel: Modules linked in: nf_conntrack_netbios_ns nf_conntrack_broadcast ipt_MASQUERADE xt_tcpudp ip6
feb 19 17:58:14 Samsung kernel:  cfg80211 psmouse evdev serio_raw pcspkr battery processor button ac thermal rfkill video micr
feb 19 17:58:14 Samsung kernel: Pid: 1191, comm: X Tainted: G           O 3.7.9-1-ck #1
feb 19 17:58:14 Samsung kernel: Call Trace:
feb 19 17:58:14 Samsung kernel:  [<ffffffff810577e1>] warn_slowpath_common+0x81/0xc0
feb 19 17:58:14 Samsung kernel:  [<ffffffff8105783b>] warn_slowpath_null+0x1b/0x20
feb 19 17:58:14 Samsung kernel:  [<ffffffffa07d0671>] psb_gtt_free_range+0x41/0x80 [gma500_gfx]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa07cfa31>] psb_gem_free_object+0x41/0x50 [gma500_gfx]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa0773eea>] drm_gem_object_free+0x2a/0x30 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa0774128>] drm_gem_handle_delete+0x118/0x140 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa07cfb80>] psb_gem_dumb_destroy+0x10/0x20 [gma500_gfx]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa0784669>] drm_mode_destroy_dumb_ioctl+0x29/0x40 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa077240c>] drm_ioctl+0x4cc/0x570 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffff8114d9b5>] ? free_pages_and_swap_cache+0x25/0xd0
feb 19 17:58:14 Samsung kernel:  [<ffffffffa0784640>] ? drm_mode_mmap_dumb_ioctl+0x40/0x40 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffff814aeb6e>] ? __mutex_lock_slowpath+0x25e/0x350
feb 19 17:58:14 Samsung kernel:  [<ffffffffa07795cb>] ? drm_vm_close_locked+0xbb/0xc0 [drm]
feb 19 17:58:14 Samsung kernel:  [<ffffffffa07d4aa3>] psb_unlocked_ioctl+0x43/0xa0 [gma500_gfx]
feb 19 17:58:14 Samsung kernel:  [<ffffffff8115d1f9>] ? kmem_cache_free+0x129/0x180
feb 19 17:58:14 Samsung kernel:  [<ffffffff81181fe7>] do_vfs_ioctl+0x97/0x530
feb 19 17:58:14 Samsung kernel:  [<ffffffff81140884>] ? remove_vma+0x54/0x70
feb 19 17:58:14 Samsung kernel:  [<ffffffff811421dc>] ? do_munmap+0x33c/0x410
feb 19 17:58:14 Samsung kernel:  [<ffffffff8118251b>] sys_ioctl+0x9b/0xb0
feb 19 17:58:14 Samsung kernel:  [<ffffffff8114230e>] ? vm_munmap+0x5e/0x80
feb 19 17:58:14 Samsung kernel:  [<ffffffff814b925d>] system_call_fastpath+0x1a/0x1f
feb 19 17:58:14 Samsung kernel: ---[ end trace 13a5003f1486333a ]---

Tests maked:
Boot in LTS: not support for gma500, nothing to do here
Boot ARCH: number of times 5; nothing happend exept the welll know

feb 19 00:02:57 Samsung kernel: gma500 0000:00:02.0: GPU: power management timed out.
feb 19 00:02:58 Samsung kernel: gma500 0000:00:02.0: trying to get vblank count for disabled pipe 0
feb 19 00:02:58 Samsung kernel: gma500 0000:00:02.0: trying to get vblank count for disabled pipe 0

all harmless (as far I know) and happen in all kernels from 3.5.1 and up in all kernel variants
Boot in ck-atom: Happen in every boot, at now I not have problem in my modesetting, but I consterned if this is really harmfull at long time or not, I want know about it

any idea or more info is nees?

Last edited by Jristz (2013-02-20 05:30:56)

graysky · 2013-02-20 11:46:41

@Jristz -

1) Does this happen with the ck-generic kernel?
2) Post the output of `cat /proc/cpuinfo`

twilight0 · 2013-02-22 16:20:55

Graysky, gpg keys are not accepted from pacman, dunno why, followed the wiki article for reseting the keys and still to no avail, I had to manually install them from /var/cache

graysky · 2013-02-22 16:26:45

@twilight0 - Odd. Can you post the error you are getting plz?

Jristz · 2013-02-22 18:28:30

yes happen to with ck-generic

➜ ~ cat /proc/cpuinfo

processor	: 0
vendor_id	: GenuineIntel
cpu family	: 6
model		: 54
model name	: Intel(R) Atom(TM) CPU N2100   @ 1.60GHz
stepping	: 1
microcode	: 0x10d
cpu MHz		: 600.000
cache size	: 512 KB
physical id	: 0
siblings	: 1
core id		: 0
cpu cores	: 1
apicid		: 0
initial apicid	: 0
fpu		: yes
fpu_exception	: yes
cpuid level	: 10
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts nopl nonstop_tsc aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 cx16 xtpr pdcm movbe lahf_lm arat dtherm
bogomips	: 3193.02
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 48 bits virtual
power management:

processor	: 1
vendor_id	: GenuineIntel
cpu family	: 6
model		: 54
model name	: Intel(R) Atom(TM) CPU N2100   @ 1.60GHz
stepping	: 1
microcode	: 0x10d
cpu MHz		: 600.000
cache size	: 512 KB
physical id	: 0
siblings	: 1
core id		: 0
cpu cores	: 0
apicid		: 1
initial apicid	: 1
fpu		: yes
fpu_exception	: yes
cpuid level	: 10
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts nopl nonstop_tsc aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 cx16 xtpr pdcm movbe lahf_lm arat dtherm
bogomips	: 3193.02
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 48 bits virtual
power management:

I forget mention that this is a Kernel Oops

twilight0 · 2013-02-22 19:33:58

graysky wrote:

@twilight0 - Odd. Can you post the error you are getting plz?

Sure, why not:

[root@DellXPS ~]# pacman -S linux-ck-sandybridge linux-ck-sandybridge-headers nvidia-ck-sandybridge
:: linux-ck-sandybridge is in IgnorePkg/IgnoreGroup. Install anyway? [Y/n] y
warning: linux-ck-sandybridge-3.7.9-1 is up to date -- reinstalling
:: linux-ck-sandybridge-headers is in IgnorePkg/IgnoreGroup. Install anyway? [Y/n] y
warning: linux-ck-sandybridge-headers-3.7.9-1 is up to date -- reinstalling
:: nvidia-ck-sandybridge is in IgnorePkg/IgnoreGroup. Install anyway? [Y/n] y
warning: nvidia-ck-sandybridge-313.18-4 is up to date -- reinstalling
resolving dependencies...
looking for inter-conflicts...

Targets (3): linux-ck-sandybridge-3.7.9-1  linux-ck-sandybridge-headers-3.7.9-1  nvidia-ck-sandybridge-313.18-4

Total Installed Size:   102.91 MiB
Net Upgrade Size:       0.00 MiB

Proceed with installation? [Y/n] y
(3/3) checking package integrity                                                           [#####################################################] 100%
error: linux-ck-sandybridge: key "88A032865EE46C4C" is unknown
:: Import PGP key 5EE46C4C, "graysky (used to sign repo-ck packages) <graysky@archlinux.us>", created 2012-12-11? [Y/n] y
(3/3) checking package integrity                                                           [#####################################################] 100%
error: linux-ck-sandybridge: signature from "graysky (used to sign repo-ck packages) <graysky@archlinux.us>" is unknown trust
error: linux-ck-sandybridge-headers: signature from "graysky (used to sign repo-ck packages) <graysky@archlinux.us>" is unknown trust
error: nvidia-ck-sandybridge: signature from "graysky (used to sign repo-ck packages) <graysky@archlinux.us>" is unknown trust
error: failed to commit transaction (invalid or corrupted package (PGP signature))
Errors occurred, no packages were upgraded.

graysky · 2013-02-22 19:56:43

@Jristz - If the bug happens with any linux-ck package as you stated, and not with the ARCH kernel, you must conclude that is caused by something in the linux-ck packages. I do not have any insight to help you debug

@twilight0 - Sign my key and try again:

# pacman-key --lsign-key 5EE46C4C

Scimmia · 2013-02-26 06:13:48

Huh? Chromium is now in this repo?

Wouldn't be a big deal, but having it as replaces=(chromium) means that everyone that has this repo enabled will get a prompt to replace chromium with chromium-scroll-pixels when they just do a simple -Syu. Is this really what you intended?

Last edited by Scimmia (2013-02-26 06:37:18)

graysky · 2013-02-26 08:50:29

Scimmia wrote:

Huh? Chromium is now in this repo?
Wouldn't be a big deal, but having it as replaces=(chromium) means that everyone that has this repo enabled will get a prompt to replace chromium with chromium-scroll-pixels when they just do a simple -Syu. Is this really what you intended?

This was not intended. Thanks for bring it to my attention. Fixed.

Related: https://wiki.archlinux.org/index.php/Ch … ouse_wheel

Last edited by graysky (2013-02-26 09:11:25)

BaloneyGeek · 2013-02-27 19:37:04

Hi Graysky, been using your kernels (core2) since the kernel26 days. I've also started using ZFS lately, and I kinda have a request - can you build the zfs package (from AUR) against your kernels? It's a real chore uninstalling the zfs package, upgrading the kernel and rebuilding zfs everytime there's a new kernel. I hope I'm not asking for too much (or am not going against some policy to not provide AUR packages or something).

graysky · 2013-02-27 19:52:34

@BG - I quickly looked in the AUR and it appears that ZFS is an add-on to the kernel, in other words, not something that has to be compiled in to use. Can you simply adapt the PKGBUILD for it to require linux-ck and linux-ck-headers and build it yourself? I don't want to take on additional packages particularly if they are not for a majority of users.

BaloneyGeek · 2013-02-28 11:34:21

@Graysky - I know, they're loadable kernel modules. I can build them of course, I just asked if you could make them so that I wouldn't be stuck without ZFS every time the kernel updated. If it's too much trouble, please don't do it. I'm cool.

graysky · 2013-03-03 19:38:18

I edited the first post in the thread with the following info. BTW, enjoy the 3.8 series (hint `pacman -Syu`):

graysky wrote:

UPDATE: 03-Mar-2013
Since some user wish to have access to 'archived' packages, I have decided to honor this request, when a new 2-pt release comes out, I will rotate the previous one into the archive on the repo. Note that this isn't meant for you to use in /etc/pacman.conf, rather, it is meant for you to manually access the old packages if you wish.
For example, 3.8.1-1 went live an hour or so ago. If from some reason you don't feel that it is meeting your needs, you may downgrade to the 3.7.10-1 release (and all associated packages like nvidia-ck, broadcom-wl-ck, etc.). Note that the archive isn't meant to be a historical source of all version. It will just hold the most recent in the last 2-pt family.

MrCode · 2013-03-03 19:51:32

It looks as though virtualbox-ck-host-modules is behind:

error: failed to prepare transaction (could not satisfy dependencies)
:: virtualbox-ck-host-modules-nehalem: requires linux-ck-nehalem<3.8

graysky · 2013-03-03 20:22:12

@MrC - Works for me... sure you're sync'ed?

% sudo pacman -S ck-nehalem
:: There are 7 members in group ck-nehalem:
:: Repository repo-ck
   1) broadcom-wl-ck-nehalem  2) linux-ck-nehalem  3) linux-ck-nehalem-headers  4) nvidia-304xx-ck-nehalem
   5) nvidia-ck-nehalem  6) virtualbox-ck-guest-modules-nehalem  7) virtualbox-ck-host-modules-nehalem

Enter a selection (default=all): 7
resolving dependencies...
looking for inter-conflicts...

Targets (2): linux-ck-nehalem-3.8.1-1  virtualbox-ck-host-modules-nehalem-4.2.8-1

Total Installed Size:   60.79 MiB

Proceed with installation? [Y/n] 
(2/2) checking package integrity                                             [############################################] 100%
(2/2) loading package files                                                  [############################################] 100%
(2/2) checking for file conflicts                                            [############################################] 100%
(1/2) installing linux-ck-nehalem                                            [############################################] 100%
>>> Updating module dependencies. Please wait ...
>>> Generating initial ramdisk, using mkinitcpio.  Please wait...
==> Building image from preset: 'default'
  -> -k /boot/vmlinuz-linux-ck -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-ck.img
==> Starting build: 3.8.1-1-ck
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating lzop initcpio image: /boot/initramfs-linux-ck.img
==> Image generation successful
==> Building image from preset: 'fallback'
  -> -k /boot/vmlinuz-linux-ck -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-ck-fallback.img -S autodetect
==> Starting build: 3.8.1-1-ck
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating lzop initcpio image: /boot/initramfs-linux-ck-fallback.img
==> Image generation successful
Optional dependencies for linux-ck-nehalem
    crda: to set the correct wireless channels of your country
    lirc-ck: Linux Infrared Remote Control kernel modules for linux-ck
    nvidia-ck: nVidia drivers for linux-ck
    nvidia-beta-ck: nVidia beta drivers for linux-ck
    modprobed_db: Keeps track of EVERY kernel module that has ever been probed - useful for those of us who make localmodconfig
(2/2) installing virtualbox-ck-host-modules-nehalem                          [############################################] 100%
===> You must load vboxdrv module before starting VirtualBox:
===> # modprobe vboxdrv

Last edited by graysky (2013-03-03 20:23:44)

Scimmia · 2013-03-03 20:25:14

I'm getting the same thing, fully synced, but with the core2 version. forget to increment the pkgrel, maybe?

Last edited by Scimmia (2013-03-03 20:29:53)

graysky · 2013-03-03 20:38:15

:puzzled:

If I look at the .PKGINFO inside virtualbox-ck-host-modules-nehalem-4.2.8-1-x86_64.pkg.tar.xz for example:

# Generated by makepkg 4.0.3
# using fakeroot version 1.18.4
# Sun Mar  3 17:36:18 UTC 2013
...
depend = linux-ck-nehalem>=3.8
depend = linux-ck-nehalem<3.9
provides = virtualbox-host-modules

I just reindexed. Please `pacman -Syyu` and let me know if this solves the problem.

Example:

repo-ck.db.tar.xz/virtualbox-ck-host-modules-nehalem-4.2.8-1/
%DEPENDS%
linux-ck-nehalem>=3.8
linux-ck-nehalem<3.9

%CONFLICTS%
virtualbox-ck-host-modules-atom
virtualbox-ck-host-modules-barcelona
virtualbox-ck-host-modules-bulldozer
virtualbox-ck-host-modules-corex
virtualbox-ck-host-modules-core2
virtualbox-ck-host-modules-haswell
virtualbox-ck-host-modules-ivybridge
virtualbox-ck-host-modules-kx
virtualbox-ck-host-modules-k10
virtualbox-ck-host-modules
virtualbox-ck-host-modules-p4
virtualbox-ck-host-modules-piledriver
virtualbox-ck-host-modules-pentm
virtualbox-ck-host-modules-sandybridge

%PROVIDES%
virtualbox-host-modules

Scimmia · 2013-03-03 20:41:30

OK, I'll be less subtle.

YOU FORGOT TO INCREMENT THE PKGREL!!!

If you already have 4.2.8-1 installed, pacman won't try to update it, so you still have the old version that was built for 3.7.10. The rebuilt one for 3.8.1 needs to be 4.2.8-2 so that pacman will update it.

graysky · 2013-03-03 20:54:38

Scimmia wrote:

OK, I'll be less subtle.
YOU FORGOT TO INCREMENT THE PKGREL!!!

Well f*ck - I sure did. Fixed and thanks.

MrCode · 2013-03-03 21:29:01

LOL, thanks graysky. All does appear to be well, now.

qinohe · 2013-03-04 10:00:47

Hi graysky,

Something went wrong after latest kernel upgrade.
I try to reinstall the following

packer -S nvidia-utils-bumblebee nvidia-ck-bumblebee

Now it installs all needed packages except 'nvidia-ck-bumblebee'
I get this error;

cat: /lib/modules/extramodules-3.7-ck/version: No such file or directory

Now I have already installed 3.8.1-1-ck and did a sync.

graysky · 2013-03-04 10:46:42

Your problem is that those bumblebee packages are out-of-date. I do not maintain them; check with whomever does and ask him/her to update.

Arch Linux

#1251 2013-02-15 23:17:25

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1252 2013-02-17 10:14:56

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1253 2013-02-17 10:32:31

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1254 2013-02-20 05:30:04

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1255 2013-02-20 11:46:41

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1256 2013-02-22 16:20:55

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1257 2013-02-22 16:26:45

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1258 2013-02-22 18:28:30

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1259 2013-02-22 19:33:58

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1260 2013-02-22 19:56:43

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1261 2013-02-26 06:13:48

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1262 2013-02-26 08:50:29

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1263 2013-02-27 19:37:04

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1264 2013-02-27 19:52:34

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1265 2013-02-28 11:34:21

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1266 2013-03-03 19:38:18

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1267 2013-03-03 19:51:32

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1268 2013-03-03 20:22:12

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1269 2013-03-03 20:25:14

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1270 2013-03-03 20:38:15

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1271 2013-03-03 20:41:30

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1272 2013-03-03 20:54:38

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1273 2013-03-03 21:29:01

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1274 2013-03-04 10:00:47

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1275 2013-03-04 10:46:42

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

Board footer