View previous topic :: View next topic |
Author |
Message |
ilikenwf Tux's lil' helper
Joined: 04 Oct 2007 Posts: 119
|
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun Dec 16, 2007 7:24 pm Post subject: |
|
|
TBM wrote: | Is dynticks better than say 1000Hz for gaming on zen kernel?
Could have been nice to see a benchmark on that enabled vs disabled.
How does dynticks compare/interact with the cpu freq control?
I guess both schemes tries to save power, and deliver power (cpu) only when needed.
Is both enabled at the same time recommended? Do they compete or interact?
I know that performance gouvernor is best regarding games in general for max FPS, so I guess that dynticks also will impact on FPS?
Just wondering,
Tom |
having dynticks enabled for the scheduler Hz is like having the ondemand cpufreq governor enabled for cpu Hz... dynticks not a performance option! _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
kriko Tux's lil' helper
Joined: 18 Mar 2006 Posts: 92 Location: Slovenia
|
Posted: Sun Dec 16, 2007 7:25 pm Post subject: |
|
|
rmh3093 wrote: | TBM wrote: | Is dynticks better than say 1000Hz for gaming on zen kernel?
Could have been nice to see a benchmark on that enabled vs disabled.
How does dynticks compare/interact with the cpu freq control?
I guess both schemes tries to save power, and deliver power (cpu) only when needed.
Is both enabled at the same time recommended? Do they compete or interact?
I know that performance gouvernor is best regarding games in general for max FPS, so I guess that dynticks also will impact on FPS?
Just wondering,
Tom |
having dynticks enabled for the scheduler Hz is like having the ondemand cpufreq governor enabled for cpu Hz... dynticks not a performance option! |
Going to test that to see how it translates to real world |
|
Back to top |
|
|
kriko Tux's lil' helper
Joined: 18 Mar 2006 Posts: 92 Location: Slovenia
|
Posted: Sun Dec 16, 2007 8:33 pm Post subject: |
|
|
Disabling CONFIG_NO_HZ leads to performance loss on my system, ut2k4 showed no difference (since it is not so cpu intensive), but q4 shows some difference:
Code: | Quake 4
2.6.24-rc5-zen0 (disabled CONFIG_NO_HZ)
------------------------------------------------------------------------------------
57.7 s (52.7 fps) | 51.1 s (59.6 fps) | 51.1 s (59.5 fps) | 51.2 s (59.5 fps) |
I was skeptic about this test, so I decided to redo it -> same thing. Then I booted into other kernel with CONFIG_NO_HZ and did the test again
Code: | Quake 4
2.6.24-rc5-zen0-adrenaline (disabled CC_OPTIMIZE_FOR_SIZE, second test - just 3 runs)
----------------------------------------------------------------
53.1 s (57.6 fps) | 47.7 s (64.0 fps) | 47.9 s (63.9 fps) | |
and now if we compare this to older test of same kernel (posted before):
Code: | Quake 4
2.6.24-rc5-zen0-adrenaline (disabled CC_OPTIMIZE_FOR_SIZE)
------------------------------------------------------------------------------------
52.8 s (57.6 fps) | 47.5 s (64.0 fps) | 47.6 s (63.9 fps) | 47.6 s (63.8 fps) |
(nr. are same, +/- 0.1 fps) we can say that it was done under same circumstances and disabling CONFIG_NO_HZ leads to 6% (~4 fps) performance loss in gaming. At least on my machine.
Surprising. |
|
Back to top |
|
|
TBM n00b
Joined: 20 Nov 2007 Posts: 28 Location: Norway
|
Posted: Sun Dec 16, 2007 8:47 pm Post subject: |
|
|
Interesting!
If you are interested, you can turn off swap and swap prefetch (if you have enough memory!) and test that
Cheers,
Tom |
|
Back to top |
|
|
eatnumber1 n00b
Joined: 13 Jan 2007 Posts: 55 Location: New York
|
Posted: Sun Dec 16, 2007 9:37 pm Post subject: |
|
|
You may want to just have the early_printk function do nothing instead and still call the function. I understand that's how embedded's removal of printk works.
EDIT: Why the dependency on DEBUG_KERNEL? Also, why put it in kernel hacking? The option for regular printk is in General Setup -> Configure standard kernel features (for small systems) |
|
Back to top |
|
|
ilikenwf Tux's lil' helper
Joined: 04 Oct 2007 Posts: 119
|
Posted: Sun Dec 16, 2007 9:45 pm Post subject: |
|
|
I didn't think of that until after the fact...I can do it, though...Just wrap the function's code with ifndef....I probably should just do that, huh?
Re:EDIT: Why the dependency on DEBUG_KERNEL? Also, why put it in kernel hacking? The option for regular printk is in General Setup -> Configure standard kernel features (for small systems)
That dependency is being removed in the patch....also, kernel hacking is the place it was originally meant to be...don't ask me.... _________________ Make The Fonts Pretty In KDE
Archlinux User |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun Dec 16, 2007 10:12 pm Post subject: |
|
|
kriko wrote: | Disabling CONFIG_NO_HZ leads to performance loss on my system, ut2k4 showed no difference (since it is not so cpu intensive), but q4 shows some difference:
Code: | Quake 4
2.6.24-rc5-zen0 (disabled CONFIG_NO_HZ)
------------------------------------------------------------------------------------
57.7 s (52.7 fps) | 51.1 s (59.6 fps) | 51.1 s (59.5 fps) | 51.2 s (59.5 fps) |
I was skeptic about this test, so I decided to redo it -> same thing. Then I booted into other kernel with CONFIG_NO_HZ and did the test again
Code: | Quake 4
2.6.24-rc5-zen0-adrenaline (disabled CC_OPTIMIZE_FOR_SIZE, second test - just 3 runs)
----------------------------------------------------------------
53.1 s (57.6 fps) | 47.7 s (64.0 fps) | 47.9 s (63.9 fps) | |
and now if we compare this to older test of same kernel (posted before):
Code: | Quake 4
2.6.24-rc5-zen0-adrenaline (disabled CC_OPTIMIZE_FOR_SIZE)
------------------------------------------------------------------------------------
52.8 s (57.6 fps) | 47.5 s (64.0 fps) | 47.6 s (63.9 fps) | 47.6 s (63.8 fps) |
(nr. are same, +/- 0.1 fps) we can say that it was done under same circumstances and disabling CONFIG_NO_HZ leads to 6% (~4 fps) performance loss in gaming. At least on my machine.
Surprising. |
you need to repeat those thoses like 9 more times for each testing codition rebooting after each test and then take an average.... those number dont really look that much different _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
satanskin Guru
Joined: 25 Apr 2005 Posts: 353
|
Posted: Mon Dec 17, 2007 1:11 am Post subject: |
|
|
When using the standard ebuilds, how does one update using git pull? I tried the following:
Code: | thor zen-sources # git pull
fatal: /usr/bin/git-pull cannot be used without a working tree.
thor zen-sources # pwd
/usr/portage/distfiles/git-src/zen-sources
|
|
|
Back to top |
|
|
goffrie n00b
Joined: 14 Mar 2007 Posts: 67
|
Posted: Mon Dec 17, 2007 1:17 am Post subject: |
|
|
Quote: | When using the standard ebuilds, how does one update using git pull? |
If you are using 'zen-sources-9999', then just re-emerge it. It will automatically update.
If using any other version, you can't. The whole point of those versions is that they don't change Install 'zen-sources-9999', or if you don't want to, wait for a new version to come out with the fix, and emerge that. |
|
Back to top |
|
|
satanskin Guru
Joined: 25 Apr 2005 Posts: 353
|
Posted: Mon Dec 17, 2007 1:28 am Post subject: |
|
|
goffrie wrote: | Quote: | When using the standard ebuilds, how does one update using git pull? |
If you are using 'zen-sources-9999', then just re-emerge it. It will automatically update.
If using any other version, you can't. The whole point of those versions is that they don't change Install 'zen-sources-9999', or if you don't want to, wait for a new version to come out with the fix, and emerge that. |
Thanks. I finally got this kernel to boot, but still don't have the alsa working. Thanks again though! |
|
Back to top |
|
|
darklegion Guru
Joined: 14 Nov 2004 Posts: 468
|
Posted: Mon Dec 17, 2007 2:26 am Post subject: |
|
|
kriko wrote: | @darklegion: I tried "Fair Desktop/Server" and I don't like it (w and w/o boosting). I noticed audio skipping in some cases (changing slideshows in impress), but this was due "task boosting".
|
I forgot to mention that I use "Verynice" to raise the priority of X to -3 (from 0).It's possible that X set to 0 could cause these problems on some machines.I don't notice any skipping on my machine with these settings and 2.6.23-kamikaze5 but this may have changed with zen-sources so YMMV.Also it could be a result of hardware differences.I only started experiencing audio skipping + vmware booting when I changed from a 939 amd-x2 machine to a core2-duo machine, so maybe these interactivity boosts don't work so well on core2-duo chips (or some other piece of hardware that was changed). |
|
Back to top |
|
|
kriko Tux's lil' helper
Joined: 18 Mar 2006 Posts: 92 Location: Slovenia
|
Posted: Mon Dec 17, 2007 2:44 am Post subject: |
|
|
There were some problems when I tried to pull git, so I deleted zen-sources dir and cloned from scratch. But there are a lot of features missing like genetic, hz_8xx, cfs tunables...
Something wrong? |
|
Back to top |
|
|
HecHacker1 Apprentice
Joined: 26 Jun 2003 Posts: 213 Location: UCSD
|
Posted: Mon Dec 17, 2007 4:53 am Post subject: |
|
|
kernelOfTruth wrote: | Quote: | Anybody want to try and port it to 2.6.24? |
waninkoko is trying
dunno how far he already has progressed - and if it is possible ... |
This would be a dream come true.. I tried porting it myself once, but its definitely too much for me. I would love to benchmark CFS vs. SD on an updated kernel.. I still firmly believe SD is superior for Gaming/OpenGL (compiz) workloads, and video/audio too. SD is (was) the fairest scheduler with only 2 tunables, /proc/sys/kernel/interactive and rr_interval. While CFS has so many undocumented tunables I don't even know how to tune the damn scheduler (I had to browse through source code and lkml posts to find answers). More tunables = more complexity = inherently unfair design that has to be corrected with tunables.
Last time I did a SD vs CFS test (2.6.22), SD won by far for my usage, even if it used twice the amount of context switches! I even tried putting the CFS interval to 0, and even then trying to force more context switches, performance was worse than SD(lower FPS).
Of course CFS today is nothing like it was then. The design document is so out of date it's useless.
About NO_HZ: This could actually enhance performance because the CPU only wakes up "when it needs to." Thus reducing context switching, and also reducing power usage. It does this by calculating when its next wakeup period will be, and then puts the processor to sleep until the event is scheduled to happen. For many programs that means instead of waking up 1000Hz*5 (a thousand times a second) for say a 5 second loop, we only have to wake up at the 5 second mark. Its all math.. there really isn't any guesswork.
The Hz options (100,250,864,1000) work in conjunction with NO_HZ. With a 1000Hz tick and NO_HZ enabled, you are basically saying "perform up to 1000 events if needed, else use exactly what is needed" |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Dec 17, 2007 5:06 am Post subject: |
|
|
HecHacker1 wrote: | kernelOfTruth wrote: | Quote: | Anybody want to try and port it to 2.6.24? |
waninkoko is trying
dunno how far he already has progressed - and if it is possible ... |
This would be a dream come true.. I tried porting it myself once, but its definitely too much for me. I would love to benchmark CFS vs. SD on an updated kernel.. I still firmly believe SD is superior for Gaming/OpenGL (compiz) workloads, and video/audio too. SD is (was) the fairest scheduler with only 2 tunables, /proc/sys/kernel/interactive and rr_interval. While CFS has so many undocumented tunables I don't even know how to tune the damn scheduler (I had to browse through source code and lkml posts to find answers). More tunables = more complexity = inherently unfair design that has to be corrected with tunables.
Last time I did a SD vs CFS test (2.6.22), SD won by far for my usage, even if it used twice the amount of context switches! I even tried putting the CFS interval to 0, and even then trying to force more context switches, performance was worse than SD(lower FPS).
Of course CFS today is nothing like it was then. The design document is so out of date it's useless.
About NO_HZ: This could actually enhance performance because the CPU only wakes up "when it needs to." Thus reducing context switching, and also reducing power usage. It does this by calculating when its next wakeup period will be, and then puts the processor to sleep until the event is scheduled to happen. For many programs that means instead of waking up 1000Hz*5 (a thousand times a second) for say a 5 second loop, we only have to wake up at the 5 second mark. Its all math.. there really isn't any guesswork.
The Hz options (100,250,864,1000) work in conjunction with NO_HZ. With a 1000Hz tick and NO_HZ enabled, you are basically saying "perform up to 1000 events if needed, else use exactly what is needed" |
me and waninkoko have both tried to port SD, my progress is in the staircase-deadline branch of zen if u want to look at it, it boots all the way into init but my computer reboots the first time something forks... it needs more work _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1179
|
Posted: Mon Dec 17, 2007 7:02 am Post subject: |
|
|
CONFLICT (content): Merge conflict in include/net/mac80211.h
Auto-merged net/ieee80211/Kconfig
Auto-merged net/mac80211/ieee80211_ioctl.c
Auto-merged net/mac80211/rx.c
CONFLICT (content): Merge conflict in net/mac80211/rx.c
Automatic merge failed; fix conflicts and then commit the result
git pull errors.
anybody? _________________ this is a strange strange world. |
|
Back to top |
|
|
TBM n00b
Joined: 20 Nov 2007 Posts: 28 Location: Norway
|
Posted: Mon Dec 17, 2007 7:35 am Post subject: |
|
|
Me too!
More than 40 wireless CONFLICTS (content and add/add) doing 'git pull' now.
I can list them all if interested... |
|
Back to top |
|
|
eatnumber1 n00b
Joined: 13 Jan 2007 Posts: 55 Location: New York
|
Posted: Mon Dec 17, 2007 7:53 am Post subject: |
|
|
The repo seems to be broken ATM... I don't know how to fix the repo's problem or I would. |
|
Back to top |
|
|
eatnumber1 n00b
Joined: 13 Jan 2007 Posts: 55 Location: New York
|
Posted: Mon Dec 17, 2007 8:28 am Post subject: |
|
|
ilikenwf wrote: | I didn't think of that until after the fact...I can do it, though...Just wrap the function's code with ifndef....I probably should just do that, huh?
Re:EDIT: Why the dependency on DEBUG_KERNEL? Also, why put it in kernel hacking? The option for regular printk is in General Setup -> Configure standard kernel features (for small systems)
That dependency is being removed in the patch....also, kernel hacking is the place it was originally meant to be...don't ask me.... |
I took a good look at your patch, and I don't really understand what you are trying to accomplish? The ability to enable and disable early_printk already exists in the kernel. |
|
Back to top |
|
|
SlashBeast Retired Dev
Joined: 23 May 2006 Posts: 2922
|
Posted: Mon Dec 17, 2007 10:32 am Post subject: |
|
|
TBM wrote: | Me too!
More than 40 wireless CONFLICTS (content and add/add) doing 'git pull' now.
I can list them all if interested... |
And me. |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Mon Dec 17, 2007 11:16 am Post subject: |
|
|
seems two merges are not doing well
Quote: | * refs/remotes/origin/wireless: forcing update to non-fast forward branch 'wireless' of git://repo.or.cz/linux-2.6/zen-sources
old...new: 09b8678...a484a9e
CONFLICT (rename/delete): Renamed drivers/net/wireless/iwlwifi/iwl-commands.h->drivers/net/wireless/iwlwifi/iwl-3945-commands.h in b050852c25a7c4524c92d4e9157f4cd9b6ac86d0 and deleted in HEAD
CONFLICT (rename/delete): Renamed drivers/net/wireless/iwlwifi/iwl-debug.h->drivers/net/wireless/iwlwifi/iwl-4965-debug.h in b050852c25a7c4524c92d4e9157f4cd9b6ac86d0 and deleted in HEAD
Auto-merged MAINTAINERS
Auto-merged drivers/char/agp/frontend.c
Auto-merged drivers/net/b44.c
Auto-merged drivers/net/wireless/Kconfig
Auto-merged drivers/net/wireless/Makefile
Auto-merged drivers/net/wireless/ath5k/ath5k.h
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/ath5k.h
Auto-merged drivers/net/wireless/ath5k/base.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/base.c
Auto-merged drivers/net/wireless/ath5k/debug.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/debug.c
Auto-merged drivers/net/wireless/ath5k/debug.h
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/debug.h
Auto-merged drivers/net/wireless/ath5k/hw.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/hw.c
Auto-merged drivers/net/wireless/ath5k/initvals.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/initvals.c
Auto-merged drivers/net/wireless/ath5k/phy.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/phy.c
Auto-merged drivers/net/wireless/ath5k/reg.h
CONFLICT (add/add): Merge conflict in drivers/net/wireless/ath5k/reg.h
Auto-merged drivers/net/wireless/b43/b43.h
CONFLICT (content): Merge conflict in drivers/net/wireless/b43/b43.h
Auto-merged drivers/net/wireless/b43/leds.c
Auto-merged drivers/net/wireless/b43/main.c
Auto-merged drivers/net/wireless/b43/phy.h
Auto-merged drivers/net/wireless/b43/tables.c
CONFLICT (content): Merge conflict in drivers/net/wireless/b43/tables.c
Auto-merged drivers/net/wireless/b43/wa.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/b43/wa.c
Auto-merged drivers/net/wireless/b43legacy/main.c
Auto-merged drivers/net/wireless/ipw2200.c
Auto-merged drivers/net/wireless/iwlwifi/Kconfig
CONFLICT (content): Merge conflict in drivers/net/wireless/iwlwifi/Kconfig
Auto-merged drivers/net/wireless/iwlwifi/iwl-3945.h
CONFLICT (content): Merge conflict in drivers/net/wireless/iwlwifi/iwl-3945.h
Auto-merged drivers/net/wireless/iwlwifi/iwl-4965.h
CONFLICT (content): Merge conflict in drivers/net/wireless/iwlwifi/iwl-4965.h
Auto-merged drivers/net/wireless/iwlwifi/iwl3945-base.c
CONFLICT (content): Merge conflict in drivers/net/wireless/iwlwifi/iwl3945-base.c
Auto-merged drivers/net/wireless/iwlwifi/iwl4965-base.c
CONFLICT (content): Merge conflict in drivers/net/wireless/iwlwifi/iwl4965-base.c
Auto-merged drivers/net/wireless/libertas/11d.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/11d.c
Auto-merged drivers/net/wireless/libertas/assoc.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/assoc.c
Auto-merged drivers/net/wireless/libertas/assoc.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/assoc.h
Auto-merged drivers/net/wireless/libertas/cmd.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/cmd.c
Auto-merged drivers/net/wireless/libertas/cmdresp.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/cmdresp.c
Auto-merged drivers/net/wireless/libertas/debugfs.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/debugfs.c
Auto-merged drivers/net/wireless/libertas/decl.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/decl.h
Auto-merged drivers/net/wireless/libertas/defs.h
Auto-merged drivers/net/wireless/libertas/dev.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/dev.h
Auto-merged drivers/net/wireless/libertas/ethtool.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/ethtool.c
Auto-merged drivers/net/wireless/libertas/host.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/host.h
Auto-merged drivers/net/wireless/libertas/hostcmd.h
Auto-merged drivers/net/wireless/libertas/if_cs.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/if_cs.c
Auto-merged drivers/net/wireless/libertas/if_sdio.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/if_sdio.c
Auto-merged drivers/net/wireless/libertas/if_usb.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/if_usb.c
Auto-merged drivers/net/wireless/libertas/if_usb.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/if_usb.h
Auto-merged drivers/net/wireless/libertas/join.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/join.c
Auto-merged drivers/net/wireless/libertas/main.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/main.c
Auto-merged drivers/net/wireless/libertas/rx.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/rx.c
Auto-merged drivers/net/wireless/libertas/scan.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/scan.c
Auto-merged drivers/net/wireless/libertas/scan.h
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/scan.h
Auto-merged drivers/net/wireless/libertas/tx.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/tx.c
Auto-merged drivers/net/wireless/libertas/wext.c
CONFLICT (content): Merge conflict in drivers/net/wireless/libertas/wext.c
Auto-merged drivers/net/wireless/orinoco.c
CONFLICT (content): Merge conflict in drivers/net/wireless/orinoco.c
Auto-merged drivers/net/wireless/rt2x00/rt2x00usb.c
CONFLICT (content): Merge conflict in drivers/net/wireless/rt2x00/rt2x00usb.c
Auto-merged drivers/net/wireless/rtl8180_dev.c
CONFLICT (add/add): Merge conflict in drivers/net/wireless/rtl8180_dev.c
Auto-merged include/linux/pci_ids.h
Auto-merged include/linux/ssb/ssb.h
Auto-merged include/net/mac80211.h
CONFLICT (content): Merge conflict in include/net/mac80211.h
Auto-merged net/ieee80211/Kconfig
Auto-merged net/mac80211/ieee80211_ioctl.c
Auto-merged net/mac80211/rx.c
CONFLICT (content): Merge conflict in net/mac80211/rx.c
Automatic merge failed; fix conflicts and then commit the result. |
_________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004
Last edited by kernelOfTruth on Mon Dec 17, 2007 11:20 am; edited 1 time in total |
|
Back to top |
|
|
Waninkoko Guru
Joined: 13 May 2005 Posts: 549
|
Posted: Mon Dec 17, 2007 11:17 am Post subject: |
|
|
eatnumber1 wrote: | ilikenwf wrote: | I didn't think of that until after the fact...I can do it, though...Just wrap the function's code with ifndef....I probably should just do that, huh?
Re:EDIT: Why the dependency on DEBUG_KERNEL? Also, why put it in kernel hacking? The option for regular printk is in General Setup -> Configure standard kernel features (for small systems)
That dependency is being removed in the patch....also, kernel hacking is the place it was originally meant to be...don't ask me.... |
I took a good look at your patch, and I don't really understand what you are trying to accomplish? The ability to enable and disable early_printk already exists in the kernel. |
You can't disable EARLY_PRINTK in a 64bits kernel. |
|
Back to top |
|
|
KenShinXF n00b
Joined: 14 Nov 2005 Posts: 8
|
Posted: Mon Dec 17, 2007 12:55 pm Post subject: |
|
|
Code: | >>> Emerging (1 of 1) sys-fs/fuse-2.7.1 to /
* fuse-2.7.1.tar.gz RMD160 SHA1 SHA256 size ;-) ... [ ok ]
* checking ebuild checksums ;-) ... [ ok ]
* checking auxfile checksums ;-) ... [ ok ]
* checking miscfile checksums ;-) ... [ ok ]
* checking fuse-2.7.1.tar.gz ;-) ... [ ok ]
* Determining the location of the kernel source code
* Found kernel source directory:
* /usr/src/linux
* Found sources for kernel version:
* 2
.6
.24
-rc4-zen1
/usr/portage/eclass/linux-info.eclass: line 200: [: too many arguments
/usr/portage/eclass/linux-info.eclass: line 200: [: too many arguments
* Checking for suitable kernel configuration options... [ ok ]
* These sources do not support loading external modules.
* to be able to use this module please enable "Loadable modules support"
* in your kernel, recompile and then try merging this module again.
|
i have this problem,
and any helps?
thanks |
|
Back to top |
|
|
kriko Tux's lil' helper
Joined: 18 Mar 2006 Posts: 92 Location: Slovenia
|
Posted: Mon Dec 17, 2007 1:14 pm Post subject: |
|
|
Fresh git clone (git clone git://repo.or.cz/linux-2.6/zen-sources.git):
.config:34:warning: trying to assign nonexistent symbol ARCH_HIBERNATION_POSSIBLE
Code: | .config:35:warning: trying to assign nonexistent symbol ARCH_SUSPEND_POSSIBLE
.config:47:warning: trying to assign nonexistent symbol CPUSCHED_CFS
.config:52:warning: trying to assign nonexistent symbol FAIR_DESKTOP
.config:53:warning: trying to assign nonexistent symbol INTERACTIVE_DESKTOP
.config:54:warning: trying to assign nonexistent symbol HIGHLY_INTERACTIVE_DESKTOP
.config:55:warning: trying to assign nonexistent symbol CUSTOM_SCHED_SETUP
.config:56:warning: trying to assign nonexistent symbol SCHED_LATENCY
.config:57:warning: trying to assign nonexistent symbol SCHED_MIN_GRANULARITY
.config:58:warning: trying to assign nonexistent symbol SCHED_WAKEUP_GRANULARITY
.config:59:warning: trying to assign nonexistent symbol SCHED_BATCH_WAKEUP_GRANULARITY
.config:72:warning: trying to assign nonexistent symbol SWAP_PREFETCH
.config:96:warning: trying to assign nonexistent symbol CC_STRIP_DEBUG
.config:98:warning: trying to assign nonexistent symbol RADIX_TREE_CONCURRENT
.config:99:warning: trying to assign nonexistent symbol RADIX_TREE_OPTIMISTIC
.config:121:warning: trying to assign nonexistent symbol PROCESSOR_SELECT
.config:122:warning: trying to assign nonexistent symbol CPU_SUP_AMD_INTEL
.config:123:warning: trying to assign nonexistent symbol CPU_SUP_CYRIX
.config:124:warning: trying to assign nonexistent symbol CPU_SUP_NSC
.config:125:warning: trying to assign nonexistent symbol CPU_SUP_CENTAUR
.config:126:warning: trying to assign nonexistent symbol CPU_SUP_TRANSMETA
.config:127:warning: trying to assign nonexistent symbol CPU_SUP_NEXGEN
.config:128:warning: trying to assign nonexistent symbol CPU_SUP_UMC
.config:156:warning: trying to assign nonexistent symbol GENETIC_IOSCHED_AS
.config:157:warning: trying to assign nonexistent symbol GENETIC_IOSCHED_DEADLINE
.config:158:warning: trying to assign nonexistent symbol GENETIC_IOSCHED_CFQ
.config:159:warning: trying to assign nonexistent symbol FINGERPRINTING
.config:208:warning: trying to assign nonexistent symbol X86_MARCH_NATIVE
.config:226:warning: trying to assign nonexistent symbol PREEMPT_SOFTIRQS
.config:227:warning: trying to assign nonexistent symbol PREEMPT_HARDIRQS
.config:229:warning: trying to assign nonexistent symbol CLASSIC_RCU
.config:230:warning: trying to assign nonexistent symbol PREEMPT_RCU
.config:231:warning: trying to assign nonexistent symbol RCU_TRACE
.config:256:warning: trying to assign nonexistent symbol BADRAM
.config:280:warning: trying to assign nonexistent symbol HZ_216
.config:281:warning: trying to assign nonexistent symbol HZ_250_NODEFAULT
.config:283:warning: trying to assign nonexistent symbol HZ_432
.config:284:warning: trying to assign nonexistent symbol HZ_864
.config:286:warning: trying to assign nonexistent symbol HZ_1500
.config:287:warning: trying to assign nonexistent symbol HZ_2000
.config:288:warning: trying to assign nonexistent symbol HZ_3000
.config:289:warning: trying to assign nonexistent symbol HZ_4000
.config:290:warning: trying to assign nonexistent symbol HZ_5000
.config:291:warning: trying to assign nonexistent symbol HZ_7500
.config:292:warning: trying to assign nonexistent symbol HZ_10000
.config:304:warning: trying to assign nonexistent symbol PERFMON
.config:331:warning: trying to assign nonexistent symbol ACPI_FUJ02B1
.config:1164:warning: trying to assign nonexistent symbol BLK_DEV_DM_BBR
.config:1380:warning: trying to assign nonexistent symbol VT_CKO
.config:1384:warning: trying to assign nonexistent symbol BASHORG_QUOTES
.config:1677:warning: trying to assign nonexistent symbol VIDEO_GSPCA
.config:1896:warning: trying to assign nonexistent symbol SND_SIS7019
.config:2129:warning: trying to assign nonexistent symbol EDAC_K8
.config:2132:warning: trying to assign nonexistent symbol EDAC_TEST_DEVICE
.config:2216:warning: trying to assign nonexistent symbol EXT2_FS_NO_PANIC
.config:2221:warning: trying to assign nonexistent symbol EXT3_FS_NO_PANIC
.config:2223:warning: trying to assign nonexistent symbol CHUNKFS_FS
.config:2227:warning: trying to assign nonexistent symbol REISER4_FS
.config:2234:warning: trying to assign nonexistent symbol BTRFS_FS
.config:2304:warning: trying to assign nonexistent symbol UNION_FS
.config:2305:warning: trying to assign nonexistent symbol AUFS
.config:2320:warning: trying to assign nonexistent symbol SQUASHFS
.config:2372:warning: trying to assign nonexistent symbol DEFAULT_RELATIME
.config:2373:warning: trying to assign nonexistent symbol DEFAULT_RELATIME_VAL
.config:2480:warning: trying to assign nonexistent symbol GCC_4003_SUPPORT
.config:2501:warning: trying to assign nonexistent symbol SECURITY_REALTIME
.config:2546:warning: trying to assign nonexistent symbol CRYPTO_LZF
.config:2569:warning: trying to assign nonexistent symbol GENETIC_LIB |
Is there something wrong with repo or on my end? |
|
Back to top |
|
|
Waninkoko Guru
Joined: 13 May 2005 Posts: 549
|
Posted: Mon Dec 17, 2007 1:45 pm Post subject: |
|
|
There's something really wrong. I'm going to fix it (current master branch is not zen-sources). |
|
Back to top |
|
|
|