View previous topic :: View next topic |
Author |
Message |
predatorfreak l33t
Joined: 13 Jan 2005 Posts: 708 Location: USA, Michigan.
|
Posted: Tue Mar 01, 2005 6:51 pm Post subject: |
|
|
fallow wrote: | Tiger683 wrote: | in fact, but it b0rks quite a lot, staircase is just not designed for smooth transitions, and preempting in realtime (...) /* snip */
T |
yeah
btw. for me O(1) since some time (2.6.10 I think [after every past mm sched patch is in vanilla]) is also very good and playing really very well. Now other schedulers are not thumping O(1) as in the past becuase he is realli nice now
cheers. |
And if you add the patches nick piggin recently submitted, it contends with staircase easly . _________________ System: predatorbox
Distro: Arch Linux x86_64
Current projects: blackhole, convmedia and anything else I cook up. |
|
Back to top |
|
|
a13x Apprentice
Joined: 05 Jan 2005 Posts: 215
|
Posted: Tue Mar 01, 2005 6:51 pm Post subject: |
|
|
My Internet connection was down. I'll download the new patch and post the results. |
|
Back to top |
|
|
Tiger683 Veteran
Joined: 08 Jan 2005 Posts: 1347 Location: Heffner's House
|
Posted: Tue Mar 01, 2005 6:59 pm Post subject: |
|
|
@ pred: u mean those that are in rc5-mm1?
im already having a run on second rt with those included..... _________________ Retired gentoo user |
|
Back to top |
|
|
NiLuJe Guru
Joined: 06 Nov 2004 Posts: 303 Location: Paris, France
|
Posted: Tue Mar 01, 2005 7:05 pm Post subject: |
|
|
Hu, I recompiled the kernel & nvidia, & .. it works! ;D Must have missed something the first time, sorry
I'll try to mount my reisser4 data partition .. cross your fingers ;D
EDIT : Reiser4 seems to works It's not my / , but it's mounted now, and seems to works fine
EDIT² : Hum, my dmesg is filled with "crq requeued" lines ... oO
Last edited by NiLuJe on Tue Mar 01, 2005 7:10 pm; edited 2 times in total |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Tue Mar 01, 2005 7:06 pm Post subject: |
|
|
but btw. about 13 Nick`s patches - they including only the multiprocessor scheduler patches
cheers. _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Tiger683 Veteran
Joined: 08 Jan 2005 Posts: 1347 Location: Heffner's House
|
Posted: Tue Mar 01, 2005 7:08 pm Post subject: |
|
|
yeah, thats what he emailed me about recently, the new smp code, but thx for the infos.
T _________________ Retired gentoo user |
|
Back to top |
|
|
Illissius Guru
Joined: 31 Jul 2004 Posts: 395 Location: Hungary
|
Posted: Tue Mar 01, 2005 7:22 pm Post subject: |
|
|
Sorry, hugely off topic, but...
Darckness wrote: | Whew. Sorry that stuff has gone to hell, but I suppose that's what happens when I get a massive fever and am clinically dead for 15 hours. |
Did you mean that last part literally? _________________ Work is punishment for failing to procrastinate effectively.
last.fm |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Tue Mar 01, 2005 7:32 pm Post subject: |
|
|
Illissius: I was unable to be roused for those 15 hours. Sure, maybe I didn't have a doctor come over here and actually declare it, but I think it was close enough. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
Zephyrus Apprentice
Joined: 01 Sep 2004 Posts: 204
|
Posted: Tue Mar 01, 2005 7:33 pm Post subject: |
|
|
Hello,
I'm trying the new nitro version with realtime enabled but i have some problems:
1]When the system loads the NVIDIA kernel module, i get this:
Code: |
Mar 1 19:52:10 tux NVRM: loading NVIDIA Linux x86 NVIDIA Kernel Module 1.0-6629 Wed Nov 3 13:12:51 PST 2004
Mar 1 19:52:10 tux BUG: modprobe/3213, lock held at task exit time!
Mar 1 19:52:10 tux [dfd115e4] {&os_sema->wait}
Mar 1 19:52:10 tux .. held by: modprobe: 3213 [df6c4110, 125]
Mar 1 19:52:10 tux ... acquired at: crq requeued
|
And then i have a lot of these:
Code: |
Mar 1 19:55:20 tux crq requeued
Mar 1 19:55:33 tux crq requeued
Mar 1 19:55:33 tux crq requeued
Mar 1 19:55:35 tux crq requeued
[...]
|
Although the error, Xorg starts and i'm able to log in ... and to post on this forum with konqueror
2]When i try to switch between framebuffer consoles i get a blank screen and no further response from the system.
I didn't find any related error in the log files.
Note:I use vesafb-tnb and fbslash.
3] I have two disks on my machine and with previous kernels the first one was registered as hda and the second one as hdb. Now i get this:
Code: |
[...]
Mar 1 19:52:10 tux hda: max request size: 1024KiB
Mar 1 19:52:10 tux hda: 80293248 sectors (41110 MB) w/2048KiB Cache, CHS=16383/255/63, UDMA(66)
Mar 1 19:52:10 tux hda: cache flushes supported
Mar 1 19:52:10 tux hda: hda1 hda2 hda3 hda4
Mar 1 19:52:10 tux hdb: max request size: 128KiB
Mar 1 19:52:10 tux hdb: 39102336 sectors (20020 MB) w/2048KiB Cache, CHS=38792/16/63, UDMA(66)
Mar 1 19:52:10 tux hdb: cache flushes not supported
Mar 1 19:52:10 tux hdq: hdq1
[...]
|
In /dev i don't have hdb (and the partition hdb1) anymore, but hdq (and hdq1), why?
Note:I use udev. |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Tue Mar 01, 2005 7:40 pm Post subject: |
|
|
Zephyrus: If it works, don't question it. The tux stuff is from tux which you have enabled in your kernel. As for the console switching, I'd say try not to switch consoles too much . I have the same issue with disk name switching on my computer. It's a bit irritating, but nothing that can't be fixed. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
a13x Apprentice
Joined: 05 Jan 2005 Posts: 215
|
Posted: Tue Mar 01, 2005 7:44 pm Post subject: |
|
|
The ebuild is FUBAR again. Look here:
Code: |
After installing a new kernel of any version, it is important
* that you have the appropriate /etc/modules.autoload.d/kernel-X.Y
* created (X.Y is the first 2 parts of your new kernel version)
* For example, this kernel will require:
* /etc/modules.autoload.d/kernel-2.6
* IMPORTANT:
* This is a experimental kernel version, I'm not responsible for breaking your system
* ALWAYS keep a second stable and bootable kernel apart in your boot manager
------------------------------------------------------------------------------------------------------------------------
QA Notice: USE Flag 'want-experimental' not in IUSE for sys-kernel/nitro-sources-2.6.11_rc5-r0
------------------------------------------------------------------------------------------------------------------------
* You have enabled experimental patches. They've been tested, but only as
* far as running on the developers' systems.
* If bad things happen, remember:
* YOU chose it.
* Now fire up your compiler, reboot, and fasten your seatbelts. You'll definitely need your seatbelts for this ride!
>>> Regenerating /etc/ld.so.cache...
* Caching service dependencies ... [ ok ]>>> sys-kernel/nitro-sources-2.6.11_rc5-r0 merged.
>>> Recording sys-kernel/nitro-sources in "world" favorites file...
>>> clean: No packages selected for removal.
>>> Auto-cleaning packages ...
>>> No outdated packages were found on your system.
* GNU info directory index is up-to-date.
|
I've had enough rc5 ebuilds for today. I'm going to do everything manually.
Last edited by a13x on Tue Mar 01, 2005 7:56 pm; edited 1 time in total |
|
Back to top |
|
|
Zephyrus Apprentice
Joined: 01 Sep 2004 Posts: 204
|
Posted: Tue Mar 01, 2005 7:53 pm Post subject: |
|
|
Darckness wrote: | Zephyrus: If it works, don't question it. The tux stuff is from tux which you have enabled in your kernel. As for the console switching, I'd say try not to switch consoles too much . I have the same issue with disk name switching on my computer. It's a bit irritating, but nothing that can't be fixed. |
Actually Tux is the name of my host.
I know, i'm not a very creative talent when we are talking about system's names .
EDIT:
I find that the "crq requeued" message appears if and only if i select the cfq scheduler, and i don't think it is linked with the NVIDIA module problem because, when i use another scheduler, i get:
[code]
Mar 1 20:55:09 tux NVRM: loading NVIDIA Linux x86 NVIDIA Kernel Module 1.0-6629 Wed Nov 3 13:12:51 PST 2004
Mar 1 20:55:09 tux BUG: modprobe/3223, lock held at task exit time!
Mar 1 20:55:09 tux [dfd0f5e4] {&os_sema->wait}
Mar 1 20:55:09 tux .. held by: modprobe: 3223 [de8d8820, 125]
Mar 1 20:55:11 tux ... acquired at: <6>eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
[code]
Somehow the following message isn't written in a new line but on the same one of "... acquired at:"
Last edited by Zephyrus on Tue Mar 01, 2005 8:17 pm; edited 2 times in total |
|
Back to top |
|
|
Tiger683 Veteran
Joined: 08 Jan 2005 Posts: 1347 Location: Heffner's House
|
Posted: Tue Mar 01, 2005 8:15 pm Post subject: |
|
|
for everyone who are annoyed about some verbose dmesg output:
i added some extra debug stuff, + rt may cause false alarms especially with spinlock debugging and preemption debugging on, as it uses different mechanisms for both tasklocking ( mutexes and such instead ) and preemption than the vanilla.
Here is my config part for kernel hacking i use for my box, and besides of dmesg messages from boot, no further trash in dmesg:
Tiger683's Kernel hacking section of .config wrote: |
#
# Kernel hacking
#
CONFIG_DEBUG_KERNEL=y
CONFIG_MAGIC_SYSRQ=y
CONFIG_LOG_BUF_SHIFT=17
# CONFIG_DETECT_SOFTLOCKUP is not set
# CONFIG_SCHEDSTATS is not set
# CONFIG_DEBUG_SLAB is not set
# CONFIG_WAKEUP_TIMING is not set
# CONFIG_CRITICAL_PREEMPT_TIMING is not set
# CONFIG_CRITICAL_IRQSOFF_TIMING is not set
# CONFIG_RT_DEADLOCK_DETECT is not set
# CONFIG_DEBUG_KOBJECT is not set
# CONFIG_DEBUG_BUGVERBOSE is not set
CONFIG_DEBUG_INFO=y
# CONFIG_PAGE_OWNER is not set
# CONFIG_DEBUG_FS is not set
# CONFIG_USE_FRAME_POINTER is not set
CONFIG_EARLY_PRINTK=y
# CONFIG_DEBUG_STACKOVERFLOW is not set
# CONFIG_KPROBES is not set
# CONFIG_DEBUG_STACK_USAGE is not set
# CONFIG_DEBUG_PAGEALLOC is not set
# CONFIG_VAR_STACKS_4 is not set
CONFIG_VAR_STACKS_8=y
# CONFIG_VAR_STACKS_16 is not set
# CONFIG_VAR_STACKS_32 is not set
# CONFIG_VAR_STACKS_64 is not set
# CONFIG_VAR_STACKS_128 is not set
# CONFIG_VAR_STACKS_256 is not set
# CONFIG_VAR_STACKS_512 is not set
# CONFIG_VAR_STACKS_1024 is not set
CONFIG_CFLAGS_EDIT=y
CONFIG_CFLAGS_STRING="-O3 -march=athlon-xp -mtune=athlon-xp -mmmx -msse -m3dnow -pipe"
# CONFIG_CFLAGS_OVERRIDE is not set
CONFIG_X86_FIND_SMP_CONFIG=y
CONFIG_X86_MPPARSE=y
|
T _________________ Retired gentoo user |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Tue Mar 01, 2005 8:16 pm Post subject: |
|
|
a13x: The ebuild is not FUBAR. That's definitely some kind of portage bug. Anyway, I get that too but it doesn't affect anything.
In other news, IT is currently 1/2 finished. I'm running IT now. You guys are gonna be sooooo surprised... _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
Tiger683 Veteran
Joined: 08 Jan 2005 Posts: 1347 Location: Heffner's House
|
Posted: Tue Mar 01, 2005 8:17 pm Post subject: |
|
|
@Zephyrus:
Your Nvidia module conflicts with the lan driver, what is your ethernet harware?
T _________________ Retired gentoo user |
|
Back to top |
|
|
a13x Apprentice
Joined: 05 Jan 2005 Posts: 215
|
Posted: Tue Mar 01, 2005 8:19 pm Post subject: |
|
|
Bad news !
This time I get an auto-reboot when I start the new rt kernel. |
|
Back to top |
|
|
Zephyrus Apprentice
Joined: 01 Sep 2004 Posts: 204
|
Posted: Tue Mar 01, 2005 8:23 pm Post subject: |
|
|
Tiger683 wrote: | @Zephyrus:
Your Nvidia module conflicts with the lan driver, what is your ethernet harware?
T |
My ethernet controller is a Realtek 8139C.
But...Are you sure? Isn't this a format error in the log?
Because with cfq i get:
Code: |
Mar 1 21:12:24 tux NVRM: loading NVIDIA Linux x86 NVIDIA Kernel Module 1.0-6629 Wed Nov 3 13:12:51 PST 2004
Mar 1 21:12:24 tux BUG: modprobe/3211, lock held at task exit time!
Mar 1 21:12:24 tux [dfd115e4] {&os_sema->wait}
Mar 1 21:12:24 tux .. held by: modprobe: 3211 [df7538f0, 125]
Mar 1 21:12:24 tux ... acquired at: crq requeued
Mar 1 21:12:24 tux crq requeued
Mar 1 21:12:26 tux eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Mar 1 21:12:27 tux crq requeued
|
with deadline:
Code: |
Mar 1 20:55:09 tux NVRM: loading NVIDIA Linux x86 NVIDIA Kernel Module 1.0-6629 Wed Nov 3 13:12:51 PST 2004
Mar 1 20:55:09 tux BUG: modprobe/3223, lock held at task exit time!
Mar 1 20:55:09 tux [dfd0f5e4] {&os_sema->wait}
Mar 1 20:55:09 tux .. held by: modprobe: 3223 [de8d8820, 125]
Mar 1 20:55:11 tux ... acquired at: <6>eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
|
Last edited by Zephyrus on Tue Mar 01, 2005 8:24 pm; edited 1 time in total |
|
Back to top |
|
|
Tiger683 Veteran
Joined: 08 Jan 2005 Posts: 1347 Location: Heffner's House
|
Posted: Tue Mar 01, 2005 8:24 pm Post subject: |
|
|
ok guys, the answer is: new rt is already on its (early ) way, ill try to reduce the amount of patches to under 100.
for now, i can only say *sigh*,
and please, report any other issues.
a13x: what errors do you get? none? not using framebuffer?
T
EDIT:
no, the task for modprobing your nvidia is unable to aquire a lock because your ethernet controller driver task is still holding it...
i dont know if its an issue with the driver itself or maybe the vector signaled interrupts or something...
i use sk98lin for my yukon w/o probs and can only say *sigh*
W will see with the next rt... i guess me and darckness are the only ones running it without issues
im puzzled....
T _________________ Retired gentoo user |
|
Back to top |
|
|
a13x Apprentice
Joined: 05 Jan 2005 Posts: 215
|
Posted: Tue Mar 01, 2005 9:22 pm Post subject: |
|
|
No framebuffer, nu logo just clean console view.
I would post the error but it just auto-reboots and there is no time to see anything. |
|
Back to top |
|
|
BlinkEye Veteran
Joined: 21 Oct 2003 Posts: 1046 Location: Gentoo Forums
|
Posted: Tue Mar 01, 2005 9:33 pm Post subject: |
|
|
well, i'm just about to compile another kernel (it has become a habit ). is realtime worth a try or is it just too bugy? _________________ Easily backup up your system? klick
Get rid of SSH Brute Force Attempts / Script Kiddies klick |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Tue Mar 01, 2005 9:43 pm Post subject: |
|
|
BlinkEye: If it works, it's worth it. If it doesn't, it isn't. Seems to be about 50/50 that it works (disable fb to bring the chance of it working up to 75%). _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
BlinkEye Veteran
Joined: 21 Oct 2003 Posts: 1046 Location: Gentoo Forums
|
Posted: Tue Mar 01, 2005 9:58 pm Post subject: |
|
|
thanks for the answer. i'll give it a try - i don't give anything on framebuffer. _________________ Easily backup up your system? klick
Get rid of SSH Brute Force Attempts / Script Kiddies klick |
|
Back to top |
|
|
Apolonius Guru
Joined: 24 Jan 2003 Posts: 325
|
Posted: Tue Mar 01, 2005 11:29 pm Post subject: |
|
|
Concerning the "ide crazy numbers bugs" in the rt sources, this patch solves the problem:
http://lkml.org/lkml/2005/2/23/214 |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Tue Mar 01, 2005 11:42 pm Post subject: |
|
|
Apolonius: Thanks; I'll pass this along to Tiger683 so he can throw it in.
Update on IT: We've just finished the final testing, and IT is fully functional now. Now all that remains is for 2.6.11 to arrive so that we can unleash IT upon unsuspecting users everywhere. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
rzZzn Tux's lil' helper
Joined: 24 Aug 2004 Posts: 96 Location: Sweden
|
Posted: Wed Mar 02, 2005 2:10 am Post subject: |
|
|
My compile fails
Here's my output
Code: |
SYSCALL arch/x86_64/ia32/vsyscall-sysenter.so
AS arch/x86_64/ia32/vsyscall-syscall.o
SYSCALL arch/x86_64/ia32/vsyscall-syscall.so
CC arch/x86_64/ia32/syscall32.o
CC arch/x86_64/ia32/ipc32.o
CC arch/x86_64/ia32/ia32_aout.o
LD arch/x86_64/ia32/built-in.o
CC kernel/sched.o
kernel/sched.c:3980: error: redefinition of '__kstrtab_task_nice'
kernel/sched.c:3972: error: previous definition of '__kstrtab_task_nice' was here
kernel/sched.c:3980: error: redefinition of '__ksymtab_task_nice'
kernel/sched.c:3972: error: previous definition of '__ksymtab_task_nice' was here
make[1]: *** [kernel/sched.o] Error 1
make: *** [kernel] Error 2
|
Any ideas
Oh I´m on amd64...
Great patchsets otherwise 2.6.10-nitro4 works just fine with both reiser4 xfs and nvidia...
But not this on
//rZn |
|
Back to top |
|
|
|