Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

Realtime kernel build quickie + bootsplash (or not)

Share your HowTo, Documentation, Tips and Tricks. Not for support questions!.
Message
Author
User avatar
mzilikazi
Forum Account
Forum Account
Posts: 3282
Joined: 2004-09-16 02:14
Location: Colorado Springs, CO

#91 Post by mzilikazi »

Scoop wrote:To build the latest rt kernel (2.6.26.8-rt13) that works with audio esp. jack and midi you need Adam Samson's patch.

I have not succeeded in building nvidia for this one yet but I am no guru!
HTH
Norv
Thanks for the tip! Did you try Nvidia from the experimental repo or are you using the Nvidia installer?
Debian Sid Laptops:
AMD Athlon(tm) 64 X2 Dual-Core Processor TK-55 / 1.5G
Intel(R) Pentium(R) Dual CPU T2390 @ 1.86GHz / 3G

Scoop
Posts: 21
Joined: 2004-09-14 20:18
Location: Australia

#92 Post by Scoop »

I just rebooted and did
sudo m-a a-i nvidia
I'm on debian sid with unstable and experimental in my sources.list - unstable pinned higher.

reakinator
Posts: 7
Joined: 2009-02-03 22:56

error: undefined references in timer.c

#93 Post by reakinator »

I'm trying to compile kernel 2.6.21.6-rt21 with this guide (it was the last one to work well for me in ubuntu - everything after had many problems with the touchpad and wireless). When running the command:

Code: Select all

time fakeroot make-kpkg --initrd -rev mz1 kernel_image kernel_headers
I get the following error:

Code: Select all

kernel/built-in.o: In function `run_timer_softirq':
timer.c:(.text+0x12405): undefined reference to `__udivdi3'
timer.c:(.text+0x12428): undefined reference to `__umoddi3'
kernel/built-in.o: In function `do_sysinfo':
(.text+0x128e2): undefined reference to `__udivdi3'
kernel/built-in.o: In function `do_sysinfo':
(.text+0x12902): undefined reference to `__umoddi3'
kernel/built-in.o: In function `do_gettimeofday':
(.text+0x12a57): undefined reference to `__udivdi3'
kernel/built-in.o: In function `do_gettimeofday':
(.text+0x12a77): undefined reference to `__umoddi3'
kernel/built-in.o: In function `getnstimeofday':
(.text+0x12bce): undefined reference to `__umoddi3'
kernel/built-in.o: In function `ktime_get_ts':
(.text+0x1e26c): undefined reference to `__umoddi3'
kernel/built-in.o: In function `hrtimer_run_queues':
(.text+0x1e588): undefined reference to `__udivdi3'
kernel/built-in.o: In function `hrtimer_run_queues':
(.text+0x1e5aa): undefined reference to `__umoddi3'
make[1]: *** [.tmp_vmlinux1] Error 1
Can anyone suggest what is going wrong? Thanks..

Rich

lurid
Posts: 1
Joined: 2009-02-26 00:09
Location: Pennsylvania, United States

#94 Post by lurid »

Thanks everyone for posting! I finally got my RT kernel built and my ATi drivers working with it. I couldn't get the kernel module for my graphics card to build because of an error that is discussed in this thread on another message board. I built a 2.6.26.8 kernel using the instructions here, the kernel build itself was smooth and easy. The only problem was the fglrx kernel module that would refuse to build even when I had all of the PARAVIRT stuff disabled.

Following the instructions in the above-posted thread, I was able to fix the problem and get the kernel and module built. I hope something gets done to avoid this problem in the future. I was up until about 3:30AM trying to get it working, but now I got it.

Code: Select all

me@my_box:~$ uname -a
Linux my_box 2.6.26.8-rt16 #1 SMP PREEMPT RT Wed Feb 25 18:23:28 EST 2009 x86_64 GNU/Linux

roooz
Posts: 143
Joined: 2008-07-27 21:01

#95 Post by roooz »

@mzilikazi and others,

There are thinking to include an RT kernel in Debian. Maybe you can offer your experience/knowledge to build, maintain or test the RT kernel:

http://www.linux-archive.org/debian-dev ... ebian.html

jongi
Posts: 472
Joined: 2007-04-15 02:41

#96 Post by jongi »

Is there a way to permanently set

Code: Select all

export CONCURRENCY_LEVEL=2 
?
Desktop: Debian (Sid) 64-bit, Gentoo 64-bit and Ubuntu Jaunty 64-bit
Laptop: Gentoo 32-bit
MythTV: Debian (Testing) 32-bit

User avatar
bugsbunny
Posts: 5354
Joined: 2008-07-06 17:04
Been thanked: 1 time

#97 Post by bugsbunny »

set it in your ~/.bashrc file.

User avatar
Hadret
Posts: 354
Joined: 2007-07-19 13:04
Location: Berlin

#98 Post by Hadret »

For quite some time I can't compile my own kernel. Here's the error I've got:

Code: Select all

  CC [M]  drivers/staging/altpciechdma/altpciechdma.o
drivers/staging/altpciechdma/altpciechdma.c: In function ‘ape_sg_to_chdma_table’:
drivers/staging/altpciechdma/altpciechdma.c:430: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 3 has type ‘dma_addr_t’
drivers/staging/altpciechdma/altpciechdma.c:434: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 3 has type ‘dma_addr_t’
drivers/staging/altpciechdma/altpciechdma.c:450: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 3 has type ‘dma_addr_t’
drivers/staging/altpciechdma/altpciechdma.c:451: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 3 has type ‘dma_addr_t’
drivers/staging/altpciechdma/altpciechdma.c: In function ‘dma_test’:
drivers/staging/altpciechdma/altpciechdma.c:548: warning: assignment makes integer from pointer without a cast
drivers/staging/altpciechdma/altpciechdma.c: In function ‘probe’:
drivers/staging/altpciechdma/altpciechdma.c:931: warning: label ‘err_rev’ defined but not used
drivers/staging/altpciechdma/altpciechdma.c:914: warning: label ‘err_cdev’ defined but not used
drivers/staging/altpciechdma/altpciechdma.c: At top level:
drivers/staging/altpciechdma/altpciechdma.c:416: warning: ‘ape_sg_to_chdma_table’ defined but not used
  CC [M]  drivers/staging/asus_oled/asus_oled.o
  CC [M]  drivers/staging/comedi/comedi_fops.o
In file included from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
drivers/staging/comedi/interrupt.h:27: error: conflicting types for ‘irqreturn_t’
include/linux/irqreturn.h:14: error: previous declaration of ‘irqreturn_t’ was here
In file included from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
drivers/staging/comedi/interrupt.h:30:1: warning: "IRQ_RETVAL" redefined
In file included from include/linux/interrupt.h:10,
                 from drivers/staging/comedi/interrupt.h:22,
                 from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
include/linux/irqreturn.h:15:1: warning: this is the location of the previous definition
make[4]: *** [drivers/staging/comedi/comedi_fops.o] BÅ‚Ä…d 1
make[3]: *** [drivers/staging/comedi] BÅ‚Ä…d 2
make[2]: *** [drivers/staging] BÅ‚Ä…d 2
make[1]: *** [drivers] BÅ‚Ä…d 2
make[1]: Opuszczenie katalogu `/home/hadret/debbuild/KERNEL/linux-2.6.29.1-rt7'
make: *** [debian/stamp/build/kernel] BÅ‚Ä…d 2
BÅ‚Ä…d == Error

It's kernel 2.6.29.1 with real time patch applied.

Any ideas why this happens?

gregg
Posts: 2
Joined: 2009-04-13 23:06
Location: michigan

#99 Post by gregg »

I recently compiled a kernel and lost eth0. Boots and runs but seems like it skipped the drivers. I used make oldconfig but I am curious, do I need to copy over the config from boot? This is a dedicated server and a real pain. :P

Any ideas?

Scoop
Posts: 21
Joined: 2004-09-14 20:18
Location: Australia

Re: Realtime kernel build quickie + bootsplash (or not)

#100 Post by Scoop »

Gregg, I had similar problem and discovered eth1 (in my case) had been renamed to eth3.
Changed all eth1 entries in /etc/network/interfaces to eth3, ran

Code: Select all

/etc/init.d/networking restart
and so far, so good .

gregg
Posts: 2
Joined: 2009-04-13 23:06
Location: michigan

Re: Realtime kernel build quickie + bootsplash (or not)

#101 Post by gregg »

Thanks for the info Scoop. I'll have support take a look at that (server in DC so no access). In my case there are two, eth0 being public network and eth1 being private, but if this is the same problem I should notice things fubar in that file.

Thanks!

Arite
Posts: 12
Joined: 2007-05-06 19:02

Re:

#102 Post by Arite »

Hadret wrote:It's kernel 2.6.29.1 with real time patch applied.

Any ideas why this happens?
I'm also getting a similar error with 2.6.29.2 with the -rt11 patch (from here) applied:

Code: Select all

make[1]: Leaving directory `/home/alan/documents/kernel/linux-2.6.29.2-rt11'
/usr/bin/make    ARCH=x86_64 \
			     modules
make[1]: Entering directory `/home/alan/documents/kernel/linux-2.6.29.2-rt11'
  CHK     include/linux/version.h
  CHK     include/linux/utsrelease.h
  SYMLINK include/asm -> include/asm-x86
  CALL    scripts/checksyscalls.sh
  CC [M]  drivers/staging/comedi/comedi_fops.o
In file included from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
drivers/staging/comedi/interrupt.h:27: error: conflicting types for ‘irqreturn_t’
include/linux/irqreturn.h:14: error: previous declaration of ‘irqreturn_t’ was here
In file included from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
drivers/staging/comedi/interrupt.h:30:1: warning: "IRQ_RETVAL" redefined
In file included from include/linux/interrupt.h:10,
                 from drivers/staging/comedi/interrupt.h:22,
                 from drivers/staging/comedi/comedidev.h:40,
                 from drivers/staging/comedi/comedi_fops.c:45:
include/linux/irqreturn.h:15:1: warning: this is the location of the previous definition
make[4]: *** [drivers/staging/comedi/comedi_fops.o] Error 1
make[3]: *** [drivers/staging/comedi] Error 2
make[2]: *** [drivers/staging] Error 2
make[1]: *** [drivers] Error 2
make[1]: Leaving directory `/home/alan/documents/kernel/linux-2.6.29.2-rt11'
make: *** [debian/stamp/build/kernel] Error 2
Any ideas?

Cheers, Arite.

Scoop
Posts: 21
Joined: 2004-09-14 20:18
Location: Australia

Re: Realtime kernel build quickie + bootsplash (or not)

#103 Post by Scoop »

The latest one I have built is the 2.6.33-2-rt13 and I got an error like yours.
I disabled all staging drivers in the kernel config and it built OK.
I have no idea what disabling the staging drivers might do tp your system ;)

User avatar
stevepusser
Posts: 12930
Joined: 2009-10-06 05:53
Has thanked: 41 times
Been thanked: 71 times

Re: Realtime kernel build quickie + bootsplash (or not)

#104 Post by stevepusser »

Best thing to do is read the help during the kernel config for each driver in staging---you most likely will not need any of them, and they are not really good quality yet anyway. Most of the realtek wifi drivers are junk, for example.
MX Linux packager and developer

Serje
Posts: 1
Joined: 2011-05-20 15:02
Location: England
Contact:

Re: Realtime kernel build quickie + bootsplash (or not)

#105 Post by Serje »

I built a 2.6.23-rt kernel and got nvidia working on Debian sid amd64.

I could not find Virtualization under Processor Type and Features in make menuconfig.
For me it was under Device IDs (I think that's it).

Also I was missing a couple of important pkgs at the start like libncurses5-dev but Googling sorted that out.
Now I'm going to recomplie with a couple of tweaks mentioned in the Alsa low-latency howto.
http://alsa-project.org/main/index.php/ ... ency_howto

Thanks)

Scoop
Posts: 21
Joined: 2004-09-14 20:18
Location: Australia

Re: Realtime kernel build quickie + bootsplash (or not)

#106 Post by Scoop »

AFAIK the realtime features are included with stock kernel as of 2.6.36, no recompiling necessary. Unless you're a tweaker. Debian sid currently running 2.6.38-2 series.

Post Reply