View previous topic :: View next topic |
Author |
Message |
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun May 29, 2005 7:28 pm Post subject: |
|
|
ldonesty wrote: | pharon wrote: |
As for the reiserfs + genetic , do you excpect people to keep jumping from one filesystem to another to work around problems? We have other things to do |
Currently I'm running on rc5-skunk1 using nicksched and genetic: no problem after moving/copying files (~ 400 mb) over 2 partitions of the same hd (notebook) and one of them is reiserfs.
What kind of problems do you refer? |
I have yet to experience a reiserfs prob on -skunk1-3, Ihave reiserfs on my root partition _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
priestjim Tux's lil' helper
Joined: 09 Jan 2005 Posts: 100 Location: Never never land
|
Posted: Sun May 29, 2005 8:24 pm Post subject: |
|
|
Could you post the exact procedure of how to apply the patchset to the source tree? _________________ My soul is my antenna, I am the instrument + the guitar is my amplifier.
-- Ottmar Liebert |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun May 29, 2005 8:49 pm Post subject: |
|
|
priestjim wrote: | Could you post the exact procedure of how to apply the patchset to the source tree? |
FOR PEOPLE NOT USING THE EBUILD
I know patching against the -mm broken out is a pain in the ass but I dont have a ton of server space now so I wanted to make use of the existing brokenouts. In the future I will just have a brokenout dir.
Code: |
For those Skunk users out there who want to help fix any bugs or hack up any work I did I supplied my broken-out patches. The patch names are identical to found on the web but some are edited so here is the trick. Install Quilt from http://savannah.nongnu.org/projects/quilt and follow these instuctions:
cd /usr/src
tar xvjf linux-2.6.11.tar.bz2
mv linux-2.6.11 linux-2.6.12-rc5-skunk2
ln -s linux-2.6.12-rc5-skunk2 linux
cd linux
tar xvjf 2.6.12-rc5-mm1-broken-out.tar.bz2
mv broken-out patches
cd patches
patch -p1 < 2.6.12-rc5-mm1-broken-out_to_2.6.12-rc5-skunk3
cd /usr/src/linux
quilt push -a (to apply all patches not commented out in series file)
Quilt looks for patches in the 'patches' dir and applies in order the files found in the series file.
Quilt Commands:
quilt push - Applies current patch then moves to next patch.
quilt pop - Removes current patch then moves to prior patch.
quipt import <patch> - Copies patch from source to patches dir and adds entry to series file.
quilt new <patch> - Creates new blank patch in patches dir and add entry to series file.
quilt edit <path/file> - Edit files to make manual changes.
quilt refresh - Injects a diff of changes made from 'quilt edit' into the current patch.
|
_________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
pharon n00b
Joined: 13 Mar 2005 Posts: 19
|
Posted: Sun May 29, 2005 10:02 pm Post subject: |
|
|
Tried rc3, and the shutdown oops is gone, good work. I still need to apply the "avoid-fragmentation" patch to fix the alsa oops. I didn't enable genetic, performance is good without anyway.
Now I can start enabling more exotic features in the kernel and see what else breaks.
Thanks this is a good, albiet souped up and overloaded kernel. I love it !! |
|
Back to top |
|
|
AaronPPC Guru
Joined: 29 May 2005 Posts: 522 Location: Tucson, AZ
|
Posted: Sun May 29, 2005 10:10 pm Post subject: |
|
|
Quote: | I haven't heard many complaints about the release, does that mean everything is ok or nobody using this. |
Since you asked...
I'm trying to make this work on a Pegasos G4 PPC. I'm currently running .11-skunk6.
-skunk3 breaks here when I'm trying to compile:
Code: | CC kernel/configs.o
CC kernel/audit.o
kernel/audit.c:386:61: macro "kthread_run" requires 5 arguments, but only 3 given
kernel/audit.c: In function `audit_receive_msg':
kernel/audit.c:386: error: `kthread_run' undeclared (first use in this function)
kernel/audit.c:386: error: (Each undeclared identifier is reported only once
kernel/audit.c:386: error: for each function it appears in.)
make[1]: *** [kernel/audit.o] Error 1
make: *** [kernel] Error 2 |
Removing auditing support allows the kernel to compile the rest of the way without problems, but the kernel doesn't boot.
This is actually am improvement. Previous .12-rc-skunks blew up here:
Code: | LD arch/ppc/lib/built-in.o
CC arch/ppc/syslib/prep_nvram.o
CC arch/ppc/syslib/prom_init.o
CC arch/ppc/syslib/prom.o
CC arch/ppc/syslib/of_device.o
arch/ppc/syslib/of_device.c: In function `of_device_suspend':
arch/ppc/syslib/of_device.c:114: error: incompatible type for argument 2 of indirect function call
arch/ppc/syslib/of_device.c: At top level:
arch/ppc/syslib/of_device.c:132: warning: initialization from incompatible pointer type
make[1]: *** [arch/ppc/syslib/of_device.o] Error 1
make: *** [arch/ppc/syslib] Error 2 |
I always suspected is was because of Suspend2 and I had this error even though I leave out Power Management Support (I don't think my hardware supports it). I was in the middle of confirming this when you released rc5-skunk1 (which compiled successfully and boots). The "943-fix-for-ppc" patch in Suspend2 seems to have fixed the problem.
Quote: | How's quilt working out for people? |
Overall, I like it.
Is there a way to navigate up and down the series without applying or removing a patch? When I was playing with -skunk2 and didn't apply the Suspend2 patches, it caused a few patches further down to error out (it was a few of the gregkh patches, but I can't remember the exact ones). So, what I wanted to do was apply all the patches and then reverse the Suspend2 ones. Quilt pop 200-agp-resume-support.patch reverses everything up to 200-agp-resume-support.patch. _________________ --Aaron |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun May 29, 2005 11:43 pm Post subject: |
|
|
[quote="AaronPPC"] Quote: |
Quote: | How's quilt working out for people? |
Overall, I like it.
Is there a way to navigate up and down the series without applying or removing a patch? When I was playing with -skunk2 and didn't apply the Suspend2 patches, it caused a few patches further down to error out (it was a few of the gregkh patches, but I can't remember the exact ones). So, what I wanted to do was apply all the patches and then reverse the Suspend2 ones. Quilt pop 200-agp-resume-support.patch reverses everything up to 200-agp-resume-support.patch. |
right what u would want to do is quilt pop <the patch right before 200-agp-resume-support.patch> then comment out all of the suspend2 patches in the series file the do a 'quilt push -a' and it will skip over those suspend2 patches and apply the rest....
if after u 'quilt push -a' lets say some patch dosent apply perfectly clean, 'quilt push -f', then use 'quilt edit <file> to clean up the reject, then quilt refresh, to fix the patch file then 'quilt push -a' to resume patching _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
AaronPPC Guru
Joined: 29 May 2005 Posts: 522 Location: Tucson, AZ
|
Posted: Sun May 29, 2005 11:51 pm Post subject: |
|
|
Quote: | if after u 'quilt push -a' lets say some patch dosent apply perfectly clean, 'quilt push -f', then use 'quilt edit <file> to clean up the reject, then quilt refresh, to fix the patch file then 'quilt push -a' to resume patching |
Ahhh...OK, I'll play with that. I kind of did some of that the first time I tried to skip Suspend2. I just didn't follow through.
I really do like this new system. It's going to make it alot easier to troubleshoot problems. _________________ --Aaron |
|
Back to top |
|
|
pharon n00b
Joined: 13 Mar 2005 Posts: 19
|
Posted: Mon May 30, 2005 6:48 am Post subject: |
|
|
didn't check yet if this broken in mm or not, but PCI express is broken:
Code: |
drivers/pci/pcie/portdrv_core.c: In function `suspend_iter':
drivers/pci/pcie/portdrv_core.c:326: error: incompatible type for argument 2 of indirect function call
drivers/pci/pcie/portdrv_core.c: At top level:
drivers/pci/pcie/portdrv_core.c:332: error: conflicting types for 'pcie_port_device_suspend'
drivers/pci/pcie/portdrv.h:34: error: previous declaration of 'pcie_port_device_suspend' was here
drivers/pci/pcie/portdrv_core.c:332: error: conflicting types for 'pcie_port_device_suspend'
drivers/pci/pcie/portdrv.h:34: error: previous declaration of 'pcie_port_device_suspend' was here
make[3]: *** [drivers/pci/pcie/portdrv_core.o] Error 1
make[2]: *** [drivers/pci/pcie] Error 2
make[1]: *** [drivers/pci] Error 2
make: *** [drivers] Error 2
|
swsup2 is broken, and also cannot be compiled as modules.
Code: |
CC kernel/power/suspend2_core/atomic_copy.o
kernel/power/suspend2_core/atomic_copy.c: In function
`suspend2_suspend_2':
kernel/power/suspend2_core/atomic_copy.c:150: warning: implicit
declaration of function `save_image_part1'
|
This is also broken, this error is typical to swsup2 I believe.
Code: |
fs/ocfs2/dlm/dlmthread.c:439:66: macro "kthread_run" requires 5 arguments, but only 3 given
fs/ocfs2/dlm/dlmthread.c: In function `dlm_launch_thread':
fs/ocfs2/dlm/dlmthread.c:439: error: `kthread_run' undeclared (first use in this function)
fs/ocfs2/dlm/dlmthread.c:439: error: (Each undeclared identifier is reported only once
fs/ocfs2/dlm/dlmthread.c:439: error: for each function it appears in.)
fs/ocfs2/dlm/dlmthread.c: At top level:
fs/ocfs2/dlm/dlmthread.c:579: warning: 'dlm_thread' defined but not used
make[3]: *** [fs/ocfs2/dlm/dlmthread.o] Error 1
make[2]: *** [fs/ocfs2/dlm] Error 2
make[1]: *** [fs/ocfs2] Error 2
make: *** [fs] Error 2
|
Distributed lock manager is also broken due to swsup2.
Code: |
drivers/dlm/ast.c:144:44: macro "kthread_run" requires 5 arguments, but only 3 given
drivers/dlm/ast.c: In function `dlm_astd_start':
drivers/dlm/ast.c:144: error: `kthread_run' undeclared (first use in this function)
drivers/dlm/ast.c:144: error: (Each undeclared identifier is reported only once
drivers/dlm/ast.c:144: error: for each function it appears in.)
drivers/dlm/ast.c: At top level:
drivers/dlm/ast.c:112: warning: 'dlm_astd' defined but not used
make[2]: *** [drivers/dlm/ast.o] Error 1
make[1]: *** [drivers/dlm] Error 2
make: *** [drivers] Error 2
|
I'll try to fix them if I can. Do you know a quick fix ? Thanks. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon May 30, 2005 1:27 pm Post subject: |
|
|
pharon wrote: | didn't check yet if this broken in mm or not, but PCI express is broken:
Code: |
drivers/pci/pcie/portdrv_core.c: In function `suspend_iter':
drivers/pci/pcie/portdrv_core.c:326: error: incompatible type for argument 2 of indirect function call
drivers/pci/pcie/portdrv_core.c: At top level:
drivers/pci/pcie/portdrv_core.c:332: error: conflicting types for 'pcie_port_device_suspend'
drivers/pci/pcie/portdrv.h:34: error: previous declaration of 'pcie_port_device_suspend' was here
drivers/pci/pcie/portdrv_core.c:332: error: conflicting types for 'pcie_port_device_suspend'
drivers/pci/pcie/portdrv.h:34: error: previous declaration of 'pcie_port_device_suspend' was here
make[3]: *** [drivers/pci/pcie/portdrv_core.o] Error 1
make[2]: *** [drivers/pci/pcie] Error 2
make[1]: *** [drivers/pci] Error 2
make: *** [drivers] Error 2
|
swsup2 is broken, and also cannot be compiled as modules.
Code: |
CC kernel/power/suspend2_core/atomic_copy.o
kernel/power/suspend2_core/atomic_copy.c: In function
`suspend2_suspend_2':
kernel/power/suspend2_core/atomic_copy.c:150: warning: implicit
declaration of function `save_image_part1'
|
This is also broken, this error is typical to swsup2 I believe.
Code: |
fs/ocfs2/dlm/dlmthread.c:439:66: macro "kthread_run" requires 5 arguments, but only 3 given
fs/ocfs2/dlm/dlmthread.c: In function `dlm_launch_thread':
fs/ocfs2/dlm/dlmthread.c:439: error: `kthread_run' undeclared (first use in this function)
fs/ocfs2/dlm/dlmthread.c:439: error: (Each undeclared identifier is reported only once
fs/ocfs2/dlm/dlmthread.c:439: error: for each function it appears in.)
fs/ocfs2/dlm/dlmthread.c: At top level:
fs/ocfs2/dlm/dlmthread.c:579: warning: 'dlm_thread' defined but not used
make[3]: *** [fs/ocfs2/dlm/dlmthread.o] Error 1
make[2]: *** [fs/ocfs2/dlm] Error 2
make[1]: *** [fs/ocfs2] Error 2
make: *** [fs] Error 2
|
Distributed lock manager is also broken due to swsup2.
Code: |
drivers/dlm/ast.c:144:44: macro "kthread_run" requires 5 arguments, but only 3 given
drivers/dlm/ast.c: In function `dlm_astd_start':
drivers/dlm/ast.c:144: error: `kthread_run' undeclared (first use in this function)
drivers/dlm/ast.c:144: error: (Each undeclared identifier is reported only once
drivers/dlm/ast.c:144: error: for each function it appears in.)
drivers/dlm/ast.c: At top level:
drivers/dlm/ast.c:112: warning: 'dlm_astd' defined but not used
make[2]: *** [drivers/dlm/ast.o] Error 1
make[1]: *** [drivers/dlm] Error 2
make: *** [drivers] Error 2
|
I'll try to fix them if I can. Do you know a quick fix ? Thanks. |
Yes Suspend2 is the big culprit in breaking things. I guess I need to start testing SMP support and fix some of those errors. As for the kthread_run errors that say 3 of 5 arguments are given. Try changing it to this:
from (a,b,c) to (a,b,0,c,NULL), that fixed the same error in hub.c for me and dosent segfault either _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
AaronPPC Guru
Joined: 29 May 2005 Posts: 522 Location: Tucson, AZ
|
Posted: Mon May 30, 2005 2:17 pm Post subject: |
|
|
Quote: | As for the kthread_run errors that say 3 of 5 arguments are given. Try changing it to this:
from (a,b,c) to (a,b,0,c,NULL) |
That did the trick for me. I'm compiling -skunk3 now.
I was able to get -skunk2 to compile after you told me how to handle quilt. It would not boot.
EDIT: -skunk3 didn't boot either. _________________ --Aaron |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon May 30, 2005 2:28 pm Post subject: |
|
|
pharon wrote: | didn't check yet if this broken in mm or not, but PCI express is broken:
swsup2 is broken, and also cannot be compiled as modules.
Code: |
CC kernel/power/suspend2_core/atomic_copy.o
kernel/power/suspend2_core/atomic_copy.c: In function
`suspend2_suspend_2':
kernel/power/suspend2_core/atomic_copy.c:150: warning: implicit
declaration of function `save_image_part1'
|
|
If you reverse suspend2/953-lowlevel-cleanup I think that fould fix any errors you get related to Suspend2 and SMP _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon May 30, 2005 2:35 pm Post subject: |
|
|
AaronPPC wrote: | Quote: | As for the kthread_run errors that say 3 of 5 arguments are given. Try changing it to this:
from (a,b,c) to (a,b,0,c,NULL) |
That did the trick for me. I'm compiling -skunk3 now.
I was able to get -skunk2 to compile after you told me how to handle quilt. It would not boot.
EDIT: -skunk3 didn't boot either. |
When/Where does it fail on boot. Does it crash immediately, do the fb drivers kick in at all? Do u get anysort of dump that says which module shit the brick? _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
AaronPPC Guru
Joined: 29 May 2005 Posts: 522 Location: Tucson, AZ
|
Posted: Mon May 30, 2005 10:37 pm Post subject: |
|
|
Quote: | When/Where does it fail on boot. Does it crash immediately, do the fb drivers kick in at all? Do u get anysort of dump that says which module shit the brick? |
Sorry, I should have posted this information sooner...
It strokes out right before the fb (radeonfb in my case) kicks in. Here is the immediate output I get, which is all normal. I don't know if it'll help or not. I imagine the immediate start up process is different on x86.
Code: | boot kernel-2.6.12-rc5-skunk3 root=/dev/hda3 video=radeonfb:1280x1024@75
chrpboot starting: loaded at 0x00800000
gunzipping (0x00010000 <-- 0x00806c88:0x0099551f)... done 3546064 bytes
46912 bytes of heap consumed, max in use 46912
start address=0x10000
instantiating rtas at 00600000... done
copying OF device tree... done
calling quiesce...
returning 0x00010000 from prom_init |
That is where the fb kicks in, only here the computer freezes. _________________ --Aaron |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon May 30, 2005 10:42 pm Post subject: |
|
|
do you use the new or old radeon driver, i know suspend2 patch 840-combined-pm.... that patch, edits the radeon fb drivers, I get a kobject segfault thing when I compile more than one fb device, if you are using radeonfb DONT use any other fb driver _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
AaronPPC Guru
Joined: 29 May 2005 Posts: 522 Location: Tucson, AZ
|
Posted: Tue May 31, 2005 4:47 am Post subject: |
|
|
Quote: | do you use the new or old radeon driver, |
I'm compiling in both the old and new drivers. It looks like I'll have to investigate and see which one I actually need.
Thanks for the advice! _________________ --Aaron |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|