Hi,
I'm highly disappointed of 12.1 and its kernel :-(
Before openSUSE 11.4 it was basically impossible to print on my local HP USB printer for several years (back to 2007 or so). With openSUSE 11.4 it finally worked (always!).
Now with 12.1 I got the very same issue as in the past.
Therefore the main question: Is it possible to use the 11.4 kernel on 12.1 (even if only for testing purposes)?
Second: What can I do to make openSUSE finally work with my printer?
Below are the log messages when the error happens.
From what I can see these are USB bus resets happening for no reason. In
the past people always claimed that some hardware is broken but I do not buy that argument as 11.4 demonstrated that it can just work with 100% same hardware. With luck I can print one page at the moment before the USB bus breaks and no single event like that happened with 11.4.
Any pointers?
Thanks, Wolfgang
Mar 6 18:37:11 Hygiea kernel: [45807.289799] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.290418] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.291544] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.292428] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.293054] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.293926] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.294541] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:11 Hygiea kernel: [45807.526456] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:35 Hygiea dbus-daemon[1228]: (packagekitd:21109): PackageKit-Zypp-DEBUG: zypp_backend_destroy Mar 6 18:37:42 Hygiea kernel: [45838.854035] usb 1-4: reset high speed USB device number 5 using ehci_hcd Mar 6 18:37:42 Hygiea hp[21391]: io/hpmud/musb.c 1043: bulk_write failed buf=0x7fff7b5aed40 size=8192 len=-34: Success Mar 6 18:37:42 Hygiea hp[21391]: io/hpmud/musb.c 1423: unable to write data hp:/usb/Officejet_6300_series?serial=CN69DCH2GW04M4: Success Mar 6 18:37:42 Hygiea kernel: [45838.969179] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:42 Hygiea kernel: [45838.969189] usb 1-4: usbfs: process 21391 (hp) did not claim interface 1 before use Mar 6 18:37:54 Hygiea kernel: [45850.923618] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:56 Hygiea kernel: [45852.924164] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:37:58 Hygiea kernel: [45854.924664] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:00 Hygiea kernel: [45856.925197] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.375039] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.375641] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.376770] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.377515] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.378140] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.379028] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.379637] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:01 Hygiea kernel: [45857.609801] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:32 Hygiea kernel: [45888.838037] usb 1-4: reset high speed USB device number 5 using ehci_hcd Mar 6 18:38:32 Hygiea hp[21613]: io/hpmud/musb.c 1043: bulk_write failed buf=0x7fffc0083ec0 size=8192 len=-34: Success Mar 6 18:38:32 Hygiea hp[21613]: io/hpmud/musb.c 1423: unable to write data hp:/usb/Officejet_6300_series?serial=CN69DCH2GW04M4: Success Mar 6 18:38:32 Hygiea kernel: [45888.953228] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:32 Hygiea kernel: [45888.953237] usb 1-4: usbfs: process 21613 (hp) did not claim interface 1 before use Mar 6 18:38:47 Hygiea kernel: [45904.055111] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:49 Hygiea kernel: [45906.055536] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:51 Hygiea kernel: [45908.056053] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:53 Hygiea kernel: [45910.056555] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.488905] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.489512] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.490387] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.491034] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.491512] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.492523] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.493139] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:38:54 Hygiea kernel: [45910.712807] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:39:25 Hygiea kernel: [45941.830032] usb 1-4: reset high speed USB device number 5 using ehci_hcd Mar 6 18:39:25 Hygiea hp[21952]: io/hpmud/musb.c 1043: bulk_write failed buf=0x7ffffb0b5d10 size=8192 len=-34: Success Mar 6 18:39:25 Hygiea hp[21952]: io/hpmud/musb.c 1423: unable to write data hp:/usb/Officejet_6300_series?serial=CN69DCH2GW04M4: Success Mar 6 18:39:25 Hygiea kernel: [45941.944987] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:39:25 Hygiea kernel: [45941.944996] usb 1-4: usbfs: process 21952 (hp) did not claim interface 1 before use Mar 6 18:39:31 Hygiea hp-toolbox: hp-toolbox[22258]: warning: Reportlab not installed. Fax coverpages disabled. Mar 6 18:39:31 Hygiea hp-toolbox: hp-toolbox[22258]: warning: Please install version 2.0+ of Reportlab for coverpage support. Mar 6 18:39:32 Hygiea hp[21952]: io/hpmud/musb.c 1043: bulk_write failed buf=0x7ffffb0b5d10 size=8192 len=-34: Success Mar 6 18:39:32 Hygiea kernel: [45948.423907] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:39:32 Hygiea hp[21952]: io/hpmud/musb.c 1423: unable to write data hp:/usb/Officejet_6300_series?serial=CN69DCH2GW04M4: Success Mar 6 18:39:32 Hygiea hp[21952]: io/hpmud/musb.c 749: invalid deviceid ret=-5: Input/output error Mar 6 18:39:32 Hygiea kernel: [45948.424048] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:39:32 Hygiea kernel: [45948.424058] usb 1-4: usbfs: process 21952 (hp) did not claim interface 1 before use Mar 6 18:39:32 Hygiea hp[21952]: prnt/backend/hp.c 625: ERROR: 5021 device communication error! Mar 6 18:39:32 Hygiea kernel: [45948.424352] usb 1-4: usbfs: process 21952 (hp) did not claim interface 1 before use Mar 6 18:39:32 Hygiea kernel: [45948.424904] Did not find alt setting 1 for intf 0, config 1 Mar 6 18:39:32 Hygiea kernel: [45948.425660] Did not find alt setting 1 for intf 0, config 1
On Tuesday, March 06, 2012 12:58 PM Wolfgang Rosenauer wrote:
Hi,
I'm highly disappointed of 12.1 and its kernel :-(
Before openSUSE 11.4 it was basically impossible to print on my local HP USB printer for several years (back to 2007 or so). With openSUSE 11.4 it finally worked (always!).
Now with 12.1 I got the very same issue as in the past.
Therefore the main question: Is it possible to use the 11.4 kernel on 12.1 (even if only for testing purposes)?
Second: What can I do to make openSUSE finally work with my printer?
Below are the log messages when the error happens.
From what I can see these are USB bus resets happening for no reason. In the past people always claimed that some hardware is broken but I do not buy that argument as 11.4 demonstrated that it can just work with 100% same hardware. With luck I can print one page at the moment before the USB bus breaks and no single event like that happened with 11.4.
Any pointers?
Thanks, Wolfgang
I have no idea what the kernel problem is, but what I've done with a hardware problem like this is set up multi-kernel and install another kernel with YaST. I've been using this repo:
http://download.opensuse.org/repositories/Kernel:/stable/standard/
On 03/06/2012 06:58 PM, Wolfgang Rosenauer wrote:
Therefore the main question: Is it possible to use the 11.4 kernel on 12.1 (even if only for testing purposes)?
Yes, that should work just fine. See Dennis's post on how to do it.
Second: What can I do to make openSUSE finally work with my printer?
I'm afraid, you have to bisect the kernel.
But first try vanilla flavor of both 11.4 and 12.1. Try the latest stable (3.2.9) and try 11-SP2 kernel (3.0). And maybe master. If they work, try also their vanilla flavors. And report the results here ;).
All at: http://download.opensuse.org/repositories/Kernel:/
regards,
Hi,
Am 06.03.2012 21:54, schrieb Jiri Slaby:
On 03/06/2012 06:58 PM, Wolfgang Rosenauer wrote:
Therefore the main question: Is it possible to use the 11.4 kernel on 12.1 (even if only for testing purposes)?
Yes, that should work just fine. See Dennis's post on how to do it.
Second: What can I do to make openSUSE finally work with my printer?
I'm afraid, you have to bisect the kernel.
But first try vanilla flavor of both 11.4 and 12.1. Try the latest stable (3.2.9) and try 11-SP2 kernel (3.0). And maybe master. If they work, try also their vanilla flavors. And report the results here ;).
Today I've tried the following kernels to verify:
kernel-desktop 3.1.10 - same issue :-( kernel-vanilla 3.1.10 - same issue kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
kernel-desktop 2.6.37.6 (from 11.4) - works like a charm!!! No issue at all with my printer
Which kernels should I try next given the above findings? (Or how to make the 3.0 kernel work on 12.1?
With kernel 2.6.37.6 there were no messages like these in the syslog: Mar 7 15:08:01 Hygiea kernel: [ 116.166116] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.168476] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.169599] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.170358] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.171093] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.171970] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.172469] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.765179] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:32 Hygiea kernel: [ 147.814041] usb 1-4: reset high speed USB device number 5 using ehci_hcd
Wolfgang
On Wednesday 07 March 2012 15.36:04 Wolfgang Rosenauer wrote:
Hi,
Am 06.03.2012 21:54, schrieb Jiri Slaby:
On 03/06/2012 06:58 PM, Wolfgang Rosenauer wrote:
Therefore the main question: Is it possible to use the 11.4 kernel on 12.1 (even if only for testing purposes)?
Yes, that should work just fine. See Dennis's post on how to do it.
Second: What can I do to make openSUSE finally work with my printer?
I'm afraid, you have to bisect the kernel.
But first try vanilla flavor of both 11.4 and 12.1. Try the latest stable (3.2.9) and try 11-SP2 kernel (3.0). And maybe master. If they work, try also their vanilla flavors. And report the results here ;).
Today I've tried the following kernels to verify:
kernel-desktop 3.1.10 - same issue :-( kernel-vanilla 3.1.10 - same issue kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
kernel-desktop 2.6.37.6 (from 11.4) - works like a charm!!! No issue at all with my printer
Which kernels should I try next given the above findings? (Or how to make the 3.0 kernel work on 12.1?
With kernel 2.6.37.6 there were no messages like these in the syslog: Mar 7 15:08:01 Hygiea kernel: [ 116.166116] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.168476] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.169599] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.170358] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.171093] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.171970] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.172469] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:01 Hygiea kernel: [ 116.765179] Did not find alt setting 1 for intf 0, config 1 Mar 7 15:08:32 Hygiea kernel: [ 147.814041] usb 1-4: reset high speed USB device number 5 using ehci_hcd
Wolfgang
To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Wolfgang perhaps the kotd which is a 3.3 version now? http://download.opensuse.org/repositories/Kernel:/HEAD/
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
regards,
On 03/07/2012 09:04 PM, Jiri Slaby wrote:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Anyway why do you need X for printing?
Am 07.03.2012 21:04, schrieb Jiri Slaby:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Actually it didn't even boot into a shell/getty. systemd just stopped somewhere and nothing worked anymore. So I had issues capturing the output. I'll check if vanilla works but I think it's more related to kernel packaging because 11.1/SLE11 was kind of different to 12.1 (kernel packaging structure).
Wolfgang
On 03/07/2012 09:09 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:04, schrieb Jiri Slaby:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Actually it didn't even boot into a shell/getty. systemd just stopped somewhere and nothing worked anymore. So I had issues capturing the output. I'll check if vanilla works but I think it's more related to kernel packaging because 11.1/SLE11 was kind of different to 12.1 (kernel packaging structure).
Yes, do you have *both* -desktop and -desktop-base installed?
Am 07.03.2012 21:14, schrieb Jiri Slaby:
On 03/07/2012 09:09 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:04, schrieb Jiri Slaby:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Actually it didn't even boot into a shell/getty. systemd just stopped somewhere and nothing worked anymore. So I had issues capturing the output. I'll check if vanilla works but I think it's more related to kernel packaging because 11.1/SLE11 was kind of different to 12.1 (kernel packaging structure).
Yes, do you have *both* -desktop and -desktop-base installed?
I had that. It's even a dependency. I can try again later or tomorrow.
Wolfgang
On 03/07/2012 09:24 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:14, schrieb Jiri Slaby:
On 03/07/2012 09:09 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:04, schrieb Jiri Slaby:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Actually it didn't even boot into a shell/getty. systemd just stopped somewhere and nothing worked anymore. So I had issues capturing the output. I'll check if vanilla works but I think it's more related to kernel packaging because 11.1/SLE11 was kind of different to 12.1 (kernel packaging structure).
Yes, do you have *both* -desktop and -desktop-base installed?
I had that. It's even a dependency. I can try again later or tomorrow.
You may need also -extra for specific controllers...
Booting with sysvinit might help as well. You need sysvinit package and specify init=/sbin/sysvinit in grub.
regards,
Am 07.03.2012 21:28, schrieb Jiri Slaby:
On 03/07/2012 09:24 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:14, schrieb Jiri Slaby:
On 03/07/2012 09:09 PM, Wolfgang Rosenauer wrote:
Am 07.03.2012 21:04, schrieb Jiri Slaby:
On 03/07/2012 03:36 PM, Wolfgang Rosenauer wrote:
kernel-desktop 3.0.22 (from SLE11-SP2) - was not able to boot into X ; no testing possible
Why? Any errors in dmesg or Xorg.log? Does SP2 -vanilla work?
Actually it didn't even boot into a shell/getty. systemd just stopped somewhere and nothing worked anymore. So I had issues capturing the output. I'll check if vanilla works but I think it's more related to kernel packaging because 11.1/SLE11 was kind of different to 12.1 (kernel packaging structure).
Yes, do you have *both* -desktop and -desktop-base installed?
I had that. It's even a dependency. I can try again later or tomorrow.
You may need also -extra for specific controllers...
Just tried again. My system is all ext4 and the SLE11 kernel only supports ext4 in read-only mode? Seriously? So nothing to check for it :-( Getting this nasty bug fixed is still my priority. What next step does help?
Wolfgang
Hi;
Am Wed 07 Mar 2012 10:39:40 PM CET schrieb Wolfgang Rosenauer wolfgang@rosenauer.org:
Just tried again. My system is all ext4 and the SLE11 kernel only supports ext4 in read-only mode? Seriously? So nothing to check for it :-( Getting this nasty bug fixed is still my priority. What next step does help?
You need ext4-writable module to be able to write into ext4 partitions.
Regards.
Hi,
Am 07.03.2012 22:54, schrieb Ismail Doenmez:
Hi;
Am Wed 07 Mar 2012 10:39:40 PM CET schrieb Wolfgang Rosenauer wolfgang@rosenauer.org:
Just tried again. My system is all ext4 and the SLE11 kernel only supports ext4 in read-only mode? Seriously? So nothing to check for it :-( Getting this nasty bug fixed is still my priority. What next step does help?
You need ext4-writable module to be able to write into ext4 partitions.
Ok, can someone please let me know what I need to do exactly to get that kernel running? Hygiea:/lib/modules/3.0.22-8-desktop # find . -name "*writable*" Hygiea:/lib/modules/3.0.22-8-desktop #
Or any other kernel I should try. Seems that SLE11 kernel is not compatible in any way with 12.1. I almost do everything to get enough debug information to you guys so that this issue can finally be fixed (again) but I'm not a kernel developer.
Wolfgang
Hi,
Am 07.03.2012 22:54, schrieb Ismail Doenmez:
Hi;
Am Wed 07 Mar 2012 10:39:40 PM CET schrieb Wolfgang Rosenauer wolfgang@rosenauer.org:
Just tried again. My system is all ext4 and the SLE11 kernel only supports ext4 in read-only mode? Seriously? So nothing to check for it :-( Getting this nasty bug fixed is still my priority. What next step does help?
You need ext4-writable module to be able to write into ext4 partitions.
ok, I got the module from someone finally. Thanks for that!
Updated list:
kernel-desktop 3.1.9 - unexpected USB reset while printing kernel-desktop 3.1.10 - unexpected USB reset while printing kernel-vanilla 3.1.10 - unexpected USB reset while printing kernel-desktop 2.6.37.6 (from 11.4) - no problems kernel-default 3.0.22 (from SLE11-SP2) - no problems
What's next?
Thanks, Wolfgang
Am Freitag, 9. März 2012, 07:44:43 schrieb Wolfgang Rosenauer:
Updated list:
kernel-desktop 3.1.9 - unexpected USB reset while printing kernel-desktop 3.1.10 - unexpected USB reset while printing kernel-vanilla 3.1.10 - unexpected USB reset while printing kernel-desktop 2.6.37.6 (from 11.4) - no problems kernel-default 3.0.22 (from SLE11-SP2) - no problems
You can verify that 3.0 vanilla works and if so bisect.
Regards Oliver
Am 09.03.2012 08:12, schrieb Oliver Neukum:
Am Freitag, 9. März 2012, 07:44:43 schrieb Wolfgang Rosenauer:
Updated list:
kernel-desktop 3.1.9 - unexpected USB reset while printing kernel-desktop 3.1.10 - unexpected USB reset while printing kernel-vanilla 3.1.10 - unexpected USB reset while printing kernel-desktop 2.6.37.6 (from 11.4) - no problems kernel-default 3.0.22 (from SLE11-SP2) - no problems
You can verify that 3.0 vanilla works and if so bisect.
I know that 3.0.22-default works I know that 3.1.0-vanilla (from 12.1 GA) does not work
I'm not sure how to minimize the regression window even further?
Wolfgang
On Friday 09 March 2012 05:51:28 pm Wolfgang Rosenauer wrote:
Am 09.03.2012 08:12, schrieb Oliver Neukum:
Am Freitag, 9. März 2012, 07:44:43 schrieb Wolfgang Rosenauer:
Updated list:
kernel-desktop 3.1.9 - unexpected USB reset while printing kernel-desktop 3.1.10 - unexpected USB reset while printing kernel-vanilla 3.1.10 - unexpected USB reset while printing kernel-desktop 2.6.37.6 (from 11.4) - no problems kernel-default 3.0.22 (from SLE11-SP2) - no problems
You can verify that 3.0 vanilla works and if so bisect.
I know that 3.0.22-default works I know that 3.1.0-vanilla (from 12.1 GA) does not work
I'm not sure how to minimize the regression window even further?
Just as Oliver told you: try to reproduce with 3.0.22-vanilla. I do expect it to work but you should never make assumptions before you start a bisection.
Then, assuming 3.0.22-vanilla works and 3.1.0-vanilla does not, you are down to an upstream kernel bisection. Best way then is to clone the upstream kernel git repository and build the kernels locally on the problematic machine, it will take some time (10-12 iterations are typically necessary between two consecutive kernel versions) but works fine in general.
Am 12.03.2012 11:20, schrieb Jean Delvare:
Just as Oliver told you: try to reproduce with 3.0.22-vanilla. I do expect it to work but you should never make assumptions before you start a bisection.
I've already confirmed that 3.0.22-vanilla works.
Then, assuming 3.0.22-vanilla works and 3.1.0-vanilla does not, you are down to an upstream kernel bisection. Best way then is to clone the upstream kernel git repository and build the kernels locally on the problematic machine, it will take some time (10-12 iterations are typically necessary between two consecutive kernel versions) but works fine in general.
The fact that I'm subscribed to opensuse-kernel does not mean that I actually know how that stuff works. I used to but my knowledge is not current anymore. So please point me to some information - how to configure the kernel correctly so I can actually boot my 12.1 system. As said I'm not totally clueless but the last time I configured and built a kernel was some years ago. (I remember make cloneconfig but not sure if it works for vanilla) - what revisions to build actually? I do not follow the kernel release workflow so what kernels were actually inbetween 3.0 and 3.1?
Thanks, Wolfgang
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/12/2012 11:27 AM, Wolfgang Rosenauer wrote:
Am 12.03.2012 11:20, schrieb Jean Delvare:
Just as Oliver told you: try to reproduce with 3.0.22-vanilla. I do expect it to work but you should never make assumptions before you start a bisection.
I've already confirmed that 3.0.22-vanilla works.
Then, assuming 3.0.22-vanilla works and 3.1.0-vanilla does not, you are down to an upstream kernel bisection. Best way then is to clone the upstream kernel git repository and build the kernels locally on the problematic machine, it will take some time (10-12 iterations are typically necessary between two consecutive kernel versions) but works fine in general.
The fact that I'm subscribed to opensuse-kernel does not mean that I actually know how that stuff works. I used to but my knowledge is not current anymore. So please point me to some information - how to configure the kernel correctly so I can actually boot my 12.1 system. As said I'm not totally clueless but the last time I configured and built a kernel was some years ago. (I remember make cloneconfig but not sure if it works for vanilla) - what revisions to build actually? I do not follow the kernel release workflow so what kernels were actually inbetween 3.0 and 3.1?
Thanks, Wolfgang
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git cd linux-2.6 git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
# bisection work: cp -a /boot/config-3.1.9-1.4-default .config sed -i s/CONFIG_DEBUG_INFO.*// .config # speeds up compile make cloneconfig # build with make -j4 # or number of your CPU cores # if it did not build, run git bisect skip # try the resulting kernel (under arch/x86/boot/bzImage) # if it worked, do git bisect good # if it failed, do git bisect bad # which will checkout another middle version # repeat above bisection work until done
even though the openSUSE config is not so good for bisecting as it contains a huge number of modules you will never need, so build will take very long. maybe make localmodconfig could help reducing build time
Ciao Bernhard M.
Am 12.03.2012 14:17, schrieb Bernhard M. Wiedemann:
[ a nice "kernel bisection for beginners" guide ]
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git cd linux-2.6 git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
# bisection work: cp -a /boot/config-3.1.9-1.4-default .config sed -i s/CONFIG_DEBUG_INFO.*// .config # speeds up compile make cloneconfig # build with make -j4 # or number of your CPU cores # if it did not build, run git bisect skip # try the resulting kernel (under arch/x86/boot/bzImage)
Something like "make install" is needed here (but I have to admit that I have not used a complete self-built kernel on openSUSE for quite some time, so it might be something else. Anyway. You need to install the built kernel. preferably without overwriting the rpm-installed one here.)
# if it worked, do git bisect good # if it failed, do git bisect bad # which will checkout another middle version # repeat above bisection work until done
On Monday 12 March 2012 02:21:18 pm Stefan Seyfried wrote:
Am 12.03.2012 14:17, schrieb Bernhard M. Wiedemann:
[ a nice "kernel bisection for beginners" guide ]
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.gi t cd linux-2.6 git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
# bisection work: cp -a /boot/config-3.1.9-1.4-default .config sed -i s/CONFIG_DEBUG_INFO.*// .config # speeds up compile
It's not just a matter of speeding up... kernels built with CONFIG_DEBUG_INFO are huge, and you don't necessarily have 4 GB of hard disk drive at hand.
I also suggest stripping LOCALVERSION in the configuration file, as it has no meaning while performing the bisection.
make cloneconfig
I do "make oldconfig" here instead... cloneconfig seems redundant as you copied the configuration file manually in the first place.
# build with make -j4 # or number of your CPU cores # if it did not build, run git bisect skip # try the resulting kernel (under arch/x86/boot/bzImage)
Something like "make install" is needed here (but I have to admit that I have not used a complete self-built kernel on openSUSE for quite some time, so it might be something else. Anyway. You need to install the built kernel. preferably without overwriting the rpm-installed one here.)
Actually: # make modules_install && make install is needed.
# if it worked, do git bisect good # if it failed, do git bisect bad # which will checkout another middle version # repeat above bisection work until done
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/12/2012 10:28 AM, Jean Delvare wrote:
On Monday 12 March 2012 02:21:18 pm Stefan Seyfried wrote:
Am 12.03.2012 14:17, schrieb Bernhard M. Wiedemann:
[ a nice "kernel bisection for beginners" guide ]
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.gi
t cd linux-2.6
git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
# bisection work: cp -a /boot/config-3.1.9-1.4-default .config sed -i s/CONFIG_DEBUG_INFO.*// .config # speeds up compile
It's not just a matter of speeding up... kernels built with CONFIG_DEBUG_INFO are huge, and you don't necessarily have 4 GB of hard disk drive at hand.
If you do and don't want huge modules in /lib, you can install with:
make modules_install INSTALL_MOD_STRIP=1
- -Jeff
I also suggest stripping LOCALVERSION in the configuration file, as it has no meaning while performing the bisection.
make cloneconfig
I do "make oldconfig" here instead... cloneconfig seems redundant as you copied the configuration file manually in the first place.
# build with make -j4 # or number of your CPU cores # if it did not build, run git bisect skip # try the resulting kernel (under arch/x86/boot/bzImage)
Something like "make install" is needed here (but I have to admit that I have not used a complete self-built kernel on openSUSE for quite some time, so it might be something else. Anyway. You need to install the built kernel. preferably without overwriting the rpm-installed one here.)
Actually: # make modules_install && make install is needed.
# if it worked, do git bisect good # if it failed, do git bisect bad # which will checkout another middle version # repeat above bisection work until done
- -- Jeff Mahoney SUSE Labs
Hi,
Am 12.03.2012 15:41, schrieb Jeff Mahoney:
On 03/12/2012 10:28 AM, Jean Delvare wrote:
On Monday 12 March 2012 02:21:18 pm Stefan Seyfried wrote:
Am 12.03.2012 14:17, schrieb Bernhard M. Wiedemann:
[ a nice "kernel bisection for beginners" guide ]
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.gi
t cd linux-2.6
git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
I've spent quite some time bisecting something. I hope I did it right because the outcome does not look obvious to me.
Please check https://bugzilla.novell.com/show_bug.cgi?id=751712 and let me know if I need to do something else. I've put a lot of work (and paper ;-) into it and I'm willing to fight to the end to get that issue finally resolved again but I need some help from kernel experts.
Thanks, Wolfgang
On 03/13/2012 04:03 PM, Wolfgang Rosenauer wrote:
Hi,
Am 12.03.2012 15:41, schrieb Jeff Mahoney:
On 03/12/2012 10:28 AM, Jean Delvare wrote:
On Monday 12 March 2012 02:21:18 pm Stefan Seyfried wrote:
Am 12.03.2012 14:17, schrieb Bernhard M. Wiedemann:
[ a nice "kernel bisection for beginners" guide ]
you do something like this:
# initial setup: git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.gi
t cd linux-2.6
git bisect start git bisect good v3.0 git bisect bad v3.1 # this will automatically checkout something from the middle
I've spent quite some time bisecting something. I hope I did it right because the outcome does not look obvious to me.
Please check https://bugzilla.novell.com/show_bug.cgi?id=751712 and let me know if I need to do something else. I've put a lot of work (and paper ;-) into it and I'm willing to fight to the end to get that issue finally resolved again but I need some help from kernel experts.
I just posted on the Bugzilla, but will repeat here. The commit ID you posted is for a merge, thus it has no associated code, and cannot be the faulty one. Please post the commit IDs (7 digits from the left is enough) for kernels near the end of the bisection that definitely work, and those that definitely fail.
Larry
Is all of this documented in the kernel bugs section of our wiki (see bugs.opensuse.org) ?
If not, could somebody do it, please?
Andreas
-
Andreas Jaeger -
Bernhard M. Wiedemann -
Bruno Friedmann -
Dennis Gallien -
Ismail Doenmez -
Jean Delvare -
Jeff Mahoney -
Jiri Slaby -
Larry Finger -
Oliver Neukum -
Stefan Seyfried -
Wolfgang Rosenauer