Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Tue, Oct 14, 2008 at 11:06:18PM +0200, Bastian Blank wrote: On Tue, Oct 14, 2008 at 04:15:29PM +0200, Aurelien Jarno wrote: The proposed patch has been applied. But maybe it is wrong. Waldi, could you please have a look? I have no Xen machine. The patch is correct. The problem is the wrong capability definition in the Etch kernel. But fixing the package to work together with the Lenny kernel is more important. Maybe I should fix the kernel in Etch also. Given how Xen works, a lot of people are not able to change their kernel. IMHO it is a really bad idea to not support Xen kernels from Etch. -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Tue, Oct 14, 2008 at 09:20:12PM +, Andy Smith wrote: On Tue, Oct 14, 2008 at 11:06:18PM +0200, Bastian Blank wrote: On Tue, Oct 14, 2008 at 04:15:29PM +0200, Aurelien Jarno wrote: The proposed patch has been applied. But maybe it is wrong. Waldi, could you please have a look? I have no Xen machine. The patch is correct. The problem is the wrong capability definition in the Etch kernel. But fixing the package to work together with the Lenny kernel is more important. Maybe I should fix the kernel in Etch also. Thanks Bastian. What is your best advice for a workaround since I can't change dom0 from the packaged Etch hypervisor and current Lenny kernels do not boot with that hypervisor? Could you try to edit /etc/ld.so.conf.d/libc6-xen.conf, and add back the line hwcap 0 nosegneg, while keeping the line hwcap 1 nosegneg? If it works we should be able to support both Etch and Lenny kernels. But that would mean one more upload. Shrug. -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
Aurelien Jarno [EMAIL PROTECTED] wrote: Could you try to edit /etc/ld.so.conf.d/libc6-xen.conf, and add back the line hwcap 0 nosegneg, while keeping the line hwcap 1 nosegneg? If it works we should be able to support both Etch and Lenny kernels. But that would mean one more upload. Shrug. That didn't seem to work, unfortunately: $ sudo ldconfig ldconfig: /etc/ld.so.conf.d/libc6-xen.conf:7: hwcap index 1 already defined as nosegneg Henrik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 01:49:11PM +0200, [EMAIL PROTECTED] wrote: Andy Smith [EMAIL PROTECTED] wrote: Unfortunately ldconfig will not allow the same hwconfig twice. For now I reverted it to 0 and rebooted; this has got rid of the warnings. I will just have to bear it in mind for when I can again use the Lenny kernel for this Lenny domU I suppose. That did it, thanks! Luckily I managed without rebooting. I just edited libc6-xen.conf, ran ldconfig and restarted several processes and services. Now only init (pid 1) complains once in a while. You can re-exec init with telinit u to fix this. Cheers, Andy signature.asc Description: Digital signature
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 12:19:49 +0200 Bastian Blank wrote: On Wed, Oct 15, 2008 at 08:55:01AM +0200, Aurelien Jarno wrote: Given how Xen works, a lot of people are not able to change their kernel. This is incorrect. There are now two possibilities to use the guest supplied kernels and the whole bunch of enterprise distributions never really supported it otherwise. One of them can even be considered safe. Could you please provide pointers to these two possibilities? And which one is considered safe? IMHO it is a really bad idea to not support Xen kernels from Etch. I don't know how. The way the kernel in Etch responds to the problem is also not pretty, but this was not really clear at that time of the Etch release. Could you please clarify what's the current (or emerging) best practice for running Xen with Debian? The latest I know of: for dom0: Lenny with the Etch kernel for domU: Lenny This libc update breaks dom0, and xen-linux-system-2.6.26-1-xen-686 from unstable doesn't boot for me either. Should it? -- Thanks, Feri. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 12:19:49PM +0200, Bastian Blank wrote: On Wed, Oct 15, 2008 at 08:55:01AM +0200, Aurelien Jarno wrote: Given how Xen works, a lot of people are not able to change their kernel. This is incorrect. There are now two possibilities to use the guest supplied kernels and the whole bunch of enterprise distributions never really supported it otherwise. One of them can even be considered safe. What I mean is that a lot of users are using Xen on a VPS, and their provider don't give them the choice about the kernel... -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
Andy Smith [EMAIL PROTECTED] wrote: Unfortunately ldconfig will not allow the same hwconfig twice. For now I reverted it to 0 and rebooted; this has got rid of the warnings. I will just have to bear it in mind for when I can again use the Lenny kernel for this Lenny domU I suppose. That did it, thanks! Luckily I managed without rebooting. I just edited libc6-xen.conf, ran ldconfig and restarted several processes and services. Now only init (pid 1) complains once in a while. Henrik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 08:56:57AM +0200, Aurelien Jarno wrote: On Tue, Oct 14, 2008 at 09:20:12PM +, Andy Smith wrote: Thanks Bastian. What is your best advice for a workaround since I can't change dom0 from the packaged Etch hypervisor and current Lenny kernels do not boot with that hypervisor? Could you try to edit /etc/ld.so.conf.d/libc6-xen.conf, and add back the line hwcap 0 nosegneg, while keeping the line hwcap 1 nosegneg? Unfortunately ldconfig will not allow the same hwconfig twice. For now I reverted it to 0 and rebooted; this has got rid of the warnings. I will just have to bear it in mind for when I can again use the Lenny kernel for this Lenny domU I suppose. Cheers, Andy signature.asc Description: Digital signature
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 08:55:01AM +0200, Aurelien Jarno wrote: Given how Xen works, a lot of people are not able to change their kernel. This is incorrect. There are now two possibilities to use the guest supplied kernels and the whole bunch of enterprise distributions never really supported it otherwise. One of them can even be considered safe. IMHO it is a really bad idea to not support Xen kernels from Etch. I don't know how. The way the kernel in Etch responds to the problem is also not pretty, but this was not really clear at that time of the Etch release. Bastian -- Without followers, evil cannot spread. -- Spock, And The Children Shall Lead, stardate 5029.5 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Wed, Oct 15, 2008 at 14:45:21 +0200 Ferenc Wagner wrote: xen-linux-system-2.6.26-1-xen-686 from unstable doesn't boot for me either. Should it? It actually does, I was misguided by the hvc vs xvc vs serial console flow control issues. Unfortunately qla2xxx is still broken, but that belongs to another bug report. So this answers my question about the preferred Debian Xen dom0 setup. Thanks for bringing things to this point! -- Feri. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Tue, Oct 14, 2008 at 04:15:29PM +0200, Aurelien Jarno wrote: The proposed patch has been applied. But maybe it is wrong. Waldi, could you please have a look? I have no Xen machine. The patch is correct. The problem is the wrong capability definition in the Etch kernel. But fixing the package to work together with the Lenny kernel is more important. Maybe I should fix the kernel in Etch also. Bastian -- Beauty is transitory. Beauty survives. -- Spock and Kirk, That Which Survives, stardate unknown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
Hi, Ever since updating to libc6-xen 2.7-14 I am seeing the 4gb seg fixup warnings: Oct 14 10:43:34 cacti kernel: printk: 11610 messages suppressed. Oct 14 10:43:34 cacti kernel: 4gb seg fixup, process head (pid 5876), cs:ip 73:b7ebb33a Oct 14 10:43:38 cacti init: Trying to re-exec init Oct 14 10:43:39 cacti kernel: printk: 1038209 messages suppressed. Oct 14 10:43:39 cacti kernel: 4gb seg fixup, process libc6.postinst (pid 6135), cs:ip 73:b7e5b948 Oct 14 10:43:44 cacti kernel: printk: 522190 messages suppressed. Oct 14 10:43:44 cacti kernel: 4gb seg fixup, process dpkg (pid 6179), cs:ip 73:b7e75eb7 Oct 14 10:43:49 cacti kernel: printk: 645077 messages suppressed. Oct 14 10:43:49 cacti kernel: 4gb seg fixup, process dpkg (pid 6356), cs:ip 73:b7e6eeb7 Oct 14 10:43:54 cacti kernel: printk: 471625 messages suppressed. Oct 14 10:43:54 cacti kernel: 4gb seg fixup, process dpkg (pid 6374), cs:ip 73:b7df1860 Oct 14 10:44:00 cacti kernel: printk: 62714 messages suppressed. [...] Oct 14 11:21:41 cacti kernel: 4gb seg fixup, process apache2 (pid 9543), cs:ip 73:b7e0450e Oct 14 11:21:59 cacti kernel: printk: 58 messages suppressed. Oct 14 11:21:59 cacti kernel: 4gb seg fixup, process apache2 (pid 9544), cs:ip 73:b7ee0598 Oct 14 11:21:59 cacti kernel: 4gb seg fixup, process apache2 (pid 9544), cs:ip 73:b7ee32e5 Oct 14 11:23:27 cacti kernel: 4gb seg fixup, process sshd (pid 9545), cs:ip 73:b7ba9eb7 Oct 14 11:24:01 cacti kernel: printk: 1972 messages suppressed. Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7e4e8c0 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7f1d810 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7e4feb7 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7f1d810 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7dba967 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7dba96a Oct 14 11:24:02 cacti kernel: printk: 663233 messages suppressed. Oct 14 11:24:02 cacti kernel: 4gb seg fixup, process php (pid 9564), cs:ip 73:b7a08f44 Restarting things doesn't help, restarting the domU doesn't help. Is there something I have missed or doesn't this update actually fix it? Cheers, Andy signature.asc Description: Digital signature
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Tue, Oct 14, 2008 at 11:06:18PM +0200, Bastian Blank wrote: On Tue, Oct 14, 2008 at 04:15:29PM +0200, Aurelien Jarno wrote: The proposed patch has been applied. But maybe it is wrong. Waldi, could you please have a look? I have no Xen machine. The patch is correct. The problem is the wrong capability definition in the Etch kernel. But fixing the package to work together with the Lenny kernel is more important. Maybe I should fix the kernel in Etch also. Thanks Bastian. What is your best advice for a workaround since I can't change dom0 from the packaged Etch hypervisor and current Lenny kernels do not boot with that hypervisor? Cheers, Andy -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499366: 4gb seg fixup not fixed in 2.7-14?
On Tue, Oct 14, 2008 at 11:47:51AM +, Andy Smith wrote: Hi, Ever since updating to libc6-xen 2.7-14 I am seeing the 4gb seg fixup warnings: Oct 14 10:43:34 cacti kernel: printk: 11610 messages suppressed. Oct 14 10:43:34 cacti kernel: 4gb seg fixup, process head (pid 5876), cs:ip 73:b7ebb33a Oct 14 10:43:38 cacti init: Trying to re-exec init Oct 14 10:43:39 cacti kernel: printk: 1038209 messages suppressed. Oct 14 10:43:39 cacti kernel: 4gb seg fixup, process libc6.postinst (pid 6135), cs:ip 73:b7e5b948 Oct 14 10:43:44 cacti kernel: printk: 522190 messages suppressed. Oct 14 10:43:44 cacti kernel: 4gb seg fixup, process dpkg (pid 6179), cs:ip 73:b7e75eb7 Oct 14 10:43:49 cacti kernel: printk: 645077 messages suppressed. Oct 14 10:43:49 cacti kernel: 4gb seg fixup, process dpkg (pid 6356), cs:ip 73:b7e6eeb7 Oct 14 10:43:54 cacti kernel: printk: 471625 messages suppressed. Oct 14 10:43:54 cacti kernel: 4gb seg fixup, process dpkg (pid 6374), cs:ip 73:b7df1860 Oct 14 10:44:00 cacti kernel: printk: 62714 messages suppressed. [...] Oct 14 11:21:41 cacti kernel: 4gb seg fixup, process apache2 (pid 9543), cs:ip 73:b7e0450e Oct 14 11:21:59 cacti kernel: printk: 58 messages suppressed. Oct 14 11:21:59 cacti kernel: 4gb seg fixup, process apache2 (pid 9544), cs:ip 73:b7ee0598 Oct 14 11:21:59 cacti kernel: 4gb seg fixup, process apache2 (pid 9544), cs:ip 73:b7ee32e5 Oct 14 11:23:27 cacti kernel: 4gb seg fixup, process sshd (pid 9545), cs:ip 73:b7ba9eb7 Oct 14 11:24:01 cacti kernel: printk: 1972 messages suppressed. Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7e4e8c0 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7f1d810 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7e4feb7 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7f1d810 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7dba967 Oct 14 11:24:01 cacti kernel: 4gb seg fixup, process sh (pid 9548), cs:ip 73:b7dba96a Oct 14 11:24:02 cacti kernel: printk: 663233 messages suppressed. Oct 14 11:24:02 cacti kernel: 4gb seg fixup, process php (pid 9564), cs:ip 73:b7a08f44 Restarting things doesn't help, restarting the domU doesn't help. Is there something I have missed or doesn't this update actually fix it? The proposed patch has been applied. But maybe it is wrong. Waldi, could you please have a look? I have no Xen machine. -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]