Hey guys. I just compiled zaptel ver 1.2.5 into a DomU on xen 3.0.1 unstable
and got a kernel bug message after I modprobe ztdummy.
The zaptel module inserts just fine, but ztdummy causes this message. I have
RTC enabled in this kernel and my timing source set to 1000.
------------[ cut here ]------------
kernel BUG at kernel/timer.c:212!
invalid opcode: 0000 [#1]
Modules linked in: ztdummy zaptel
CPU: 0
EIP: 0061:[<c01224c3>] Not tainted VLI
EFLAGS: 00010046 (2.6.16-xenU #16)
EIP is at __mod_timer+0x83/0xa0
eax: 00000000 ebx: c04ec250 ecx: fbfef000 edx: 00000000
esi: 00000000 edi: 00000001 ebp: ccf33e70 esp: ccf33e5c
ds: 007b es: 007b ss: 0069
Process modprobe (pid: 1422, threadinfo=ccf32000 task=c13b7570)
Stack: <0>ccf33e60 00000000 cf6888c0 00000000 00000000 ccf33f4c c0232d15
c04ec250
00054f56 00000000 00000000 c01896c2 c04e3890 00000000 ccf33e98 00000412
ccf33eac c0189990 00008000 cf6888c0 ccf33ecc c0189d9f cf6883c0 cf6888c0
Call Trace:
[<c010512a>] show_stack_log_lvl+0xaa/0xe0
[<c010533e>] show_registers+0x18e/0x210
[<c0105539>] die+0xd9/0x180
[<c0105669>] do_trap+0x89/0xd0
[<c01059e8>] do_invalid_op+0xb8/0xd0
[<c0104d4f>] error_code+0x2b/0x30
[<c0232d15>] rtc_do_ioctl+0x7c5/0xa40
[<c0233347>] rtc_control+0x67/0xb0
[<d103519d>] init_module+0x19d/0x210 [ztdummy]
[<c0134eb5>] sys_init_module+0x135/0x1c0
[<c0104bd9>] syscall_call+0x7/0xb
Code: 08 89 5c 24 04 89 14 24 e8 2b fe ff ff 8b 45 f4 8b 0d 3c 40 42 c0 84 c0
88 41 01 75 05 80 39 00 75 17 83 c4 0c 89 f0 5b 5e 5d c3 <0f> 0b d4 00 f3
99 3d
c0 eb 87 89 ca eb c2 e8 ba c7 16 00 eb e2
--
Jason
The place where you made your stand never mattered,
only that you were there... and still on your feet
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
On 8 May 2006, at 07:37, Jason wrote:> Hey guys. I just compiled zaptel ver 1.2.5 into a DomU on xen 3.0.1 > unstable and got a kernel bug message after I modprobe ztdummy. > The zaptel module inserts just fine, but ztdummy causes this message. > I have > RTC enabled in this kernel and my timing source set to 1000.Do you get the same thing on current unstable (3.0.3 development tree)? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
I have similar problems with Xen 3.0.2 and a 2.6.16 xenU kernel. I have have the kernel bug issue as well if I try to load RTC it fails with: rtc: IRQ 8 is not free. I could get it to work fine on the 3.0.1 version though with a 2.6.12.6 XenU kernel. Thanks John -----Original Message----- From: xen-devel-bounces@lists.xensource.com [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Keir Fraser Sent: Monday, May 08, 2006 5:05 AM To: Jason Cc: xen-devel@lists.xensource.com Subject: Re: [Xen-devel] Kernel Bug asterisk On 8 May 2006, at 07:37, Jason wrote:> Hey guys. I just compiled zaptel ver 1.2.5 into a DomU on xen 3.0.1 > unstable and got a kernel bug message after I modprobe ztdummy. > The zaptel module inserts just fine, but ztdummy causes this message. > I have > RTC enabled in this kernel and my timing source set to 1000.Do you get the same thing on current unstable (3.0.3 development tree)? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
I''ll give it a shot and see what happens. Does anyone else have a 3.0.3 box that they would be willing to try out the zaptel modules on? I am moving this week and won''t be able to test until sometime early next week. -- Jason The place where you made your stand never mattered, only that you were there... and still on your feet On Mon, 8 May 2006, Keir Fraser wrote:> > On 8 May 2006, at 07:37, Jason wrote: > >> Hey guys. I just compiled zaptel ver 1.2.5 into a DomU on xen 3.0.1 >> unstable and got a kernel bug message after I modprobe ztdummy. >> The zaptel module inserts just fine, but ztdummy causes this message. I >> have >> RTC enabled in this kernel and my timing source set to 1000. > > Do you get the same thing on current unstable (3.0.3 development tree)? > > -- Keir > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 8 May 2006, at 19:34, John Cianfarani wrote:> I have similar problems with Xen 3.0.2 and a 2.6.16 xenU kernel. > I have have the kernel bug issue as well if I try to load RTC it fails > with: > rtc: IRQ 8 is not free. > > I could get it to work fine on the 3.0.1 version though with a 2.6.12.6 > XenU kernel.Only one domain can own the RTC. We had problems with some user-space programs hanging if the RTC driver was loaded but unable to bind to IRQ 8 or access the RTC I/O ports. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Am Dienstag, 9. Mai 2006 10:18 schrieb Keir Fraser:> > On 8 May 2006, at 19:34, John Cianfarani wrote: > > > I have similar problems with Xen 3.0.2 and a 2.6.16 xenU kernel. > > I have have the kernel bug issue as well if I try to load RTC it fails > > with: > > rtc: IRQ 8 is not free. > > > > I could get it to work fine on the 3.0.1 version though with a 2.6.12.6 > > XenU kernel. > > Only one domain can own the RTC. We had problems with some user-space > programs hanging if the RTC driver was loaded but unable to bind to IRQ > 8 or access the RTC I/O ports.I am having the problem, that hwclock hangs, if rtc is compiled into the kernel. Is there a easy fix possible, to avoid systems hanging after a dist-upgrade? Johnny> > -- Keir > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 9 May 2006, at 16:30, Hans-Christian Armingeon wrote:>> Only one domain can own the RTC. We had problems with some user-space >> programs hanging if the RTC driver was loaded but unable to bind to >> IRQ >> 8 or access the RTC I/O ports. > > I am having the problem, that hwclock hangs, if rtc is compiled into > the kernel. > > Is there a easy fix possible, to avoid systems hanging after a > dist-upgrade?Not allowing RTC to bind to IRQ8 is our current workaround. Are you seeing the hang in domain0 or in other domains? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Am Mittwoch, 10. Mai 2006 14:20 schrieb Keir Fraser:> > On 9 May 2006, at 16:30, Hans-Christian Armingeon wrote: > > >> Only one domain can own the RTC. We had problems with some user-space > >> programs hanging if the RTC driver was loaded but unable to bind to > >> IRQ > >> 8 or access the RTC I/O ports. > > > > I am having the problem, that hwclock hangs, if rtc is compiled into > > the kernel. > > > > Is there a easy fix possible, to avoid systems hanging after a > > dist-upgrade? > > Not allowing RTC to bind to IRQ8 is our current workaround. Are you > seeing the hang in domain0 or in other domains?I only see the hang in domU. Dom0 has physical access to the rtc. I had the rtc driver built into the kernel, not as a module. It was a bit annoying, especially on ubuntu dapper domains, because after a dist-upgrade, /sbin/hwclock usually got restored, and after a reboot, the domain was hanging. johnny> > -- Keir >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 10 May 2006, at 14:49, Hans-Christian Armingeon wrote:> I only see the hang in domU. Dom0 has physical access to the rtc. > I had the rtc driver built into the kernel, not as a module. > > It was a bit annoying, especially on ubuntu dapper domains, because > after a dist-upgrade, /sbin/hwclock usually got restored, and after a > reboot, the domain was hanging.Unfortunately the problem will need root-causing. Clearly the driver is ignoring the fact it cannot bind to the IRQ line. Maybe there''s some other way we can clobber the driver. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Mer, 2006-05-10 at 15:15 +0100, Keir Fraser wrote:> Unfortunately the problem will need root-causing. Clearly the driver is > ignoring the fact it cannot bind to the IRQ line. Maybe there''s some > other way we can clobber the driver.Current drivers/char/rtc checks the IRQ line is available and does not load otherwise in the normal x86 setup. It is possible to build the driver for "IRQless" operation but currently that is only done for non x86 platforms (eg Sparc). So if it gets loaded on x86, IRQ 8 was free. But /sbin/hwclock depending on the version has some other uglies too. Some versions of it if they find no /dev/rtc will iopl() as root and then proceed to whack the hardware directly. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel