Long Wang
2010-Jul-23 21:19 UTC
[Xen-devel] how I can set up an environment to debug PV guest and xen hypervisor
Hi, I am working on a project (a research project in my PhD thesis). The project involves instrumentation into the source code of xen hypervisor and Dom0. But now PV guests crash again and again. I need to debug this project. I really need your help on how setting up an environment to debug PV guest and xen hypervisor. 1) Here is the basic information about my testbed: x86 32bit CPU PV guest Xen 3.3.1 + Linux 2.6.18 (comes with xen) I built xen, Dom0 kernel, and DomU kernel by running make world KERNELS=”linux-2.6-xen0 linux-2.6-xenU” 2) Here is more information on my problem:> - Does dom0 start?Yes.> - Is the problem only with domU ?The current failure is with domU. But because in my project I added some code in the hypervisor, I expect that there might be some hypervisor error in the future. I need a way to debug both the domU and the hypervisor, or two separate ways: one for debugging the domU and the other for debugging the xen hypervisor.> - What''s the problem/error?I added code in the hypervisor to do some monitoring work in my project (my code monitors some behavior of the hypervisor). Now after the domU boots and I run a program in the domU, the domU crashes immediately. So I really want to know how you guys set up an environment and debug the domU (PV guest) and the xen hypervisor.> - Please provide full logs showing the error messages.The domU reports the following error log (dom0 does not report any error): BUG: scheduling while atomic: swapper/0x00010000/0 [<c0105b2b>] show_trace+0x1b/0x20 [<c01061c6>] dump_stack+0x26/0x30 [<c02daca6>] schedule+0x576/0x8a0 [<c010312a>] cpu_idle+0x7a/0xc0 [<c0102405>] rest_init+0x35/0x40 [<c0360891>] start_kernel+0x371/0x410 [<00000000>] _stext+0x3fefe000/0x60 BUG: unable to handle kernel paging request at virtual address 00003101 printing eip: c035bf27 00691000 -> *pde = 00000000:0102a001 03fc7000 -> *pme = 00000000:00000000 Oops: 0002 [#1] SMP Modules linked in: CPU: 0 EIP: 0061:[<c035bf27>] Not tainted VLI EFLAGS: 00010096 (2.6.18 #5) EIP is at 0xc035bf27 eax: 00003101 ebx: c01206d8 ecx: c031b2c0 edx: c01013a7 esi: c038d420 edi: 00000400 ebp: c02ee850 esp: c035be9f ds: 007b es: 007b ss: 0069 Process swapper (pid: 0, ti=c035a000 task=c031b2c0 task.ti=c035a000) Stack: 000100c0 00000000 35bf3c00 3551a0c0 00010001 00000100 35bee800 141d6fc0 000100c0 35bf3c00 0bf580c0 000000c1 3551c800 35bf3cc0 0bf580c0 000100c1 35bf3c00 000000c0 35bf0400 1071d2c0 bf221cc0 00002393 09100000 000000c1 Call Trace: [<c0105ab1>] show_stack_log_lvl+0xb1/0xe0 [<c0105ce6>] show_registers+0x1b6/0x240 [<c0105ead>] die+0x13d/0x310 [<c011433d>] do_page_fault+0x7ed/0xf51 [<c01055db>] error_code+0x2b/0x30 Code: 00 00 00 00 00 00 00 00 18 bf 35 c0 bb 0a 12 c0 50 e8 2e c0 24 bf long _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2010-Jul-23 21:46 UTC
Re: [Xen-devel] how I can set up an environment to debug PV guest and xen hypervisor
On 07/23/2010 02:19 PM, Long Wang wrote:>> - What''s the problem/error? >> > I added code in the hypervisor to do some monitoring work in my > project (my code monitors some behavior of the hypervisor). Now after > the domU boots and I run a program in the domU, the domU crashes > immediately. So I really want to know how you guys set up an > environment and debug the domU (PV guest) and the xen hypervisor. >You can use gdbsx to allow you to attach to the domain and poke around in it with gdb.>> - Please provide full logs showing the error messages. >> > The domU reports the following error log (dom0 does not report any error): >If you''re going to be modifying the kernel and getting fairly straightforward-looking crashes such as this, the least you could do is post the changes you''ve made.> BUG: scheduling while atomic: swapper/0x00010000/0 >This probably means you''ve called a blocking call while holding a spinlock or with interrupts disabled.> [<c0105b2b>] show_trace+0x1b/0x20 > [<c01061c6>] dump_stack+0x26/0x30 > [<c02daca6>] schedule+0x576/0x8a0 > [<c010312a>] cpu_idle+0x7a/0xc0 > [<c0102405>] rest_init+0x35/0x40 > [<c0360891>] start_kernel+0x371/0x410 > [<00000000>] _stext+0x3fefe000/0x60 > BUG: unable to handle kernel paging request at virtual address 00003101 > printing eip: > c035bf27 > 00691000 -> *pde = 00000000:0102a001 > 03fc7000 -> *pme = 00000000:00000000 > Oops: 0002 [#1] > SMP > Modules linked in: > CPU: 0 > EIP: 0061:[<c035bf27>] Not tainted VLI > EFLAGS: 00010096 (2.6.18 #5) > EIP is at 0xc035bf27 > eax: 00003101 ebx: c01206d8 ecx: c031b2c0 edx: c01013a7 > esi: c038d420 edi: 00000400 ebp: c02ee850 esp: c035be9f > ds: 007b es: 007b ss: 0069 > Process swapper (pid: 0, ti=c035a000 task=c031b2c0 task.ti=c035a000) > Stack: 000100c0 00000000 35bf3c00 3551a0c0 00010001 00000100 35bee800 141d6fc0 > 000100c0 35bf3c00 0bf580c0 000000c1 3551c800 35bf3cc0 0bf580c0 000100c1 > 35bf3c00 000000c0 35bf0400 1071d2c0 bf221cc0 00002393 09100000 000000c1 > Call Trace: > [<c0105ab1>] show_stack_log_lvl+0xb1/0xe0 > [<c0105ce6>] show_registers+0x1b6/0x240 > [<c0105ead>] die+0x13d/0x310 > [<c011433d>] do_page_fault+0x7ed/0xf51 > [<c01055db>] error_code+0x2b/0x30 > Code: 00 00 00 00 00 00 00 00 18 bf 35 c0 bb 0a 12 c0 50 e8 2e c0 24 bf >J _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel