[Xen-ia64-devel] [PATCH] fix spinlock initializer
fix spinlock initializer. It's very surprising this kind of bug remained. I'm wondering why noone else encounter this so far. -- yamahata # HG changeset patch # User [EMAIL PROTECTED] # Date 1194430183 -32400 # Node ID ae8b895f0c553d1b2a6aa81fd2d249a2b427704c # Parent 1abbc4b58494d79905ece072cf1ca66d1e616817 fix spinlock initializer. PATCHNAME: fix_spinlock_initializer Signed-off-by: Isaku Yamahata [EMAIL PROTECTED] diff -r 1abbc4b58494 -r ae8b895f0c55 xen/include/asm-ia64/linux-xen/asm/spinlock.h --- a/xen/include/asm-ia64/linux-xen/asm/spinlock.h Wed Nov 07 18:16:08 2007 +0900 +++ b/xen/include/asm-ia64/linux-xen/asm/spinlock.h Wed Nov 07 19:09:43 2007 +0900 @@ -33,8 +33,20 @@ typedef struct { #endif } spinlock_t; +#ifdef XEN +#ifdef DEBUG_SPINLOCK +#define SPIN_LOCK_UNLOCKED /*(spinlock_t)*/ { 0, NULL, -1, 0 } +#else +#define SPIN_LOCK_UNLOCKED /*(spinlock_t)*/ { 0, -1, 0 } +#endif +static inline void spin_lock_init(spinlock_t *lock) +{ + *lock = ((spinlock_t)SPIN_LOCK_UNLOCKED); +} +#else #define SPIN_LOCK_UNLOCKED /*(spinlock_t)*/ { 0 } #define spin_lock_init(x) ((x)-lock = 0) +#endif #ifdef ASM_SUPPORTED /* ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] [PATCH] clean up vcpu_init_regs()
clean up of vcpu_init_regs(). use symbolical macro value instead of magic number. -- yamahata # HG changeset patch # User [EMAIL PROTECTED] # Date 1194431346 -32400 # Node ID 07e9f799e287f6e39351c5b9e632b0725aec8533 # Parent 2f8c2eb73f50cc68bde19d07fca12c44496f3c71 clean up of vcpu_init_regs(). use symbolical macro value instead of magic number. PATCHNAME: vcpu_init_regs_clean_up Signed-off-by: Isaku Yamahata [EMAIL PROTECTED] diff -r 2f8c2eb73f50 -r 07e9f799e287 xen/arch/ia64/xen/vcpu.c --- a/xen/arch/ia64/xen/vcpu.c Wed Nov 07 19:48:46 2007 +0900 +++ b/xen/arch/ia64/xen/vcpu.c Wed Nov 07 19:29:06 2007 +0900 @@ -146,8 +146,9 @@ void vcpu_init_regs(struct vcpu *v) regs = vcpu_regs(v); if (VMX_DOMAIN(v)) { /* dt/rt/it:1;i/ic:1, si:1, vm/bn:1, ac:1 */ - /* Need to be expanded as macro */ - regs-cr_ipsr = 0x501008826008; + regs-cr_ipsr = IA64_PSR_DT | IA64_PSR_RT | IA64_PSR_IT | + IA64_PSR_I | IA64_PSR_IC | + IA64_PSR_SI | IA64_PSR_AC | IA64_PSR_BN | IA64_PSR_VM; /* lazy fp */ FP_PSR(v) = IA64_PSR_DFH; regs-cr_ipsr |= IA64_PSR_DFH; ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [PATCH 0/7] vti save-restore
Hi Alex. Thank you for the register dump. Otherwise it had been very hard to track it down without the info. The culprit is the patch, 1/7 fix ia64_leave_hypervisor. Although I haven't figured out why the patch causes hang yet, just omitting it is okay for now because it isn't directly related to HVM domain save/restore, I updated and attached the patches. I can boot linux with GFW of - the precompiled binary added by c/s 35:ad52e9f19cbc - c/s 32 with unaligned patch by Kouya. (Withouth Kouya's patch, it results in domain panic.) - efi-vfirmware-cset28.bin you offered thanks. On Tue, Oct 23, 2007 at 09:38:25AM -0600, Alex Williamson wrote: On Tue, 2007-10-23 at 09:29 -0600, Alex Williamson wrote: On Tue, 2007-10-23 at 14:29 +0900, Isaku Yamahata wrote: I haven't found anything wrong with your bundle. I tested xen-unstable.hg with the bundle. With the binary GFW in efi-vfirmware.hg which is somewhat old, I can boot Linux in VTi domain whose memory is more than or equal to 256MB. However with 128MB memory, it reaches to start_kernel(), but it loops at somewhere. (I haven't tracked it down yet.) Your case seems that the GFW is unable to load kernel image because of the IdeController driver related issue, right? So far I've avoided building the GFW, I have to do this time. I'll continue to look into this issue. Right, I never make it to the point of booting the guest OS. I've placed a efi-vfirmware binary here: http://free.linux.hp.com/~awilliam/efi-vfirmware-cset28.bin.bz2 I confirmed that the older Intel GFW does work with this patch series, but the open source GFW hangs. Here's the last output from the console: InstallProtocolInterface: 18A031AB-B443-4D1A-A5C0-0C09261E9F71 13ECB4830 InstallProtocolInterface: 107A772C-D5E1-11D4-9A46-0090273FC14D 13ECB47D0 InstallProtocolInterface: 5B1B31A1-9562-11D2-8E3F-00A0C969723B 13ED07C40 Relocating driver at 0x013EB3A000 /home/awilliam/vefi-build/20071023/edk2-xen/Build/Xen/DEBUG_UNIXGCC/IPF/EdkModulePkg/Bus/Pci/IdeBus/Dxe/idebus/DEBUG/./IdeBus.dll Loading driver at 0x013EB3A000 EntryPoint=0x013EB5EEA0 /home/awilliam/vefi-build/20071023/edk2-xen/Build/Xen/DEBUG_UNIXGCC/IPF/EdkModulePkg/Bus/Pci/IdeBus/Dxe/idebus/DEBUG/.efi InstallProtocolInterface: 18A031AB-B443-4D1A-A5C0-0C09261E9F71 13EB5F2D8 InstallProtocolInterface: 107A772C-D5E1-11D4-9A46-0090273FC14D 13EB5F230 InstallProtocolInterface: 107A772B-D5E1-11D4-9A46-0090273FC14D 13EB5F288 InstallProtocolInterface: 0784924F-E296-11D4-9A49-0090273FC14D 13EB5F2C8 InstallProtocolInterface: 5B1B31A1-9562-11D2-8E3F-00A0C969723B 13ED07A40 Relocating driver at 0x013EC3D000 /home/awilliam/vefi-build/20071023/edk2-xen/Build/Xen/DEBUG_UNIXGCC/IPF/EdkQemuPkg/Bus/Pci/IdeController/Dxe/IdeController/DEBUG/./IdeController.dll Loading driver at 0x013EC3D000 EntryPoint=0x013EC47120 /home/awilliam/vefi-build/20071023/edk2-xen/Build/Xen/DEBUG_UNIXGCC/IPF/EdkQemuPkg/Bus/Pci/IdeController/Dxe/IdeController/DEBUG/.efi InstallProtocolInterface: 18A031AB-B443-4D1A-A5C0-0C09261E9F71 13EC473D0 InstallProtocolInterface: 107A772C-D5E1-11D4-9A46-0090273FC14D 13EC47410 hang The whole machine seems hung at this point. Thanks, Here's the most interesting bit of the INIT dump at this point: (XEN) [2007-10-23 15:22:56] NaT bits (XEN) [2007-10-23 15:22:56] pr 0002aac3 (XEN) [2007-10-23 15:22:56] b0 f409c970 vmx_rse_clear_invalid+0x40/0x130 (XEN) [2007-10-23 15:22:56] ar.rsc (XEN) [2007-10-23 15:22:56] cr.iip f40a0080 vmx_panic+0x0/0x10 (XEN) [2007-10-23 15:22:56] cr.ipsr 001008220030 (XEN) [2007-10-23 15:22:56] cr.ifs 8000 (XEN) [2007-10-23 15:22:56] xip 00013f121b42 ??? (XEN) [2007-10-23 15:22:56] xpsr5410088a6038 (XEN) [2007-10-23 15:22:56] xfs 8000 (XEN) [2007-10-23 15:22:56] b1 fffef540 ??? (XEN) [2007-10-23 15:22:56] (XEN) [2007-10-23 15:22:56] static registers r0-r15: (XEN) [2007-10-23 15:22:56] r0- 3 00013f138800 0011 (XEN) [2007-10-23 15:22:56] r4- 7 00100661 (XEN) [2007-10-23 15:22:56] r8-11 000e 0001 (XEN) [2007-10-23 15:22:56] r12-15 00013ff74e50 f7bb8f71 (XEN) [2007-10-23 15:22:56] (XEN) [2007-10-23 15:22:56] bank 0: (XEN) [2007-10-23 15:22:56] r16-19 f7bb8000 1000 00800030 0018 (XEN) [2007-10-23 15:22:56] r20-23 00013f138800 f7bb8000 00013f121ec0 00029a01 (XEN) [2007-10-23 15:22:56] r24-27 0710 (XEN) [2007-10-23
[Xen-ia64-devel] [Patch] remove panic_domain in guest_write_eoi() for not in-service
Hi, http://lists.xensource.com/archives/html/xen-ia64-devel/2007-10/msg00047.html I'm sorry for my delay response. I remade the patch. This patch is for kexec on HVM domain. Signed-off-by: Akio Takebe [EMAIL PROTECTED] Best Regards, Akio Takebe fix_ia64_eoi_of_null_vector.patch Description: Binary data ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
Keir Fraser wrote: [Wed Nov 07 2007, 11:31:35AM EST] On 7/11/07 16:02, Aron Griffis [EMAIL PROTECTED] wrote: You need to reinstall the tools. I have, multiple times, including make uninstall first. What are you anticipating that a tools re-install will fix? The sysctl is almost certainly failing because of a interface version mismatch. Could you be building the tools and Xen against different versions of the public sysctl.h header file? Yeah, I think that's it. I'm building on one system, then tar/untar to the test system to do the final make install-tools Maybe it's rebuilding something it doesn't need to? I'd probably rather fix that than worry about keeping the userlands in perfect sync, unless this is going to crop up at runtime regardless. Thanks, Aron ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
On 7/11/07 16:02, Aron Griffis [EMAIL PROTECTED] wrote: You need to reinstall the tools. I have, multiple times, including make uninstall first. What are you anticipating that a tools re-install will fix? The sysctl is almost certainly failing because of a interface version mismatch. Could you be building the tools and Xen against different versions of the public sysctl.h header file? Try instrumenting xen/common/sysctl.c in the hypervisor and see if you are indeed failing on the obvious version check at the top of the do_sysctl() function. -- Keir ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [PATCH] Simplify thash_purge_and_insert()
On Wed, 2007-11-07 at 13:24 +0900, Kouya Shimura wrote: Hi, This patch simplifies thash_purge_and_insert() for readability. PV domain never use this function. Applied. Thanks, Alex -- Alex Williamson HP Open Source Linux Org. ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [PATCH] fix spinlock initializer
On Wed, 2007-11-07 at 19:12 +0900, Isaku Yamahata wrote: fix spinlock initializer. It's very surprising this kind of bug remained. I'm wondering why noone else encounter this so far. Yes, odd we haven't seen any problems from it. Applied. Thanks, Alex -- Alex Williamson HP Open Source Linux Org. ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [PATCH 0/7] vti save-restore
On Wed, 2007-11-07 at 20:05 +0900, Isaku Yamahata wrote: Hi Alex. Thank you for the register dump. Otherwise it had been very hard to track it down without the info. The culprit is the patch, 1/7 fix ia64_leave_hypervisor. Although I haven't figured out why the patch causes hang yet, just omitting it is okay for now because it isn't directly related to HVM domain save/restore, I updated and attached the patches. I can boot linux with GFW of - the precompiled binary added by c/s 35:ad52e9f19cbc - c/s 32 with unaligned patch by Kouya. (Withouth Kouya's patch, it results in domain panic.) - efi-vfirmware-cset28.bin you offered Great, works for me know too. Applied. Thanks, Alex -- Alex Williamson HP Open Source Linux Org. ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [PATCH] clean up vcpu_init_regs()
On Wed, 2007-11-07 at 20:07 +0900, Isaku Yamahata wrote: clean up of vcpu_init_regs(). use symbolical macro value instead of magic number. Applied. Thanks, Alex -- Alex Williamson HP Open Source Linux Org. ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [Patch] remove panic_domain in guest_write_eoi() for not in-service
On Wed, 2007-11-07 at 21:12 +0900, Akio Takebe wrote: Hi, http://lists.xensource.com/archives/html/xen-ia64-devel/2007-10/msg00047.html I'm sorry for my delay response. I remade the patch. This patch is for kexec on HVM domain. Applied. Thanks, Alex -- Alex Williamson HP Open Source Linux Org. ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
Keir Fraser wrote: [Wed Nov 07 2007, 12:30:29PM EST] Well, now you know the likely problem it doesn't sound like it will be very hard to track down! Actually, it wasn't the problem at all... My script was trying to copy the kernel from dist/install/boot/vmlinuz but recently it's changed to dist/install/boot/vmlinuz-2.6.18.8-xen. Unfortunately I wrote that part of the script to fail silently, so I wasn't aware that my recent testing has been mismatched. Thanks for the clues, they helped me finally track this down. Aron ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
On 7/11/07 17:26, Aron Griffis [EMAIL PROTECTED] wrote: The sysctl is almost certainly failing because of a interface version mismatch. Could you be building the tools and Xen against different versions of the public sysctl.h header file? Yeah, I think that's it. I'm building on one system, then tar/untar to the test system to do the final make install-tools Maybe it's rebuilding something it doesn't need to? I'd probably rather fix that than worry about keeping the userlands in perfect sync, unless this is going to crop up at runtime regardless. Well, now you know the likely problem it doesn't sound like it will be very hard to track down! -- Keir ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
On 7/11/07 15:55, Aron Griffis [EMAIL PROTECTED] wrote: Anybody know what might be the problem here? I'm sure the build is good and matches the running hypervisor and kernel, which are csets ia64/linux-2.6.18-xen.hg-210 and ia64/xen-unstable.hg-16314 You need to reinstall the tools. -- Keir ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Re: [Xen-devel] xend sysctl operation failed
Keir Fraser wrote: [Wed Nov 07 2007, 10:57:09AM EST] On 7/11/07 15:55, Aron Griffis [EMAIL PROTECTED] wrote: Anybody know what might be the problem here? I'm sure the build is good and matches the running hypervisor and kernel, which are csets ia64/linux-2.6.18-xen.hg-210 and ia64/xen-unstable.hg-16314 You need to reinstall the tools. I have, multiple times, including make uninstall first. What are you anticipating that a tools re-install will fix? Thanks, Aron ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] xend sysctl operation failed
I'm seeing this failure trying to start xend: == xend-debug.log == Xend started at Wed Nov 7 09:30:01 2007. sysctl operation failed -- need to rebuild the user-space tool set? Exception starting xend: (13, 'Permission denied') == xend.log == [2007-11-07 09:30:01 3723] INFO (__init__:1072) Xend Daemon started [2007-11-07 09:30:01 3723] INFO (__init__:1072) Xend changeset: Fri Nov 02 16:38:11 2007 + 16314:650cadd1b283. [2007-11-07 09:30:01 3723] INFO (__init__:1072) Xend version: Unknown. [2007-11-07 09:30:01 3723] ERROR (__init__:1072) Exception starting xend ((13, 'Permission denied')) Traceback (most recent call last): File //usr/lib/python/xen/xend/server/SrvDaemon.py, line 345, in run servers = SrvServer.create() File //usr/lib/python/xen/xend/server/SrvServer.py, line 251, in create root.putChild('xend', SrvRoot()) File //usr/lib/python/xen/xend/server/SrvRoot.py, line 40, in __init__ self.get(name) File //usr/lib/python/xen/web/SrvDir.py, line 82, in get val = val.getobj() File //usr/lib/python/xen/web/SrvDir.py, line 52, in getobj self.obj = klassobj() File //usr/lib/python/xen/xend/server/SrvNode.py, line 30, in __init__ self.xn = XendNode.instance() File //usr/lib/python/xen/xend/XendNode.py, line 703, in instance inst = XendNode() File //usr/lib/python/xen/xend/XendNode.py, line 83, in __init__ self.other_config[xen_pagesize] = self.xeninfo_dict()[xen_pagesize] File //usr/lib/python/xen/xend/XendNode.py, line 667, in xeninfo_dict return dict(self.xeninfo()) File //usr/lib/python/xen/xend/XendNode.py, line 617, in xeninfo info['xen_scheduler'] = self.xenschedinfo() File //usr/lib/python/xen/xend/XendNode.py, line 607, in xenschedinfo sched_id = self.xc.sched_id_get() Error: (13, 'Permission denied') Anybody know what might be the problem here? I'm sure the build is good and matches the running hypervisor and kernel, which are csets ia64/linux-2.6.18-xen.hg-210 and ia64/xen-unstable.hg-16314 Thanks, Aron ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] Xen summit
On Wed, Nov 07, 2007 at 11:24:40PM -0500, Aron Griffis wrote: Hi All, As I think you know, I volunteered to do the xen/ia64 status update since Alex won't be at the summit next week. I'm planning to talk about the ia64 work that's been accomplished since April, plus talk about the roadmap that has been discussed recently on this list. Who's coming? In the past we've held an informal BOF to talk about issues, progress, and just get to know each other better. Would you like to do that again? Yamahata-san and I will be there. A BOF sounds like a great idea. -- Horms H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/ ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel] Xen summit
Hi All, As I think you know, I volunteered to do the xen/ia64 status update since Alex won't be at the summit next week. I'm planning to talk about the ia64 work that's been accomplished since April, plus talk about the roadmap that has been discussed recently on this list. Who's coming? In the past we've held an informal BOF to talk about issues, progress, and just get to know each other better. Would you like to do that again? Thanks, Aron ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] Xen summit
Hi, Aron As I think you know, I volunteered to do the xen/ia64 status update since Alex won't be at the summit next week. I'm planning to talk about the ia64 work that's been accomplished since April, plus talk about the roadmap that has been discussed recently on this list. Who's coming? In the past we've held an informal BOF to talk about issues, progress, and just get to know each other better. Would you like to do that again? Yoshi, Matsumoto, Ezaki, Kama and Akio will be there. I also hope the BOF. Best Regards, Akio Takebe ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel