RE: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang
Hi Alex, Seems this issue was caused by Cset 10688. In vcpu_itr_d, the current logic purges vhpt with cpu_flush_vhpt_range but it is very heavy to xen0. When creating VTi domain @ early stage, IO operation is very excessive, so qemu was scheduled out and in very frequently and this logic was executed every time. In addition, cpu_flush_vhpt_range using identity map to purge vhpt may cause more tlb miss due to no TR map. If remove vcpu_flush_tlb_vhpt_range logic although it definitely needed, seems VTi becomes healthy. Maybe potential bugs exist there.:) Thanks Best Regards -Xiantao OTC,Intel Corporation -Original Message- From: Alex Williamson [mailto:[EMAIL PROTECTED] Sent: 2006年7月11日 11:28 To: Zhang, Jingke Cc: xen-ia64-devel@lists.xensource.com; Zhang, Xiantao; [EMAIL PROTECTED] Subject: RE: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang On Tue, 2006-07-11 at 10:13 +0800, Zhang, Jingke wrote: Just add comments to this problem: With Cset 10684, VTI domain creating does well. Cset 10685 emulate PAL_HALT_LIGHT on domU seems to play a significant part of this. On cset 10690 it takes ~30 seconds to get to an EFI shell prompt for a VTI domain. If I 'hg export 10685 | patch -p1 -R' (ie. back out cset 10685), VTI domains start in ~5 seconds. 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] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang
On Tue, 2006-07-11 at 19:42 +0800, Zhang, Xiantao wrote: Hi Alex, Seems this issue was caused by Cset 10688. In vcpu_itr_d, the current logic purges vhpt with cpu_flush_vhpt_range but it is very heavy to xen0. When creating VTi domain @ early stage, IO operation is very excessive, so qemu was scheduled out and in very frequently and this logic was executed every time. In addition, cpu_flush_vhpt_range using identity map to purge vhpt may cause more tlb miss due to no TR map. If remove vcpu_flush_tlb_vhpt_range logic although it definitely needed, seems VTi becomes healthy. Maybe potential bugs exist there.:) Thanks for investigating Xiantao. Isaku, any thoughts on how to regain VTI performance? 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: PATCH: tools/libxc/ia64 directory
On Mon, 2006-07-10 at 14:44 +0200, Tristan Gingold wrote: Hi in order to help integration of the saverestore patch, I submit this patch. It creates the tools/libxc/ia64 directory and moves xc_ia64_stubs.c into it. I have also added a Makefile in it. Applied. -- 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: PATCH: save restore
On Tue, 2006-07-11 at 11:21 +0200, Tristan Gingold wrote: Hi, I have updated my patch taking into account your comments (#if/#endif). This patch applies on top of my previous patch which creates tools/libxc/ia64. Also saverestore of SMP is now working. Applied. Thanks for your hard work in providing yet another huge piece of functionality to Xen/ia64! 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] NOTICE: console changes
Please note, as of xen-ia64-unstable cset 10694 xencons is now turned on by default. If you are currently using xencons= and console= in your elilo, you should now be able to remove them. Xen/ia64 should behave identical to Xen/x86. If there are any problems, please report them to the list. 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] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang
Hi Alex, Our QA team reported that xen0 would hang after creating VTi domain @CSet 10688-10694. Should we reverse the logic of vcpu_itr_d first ? Maybe we can find out the real bottleneck later.:) Thanks Best Regards -Xiantao OTC,Intel Corporation -Original Message- From: Atsushi SAKAI [mailto:[EMAIL PROTECTED] Sent: 2006年7月11日 21:46 To: Alex Williamson; Zhang, Xiantao Cc: Isaku Yamahata; xen-ia64-devel@lists.xensource.com Subject: Re: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang Hi, Alex Sorry for late. I found your problem(boot time difference w/ PAL_HALT_LIGHT emulation patch) occurred in SMP(credit). But, it does not occurred in UP, SMP(bvt) and SMP(credit w/ affinity). I think the emulation of pal_halt_light for domU does not good work for DomVTI boot up under credit scheduling w/o affinity. And consider the Xiantao survey, qemu make heavy I/O operations at the boot up. Consider the above two conditions, I think credit scheduler algorithm does not consider the block state.(caused by pal_halt_light emulation) So I want to switch off the vcpu migration at heavy load I planned as follows. 1)In the short term, I want to avoid this problem by HALT the PAL_HALT_LIGHT emulation while DomVTI boot up. or Lock VCPUs migrations while DomVTI boot up. (when Credit scheduler runs) 2)In the long term, I will make a patch to avoid this problem. (Consider the heavy io w/ vcpu migration) N.B. I checked under CS:10559.(original patch made) Thanks, Atsushi SAKAI On Tue, 2006-07-11 at 19:42 +0800, Zhang, Xiantao wrote: Hi Alex, Seems this issue was caused by Cset 10688. In vcpu_itr_d, the current logic purges vhpt with cpu_flush_vhpt_range but it is very heavy to xen0. When creating VTi domain @ early stage, IO operation is very excessive, so qemu was scheduled out and in very frequently and this logic was executed every time. In addition, cpu_flush_vhpt_range using identity map to purge vhpt may cause more tlb miss due to no TR map. If remove vcpu_flush_tlb_vhpt_range logic although it definitely needed, seems VTi becomes healthy. Maybe potential bugs exist there.:) Thanks for investigating Xiantao. Isaku, any thoughts on how to regain VTI performance? 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 mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
RE: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImakexen0 hang
Sakai-san, Another short term approach, When emulating PAL_HALT_LIGHT If(dom0) Ia64_pal_halt_hight(); Else //domU do_block; Thanks, anthony -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Atsushi SAKAI Sent: 2006?7?11? 21:46 To: Alex Williamson; Zhang, Xiantao Cc: Isaku Yamahata; xen-ia64-devel@lists.xensource.com Subject: Re: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImakexen0 hang Hi, Alex Sorry for late. I found your problem(boot time difference w/ PAL_HALT_LIGHT emulation patch) occurred in SMP(credit). But, it does not occurred in UP, SMP(bvt) and SMP(credit w/ affinity). I think the emulation of pal_halt_light for domU does not good work for DomVTI boot up under credit scheduling w/o affinity. And consider the Xiantao survey, qemu make heavy I/O operations at the boot up. Consider the above two conditions, I think credit scheduler algorithm does not consider the block state.(caused by pal_halt_light emulation) So I want to switch off the vcpu migration at heavy load I planned as follows. 1)In the short term, I want to avoid this problem by HALT the PAL_HALT_LIGHT emulation while DomVTI boot up. or Lock VCPUs migrations while DomVTI boot up. (when Credit scheduler runs) 2)In the long term, I will make a patch to avoid this problem. (Consider the heavy io w/ vcpu migration) N.B. I checked under CS:10559.(original patch made) Thanks, Atsushi SAKAI On Tue, 2006-07-11 at 19:42 +0800, Zhang, Xiantao wrote: Hi Alex, Seems this issue was caused by Cset 10688. In vcpu_itr_d, the current logic purges vhpt with cpu_flush_vhpt_range but it is very heavy to xen0. When creating VTi domain @ early stage, IO operation is very excessive, so qemu was scheduled out and in very frequently and this logic was executed every time. In addition, cpu_flush_vhpt_range using identity map to purge vhpt may cause more tlb miss due to no TR map. If remove vcpu_flush_tlb_vhpt_range logic although it definitely needed, seems VTi becomes healthy. Maybe potential bugs exist there.:) Thanks for investigating Xiantao. Isaku, any thoughts on how to regain VTI performance? 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 mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang
On Wed, Jul 12, 2006 at 11:12:00AM +0800, Xu, Anthony wrote: Hi Alex, When VTI-domain does IO operation, an event is sent to dom0, then Qemu is woken up to handle this request. Process switch on dom0 may call itr_d, if task_struct of qemu is not covered by TR. But the emulation of itr_d is very slow, due to purge vhpt. This is the root cause of slowness of vti-domain boot on your side, and vti-domain can't boot on my side. I think purging vhpt in emulating itr_d also slow down dom0 and domU, So the question is, is purging vhpt in emulating itr_d necessary? For correctness, yes. (SDM vol2 4.1.1.4 and table 4-1.) For Linux working, no (at least linux 2.6.16.13). I don't know other OSes other than Linux. Thanks. -- yamahata ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
RE: [Xen-ia64-devel] [IPF-ia64] with Cset 10690, creating a VTImake xen0 hang
From: Isaku Yamahata For correctness, yes. (SDM vol2 4.1.1.4 and table 4-1.) For Linux working, no (at least linux 2.6.16.13). I don't know other OSes other than Linux. If PAGE_SIZE is 16M, thing may becomes simple. Thanks. -- yamahata ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
[Xen-ia64-devel][PATCH] enable acceleration of external interrupt
This patch is to enable acceleration of external interrupt which is described in VTI spec. Signed-off-by: Anthony Xu [EMAIL PROTECTED] Thanks, Anthony enable_vhpi.patch Description: enable_vhpi.patch ___ Xen-ia64-devel mailing list Xen-ia64-devel@lists.xensource.com http://lists.xensource.com/xen-ia64-devel
