similar to: [PATCH] x86/mm: fix invalid unlinking of nested p2m tables

This patch worked on top of paging interface patches posted today. It create common interface for P2M table, which handles guest physical address to machine physical address translation. Tested on the following platforms: 1. AMD SVM boxes * 64-bit Xen: 32-bit WinXP SP2, 32-bit SUSE10, 32-bit SUSE 10 PAE BigSMP, and 64-bit RHEL4 * 32-bit PAE Xen: 32-bit WinXP SP2, 32-bit SUSE10, 32-bit SUSE 10

We would the full P2M top directory from 0->MAX_DOMAIN_PAGES (inclusive). Which meant that if the kernel was compiled with MAX_DOMAIN_PAGES=512 we would try to use the 512th entry. Fortunately for us the p2m_top_index has a check for this: BUG_ON(pfn >= MAX_P2M_PFN); which we hit and saw this: (XEN) domain_crash_sync called from entry.S (XEN) Domain 0 (vcpu#0) crashed on cpu#0: (XEN)

The VMID field is 8 bits. Rather than allowing only up to 256 VMs per host reboot before things start "acting strange" instead maintain a simple bitmap of used VMIDs and allocate them statically to guests upon creation. This limits us to 256 concurrent VMs which is a reasonable improvement. Eventually we will want a proper scheme to allocate VMIDs on context switch. The existing code

> # HG changeset patch > # User George Dunlap <george.dunlap@eu.citrix.com> > # Date 1340815812 -3600 > # Node ID ea827c449088a1017b6e5a9564eb33df70f8a9c6 > # Parent b4e1fec1c98f6cbad666a972f473854518c25500 > xen,pod: Zero-check recently populated pages (checklast) > > When demand-populating pages due to guest accesses, check recently > populated > pages to see

Hello, how can a gfn become 0x555555 as checked in p2m.c:guest_physmap_add_entry() and p2m_alloc_table()? Looking further in p2m.c, audit_p2m() checks only for a 32bit value. So where is that magic number set, and why is it not a define to simplify grepping for users of that value? Olaf _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com

minios: fix p2m on x86 15675:bf85b467ee89 Fixed p2m on x86_64, but instead broke it on x86. Signed-off-by: Samuel Thibault <samuel.thibaut@ens-lyon.org> diff -r 8df3c145923f extras/mini-os/arch/x86/mm.c --- a/extras/mini-os/arch/x86/mm.c Mon Jan 19 17:40:28 2009 +0000 +++ b/extras/mini-os/arch/x86/mm.c Sat Feb 07 03:44:41 2009 +0100 @@ -550,9 +550,15 @@ void arch_init_p2m(unsigned long

build p2m mapping according to m2p mapping. Signed-off-by: Xin Li <xin.b.li@intel.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel

In p2m.c (line 197 and line 550), the code assumes the number of L3 P2M table entries is 8 (under PAE mode). According to Intel and AMD specs, it is 4. Could someone explain this discrepancy? Is it a bug? -Wei _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel

In commit cd9151e2: xen/balloon: set a mapping for ballooned out pages we have the ballooned out page''s mapping set to a scratch page. That commit also sets the P2M entry of ballooned out page to the scratch page''s MFN. This is necessary for PV guest but not for HVM guest. On the other hand, setting the P2M entry would trigger BUG_ON in __set_phys_to_machine. The correct thing

Hi! Attached patch removes gfn_to_mfn_current(). It is redundant and effectively unused. All remaining users now use proper replacement function. => No functional change. While here, silence stdvga and use page_mode_hap() consequently. Signed-off-by: Christoph Egger <Christoph.Egger@amd.com> -- ---to satisfy European Law for business letters: Advanced Micro Devices GmbH

And can we make p2m memory allocation from domain heap directly? -Xin _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel

Not just dom0. Signed-off-by: Ian Campbell <ian.campbell@citrix.com> --- xen/arch/arm/domain.c | 3 +++ xen/arch/arm/domain_build.c | 3 --- xen/arch/arm/p2m.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/xen/arch/arm/domain.c b/xen/arch/arm/domain.c index 5702399..4b38790 100644 --- a/xen/arch/arm/domain.c +++ b/xen/arch/arm/domain.c @@

Hi Xen developers and enthousiasts, Recently I have been having a lot of problems with HVM hosts on a dual Xeon E5606 box. Problem is that HVM hosts don''t work on it at all and crash the server. Today I got the logging to work over an IPMI console so then I compiled a fresh xen-unstable. Apart from the message I posted in the subject the call trace reads as follows: (XEN) Xen call

-- Advanced Micro Devices GmbH Sitz: Dornach, Gemeinde Aschheim, Landkreis München Registergericht München, HRB Nr. 43632 WEEE-Reg-Nr: DE 12919551 Geschäftsführer: Alberto Bozzo, Andrew Bowd _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel

From: Zhang Xiantao <xiantao.zhang@intel.com> With virtual EPT support, L1 hyerpvisor can use EPT hardware for L2 guest''s memory virtualization. In this way, L2 guest''s performance can be improved sharply. According to our testing, some benchmarks can show > 5x performance gain. Signed-off-by: Zhang Xiantao <xiantao.zhang@intel.com> Zhang Xiantao (11):

This patch series tidies up a few bits ofthe nested p2m code. The main thing it does is reorganize the locking so that most of the changes to nested p2m tables happen only under the p2m lock, and the nestedp2m lock is only needed to reassign p2m tables to new cr3 values. Changes since v1: - a few minor fixes - more sensible flushing policy in p2m_get_nestedp2m() - smoke-tested this time!

Hi, here what i have: [2001]--[Asterisk]---[ISDN-Trunk]---[PBX]--[8004] Eicon Diva 4BRI Card to a PBX. Asterisk is running in version 1.0.0 on RedHat Enterprise Linux 3AS with kernel 2.4.21-4.EL. Dialing from Astersik extension 2001 to PBX extension 8004 via ISDN Trunk gives me the following error, -- Executing Dial("SIP/2001-8a8e", "Modem/ttyI0:998004|20|r") in new stack

Ok, thanks. I removed --inplace and --append-verify and kept --link-dest and --partial. And now the script works exactly as I want: hard-links are not updated, the script is still robust and can copy large files over unstable links etc, etc. == Askar Safin http://vk.com/safinaskar Kazan, Russia

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Are you specifying --partial-dir? Seems like that would be needed or else you will end up would end up with an incomplete copy in the previous target? On 12/26/2014 06:50 PM, Askar Safin wrote: > Ok, thanks. I removed --inplace and --append-verify and kept > --link-dest and --partial. And now the script works exactly as I > want: hard-links

I don't specify --partial-dir. As you can see from the script, rsync at first copies to "in-progress", and then renames this to (for example) 2014-12-01-000000. So, if rsync interrupts, then at the next run the script will end "in-progress" (all partial files will be done) and then will rename this dir to 2014-12-01-000000. So, there never will be partial files in finished