From: Paolo Bonzini <pbonzini@redhat.com> To: "Gonglei (Arei)" <arei.gonglei@huawei.com>, Igor Mammedov <imammedo@redhat.com>, xuyandong <xuyandong2@huawei.com> Cc: "Huangweidong \(C\)" <weidong.huang@huawei.com>, Zhanghailiang <zhang.zhanghailiang@huawei.com>, "kvm@vger.kernel.org" <kvm@vger.kernel.org>, "wangxin \(U\)" <wangxinxin.wang@huawei.com>, "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>, lidonglin <lidonglin@huawei.com> Subject: Re: An emulation failure occurs, if I hotplug vcpus immediately after the VM start Date: Wed, 6 Jun 2018 15:57:42 +0200 [thread overview] Message-ID: <50481bea-bb5b-dd71-b712-6418c3bb29ac@redhat.com> (raw) In-Reply-To: <33183CC9F5247A488A2544077AF19020DB00F4E4@dggeml511-mbx.china.huawei.com> On 06/06/2018 15:28, Gonglei (Arei) wrote: > gonglei********: mem.slot: 3, mem.guest_phys_addr=0xc0000, > mem.userspace_addr=0x7fc343ec0000, mem.flags=0, memory_size=0x0 > gonglei********: mem.slot: 3, mem.guest_phys_addr=0xc0000, > mem.userspace_addr=0x7fc343ec0000, mem.flags=0, memory_size=0x9000 > > When the memory region is cleared, the KVM will tell the slot to be > invalid (which it is set to KVM_MEMSLOT_INVALID). > > If SeaBIOS accesses this memory and cause page fault, it will find an > invalid value according to gfn (by __gfn_to_pfn_memslot), and finally > it will return an invalid value, and finally it will return a > failure. > > So, My questions are: > > 1) Why don't we hold kvm->slots_lock during page fault processing? Because it's protected by SRCU. We don't need kvm->slots_lock on the read side. > 2) How do we assure that vcpus will not access the corresponding > region when deleting an memory slot? We don't. It's generally a guest bug if they do, but the problem here is that QEMU is splitting a memory region in two parts and that is not atomic. One fix could be to add a KVM_SET_USER_MEMORY_REGIONS ioctl that replaces the entire memory map atomically. Paolo
WARNING: multiple messages have this Message-ID (diff)
From: Paolo Bonzini <pbonzini@redhat.com> To: "Gonglei (Arei)" <arei.gonglei@huawei.com>, Igor Mammedov <imammedo@redhat.com>, xuyandong <xuyandong2@huawei.com> Cc: Zhanghailiang <zhang.zhanghailiang@huawei.com>, "wangxin (U)" <wangxinxin.wang@huawei.com>, lidonglin <lidonglin@huawei.com>, "kvm@vger.kernel.org" <kvm@vger.kernel.org>, "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>, "Huangweidong (C)" <weidong.huang@huawei.com> Subject: Re: [Qemu-devel] An emulation failure occurs, if I hotplug vcpus immediately after the VM start Date: Wed, 6 Jun 2018 15:57:42 +0200 [thread overview] Message-ID: <50481bea-bb5b-dd71-b712-6418c3bb29ac@redhat.com> (raw) In-Reply-To: <33183CC9F5247A488A2544077AF19020DB00F4E4@dggeml511-mbx.china.huawei.com> On 06/06/2018 15:28, Gonglei (Arei) wrote: > gonglei********: mem.slot: 3, mem.guest_phys_addr=0xc0000, > mem.userspace_addr=0x7fc343ec0000, mem.flags=0, memory_size=0x0 > gonglei********: mem.slot: 3, mem.guest_phys_addr=0xc0000, > mem.userspace_addr=0x7fc343ec0000, mem.flags=0, memory_size=0x9000 > > When the memory region is cleared, the KVM will tell the slot to be > invalid (which it is set to KVM_MEMSLOT_INVALID). > > If SeaBIOS accesses this memory and cause page fault, it will find an > invalid value according to gfn (by __gfn_to_pfn_memslot), and finally > it will return an invalid value, and finally it will return a > failure. > > So, My questions are: > > 1) Why don't we hold kvm->slots_lock during page fault processing? Because it's protected by SRCU. We don't need kvm->slots_lock on the read side. > 2) How do we assure that vcpus will not access the corresponding > region when deleting an memory slot? We don't. It's generally a guest bug if they do, but the problem here is that QEMU is splitting a memory region in two parts and that is not atomic. One fix could be to add a KVM_SET_USER_MEMORY_REGIONS ioctl that replaces the entire memory map atomically. Paolo
next prev parent reply other threads:[~2018-06-06 13:57 UTC|newest] Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-06-01 8:17 An emulation failure occurs, if I hotplug vcpus immediately after the VM start xuyandong 2018-06-01 8:17 ` [Qemu-devel] " xuyandong 2018-06-01 10:23 ` Igor Mammedov 2018-06-01 10:23 ` [Qemu-devel] " Igor Mammedov 2018-06-06 13:28 ` Gonglei (Arei) 2018-06-06 13:28 ` [Qemu-devel] " Gonglei (Arei) 2018-06-06 13:57 ` Paolo Bonzini [this message] 2018-06-06 13:57 ` Paolo Bonzini 2018-06-06 14:18 ` xuyandong 2018-06-06 14:18 ` [Qemu-devel] " xuyandong 2018-06-06 14:23 ` Paolo Bonzini 2018-06-06 14:23 ` [Qemu-devel] " Paolo Bonzini 2018-06-07 10:37 ` David Hildenbrand 2018-06-07 10:37 ` [Qemu-devel] " David Hildenbrand 2018-06-07 11:02 ` Paolo Bonzini 2018-06-07 11:02 ` [Qemu-devel] " Paolo Bonzini 2018-06-07 11:36 ` David Hildenbrand 2018-06-07 11:36 ` [Qemu-devel] " David Hildenbrand 2018-06-07 12:36 ` Paolo Bonzini 2018-06-07 12:36 ` [Qemu-devel] " Paolo Bonzini 2018-06-07 12:55 ` David Hildenbrand 2018-06-07 12:55 ` [Qemu-devel] " David Hildenbrand 2018-06-07 16:03 ` 浙大邮箱 2018-06-07 16:03 ` [Qemu-devel] " 浙大邮箱 2018-06-11 10:44 ` David Hildenbrand 2018-06-11 10:44 ` [Qemu-devel] " David Hildenbrand 2018-06-11 12:25 ` Gonglei (Arei) 2018-06-11 12:25 ` [Qemu-devel] " Gonglei (Arei) 2018-06-11 12:36 ` David Hildenbrand 2018-06-11 12:36 ` [Qemu-devel] " David Hildenbrand 2018-06-11 13:25 ` Gonglei (Arei) 2018-06-11 13:25 ` [Qemu-devel] " Gonglei (Arei) 2018-06-07 10:39 ` David Hildenbrand 2018-06-07 10:39 ` [Qemu-devel] " David Hildenbrand 2018-06-07 11:13 ` Gonglei (Arei) 2018-06-07 11:13 ` [Qemu-devel] " Gonglei (Arei) 2018-06-07 11:43 ` David Hildenbrand 2018-06-07 11:43 ` [Qemu-devel] " David Hildenbrand
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=50481bea-bb5b-dd71-b712-6418c3bb29ac@redhat.com \ --to=pbonzini@redhat.com \ --cc=arei.gonglei@huawei.com \ --cc=imammedo@redhat.com \ --cc=kvm@vger.kernel.org \ --cc=lidonglin@huawei.com \ --cc=qemu-devel@nongnu.org \ --cc=wangxinxin.wang@huawei.com \ --cc=weidong.huang@huawei.com \ --cc=xuyandong2@huawei.com \ --cc=zhang.zhanghailiang@huawei.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.