From: Ben Gardon <bgardon@google.com>
To: LKML <linux-kernel@vger.kernel.org>, kvm <kvm@vger.kernel.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>, Peter Xu <peterx@redhat.com>,
Sean Christopherson <seanjc@google.com>,
Peter Shier <pshier@google.com>,
Yulei Zhang <yulei.kernel@gmail.com>,
Wanpeng Li <kernellwp@gmail.com>,
Xiao Guangrong <xiaoguangrong.eric@gmail.com>,
Kai Huang <kai.huang@intel.com>,
Keqian Zhu <zhukeqian1@huawei.com>
Subject: Re: [PATCH v3 5/8] KVM: x86/mmu: Add a field to control memslot rmap allocation
Date: Thu, 6 May 2021 16:44:19 -0700 [thread overview]
Message-ID: <CANgfPd-eJsHRYARTa0tm4EUVQyXvdQxGQfGfj=qLi5vkLTG6pw@mail.gmail.com> (raw)
In-Reply-To: <20210506184241.618958-6-bgardon@google.com>
On Thu, May 6, 2021 at 11:43 AM Ben Gardon <bgardon@google.com> wrote:
>
> Add a field to control whether new memslots should have rmaps allocated
> for them. As of this change, it's not safe to skip allocating rmaps, so
> the field is always set to allocate rmaps. Future changes will make it
> safe to operate without rmaps, using the TDP MMU. Then further changes
> will allow the rmaps to be allocated lazily when needed for nested
> oprtation.
>
> No functional change expected.
>
> Signed-off-by: Ben Gardon <bgardon@google.com>
> ---
> arch/x86/include/asm/kvm_host.h | 8 ++++++++
> arch/x86/kvm/mmu/mmu.c | 2 ++
> arch/x86/kvm/x86.c | 18 +++++++++++++-----
> 3 files changed, 23 insertions(+), 5 deletions(-)
>
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index ad22d4839bcc..00065f9bbc5e 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1122,6 +1122,12 @@ struct kvm_arch {
> */
> spinlock_t tdp_mmu_pages_lock;
> #endif /* CONFIG_X86_64 */
> +
> + /*
> + * If set, rmaps have been allocated for all memslots and should be
> + * allocated for any newly created or modified memslots.
> + */
> + bool memslots_have_rmaps;
> };
>
> struct kvm_vm_stat {
> @@ -1853,4 +1859,6 @@ static inline int kvm_cpu_get_apicid(int mps_cpu)
>
> int kvm_cpu_dirty_log_size(void);
>
> +inline bool kvm_memslots_have_rmaps(struct kvm *kvm);
Woops, this shouldn't be marked inline as it creates build problems
for the next patch with some configs.
> +
> #endif /* _ASM_X86_KVM_HOST_H */
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 930ac8a7e7c9..8761b4925755 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -5469,6 +5469,8 @@ void kvm_mmu_init_vm(struct kvm *kvm)
>
> kvm_mmu_init_tdp_mmu(kvm);
>
> + kvm->arch.memslots_have_rmaps = true;
> +
> node->track_write = kvm_mmu_pte_write;
> node->track_flush_slot = kvm_mmu_invalidate_zap_pages_in_memslot;
> kvm_page_track_register_notifier(kvm, node);
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index fc32a7dbe4c4..d7a40ce342cc 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -10868,7 +10868,13 @@ static int alloc_memslot_rmap(struct kvm_memory_slot *slot,
> return -ENOMEM;
> }
>
> -static int kvm_alloc_memslot_metadata(struct kvm_memory_slot *slot,
> +bool kvm_memslots_have_rmaps(struct kvm *kvm)
> +{
> + return kvm->arch.memslots_have_rmaps;
> +}
> +
> +static int kvm_alloc_memslot_metadata(struct kvm *kvm,
> + struct kvm_memory_slot *slot,
> unsigned long npages)
> {
> int i;
> @@ -10881,9 +10887,11 @@ static int kvm_alloc_memslot_metadata(struct kvm_memory_slot *slot,
> */
> memset(&slot->arch, 0, sizeof(slot->arch));
>
> - r = alloc_memslot_rmap(slot, npages);
> - if (r)
> - return r;
> + if (kvm_memslots_have_rmaps(kvm)) {
> + r = alloc_memslot_rmap(slot, npages);
> + if (r)
> + return r;
> + }
>
> for (i = 1; i < KVM_NR_PAGE_SIZES; ++i) {
> struct kvm_lpage_info *linfo;
> @@ -10954,7 +10962,7 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
> enum kvm_mr_change change)
> {
> if (change == KVM_MR_CREATE || change == KVM_MR_MOVE)
> - return kvm_alloc_memslot_metadata(memslot,
> + return kvm_alloc_memslot_metadata(kvm, memslot,
> mem->memory_size >> PAGE_SHIFT);
> return 0;
> }
> --
> 2.31.1.607.g51e8a6a459-goog
>
next prev parent reply other threads:[~2021-05-06 23:44 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-06 18:42 [PATCH v3 0/8] Lazily allocate memslot rmaps Ben Gardon
2021-05-06 18:42 ` [PATCH v3 1/8] KVM: x86/mmu: Deduplicate rmap freeing Ben Gardon
2021-05-07 7:42 ` David Hildenbrand
2021-05-06 18:42 ` [PATCH v3 2/8] KVM: x86/mmu: Factor out allocating memslot rmap Ben Gardon
2021-05-07 7:46 ` David Hildenbrand
2021-05-10 16:29 ` Ben Gardon
2021-05-06 18:42 ` [PATCH v3 3/8] KVM: mmu: Refactor memslot copy Ben Gardon
2021-05-07 7:48 ` David Hildenbrand
2021-05-06 18:42 ` [PATCH v3 4/8] KVM: mmu: Add slots_arch_lock for memslot arch fields Ben Gardon
2021-05-06 18:42 ` [PATCH v3 5/8] KVM: x86/mmu: Add a field to control memslot rmap allocation Ben Gardon
2021-05-06 23:44 ` Ben Gardon [this message]
2021-05-07 7:50 ` David Hildenbrand
2021-05-07 8:28 ` Paolo Bonzini
2021-05-10 16:14 ` Ben Gardon
2021-05-10 16:33 ` Paolo Bonzini
2021-05-10 16:37 ` Ben Gardon
2021-05-06 18:42 ` [PATCH v3 6/8] KVM: x86/mmu: Skip rmap operations if rmaps not allocated Ben Gardon
2021-05-06 23:07 ` kernel test robot
2021-05-06 18:42 ` [PATCH v3 7/8] KVM: x86/mmu: Protect rmaps independently with SRCU Ben Gardon
2021-05-06 23:58 ` kernel test robot
2021-05-07 0:56 ` kernel test robot
2021-05-07 8:42 ` Paolo Bonzini
2021-05-10 17:45 ` Sean Christopherson
2021-05-10 17:53 ` Paolo Bonzini
2021-05-10 18:28 ` Sean Christopherson
2021-05-11 16:22 ` Ben Gardon
2021-05-11 16:45 ` Paolo Bonzini
2021-05-06 18:42 ` [PATCH v3 8/8] KVM: x86/mmu: Lazily allocate memslot rmaps Ben Gardon
2021-05-07 1:10 ` kernel test robot
2021-05-07 8:28 ` Paolo Bonzini
2021-05-07 7:40 ` [PATCH v3 0/8] " 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='CANgfPd-eJsHRYARTa0tm4EUVQyXvdQxGQfGfj=qLi5vkLTG6pw@mail.gmail.com' \
--to=bgardon@google.com \
--cc=kai.huang@intel.com \
--cc=kernellwp@gmail.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterx@redhat.com \
--cc=pshier@google.com \
--cc=seanjc@google.com \
--cc=xiaoguangrong.eric@gmail.com \
--cc=yulei.kernel@gmail.com \
--cc=zhukeqian1@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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).