From: Sean Christopherson <seanjc@google.com>
To: Sean Christopherson <seanjc@google.com>,
Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
Michal Luczaj <mhal@rbox.co>,
David Woodhouse <dwmw2@infradead.org>
Subject: [PATCH v2 11/16] KVM: Drop KVM's API to allow temprorarily unmapping gfn=>pfn cache
Date: Thu, 13 Oct 2022 21:12:29 +0000 [thread overview]
Message-ID: <20221013211234.1318131-12-seanjc@google.com> (raw)
In-Reply-To: <20221013211234.1318131-1-seanjc@google.com>
Drop kvm_gpc_unmap() as it has no users and unclear requirements. The
API was added as part of the original gfn_to_pfn_cache support, but its
sole usage[*] was never merged. Fold the guts of kvm_gpc_unmap() into
the deactivate path and drop the API. Omit acquiring refresh_lock as
as concurrent calls to kvm_gpc_deactivate() are not allowed (this is
not enforced, e.g. via lockdep. due to it being called during vCPU
destruction).
If/when temporary unmapping makes a comeback, the desirable behavior is
likely to restrict temporary unmapping to vCPU-exclusive mappings and
require the vcpu->mutex be held to serialize unmap. Use of the
refresh_lock to protect unmapping was somewhat specuatively added by
commit 93984f19e7bc ("KVM: Fully serialize gfn=>pfn cache refresh via
mutex") to guard against concurrent unmaps, but the primary use case of
the temporary unmap, nested virtualization[*], doesn't actually need or
want concurrent unmaps.
[*] https://lore.kernel.org/all/20211210163625.2886-7-dwmw2@infradead.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
---
include/linux/kvm_host.h | 12 -----------
virt/kvm/pfncache.c | 44 +++++++++++++++-------------------------
2 files changed, 16 insertions(+), 40 deletions(-)
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index b63d2abbef56..22cf43389954 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1315,18 +1315,6 @@ bool kvm_gpc_check(struct gfn_to_pfn_cache *gpc, gpa_t gpa);
*/
int kvm_gpc_refresh(struct gfn_to_pfn_cache *gpc, gpa_t gpa);
-/**
- * kvm_gpc_unmap - temporarily unmap a gfn_to_pfn_cache.
- *
- * @kvm: pointer to kvm instance.
- * @gpc: struct gfn_to_pfn_cache object.
- *
- * This unmaps the referenced page. The cache is left in the invalid state
- * but at least the mapping from GPA to userspace HVA will remain cached
- * and can be reused on a subsequent refresh.
- */
-void kvm_gpc_unmap(struct kvm *kvm, struct gfn_to_pfn_cache *gpc);
-
/**
* kvm_gpc_deactivate - deactivate and unlink a gfn_to_pfn_cache.
*
diff --git a/virt/kvm/pfncache.c b/virt/kvm/pfncache.c
index 432b150bd9f1..62b47feed36c 100644
--- a/virt/kvm/pfncache.c
+++ b/virt/kvm/pfncache.c
@@ -328,33 +328,6 @@ int kvm_gpc_refresh(struct gfn_to_pfn_cache *gpc, gpa_t gpa)
}
EXPORT_SYMBOL_GPL(kvm_gpc_refresh);
-void kvm_gpc_unmap(struct kvm *kvm, struct gfn_to_pfn_cache *gpc)
-{
- void *old_khva;
- kvm_pfn_t old_pfn;
-
- mutex_lock(&gpc->refresh_lock);
- write_lock_irq(&gpc->lock);
-
- gpc->valid = false;
-
- old_khva = gpc->khva - offset_in_page(gpc->khva);
- old_pfn = gpc->pfn;
-
- /*
- * We can leave the GPA → uHVA map cache intact but the PFN
- * lookup will need to be redone even for the same page.
- */
- gpc->khva = NULL;
- gpc->pfn = KVM_PFN_ERR_FAULT;
-
- write_unlock_irq(&gpc->lock);
- mutex_unlock(&gpc->refresh_lock);
-
- gpc_unmap_khva(old_pfn, old_khva);
-}
-EXPORT_SYMBOL_GPL(kvm_gpc_unmap);
-
void kvm_gpc_init(struct gfn_to_pfn_cache *gpc, struct kvm *kvm,
struct kvm_vcpu *vcpu, enum pfn_cache_usage usage,
unsigned long len)
@@ -402,6 +375,8 @@ EXPORT_SYMBOL_GPL(kvm_gpc_activate);
void kvm_gpc_deactivate(struct gfn_to_pfn_cache *gpc)
{
struct kvm *kvm = gpc->kvm;
+ kvm_pfn_t old_pfn;
+ void *old_khva;
if (gpc->active) {
/*
@@ -411,13 +386,26 @@ void kvm_gpc_deactivate(struct gfn_to_pfn_cache *gpc)
*/
write_lock_irq(&gpc->lock);
gpc->active = false;
+ gpc->valid = false;
+
+ /*
+ * Leave the GPA => uHVA cache intact, it's protected by the
+ * memslot generation. The PFN lookup needs to be redone every
+ * time as mmu_notifier protection is lost when the cache is
+ * removed from the VM's gpc_list.
+ */
+ old_khva = gpc->khva - offset_in_page(gpc->khva);
+ gpc->khva = NULL;
+
+ old_pfn = gpc->pfn;
+ gpc->pfn = KVM_PFN_ERR_FAULT;
write_unlock_irq(&gpc->lock);
spin_lock(&kvm->gpc_lock);
list_del(&gpc->list);
spin_unlock(&kvm->gpc_lock);
- kvm_gpc_unmap(kvm, gpc);
+ gpc_unmap_khva(old_pfn, old_khva);
}
}
EXPORT_SYMBOL_GPL(kvm_gpc_deactivate);
--
2.38.0.413.g74048e4d9e-goog
next prev parent reply other threads:[~2022-10-13 21:14 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-13 21:12 [PATCH v2 00/16] KVM: x86: gfn_to_pfn_cache fixes and cleanups Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 01/16] KVM: Initialize gfn_to_pfn_cache locks in dedicated helper Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 02/16] KVM: Reject attempts to consume or refresh inactive gfn_to_pfn_cache Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 03/16] KVM: x86: Always use non-compat vcpu_runstate_info size for gfn=>pfn cache Sean Christopherson
2022-10-27 11:11 ` Paolo Bonzini
2022-10-27 14:44 ` Sean Christopherson
2022-10-27 15:03 ` Paolo Bonzini
2022-10-27 15:10 ` Sean Christopherson
2022-10-27 16:56 ` Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 04/16] KVM: Shorten gfn_to_pfn_cache function names Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 05/16] KVM: x86: Remove unused argument in gpc_unmap_khva() Sean Christopherson
2022-12-02 9:28 ` Like Xu
2022-12-02 10:57 ` Michal Luczaj
2022-12-02 14:21 ` David Woodhouse
2022-12-02 17:01 ` Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 06/16] KVM: Store immutable gfn_to_pfn_cache properties Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 07/16] KVM: Store gfn_to_pfn_cache length as an immutable property Sean Christopherson
2022-11-21 14:26 ` David Woodhouse
2022-11-21 19:11 ` Sean Christopherson
2022-11-21 20:02 ` David Woodhouse
2022-11-22 18:59 ` Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 08/16] KVM: Use gfn_to_pfn_cache's immutable "kvm" in kvm_gpc_check() Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 09/16] KVM: Clean up hva_to_pfn_retry() Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 10/16] KVM: Use gfn_to_pfn_cache's immutable "kvm" in kvm_gpc_refresh() Sean Christopherson
2022-10-13 21:12 ` Sean Christopherson [this message]
2022-10-13 21:12 ` [PATCH v2 12/16] KVM: Do not partially reinitialize gfn=>pfn cache during activation Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 13/16] KVM: Drop @gpa from exported gfn=>pfn cache check() and refresh() helpers Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 14/16] KVM: Skip unnecessary "unmap" if gpc is already valid during refresh Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 15/16] KVM: selftests: Add tests in xen_shinfo_test to detect lock races Sean Christopherson
2022-10-13 21:12 ` [PATCH v2 16/16] KVM: selftests: Mark "guest_saw_irq" as volatile in xen_shinfo_test Sean Christopherson
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=20221013211234.1318131-12-seanjc@google.com \
--to=seanjc@google.com \
--cc=dwmw2@infradead.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mhal@rbox.co \
--cc=pbonzini@redhat.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