[Xen-devel] [PATCH v2 0/3] purge free_shared_domheap_page()

[Xen-devel] [PATCH v2 0/3] purge free_shared_domheap_page()

[Paul Durrant]

Paul Durrant (3):
  x86 / vmx: make apic_access_mfn type-safe
  x86 / hvm: add domain_relinquish_resources() method
  x86 / vmx: use a 'normal' domheap page for APIC_DEFAULT_PHYS_BASE

 xen/arch/x86/hvm/hvm.c             |  7 +++++-
 xen/arch/x86/hvm/mtrr.c            |  2 +-
 xen/arch/x86/hvm/svm/svm.c         |  5 ----
 xen/arch/x86/hvm/vmx/vmx.c         | 38 +++++++++++++++++++++---------
 xen/arch/x86/mm.c                  | 10 --------
 xen/common/domain.c                |  8 +++++++
 xen/include/asm-x86/hvm/hvm.h      |  1 +
 xen/include/asm-x86/hvm/vmx/vmcs.h |  2 +-
 xen/include/asm-x86/mm.h           |  2 --
 9 files changed, 44 insertions(+), 31 deletions(-)

-- 
2.20.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

[Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe

[Paul Durrant]

Use mfn_t rather than unsigned long and change previous tests against 0 to
tests against INVALID_MFN (also introducing initialization to that value).

Signed-off-by: Paul Durrant <pdurrant@amazon.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: Wei Liu <wl@xen.org>
Cc: "Roger Pau Monné" <roger.pau@citrix.com>
Cc: Jun Nakajima <jun.nakajima@intel.com>

v2:
 - Set apic_access_mfn to INVALID_MFN in vmx_free_vlapic_mapping() to make
   the function idempotent
---
 xen/arch/x86/hvm/mtrr.c            |  2 +-
 xen/arch/x86/hvm/vmx/vmx.c         | 15 ++++++++-------
 xen/include/asm-x86/hvm/vmx/vmcs.h |  2 +-
 3 files changed, 10 insertions(+), 9 deletions(-)

diff --git a/xen/arch/x86/hvm/mtrr.c b/xen/arch/x86/hvm/mtrr.c
index 5ad15eafe0..8356e8de3d 100644
--- a/xen/arch/x86/hvm/mtrr.c
+++ b/xen/arch/x86/hvm/mtrr.c
@@ -818,7 +818,7 @@ int epte_get_entry_emt(struct domain *d, unsigned long gfn, mfn_t mfn,
 
     if ( direct_mmio )
     {
-        if ( (mfn_x(mfn) ^ d->arch.hvm.vmx.apic_access_mfn) >> order )
+        if ( (mfn_x(mfn) ^ mfn_x(d->arch.hvm.vmx.apic_access_mfn)) >> order )
             return MTRR_TYPE_UNCACHABLE;
         if ( order )
             return -1;
diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index f83f102638..8706954d73 100644
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -413,6 +413,7 @@ static int vmx_domain_initialise(struct domain *d)
     if ( !has_vlapic(d) )
         return 0;
 
+    d->arch.hvm.vmx.apic_access_mfn = INVALID_MFN;
     if ( (rc = vmx_alloc_vlapic_mapping(d)) != 0 )
         return rc;
 
@@ -3034,7 +3035,7 @@ static int vmx_alloc_vlapic_mapping(struct domain *d)
     mfn = page_to_mfn(pg);
     clear_domain_page(mfn);
     share_xen_page_with_guest(pg, d, SHARE_rw);
-    d->arch.hvm.vmx.apic_access_mfn = mfn_x(mfn);
+    d->arch.hvm.vmx.apic_access_mfn = mfn;
 
     return set_mmio_p2m_entry(d, paddr_to_pfn(APIC_DEFAULT_PHYS_BASE), mfn,
                               PAGE_ORDER_4K,
@@ -3043,24 +3044,24 @@ static int vmx_alloc_vlapic_mapping(struct domain *d)
 
 static void vmx_free_vlapic_mapping(struct domain *d)
 {
-    unsigned long mfn = d->arch.hvm.vmx.apic_access_mfn;
+    mfn_t mfn = d->arch.hvm.vmx.apic_access_mfn;
 
-    if ( mfn != 0 )
-        free_shared_domheap_page(mfn_to_page(_mfn(mfn)));
+    d->arch.hvm.vmx.apic_access_mfn = INVALID_MFN;
+    if ( !mfn_eq(mfn, INVALID_MFN) )
+        free_shared_domheap_page(mfn_to_page(mfn));
 }
 
 static void vmx_install_vlapic_mapping(struct vcpu *v)
 {
     paddr_t virt_page_ma, apic_page_ma;
 
-    if ( v->domain->arch.hvm.vmx.apic_access_mfn == 0 )
+    if ( mfn_eq(v->domain->arch.hvm.vmx.apic_access_mfn, INVALID_MFN) )
         return;
 
     ASSERT(cpu_has_vmx_virtualize_apic_accesses);
 
     virt_page_ma = page_to_maddr(vcpu_vlapic(v)->regs_page);
-    apic_page_ma = v->domain->arch.hvm.vmx.apic_access_mfn;
-    apic_page_ma <<= PAGE_SHIFT;
+    apic_page_ma = mfn_to_maddr(v->domain->arch.hvm.vmx.apic_access_mfn);
 
     vmx_vmcs_enter(v);
     __vmwrite(VIRTUAL_APIC_PAGE_ADDR, virt_page_ma);
diff --git a/xen/include/asm-x86/hvm/vmx/vmcs.h b/xen/include/asm-x86/hvm/vmx/vmcs.h
index a514299144..be4661a929 100644
--- a/xen/include/asm-x86/hvm/vmx/vmcs.h
+++ b/xen/include/asm-x86/hvm/vmx/vmcs.h
@@ -59,7 +59,7 @@ struct ept_data {
 #define _VMX_DOMAIN_PML_ENABLED    0
 #define VMX_DOMAIN_PML_ENABLED     (1ul << _VMX_DOMAIN_PML_ENABLED)
 struct vmx_domain {
-    unsigned long apic_access_mfn;
+    mfn_t apic_access_mfn;
     /* VMX_DOMAIN_* */
     unsigned int status;
 
-- 
2.20.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

[Xen-devel] [PATCH v2 2/3] x86 / hvm: add domain_relinquish_resources() method

[Paul Durrant]

There are two functions in hvm.c to deal with tear-down and a domain:
hvm_domain_relinquish_resources() and hvm_domain_destroy(). However, only
the latter has an associated method in 'hvm_funcs'. This patch adds
a method for the former.

A subsequent patch will define a VMX implementation.

Signed-off-by: Paul Durrant <pdurrant@amazon.com>
---
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: Wei Liu <wl@xen.org>
Cc: "Roger Pau Monné" <roger.pau@citrix.com>
Cc: Jun Nakajima <jun.nakajima@intel.com>
Cc: Kevin Tian <kevin.tian@intel.com>

v2:
 - Make the new method optional and make it an alternative_vcall
---
 xen/arch/x86/hvm/hvm.c        | 3 +++
 xen/include/asm-x86/hvm/hvm.h | 1 +
 2 files changed, 4 insertions(+)

diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
index 4723f5d09c..e51c077269 100644
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -715,6 +715,9 @@ int hvm_domain_initialise(struct domain *d)
 
 void hvm_domain_relinquish_resources(struct domain *d)
 {
+    if ( hvm_funcs.domain_relinquish_resources )
+        alternative_vcall(hvm_funcs.domain_relinquish_resources, d);
+
     if ( hvm_funcs.nhvm_domain_relinquish_resources )
         hvm_funcs.nhvm_domain_relinquish_resources(d);
 
diff --git a/xen/include/asm-x86/hvm/hvm.h b/xen/include/asm-x86/hvm/hvm.h
index 09793c12e9..9eab1d7493 100644
--- a/xen/include/asm-x86/hvm/hvm.h
+++ b/xen/include/asm-x86/hvm/hvm.h
@@ -107,6 +107,7 @@ struct hvm_function_table {
      * Initialise/destroy HVM domain/vcpu resources
      */
     int  (*domain_initialise)(struct domain *d);
+    void (*domain_relinquish_resources)(struct domain *d);
     void (*domain_destroy)(struct domain *d);
     int  (*vcpu_initialise)(struct vcpu *v);
     void (*vcpu_destroy)(struct vcpu *v);
-- 
2.20.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

[Xen-devel] [PATCH v2 3/3] x86 / vmx: use a 'normal' domheap page for APIC_DEFAULT_PHYS_BASE

[Paul Durrant]

vmx_alloc_vlapic_mapping() currently contains some very odd looking code
that allocates a MEMF_no_owner domheap page and then shares with the guest
as if it were a xenheap page. This then requires vmx_free_vlapic_mapping()
to call a special function in the mm code: free_shared_domheap_page().

By using a 'normal' domheap page (i.e. by not passing MEMF_no_owner to
alloc_domheap_page()), the odd looking code in vmx_alloc_vlapic_mapping()
can simply use get_page_and_type() to set up a writable mapping before
insertion in the P2M and vmx_free_vlapic_mapping() can simply release the
page using put_page_alloc_ref() followed by put_page_and_type(). This
then allows free_shared_domheap_page() to be purged.

There is, however, some fall-out from this simplification:

- alloc_domheap_page() will now call assign_pages() and run into the fact
  that 'max_pages' is not set until some time after domain_create(). To
  avoid an allocation failure, domain_create() is modified to set
  max_pages to an initial value, sufficient to cover any domheap
  allocations required to complete domain creation. The value will be
  set to the 'real' max_pages when the tool-stack later performs the
  XEN_DOMCTL_max_mem operation, thus allowing the rest of the domain's
  memory to be allocated.

- Because the domheap page is no longer a pseudo-xenheap page, the
  reference counting will prevent the domain from being destroyed. Thus
  the call to vmx_free_vlapic_mapping() is moved from the
  domain_destroy() method into the domain_relinquish_resources() method.
  Whilst in the area, make the domain_destroy() method an optional
  alternative_vcall() (since it will no longer peform any function in VMX
  and is stubbed in SVM anyway).

Signed-off-by: Paul Durrant <pdurrant@amazon.com>
---
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: Wei Liu <wl@xen.org>
Cc: "Roger Pau Monné" <roger.pau@citrix.com>
Cc: George Dunlap <George.Dunlap@eu.citrix.com>
Cc: Ian Jackson <ian.jackson@eu.citrix.com>
Cc: Julien Grall <julien@xen.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Jun Nakajima <jun.nakajima@intel.com>
Cc: Kevin Tian <kevin.tian@intel.com>

v2:
 - Set an initial value for max_pages rather than avoiding the check in
   assign_pages()
 - Make domain_destroy() optional
---
 xen/arch/x86/hvm/hvm.c     |  4 +++-
 xen/arch/x86/hvm/svm/svm.c |  5 -----
 xen/arch/x86/hvm/vmx/vmx.c | 25 ++++++++++++++++++++-----
 xen/arch/x86/mm.c          | 10 ----------
 xen/common/domain.c        |  8 ++++++++
 xen/include/asm-x86/mm.h   |  2 --
 6 files changed, 31 insertions(+), 23 deletions(-)

diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
index e51c077269..d2610f5f01 100644
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -746,7 +746,9 @@ void hvm_domain_destroy(struct domain *d)
 
     hvm_destroy_cacheattr_region_list(d);
 
-    hvm_funcs.domain_destroy(d);
+    if ( hvm_funcs.domain_destroy )
+        alternative_vcall(hvm_funcs.domain_destroy, d);
+
     rtc_deinit(d);
     stdvga_deinit(d);
     vioapic_deinit(d);
diff --git a/xen/arch/x86/hvm/svm/svm.c b/xen/arch/x86/hvm/svm/svm.c
index b1c376d455..b7f67f9f03 100644
--- a/xen/arch/x86/hvm/svm/svm.c
+++ b/xen/arch/x86/hvm/svm/svm.c
@@ -1155,10 +1155,6 @@ static int svm_domain_initialise(struct domain *d)
     return 0;
 }
 
-static void svm_domain_destroy(struct domain *d)
-{
-}
-
 static int svm_vcpu_initialise(struct vcpu *v)
 {
     int rc;
@@ -2425,7 +2421,6 @@ static struct hvm_function_table __initdata svm_function_table = {
     .cpu_up               = svm_cpu_up,
     .cpu_down             = svm_cpu_down,
     .domain_initialise    = svm_domain_initialise,
-    .domain_destroy       = svm_domain_destroy,
     .vcpu_initialise      = svm_vcpu_initialise,
     .vcpu_destroy         = svm_vcpu_destroy,
     .save_cpu_ctxt        = svm_save_vmcb_ctxt,
diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index 8706954d73..f76fdd4f96 100644
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -420,7 +420,7 @@ static int vmx_domain_initialise(struct domain *d)
     return 0;
 }
 
-static void vmx_domain_destroy(struct domain *d)
+static void vmx_domain_relinquish_resources(struct domain *d)
 {
     if ( !has_vlapic(d) )
         return;
@@ -2241,7 +2241,7 @@ static struct hvm_function_table __initdata vmx_function_table = {
     .cpu_up_prepare       = vmx_cpu_up_prepare,
     .cpu_dead             = vmx_cpu_dead,
     .domain_initialise    = vmx_domain_initialise,
-    .domain_destroy       = vmx_domain_destroy,
+    .domain_relinquish_resources = vmx_domain_relinquish_resources,
     .vcpu_initialise      = vmx_vcpu_initialise,
     .vcpu_destroy         = vmx_vcpu_destroy,
     .save_cpu_ctxt        = vmx_save_vmcs_ctxt,
@@ -3029,12 +3029,22 @@ static int vmx_alloc_vlapic_mapping(struct domain *d)
     if ( !cpu_has_vmx_virtualize_apic_accesses )
         return 0;
 
-    pg = alloc_domheap_page(d, MEMF_no_owner);
+    pg = alloc_domheap_page(d, 0);
     if ( !pg )
         return -ENOMEM;
+
+    if ( !get_page_and_type(pg, d, PGT_writable_page) )
+    {
+        /*
+         * The domain can't possibly know about this page yet, so failure
+         * here is a clear indication of something fishy going on.
+         */
+        domain_crash(d);
+        return -ENODATA;
+    }
+
     mfn = page_to_mfn(pg);
     clear_domain_page(mfn);
-    share_xen_page_with_guest(pg, d, SHARE_rw);
     d->arch.hvm.vmx.apic_access_mfn = mfn;
 
     return set_mmio_p2m_entry(d, paddr_to_pfn(APIC_DEFAULT_PHYS_BASE), mfn,
@@ -3048,7 +3058,12 @@ static void vmx_free_vlapic_mapping(struct domain *d)
 
     d->arch.hvm.vmx.apic_access_mfn = INVALID_MFN;
     if ( !mfn_eq(mfn, INVALID_MFN) )
-        free_shared_domheap_page(mfn_to_page(mfn));
+    {
+        struct page_info *pg = mfn_to_page(mfn);
+
+        put_page_alloc_ref(pg);
+        put_page_and_type(pg);
+    }
 }
 
 static void vmx_install_vlapic_mapping(struct vcpu *v)
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index 654190e9e9..2a6d2e8af9 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -496,16 +496,6 @@ void share_xen_page_with_guest(struct page_info *page, struct domain *d,
     spin_unlock(&d->page_alloc_lock);
 }
 
-void free_shared_domheap_page(struct page_info *page)
-{
-    put_page_alloc_ref(page);
-    if ( !test_and_clear_bit(_PGC_xen_heap, &page->count_info) )
-        ASSERT_UNREACHABLE();
-    page->u.inuse.type_info = 0;
-    page_set_owner(page, NULL);
-    free_domheap_page(page);
-}
-
 void make_cr3(struct vcpu *v, mfn_t mfn)
 {
     struct domain *d = v->domain;
diff --git a/xen/common/domain.c b/xen/common/domain.c
index ee3f9ffd3e..30c777acb8 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -339,6 +339,8 @@ static int sanitise_domain_config(struct xen_domctl_createdomain *config)
     return arch_sanitise_domain_config(config);
 }
 
+#define DOMAIN_INIT_PAGES 1
+
 struct domain *domain_create(domid_t domid,
                              struct xen_domctl_createdomain *config,
                              bool is_priv)
@@ -441,6 +443,12 @@ struct domain *domain_create(domid_t domid,
         radix_tree_init(&d->pirq_tree);
     }
 
+    /*
+     * Allow a limited number of special pages to be allocated for the
+     * domain
+     */
+    d->max_pages = DOMAIN_INIT_PAGES;
+
     if ( (err = arch_domain_create(d, config)) != 0 )
         goto fail;
     init_status |= INIT_arch;
diff --git a/xen/include/asm-x86/mm.h b/xen/include/asm-x86/mm.h
index 2ca8882ad0..e429f38228 100644
--- a/xen/include/asm-x86/mm.h
+++ b/xen/include/asm-x86/mm.h
@@ -317,8 +317,6 @@ struct page_info
 
 #define maddr_get_owner(ma)   (page_get_owner(maddr_to_page((ma))))
 
-extern void free_shared_domheap_page(struct page_info *page);
-
 #define frame_table ((struct page_info *)FRAMETABLE_VIRT_START)
 extern unsigned long max_page;
 extern unsigned long total_pages;
-- 
2.20.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe

[Jan Beulich]

On 23.01.2020 13:21, Paul Durrant wrote:
> Use mfn_t rather than unsigned long and change previous tests against 0 to
> tests against INVALID_MFN (also introducing initialization to that value).
> 
> Signed-off-by: Paul Durrant <pdurrant@amazon.com>
> Acked-by: Kevin Tian <kevin.tian@intel.com>
> Reviewed-by: Jan Beulich <jbeulich@suse.com>

No, this isn't what the R-b was given for.

> v2:
>  - Set apic_access_mfn to INVALID_MFN in vmx_free_vlapic_mapping() to make
>    the function idempotent

Andrew had suggested to use 0 instead of INVALID_MFN. I don't see
how you achieved idempotency with this adjustment. Aiui
vmx_free_vlapic_mapping() is supposed to also run correctly if
vmx_alloc_vlapic_mapping() was never called.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe

[Durrant, Paul]

> -----Original Message-----
> From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of Jan
> Beulich
> Sent: 23 January 2020 12:45
> To: Durrant, Paul <pdurrant@amazon.co.uk>
> Cc: Kevin Tian <kevin.tian@intel.com>; Wei Liu <wl@xen.org>; Andrew Cooper
> <andrew.cooper3@citrix.com>; Jun Nakajima <jun.nakajima@intel.com>; xen-
> devel@lists.xenproject.org; Roger Pau Monné <roger.pau@citrix.com>
> Subject: Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn
> type-safe
> 
> On 23.01.2020 13:21, Paul Durrant wrote:
> > Use mfn_t rather than unsigned long and change previous tests against 0
> to
> > tests against INVALID_MFN (also introducing initialization to that
> value).
> >
> > Signed-off-by: Paul Durrant <pdurrant@amazon.com>
> > Acked-by: Kevin Tian <kevin.tian@intel.com>
> > Reviewed-by: Jan Beulich <jbeulich@suse.com>
> 
> No, this isn't what the R-b was given for.

Oh, sorry, I misunderstood; I thought the R-b was good as long as idempotency was ensured.

> 
> > v2:
> >  - Set apic_access_mfn to INVALID_MFN in vmx_free_vlapic_mapping() to
> make
> >    the function idempotent
> 
> Andrew had suggested to use 0 instead of INVALID_MFN. I don't see
> how you achieved idempotency with this adjustment. Aiui
> vmx_free_vlapic_mapping() is supposed to also run correctly if
> vmx_alloc_vlapic_mapping() was never called.

It will. vmx_domain_initialise() will set apic_access_mfn to INVALID_MFN so vmx_free_vlapic_mapping() will do nothing.

  Paul
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe

[Jan Beulich]

On 23.01.2020 14:09, Durrant, Paul wrote:
>> -----Original Message-----
>> From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of Jan
>> Beulich
>> Sent: 23 January 2020 12:45
>> To: Durrant, Paul <pdurrant@amazon.co.uk>
>> Cc: Kevin Tian <kevin.tian@intel.com>; Wei Liu <wl@xen.org>; Andrew Cooper
>> <andrew.cooper3@citrix.com>; Jun Nakajima <jun.nakajima@intel.com>; xen-
>> devel@lists.xenproject.org; Roger Pau Monné <roger.pau@citrix.com>
>> Subject: Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn
>> type-safe
>>
>> On 23.01.2020 13:21, Paul Durrant wrote:
>>> Use mfn_t rather than unsigned long and change previous tests against 0
>> to
>>> tests against INVALID_MFN (also introducing initialization to that
>> value).
>>>
>>> Signed-off-by: Paul Durrant <pdurrant@amazon.com>
>>> Acked-by: Kevin Tian <kevin.tian@intel.com>
>>> Reviewed-by: Jan Beulich <jbeulich@suse.com>
>>
>> No, this isn't what the R-b was given for.
> 
> Oh, sorry, I misunderstood; I thought the R-b was good as long as idempotency was ensured.
> 
>>
>>> v2:
>>>  - Set apic_access_mfn to INVALID_MFN in vmx_free_vlapic_mapping() to
>> make
>>>    the function idempotent
>>
>> Andrew had suggested to use 0 instead of INVALID_MFN. I don't see
>> how you achieved idempotency with this adjustment. Aiui
>> vmx_free_vlapic_mapping() is supposed to also run correctly if
>> vmx_alloc_vlapic_mapping() was never called.
> 
> It will. vmx_domain_initialise() will set apic_access_mfn to INVALID_MFN
> so vmx_free_vlapic_mapping() will do nothing.

I'm sorry, it was implied that it also needs to work if
vmx_domain_initialise() was never called. Andrew's goal after
all is, aiui, to be able to call "destroy" functions on error
paths irrespective of how far "create" had managed to progress.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe

[Durrant, Paul]

> -----Original Message-----
> From: Jan Beulich <jbeulich@suse.com>
> Sent: 23 January 2020 13:30
> To: Durrant, Paul <pdurrant@amazon.co.uk>
> Cc: Kevin Tian <kevin.tian@intel.com>; Wei Liu <wl@xen.org>; Andrew Cooper
> <andrew.cooper3@citrix.com>; Jun Nakajima <jun.nakajima@intel.com>; xen-
> devel@lists.xenproject.org; Roger Pau Monné <roger.pau@citrix.com>
> Subject: Re: [PATCH v2 1/3] x86 / vmx: make apic_access_mfn type-safe
> 
> On 23.01.2020 14:09, Durrant, Paul wrote:
> >> -----Original Message-----
> >> From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of
> Jan
> >> Beulich
> >> Sent: 23 January 2020 12:45
> >> To: Durrant, Paul <pdurrant@amazon.co.uk>
> >> Cc: Kevin Tian <kevin.tian@intel.com>; Wei Liu <wl@xen.org>; Andrew
> Cooper
> >> <andrew.cooper3@citrix.com>; Jun Nakajima <jun.nakajima@intel.com>;
> xen-
> >> devel@lists.xenproject.org; Roger Pau Monné <roger.pau@citrix.com>
> >> Subject: Re: [Xen-devel] [PATCH v2 1/3] x86 / vmx: make apic_access_mfn
> >> type-safe
> >>
> >> On 23.01.2020 13:21, Paul Durrant wrote:
> >>> Use mfn_t rather than unsigned long and change previous tests against
> 0
> >> to
> >>> tests against INVALID_MFN (also introducing initialization to that
> >> value).
> >>>
> >>> Signed-off-by: Paul Durrant <pdurrant@amazon.com>
> >>> Acked-by: Kevin Tian <kevin.tian@intel.com>
> >>> Reviewed-by: Jan Beulich <jbeulich@suse.com>
> >>
> >> No, this isn't what the R-b was given for.
> >
> > Oh, sorry, I misunderstood; I thought the R-b was good as long as
> idempotency was ensured.
> >
> >>
> >>> v2:
> >>>  - Set apic_access_mfn to INVALID_MFN in vmx_free_vlapic_mapping() to
> >> make
> >>>    the function idempotent
> >>
> >> Andrew had suggested to use 0 instead of INVALID_MFN. I don't see
> >> how you achieved idempotency with this adjustment. Aiui
> >> vmx_free_vlapic_mapping() is supposed to also run correctly if
> >> vmx_alloc_vlapic_mapping() was never called.
> >
> > It will. vmx_domain_initialise() will set apic_access_mfn to INVALID_MFN
> > so vmx_free_vlapic_mapping() will do nothing.
> 
> I'm sorry, it was implied that it also needs to work if
> vmx_domain_initialise() was never called. Andrew's goal after
> all is, aiui, to be able to call "destroy" functions on error
> paths irrespective of how far "create" had managed to progress.
> 

Oh, I see. That implication was not at all obvious to me. I thought he was just after being able to 'destroy' as many times as it took to finish, in which case our choice for the value of INVALID_MFN is indeed unfortunate. If that's the goal I'll switch to use _mfn(0) as a comparator.

  Paul

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel