[PATCH 0/2] x86: fix memory leaks

[PATCH 0/2] x86: fix memory leaks

[Jan Beulich]

1: p2m: clean up altp2m
2: fix domain cleanup

Signed-off-by: Jan Beulich <jbeulich@suse.com>


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

[PATCH 1/2] x86/p2m: clean up altp2m

[Jan Beulich]

[-- Attachment #1: Type: text/plain, Size: 522 bytes --]

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/mm/p2m.c
+++ b/xen/arch/x86/mm/p2m.c
@@ -618,9 +618,11 @@ void p2m_teardown(struct p2m_domain *p2m
 
 void p2m_final_teardown(struct domain *d)
 {
-    /* We must teardown unconditionally because
+    /*
+     * We must teardown both of them unconditionally because
      * we initialise them unconditionally.
      */
+    p2m_teardown_altp2m(d);
     p2m_teardown_nestedp2m(d);
 
     /* Iterate over all p2m tables per domain */




[-- Attachment #2: x86-altp2m-teardown.patch --]
[-- Type: text/plain, Size: 544 bytes --]

x86/p2m: clean up altp2m

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/mm/p2m.c
+++ b/xen/arch/x86/mm/p2m.c
@@ -618,9 +618,11 @@ void p2m_teardown(struct p2m_domain *p2m
 
 void p2m_final_teardown(struct domain *d)
 {
-    /* We must teardown unconditionally because
+    /*
+     * We must teardown both of them unconditionally because
      * we initialise them unconditionally.
      */
+    p2m_teardown_altp2m(d);
     p2m_teardown_nestedp2m(d);
 
     /* Iterate over all p2m tables per domain */

[-- Attachment #3: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

[PATCH 2/2] x86: fix domain cleanup

[Jan Beulich]

[-- Attachment #1: Type: text/plain, Size: 2054 bytes --]

Free d->arch.cpuids on both the creation error path and during
destruction.

Don't bypass iommu_domain_destroy().

Move psr_domain_init() up so that hvm_domain_initialise() again is the
last thing which can fail, and hence doesn't require undoing later on.

Move psr_domain_free() up on the creation error path, so that cleanup
once again gets done in reverse order of setup.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/domain.c
+++ b/xen/arch/x86/domain.c
@@ -642,21 +642,18 @@ int arch_domain_create(struct domain *d,
     }
     spin_lock_init(&d->arch.e820_lock);
 
+    if ( (rc = psr_domain_init(d)) != 0 )
+        goto fail;
+
     if ( has_hvm_container_domain(d) )
     {
         if ( (rc = hvm_domain_initialise(d)) != 0 )
-        {
-            iommu_domain_destroy(d);
             goto fail;
-        }
     }
     else
         /* 64-bit PV guest by default. */
         d->arch.is_32bit_pv = d->arch.has_32bit_shinfo = 0;
 
-    if ( (rc = psr_domain_init(d)) != 0 )
-        goto fail;
-
     /* initialize default tsc behavior in case tools don't */
     tsc_set_info(d, TSC_MODE_DEFAULT, 0UL, 0, 0);
     spin_lock_init(&d->arch.vtsc_lock);
@@ -674,8 +671,11 @@ int arch_domain_create(struct domain *d,
 
  fail:
     d->is_dying = DOMDYING_dead;
+    psr_domain_free(d);
+    iommu_domain_destroy(d);
     cleanup_domain_irq_mapping(d);
     free_xenheap_page(d->shared_info);
+    xfree(d->arch.cpuids);
     if ( paging_initialised )
         paging_final_teardown(d);
     free_perdomain_mappings(d);
@@ -684,7 +684,6 @@ int arch_domain_create(struct domain *d,
         xfree(d->arch.pv_domain.cpuidmasks);
         free_xenheap_page(d->arch.pv_domain.gdt_ldt_l1tab);
     }
-    psr_domain_free(d);
     return rc;
 }
 
@@ -694,6 +693,7 @@ void arch_domain_destroy(struct domain *
         hvm_domain_destroy(d);
 
     xfree(d->arch.e820);
+    xfree(d->arch.cpuids);
 
     free_domain_pirqs(d);
     if ( !is_idle_domain(d) )




[-- Attachment #2: x86-free-cpuids.patch --]
[-- Type: text/plain, Size: 2075 bytes --]

x86: fix domain cleanup

Free d->arch.cpuids on both the creation error path and during
destruction.

Don't bypass iommu_domain_destroy().

Move psr_domain_init() up so that hvm_domain_initialise() again is the
last thing which can fail, and hence doesn't require undoing later on.

Move psr_domain_free() up on the creation error path, so that cleanup
once again gets done in reverse order of setup.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/domain.c
+++ b/xen/arch/x86/domain.c
@@ -642,21 +642,18 @@ int arch_domain_create(struct domain *d,
     }
     spin_lock_init(&d->arch.e820_lock);
 
+    if ( (rc = psr_domain_init(d)) != 0 )
+        goto fail;
+
     if ( has_hvm_container_domain(d) )
     {
         if ( (rc = hvm_domain_initialise(d)) != 0 )
-        {
-            iommu_domain_destroy(d);
             goto fail;
-        }
     }
     else
         /* 64-bit PV guest by default. */
         d->arch.is_32bit_pv = d->arch.has_32bit_shinfo = 0;
 
-    if ( (rc = psr_domain_init(d)) != 0 )
-        goto fail;
-
     /* initialize default tsc behavior in case tools don't */
     tsc_set_info(d, TSC_MODE_DEFAULT, 0UL, 0, 0);
     spin_lock_init(&d->arch.vtsc_lock);
@@ -674,8 +671,11 @@ int arch_domain_create(struct domain *d,
 
  fail:
     d->is_dying = DOMDYING_dead;
+    psr_domain_free(d);
+    iommu_domain_destroy(d);
     cleanup_domain_irq_mapping(d);
     free_xenheap_page(d->shared_info);
+    xfree(d->arch.cpuids);
     if ( paging_initialised )
         paging_final_teardown(d);
     free_perdomain_mappings(d);
@@ -684,7 +684,6 @@ int arch_domain_create(struct domain *d,
         xfree(d->arch.pv_domain.cpuidmasks);
         free_xenheap_page(d->arch.pv_domain.gdt_ldt_l1tab);
     }
-    psr_domain_free(d);
     return rc;
 }
 
@@ -694,6 +693,7 @@ void arch_domain_destroy(struct domain *
         hvm_domain_destroy(d);
 
     xfree(d->arch.e820);
+    xfree(d->arch.cpuids);
 
     free_domain_pirqs(d);
     if ( !is_idle_domain(d) )

[-- Attachment #3: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 1/2] x86/p2m: clean up altp2m

[Andrew Cooper]

On 29/04/16 15:49, Jan Beulich wrote:
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 2/2] x86: fix domain cleanup

[Andrew Cooper]

On 29/04/16 15:49, Jan Beulich wrote:
> Free d->arch.cpuids on both the creation error path and during
> destruction.
>
> Don't bypass iommu_domain_destroy().
>
> Move psr_domain_init() up so that hvm_domain_initialise() again is the
> last thing which can fail, and hence doesn't require undoing later on.
>
> Move psr_domain_free() up on the creation error path, so that cleanup
> once again gets done in reverse order of setup.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 0/2] x86: fix memory leaks

[Wei Liu]

On Fri, Apr 29, 2016 at 08:41:14AM -0600, Jan Beulich wrote:
> 1: p2m: clean up altp2m
> 2: fix domain cleanup
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> 

Release-acked-by: Wei Liu <wei.liu2@citrix.com>

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 1/2] x86/p2m: clean up altp2m

[George Dunlap]

On Fri, Apr 29, 2016 at 3:49 PM, Jan Beulich <JBeulich@suse.com> wrote:
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Acked-by: George Dunlap <george.dunlap@citrix.com>

>
> --- a/xen/arch/x86/mm/p2m.c
> +++ b/xen/arch/x86/mm/p2m.c
> @@ -618,9 +618,11 @@ void p2m_teardown(struct p2m_domain *p2m
>
>  void p2m_final_teardown(struct domain *d)
>  {
> -    /* We must teardown unconditionally because
> +    /*
> +     * We must teardown both of them unconditionally because
>       * we initialise them unconditionally.
>       */
> +    p2m_teardown_altp2m(d);
>      p2m_teardown_nestedp2m(d);
>
>      /* Iterate over all p2m tables per domain */
>
>
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
>

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 1/2] x86/p2m: clean up altp2m

[George Dunlap]

On Tue, May 3, 2016 at 5:44 PM, George Dunlap
<George.Dunlap@eu.citrix.com> wrote:
> On Fri, Apr 29, 2016 at 3:49 PM, Jan Beulich <JBeulich@suse.com> wrote:
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>
> Acked-by: George Dunlap <george.dunlap@citrix.com>

And I guess this is a candidate for backport to 4.6 too?

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH 1/2] x86/p2m: clean up altp2m

[Jan Beulich]

>>> On 03.05.16 at 18:45, <George.Dunlap@eu.citrix.com> wrote:
> On Tue, May 3, 2016 at 5:44 PM, George Dunlap
> <George.Dunlap@eu.citrix.com> wrote:
>> On Fri, Apr 29, 2016 at 3:49 PM, Jan Beulich <JBeulich@suse.com> wrote:
>>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>>
>> Acked-by: George Dunlap <george.dunlap@citrix.com>
> 
> And I guess this is a candidate for backport to 4.6 too?

Definitely (and I have it in my queue already).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel