xen/vcpu: Improve sanity checks in vcpu_create()
diff mbox series

Message ID 20200115184418.21365-1-andrew.cooper3@citrix.com
State New, archived
Headers show
Series
  • xen/vcpu: Improve sanity checks in vcpu_create()
Related show

Commit Message

Andrew Cooper Jan. 15, 2020, 6:44 p.m. UTC
The BUG_ON() is confusing to follow.  The (!is_idle_domain(d) || vcpu_id) part
is a vestigial remnant of architectures poisioning idle_vcpu[0] with non-NULL
pointers.

Now that idle_vcpu[0] is NULL on all architectures, and d->max_vcpus specified
before vcpu_create() is called, we can properly range check the requested
vcpu_id.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
CC: Wei Liu <wl@xen.org>
CC: Roger Pau Monné <roger.pau@citrix.com>
CC: Stefano Stabellini <sstabellini@kernel.org>
CC: Julien Grall <julien@xen.org>
CC: Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
---
 xen/common/domain.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

Comments

Jan Beulich Jan. 16, 2020, 9:55 a.m. UTC | #1
On 15.01.2020 19:44, Andrew Cooper wrote:
> The BUG_ON() is confusing to follow.  The (!is_idle_domain(d) || vcpu_id) part
> is a vestigial remnant of architectures poisioning idle_vcpu[0] with non-NULL
> pointers.
> 
> Now that idle_vcpu[0] is NULL on all architectures, and d->max_vcpus specified
> before vcpu_create() is called, we can properly range check the requested
> vcpu_id.

I guess this is meant to be true on top of your Arm side change
which hasn't been committed yet? And perhaps better "... starts
out as NULL on all ..."?

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

Reviewed-by: Jan Beulich <jbeulich@suse.com>
Julien Grall Jan. 16, 2020, 10:04 a.m. UTC | #2
Hi,

On 16/01/2020 09:55, Jan Beulich wrote:
> On 15.01.2020 19:44, Andrew Cooper wrote:
>> The BUG_ON() is confusing to follow.  The (!is_idle_domain(d) || vcpu_id) part
>> is a vestigial remnant of architectures poisioning idle_vcpu[0] with non-NULL
>> pointers.
>>
>> Now that idle_vcpu[0] is NULL on all architectures, and d->max_vcpus specified
>> before vcpu_create() is called, we can properly range check the requested
>> vcpu_id.
> 
> I guess this is meant to be true on top of your Arm side change
> which hasn't been committed yet? And perhaps better "... starts
> out as NULL on all ..."?

I can commit the two together.

> 
>> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> 
> Reviewed-by: Jan Beulich <jbeulich@suse.com>

Acked-by: Julien Grall <julien@xen.org>

Cheers,

Patch
diff mbox series

diff --git a/xen/common/domain.c b/xen/common/domain.c
index 0b1103fdb2..ee3f9ffd3e 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -139,7 +139,19 @@  struct vcpu *vcpu_create(struct domain *d, unsigned int vcpu_id)
 {
     struct vcpu *v;
 
-    BUG_ON((!is_idle_domain(d) || vcpu_id) && d->vcpu[vcpu_id]);
+    /*
+     * Sanity check some input expectations:
+     * - vcpu_id should be bounded by d->max_vcpus, and not previously
+     *   allocated.
+     * - VCPUs should be tightly packed and allocated in ascending order,
+     *   except for the idle domain which may vary based on PCPU numbering.
+     */
+    if ( vcpu_id >= d->max_vcpus || d->vcpu[vcpu_id] ||
+         (!is_idle_domain(d) && vcpu_id && !d->vcpu[vcpu_id - 1]) )
+    {
+        ASSERT_UNREACHABLE();
+        return NULL;
+    }
 
     if ( (v = alloc_vcpu_struct(d)) == NULL )
         return NULL;