x86/vMSI-X: avoid missing first unmask of vectors

Recent changes to Linux result in there just being a single unmask
operation prior to expecting the first interrupts to arrive. However,
we've had a chicken-and-egg problem here: Qemu invokes
xc_domain_update_msi_irq(), ultimately leading to
msixtbl_pt_register(), upon seeing that first unmask operation. Yet
for msixtbl_range() to return true (in order to msixtbl_write() to get
invoked at all) msixtbl_pt_register() must have completed.

Deal with this by snooping suitable writes in msixtbl_range() and
triggering the invocation of msix_write_completion() from
msixtbl_pt_register() when that happens in the context of a still in
progress vector control field write.

Note that the seemingly unrelated deletion of the redundant
irq_desc->msi_desc check in msixtbl_pt_register() is to make clear to
any compiler version used that the "msi_desc" local variable isn't
being used uninitialized. (Doing the same in msixtbl_pt_unregister() is
just for consistency reasons.)

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
Release-acked-by: Wei Liu <wei.liu2@citrix.com>

diff --git a/xen/arch/x86/hvm/vmsi.c b/xen/arch/x86/hvm/vmsi.c
index d450a3c6bbe6458c59611d49158f6a1e1f3ecd0d..805e560701c5b5b82132d88c5d256e55b0a8fc8b 100644 (file)
--- a/xen/arch/x86/hvm/vmsi.c
+++ b/xen/arch/x86/hvm/vmsi.c
@@ -341,7 +341,21 @@ static int msixtbl_range(struct vcpu *v, unsigned long addr)
     desc = msixtbl_addr_to_desc(msixtbl_find_entry(v, addr), addr);
     rcu_read_unlock(&msixtbl_rcu_lock);
 
-    return !!desc;
+    if ( desc )
+        return 1;
+
+    if ( (addr & (PCI_MSIX_ENTRY_SIZE - 1)) ==
+         PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET )
+    {
+        const ioreq_t *r = &v->arch.hvm_vcpu.hvm_io.io_req;
+
+        ASSERT(r->type == IOREQ_TYPE_COPY);
+        if ( r->dir == IOREQ_WRITE && r->size == 4 && !r->data_is_ptr
+             && !(r->data & PCI_MSIX_VECTOR_BITMASK) )
+            v->arch.hvm_vcpu.hvm_io.msix_snoop_address = addr;
+    }
+
+    return 0;
 }
 
 static const struct hvm_mmio_ops msixtbl_mmio_ops = {
@@ -410,9 +424,6 @@ int msixtbl_pt_register(struct domain *d, struct pirq *pirq, uint64_t gtable)
         return r;
     }
 
-    if ( !irq_desc->msi_desc )
-        goto out;
-
     msi_desc = irq_desc->msi_desc;
     if ( !msi_desc )
         goto out;
@@ -437,6 +448,22 @@ found:
 out:
     spin_unlock_irq(&irq_desc->lock);
     xfree(new_entry);
+
+    if ( !r )
+    {
+        struct vcpu *v;
+
+        for_each_vcpu ( d, v )
+        {
+            if ( v->arch.hvm_vcpu.hvm_io.msix_snoop_address ==
+                 (gtable + msi_desc->msi_attrib.entry_nr *
+                           PCI_MSIX_ENTRY_SIZE +
+                  PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET) )
+                v->arch.hvm_vcpu.hvm_io.msix_unmask_address =
+                    v->arch.hvm_vcpu.hvm_io.msix_snoop_address;
+        }
+    }
+
     return r;
 }
 
@@ -457,9 +484,6 @@ void msixtbl_pt_unregister(struct domain *d, struct pirq *pirq)
     if ( !irq_desc )
         return;
 
-    if ( !irq_desc->msi_desc )
-        goto out;
-
     msi_desc = irq_desc->msi_desc;
     if ( !msi_desc )
         goto out;
@@ -522,6 +546,8 @@ void msix_write_completion(struct vcpu *v)
 {
     unsigned long ctrl_address = v->arch.hvm_vcpu.hvm_io.msix_unmask_address;
 
+    v->arch.hvm_vcpu.hvm_io.msix_snoop_address = 0;
+
     if ( !ctrl_address )
         return;
 

diff --git a/xen/include/asm-x86/hvm/vcpu.h b/xen/include/asm-x86/hvm/vcpu.h
index 152d9f3ee8cec700f106276edf7c5deb83afe016..06778eb5afce404d96d7036dc6a7ee13ca47d4a9 100644 (file)
--- a/xen/include/asm-x86/hvm/vcpu.h
+++ b/xen/include/asm-x86/hvm/vcpu.h
@@ -85,6 +85,7 @@ struct hvm_vcpu_io {
     bool_t mmio_retry;
 
     unsigned long msix_unmask_address;
+    unsigned long msix_snoop_address;
 
     const struct g2m_ioport *g2m_ioport;
 };