Change semantics of grant transfers for vp guests so that the

operation automatically gets you a fresh page at the same
pseudo-physical address as Keir suggested.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

diff --git a/linux-2.6-xen-sparse/drivers/xen/netback/netback.c b/linux-2.6-xen-sparse/drivers/xen/netback/netback.c
index c5ea3a8b24da4c6b808a2a20df7f78983228bdbf..a58e215163d29003e2d428311254a8fc04552587 100644 (file)
--- a/linux-2.6-xen-sparse/drivers/xen/netback/netback.c
+++ b/linux-2.6-xen-sparse/drivers/xen/netback/netback.c
@@ -235,23 +235,35 @@ static void net_rx_action(unsigned long unused)
                vdata   = (unsigned long)skb->data;
                old_mfn = virt_to_mfn(vdata);
 
-               /* Memory squeeze? Back off for an arbitrary while. */
-               if ((new_mfn = alloc_mfn()) == 0) {
-                       if ( net_ratelimit() )
-                               WPRINTK("Memory squeeze in netback driver.\n");
-                       mod_timer(&net_timer, jiffies + HZ);
-                       skb_queue_head(&rx_queue, skb);
-                       break;
-               }
-               /*
-                * Set the new P2M table entry before reassigning the old data
-                * page. Heed the comment in pgtable-2level.h:pte_page(). :-)
-                */
-               set_phys_to_machine(__pa(skb->data) >> PAGE_SHIFT, new_mfn);
+               if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+                       /* Memory squeeze? Back off for an arbitrary while. */
+                       if ((new_mfn = alloc_mfn()) == 0) {
+                               if ( net_ratelimit() )
+                                       WPRINTK("Memory squeeze in netback "
+                                               "driver.\n");
+                               mod_timer(&net_timer, jiffies + HZ);
+                               skb_queue_head(&rx_queue, skb);
+                               break;
+                       }
+                       /*
+                        * Set the new P2M table entry before reassigning
+                        * the old data page. Heed the comment in
+                        * pgtable-2level.h:pte_page(). :-)
+                        */
+                       set_phys_to_machine(
+                               __pa(skb->data) >> PAGE_SHIFT,
+                               new_mfn);
+
+                       MULTI_update_va_mapping(mcl, vdata,
+                                               pfn_pte_ma(new_mfn,
+                                                          PAGE_KERNEL), 0);
+                       mcl++;
 
-               MULTI_update_va_mapping(mcl, vdata,
-                                       pfn_pte_ma(new_mfn, PAGE_KERNEL), 0);
-               mcl++;
+                       mmu->ptr = ((maddr_t)new_mfn << PAGE_SHIFT) |
+                               MMU_MACHPHYS_UPDATE;
+                       mmu->val = __pa(vdata) >> PAGE_SHIFT;
+                       mmu++;
+               }
 
                gop->mfn = old_mfn;
                gop->domid = netif->domid;
@@ -260,13 +272,6 @@ static void net_rx_action(unsigned long unused)
                netif->rx.req_cons++;
                gop++;
 
-               if (!xen_feature(XENFEAT_auto_translated_physmap)) {
-                       mmu->ptr = ((maddr_t)new_mfn << PAGE_SHIFT) |
-                               MMU_MACHPHYS_UPDATE;
-                       mmu->val = __pa(vdata) >> PAGE_SHIFT;
-                       mmu++;
-               }
-
                __skb_queue_tail(&rxq, skb);
 
                /* Filled the batch queue? */
@@ -274,22 +279,24 @@ static void net_rx_action(unsigned long unused)
                        break;
        }
 
-       if (mcl == rx_mcl)
-               return;
+       if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+               if (mcl == rx_mcl)
+                       return;
 
-       mcl[-1].args[MULTI_UVMFLAGS_INDEX] = UVMF_TLB_FLUSH|UVMF_ALL;
+               mcl[-1].args[MULTI_UVMFLAGS_INDEX] = UVMF_TLB_FLUSH|UVMF_ALL;
 
-       if (mmu - rx_mmu) {
-               mcl->op = __HYPERVISOR_mmu_update;
-               mcl->args[0] = (unsigned long)rx_mmu;
-               mcl->args[1] = mmu - rx_mmu;
-               mcl->args[2] = 0;
-               mcl->args[3] = DOMID_SELF;
-               mcl++;
-       }
+               if (mmu - rx_mmu) {
+                       mcl->op = __HYPERVISOR_mmu_update;
+                       mcl->args[0] = (unsigned long)rx_mmu;
+                       mcl->args[1] = mmu - rx_mmu;
+                       mcl->args[2] = 0;
+                       mcl->args[3] = DOMID_SELF;
+                       mcl++;
+               }
 
-       ret = HYPERVISOR_multicall(rx_mcl, mcl - rx_mcl);
-       BUG_ON(ret != 0);
+               ret = HYPERVISOR_multicall(rx_mcl, mcl - rx_mcl);
+               BUG_ON(ret != 0);
+       }
 
        ret = HYPERVISOR_grant_table_op(GNTTABOP_transfer, grant_rx_op, 
                                        gop - grant_rx_op);
@@ -308,8 +315,11 @@ static void net_rx_action(unsigned long unused)
                netif->stats.tx_bytes += size;
                netif->stats.tx_packets++;
 
-               /* The update_va_mapping() must not fail. */
-               BUG_ON(mcl->result != 0);
+               if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+                       /* The update_va_mapping() must not fail. */
+                       BUG_ON(mcl->result != 0);
+                       mcl++;
+               }
 
                /* Check the reassignment error code. */
                status = NETIF_RSP_OKAY;
@@ -340,7 +350,6 @@ static void net_rx_action(unsigned long unused)
 
                netif_put(netif);
                dev_kfree_skb(skb);
-               mcl++;
                gop++;
        }