spapr_pci/spapr_pci_vfio: Support Dynamic DMA Windows (DDW)

This adds support for Dynamic DMA Windows (DDW) option defined by
the SPAPR specification which allows to have additional DMA window(s)

The "ddw" property is enabled by default on a PHB but for compatibility
the pseries-2.6 machine and older disable it.
This also creates a single DMA window for the older machines to
maintain backward migration.

This implements DDW for PHB with emulated and VFIO devices. The host
kernel support is required. The advertised IOMMU page sizes are 4K and
64K; 16M pages are supported but not advertised by default, in order to
enable them, the user has to specify "pgsz" property for PHB and
enable huge pages for RAM.

The existing linux guests try creating one additional huge DMA window
with 64K or 16MB pages and map the entire guest RAM to. If succeeded,
the guest switches to dma_direct_ops and never calls TCE hypercalls
(H_PUT_TCE,...) again. This enables VFIO devices to use the entire RAM
and not waste time on map/unmap later. This adds a "dma64_win_addr"
property which is a bus address for the 64bit window and by default
set to 0x800.0000.0000.0000 as this is what the modern POWER8 hardware
uses and this allows having emulated and VFIO devices on the same bus.

This adds 4 RTAS handlers:
* ibm,query-pe-dma-window
* ibm,create-pe-dma-window
* ibm,remove-pe-dma-window
* ibm,reset-pe-dma-window
These are registered from type_init() callback.

These RTAS handlers are implemented in a separate file to avoid polluting
spapr_iommu.c with PCI.

This changes sPAPRPHBState::dma_liobn to an array to allow 2 LIOBNs
and updates all references to dma_liobn. However this does not add
64bit LIOBN to the migration stream as in fact even 32bit LIOBN is
rather pointless there (as it is a PHB property and the management
software can/should pass LIOBNs via CLI) but we keep it for the backward
migration support.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

1 files changed, 57 insertions, 18 deletions

diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
index cbb7cdd774..949c44fec8 100644
--- a/hw/ppc/spapr_pci.c
+++ b/hw/ppc/spapr_pci.c
@@ -35,6 +35,7 @@
 #include "hw/ppc/spapr.h"
 #include "hw/pci-host/spapr.h"
 #include "exec/address-spaces.h"
+#include "exec/ram_addr.h"
 #include <libfdt.h>
 #include "trace.h"
 #include "qemu/error-report.h"
@@ -45,6 +46,7 @@
 #include "hw/ppc/spapr_drc.h"
 #include "sysemu/device_tree.h"
 #include "sysemu/kvm.h"
+#include "sysemu/hostmem.h"
 
 #include "hw/vfio/vfio.h"
 
@@ -1304,11 +1306,14 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp)
     PCIBus *bus;
     uint64_t msi_window_size = 4096;
     sPAPRTCETable *tcet;
+    const unsigned windows_supported =
+        sphb->ddw_enabled ? SPAPR_PCI_DMA_MAX_WINDOWS : 1;
 
     if (sphb->index != (uint32_t)-1) {
         hwaddr windows_base;
 
-        if ((sphb->buid != (uint64_t)-1) || (sphb->dma_liobn != (uint32_t)-1)
+        if ((sphb->buid != (uint64_t)-1) || (sphb->dma_liobn[0] != (uint32_t)-1)
+            || (sphb->dma_liobn[1] != (uint32_t)-1 && windows_supported == 2)
             || (sphb->mem_win_addr != (hwaddr)-1)
             || (sphb->io_win_addr != (hwaddr)-1)) {
             error_setg(errp, "Either \"index\" or other parameters must"
@@ -1323,7 +1328,9 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp)
         }
 
         sphb->buid = SPAPR_PCI_BASE_BUID + sphb->index;
-        sphb->dma_liobn = SPAPR_PCI_LIOBN(sphb->index, 0);
+        for (i = 0; i < windows_supported; ++i) {
+            sphb->dma_liobn[i] = SPAPR_PCI_LIOBN(sphb->index, i);
+        }
 
         windows_base = SPAPR_PCI_WINDOW_BASE
             + sphb->index * SPAPR_PCI_WINDOW_SPACING;
@@ -1336,8 +1343,9 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp)
         return;
     }
 
-    if (sphb->dma_liobn == (uint32_t)-1) {
-        error_setg(errp, "LIOBN not specified for PHB");
+    if ((sphb->dma_liobn[0] == (uint32_t)-1) ||
+        ((sphb->dma_liobn[1] == (uint32_t)-1) && (windows_supported > 1))) {
+        error_setg(errp, "LIOBN(s) not specified for PHB");
         return;
     }
 
@@ -1456,16 +1464,18 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp)
         }
     }
 
-    tcet = spapr_tce_new_table(DEVICE(sphb), sphb->dma_liobn);
-    if (!tcet) {
-        error_setg(errp, "Unable to create TCE table for %s",
-                   sphb->dtbusname);
-        return;
+    /* DMA setup */
+    for (i = 0; i < windows_supported; ++i) {
+        tcet = spapr_tce_new_table(DEVICE(sphb), sphb->dma_liobn[i]);
+        if (!tcet) {
+            error_setg(errp, "Creating window#%d failed for %s",
+                       i, sphb->dtbusname);
+            return;
+        }
+        memory_region_add_subregion_overlap(&sphb->iommu_root, 0,
+                                            spapr_tce_get_iommu(tcet), 0);
     }
 
-    memory_region_add_subregion_overlap(&sphb->iommu_root, 0,
-                                        spapr_tce_get_iommu(tcet), 0);
-
     sphb->msi = g_hash_table_new_full(g_int_hash, g_int_equal, g_free, g_free);
 }
 
@@ -1482,13 +1492,19 @@ static int spapr_phb_children_reset(Object *child, void *opaque)
 
 void spapr_phb_dma_reset(sPAPRPHBState *sphb)
 {
-    sPAPRTCETable *tcet = spapr_tce_find_by_liobn(sphb->dma_liobn);
+    int i;
+    sPAPRTCETable *tcet;
+
+    for (i = 0; i < SPAPR_PCI_DMA_MAX_WINDOWS; ++i) {
+        tcet = spapr_tce_find_by_liobn(sphb->dma_liobn[i]);
 
-    if (tcet && tcet->nb_table) {
-        spapr_tce_table_disable(tcet);
+        if (tcet && tcet->nb_table) {
+            spapr_tce_table_disable(tcet);
+        }
     }
 
     /* Register default 32bit DMA window */
+    tcet = spapr_tce_find_by_liobn(sphb->dma_liobn[0]);
     spapr_tce_table_enable(tcet, SPAPR_TCE_PAGE_SHIFT, sphb->dma_win_addr,
                            sphb->dma_win_size >> SPAPR_TCE_PAGE_SHIFT);
 }
@@ -1510,7 +1526,8 @@ static void spapr_phb_reset(DeviceState *qdev)
 static Property spapr_phb_properties[] = {
     DEFINE_PROP_UINT32("index", sPAPRPHBState, index, -1),
     DEFINE_PROP_UINT64("buid", sPAPRPHBState, buid, -1),
-    DEFINE_PROP_UINT32("liobn", sPAPRPHBState, dma_liobn, -1),
+    DEFINE_PROP_UINT32("liobn", sPAPRPHBState, dma_liobn[0], -1),
+    DEFINE_PROP_UINT32("liobn64", sPAPRPHBState, dma_liobn[1], -1),
     DEFINE_PROP_UINT64("mem_win_addr", sPAPRPHBState, mem_win_addr, -1),
     DEFINE_PROP_UINT64("mem_win_size", sPAPRPHBState, mem_win_size,
                        SPAPR_PCI_MMIO_WIN_SIZE),
@@ -1522,6 +1539,11 @@ static Property spapr_phb_properties[] = {
     /* Default DMA window is 0..1GB */
     DEFINE_PROP_UINT64("dma_win_addr", sPAPRPHBState, dma_win_addr, 0),
     DEFINE_PROP_UINT64("dma_win_size", sPAPRPHBState, dma_win_size, 0x40000000),
+    DEFINE_PROP_UINT64("dma64_win_addr", sPAPRPHBState, dma64_win_addr,
+                       0x800000000000000ULL),
+    DEFINE_PROP_BOOL("ddw", sPAPRPHBState, ddw_enabled, true),
+    DEFINE_PROP_UINT64("pgsz", sPAPRPHBState, page_size_mask,
+                       (1ULL << 12) | (1ULL << 16)),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -1598,7 +1620,7 @@ static const VMStateDescription vmstate_spapr_pci = {
     .post_load = spapr_pci_post_load,
     .fields = (VMStateField[]) {
         VMSTATE_UINT64_EQUAL(buid, sPAPRPHBState),
-        VMSTATE_UINT32_EQUAL(dma_liobn, sPAPRPHBState),
+        VMSTATE_UINT32_EQUAL(dma_liobn[0], sPAPRPHBState),
         VMSTATE_UINT64_EQUAL(mem_win_addr, sPAPRPHBState),
         VMSTATE_UINT64_EQUAL(mem_win_size, sPAPRPHBState),
         VMSTATE_UINT64_EQUAL(io_win_addr, sPAPRPHBState),
@@ -1774,6 +1796,15 @@ int spapr_populate_pci_dt(sPAPRPHBState *phb,
     uint32_t interrupt_map_mask[] = {
         cpu_to_be32(b_ddddd(-1)|b_fff(0)), 0x0, 0x0, cpu_to_be32(-1)};
     uint32_t interrupt_map[PCI_SLOT_MAX * PCI_NUM_PINS][7];
+    uint32_t ddw_applicable[] = {
+        cpu_to_be32(RTAS_IBM_QUERY_PE_DMA_WINDOW),
+        cpu_to_be32(RTAS_IBM_CREATE_PE_DMA_WINDOW),
+        cpu_to_be32(RTAS_IBM_REMOVE_PE_DMA_WINDOW)
+    };
+    uint32_t ddw_extensions[] = {
+        cpu_to_be32(1),
+        cpu_to_be32(RTAS_IBM_RESET_PE_DMA_WINDOW)
+    };
     sPAPRTCETable *tcet;
     PCIBus *bus = PCI_HOST_BRIDGE(phb)->bus;
     sPAPRFDT s_fdt;
@@ -1798,6 +1829,14 @@ int spapr_populate_pci_dt(sPAPRPHBState *phb,
     _FDT(fdt_setprop_cell(fdt, bus_off, "ibm,pci-config-space-type", 0x1));
     _FDT(fdt_setprop_cell(fdt, bus_off, "ibm,pe-total-#msi", XICS_IRQS_SPAPR));
 
+    /* Dynamic DMA window */
+    if (phb->ddw_enabled) {
+        _FDT(fdt_setprop(fdt, bus_off, "ibm,ddw-applicable", &ddw_applicable,
+                         sizeof(ddw_applicable)));
+        _FDT(fdt_setprop(fdt, bus_off, "ibm,ddw-extensions",
+                         &ddw_extensions, sizeof(ddw_extensions)));
+    }
+
     /* Build the interrupt-map, this must matches what is done
      * in pci_spapr_map_irq
      */
@@ -1821,7 +1860,7 @@ int spapr_populate_pci_dt(sPAPRPHBState *phb,
     _FDT(fdt_setprop(fdt, bus_off, "interrupt-map", &interrupt_map,
                      sizeof(interrupt_map)));
 
-    tcet = spapr_tce_find_by_liobn(phb->dma_liobn);
+    tcet = spapr_tce_find_by_liobn(phb->dma_liobn[0]);
     if (!tcet) {
         return -1;
     }