[Qemu-devel] [PULL 0/8] ppc patch queue 2012-05-01

[Qemu-devel] [PULL 0/8] ppc patch queue 2012-05-01

[Alexander Graf]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1626 bytes --]

Hi Anthony,

This is my current patch queue for ppc. Please pull.

Alex


The following changes since commit 42fe1c245f0239ebcdc084740a1777ac3699d071:
  Stefan Weil (1):
        main-loop: Fix build for w32 and w64

are available in the git repository at:

  git://repo.or.cz/qemu/agraf.git ppc-for-upstream

Alexander Graf (2):
      PPC: Fix up e500 cache size setting
      linux-user: Fix undefined HOST_LONG_SIZE on PPC hosts

Bharat Bhushan (1):
      booke:Use MMU API for creating initial mapping for secondary cpus

David Gibson (2):
      pseries: Implement automatic PAPR VIO address allocation
      pseries: Use the same interrupt swizzling for host bridges as p2p bridges

François Revol (1):
      target-ppc: Some support for dumping TLB_EMB TLBs

Peter Portante (1):
      pseries: Fix use of global CPU state

Stefan Weil (1):
      ppce500_spin: Replace assert by hw_error (fixes compiler warning)

 hw/ppce500_spin.c           |    3 +-
 hw/spapr.c                  |    7 ++---
 hw/spapr_hcall.c            |    2 +-
 hw/spapr_llan.c             |    5 +--
 hw/spapr_pci.c              |   49 ++++++++++++++++++++++----------------
 hw/spapr_pci.h              |    5 +--
 hw/spapr_vio.c              |   54 +++++++++++++++++++++++++++----------------
 hw/spapr_vio.h              |   13 ++++-----
 hw/spapr_vscsi.c            |    5 +--
 hw/spapr_vty.c              |    5 +--
 target-ppc/helper.c         |   50 +++++++++++++++++++++++++++++++++++++++
 target-ppc/translate_init.c |   26 +++++++++++---------
 thunk.h                     |    2 +-
 13 files changed, 147 insertions(+), 79 deletions(-)

[Qemu-devel] [PULL 0/8] ppc patch queue 2012-05-01

[Alexander Graf]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1626 bytes --]

Hi Anthony,

This is my current patch queue for ppc. Please pull.

Alex


The following changes since commit 42fe1c245f0239ebcdc084740a1777ac3699d071:
  Stefan Weil (1):
        main-loop: Fix build for w32 and w64

are available in the git repository at:

  git://repo.or.cz/qemu/agraf.git ppc-for-upstream

Alexander Graf (2):
      PPC: Fix up e500 cache size setting
      linux-user: Fix undefined HOST_LONG_SIZE on PPC hosts

Bharat Bhushan (1):
      booke:Use MMU API for creating initial mapping for secondary cpus

David Gibson (2):
      pseries: Implement automatic PAPR VIO address allocation
      pseries: Use the same interrupt swizzling for host bridges as p2p bridges

François Revol (1):
      target-ppc: Some support for dumping TLB_EMB TLBs

Peter Portante (1):
      pseries: Fix use of global CPU state

Stefan Weil (1):
      ppce500_spin: Replace assert by hw_error (fixes compiler warning)

 hw/ppce500_spin.c           |    3 +-
 hw/spapr.c                  |    7 ++---
 hw/spapr_hcall.c            |    2 +-
 hw/spapr_llan.c             |    5 +--
 hw/spapr_pci.c              |   49 ++++++++++++++++++++++----------------
 hw/spapr_pci.h              |    5 +--
 hw/spapr_vio.c              |   54 +++++++++++++++++++++++++++----------------
 hw/spapr_vio.h              |   13 ++++-----
 hw/spapr_vscsi.c            |    5 +--
 hw/spapr_vty.c              |    5 +--
 target-ppc/helper.c         |   50 +++++++++++++++++++++++++++++++++++++++
 target-ppc/translate_init.c |   26 +++++++++++---------
 thunk.h                     |    2 +-
 13 files changed, 147 insertions(+), 79 deletions(-)

[Qemu-devel] [PATCH 1/8] booke:Use MMU API for creating initial mapping for secondary cpus

[Alexander Graf]

From: Bharat Bhushan <r65777@freescale.com>

Initial Mapping creation for secondary CPU in SMP was missing new MMU API.

Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/ppce500_spin.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/hw/ppce500_spin.c b/hw/ppce500_spin.c
index 960b7b0..95a2825 100644
--- a/hw/ppce500_spin.c
+++ b/hw/ppce500_spin.c
@@ -86,6 +86,7 @@ static void mmubooke_create_initial_mapping(CPUPPCState *env,
     tlb->mas2 = (va & TARGET_PAGE_MASK) | MAS2_M;
     tlb->mas7_3 = pa & TARGET_PAGE_MASK;
     tlb->mas7_3 |= MAS3_UR | MAS3_UW | MAS3_UX | MAS3_SR | MAS3_SW | MAS3_SX;
+    env->tlb_dirty = true;
 }
 
 static void spin_kick(void *data)
-- 
1.6.0.2

[Qemu-devel] [PATCH 2/8] PPC: Fix up e500 cache size setting

[Alexander Graf]

When initializing the e500 code, we need to expose its
cache line size for user and system mode, while the mmu
details are only interesting for system emulation.

Split the 2 switch statements apart, allowing us to #ifdef
out the mmu parts for user mode emulation while keeping all
cache information consistent.

Signed-off-by: Alexander Graf <agraf@suse.de>
---
 target-ppc/translate_init.c |   26 ++++++++++++++------------
 1 files changed, 14 insertions(+), 12 deletions(-)

diff --git a/target-ppc/translate_init.c b/target-ppc/translate_init.c
index ba4b84d..6f61175 100644
--- a/target-ppc/translate_init.c
+++ b/target-ppc/translate_init.c
@@ -4461,33 +4461,36 @@ static void init_proc_e500 (CPUPPCState *env, int version)
                  &spr_read_spefscr, &spr_write_spefscr,
                  &spr_read_spefscr, &spr_write_spefscr,
                  0x00000000);
+#if !defined(CONFIG_USER_ONLY)
     /* Memory management */
-#if defined(CONFIG_USER_ONLY)
-    env->dcache_line_size = 32;
-    env->icache_line_size = 32;
-#else /* !defined(CONFIG_USER_ONLY) */
     env->nb_pids = 3;
     env->nb_ways = 2;
     env->id_tlbs = 0;
     switch (version) {
     case fsl_e500v1:
-        /* e500v1 */
         tlbncfg[0] = gen_tlbncfg(2, 1, 1, 0, 256);
         tlbncfg[1] = gen_tlbncfg(16, 1, 9, TLBnCFG_AVAIL | TLBnCFG_IPROT, 16);
-        env->dcache_line_size = 32;
-        env->icache_line_size = 32;
         break;
     case fsl_e500v2:
-        /* e500v2 */
         tlbncfg[0] = gen_tlbncfg(4, 1, 1, 0, 512);
         tlbncfg[1] = gen_tlbncfg(16, 1, 12, TLBnCFG_AVAIL | TLBnCFG_IPROT, 16);
-        env->dcache_line_size = 32;
-        env->icache_line_size = 32;
         break;
     case fsl_e500mc:
-        /* e500mc */
         tlbncfg[0] = gen_tlbncfg(4, 1, 1, 0, 512);
         tlbncfg[1] = gen_tlbncfg(64, 1, 12, TLBnCFG_AVAIL | TLBnCFG_IPROT, 64);
+        break;
+    default:
+        cpu_abort(env, "Unknown CPU: " TARGET_FMT_lx "\n", env->spr[SPR_PVR]);
+    }
+#endif
+    /* Cache sizes */
+    switch (version) {
+    case fsl_e500v1:
+    case fsl_e500v2:
+        env->dcache_line_size = 32;
+        env->icache_line_size = 32;
+        break;
+    case fsl_e500mc:
         env->dcache_line_size = 64;
         env->icache_line_size = 64;
         l1cfg0 |= 0x1000000; /* 64 byte cache block size */
@@ -4495,7 +4498,6 @@ static void init_proc_e500 (CPUPPCState *env, int version)
     default:
         cpu_abort(env, "Unknown CPU: " TARGET_FMT_lx "\n", env->spr[SPR_PVR]);
     }
-#endif
     gen_spr_BookE206(env, 0x000000DF, tlbncfg);
     /* XXX : not implemented */
     spr_register(env, SPR_HID0, "HID0",
-- 
1.6.0.2

[Qemu-devel] [PATCH 3/8] linux-user: Fix undefined HOST_LONG_SIZE on PPC hosts

[Alexander Graf]

On my PPC host, HOST_LONG_SIZE is not defined even after
running configure. Use the normal C way of determining the
long size instead.

Signed-off-by: Alexander Graf <agraf@suse.de>
---
 thunk.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/thunk.h b/thunk.h
index 5be8f91..87025c3 100644
--- a/thunk.h
+++ b/thunk.h
@@ -113,7 +113,7 @@ static inline int thunk_type_size(const argtype *type_ptr, int is_host)
       defined(HOST_PARISC) || defined(HOST_SPARC64)
             return 4;
 #elif defined(HOST_PPC)
-            return HOST_LONG_SIZE;
+            return sizeof(void *);
 #else
             return 2;
 #endif
-- 
1.6.0.2

[Qemu-devel] [PATCH 4/8] pseries: Implement automatic PAPR VIO address allocation

[Alexander Graf]

From: David Gibson <david@gibson.dropbear.id.au>

PAPR virtual IO (VIO) devices require a unique, but otherwise arbitrary,
"address" used as a token to the hypercalls which manipulate them.

Currently the pseries machine code does an ok job of allocating these
addresses when the legacy -net nic / -serial and so forth options are used
but will fail to allocate them properly when using -device.

Specifically, you can use -device if all addresses are explicitly assigned.
Without explicit assignment, only one VIO device of each type (network,
console, SCSI) will be assigned properly, any further ones will attempt
to take the same address leading to a fatal error.

This patch fixes the situation by adding a proper address allocator to the
VIO "bus" code.  This is used both by -device and the legacy options and
default devices.  Addresses can still be explicitly assigned with -device
options if desired.

This patch changes the (guest visible) numbering of VIO devices, but since
their addresses are discovered using the device tree and already differ
from the numbering found on existing PowerVM systems, this does not break
compatibility.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/spapr.c       |    7 +++----
 hw/spapr_llan.c  |    5 ++---
 hw/spapr_vio.c   |   54 ++++++++++++++++++++++++++++++++++--------------------
 hw/spapr_vio.h   |   13 ++++++-------
 hw/spapr_vscsi.c |    5 ++---
 hw/spapr_vty.c   |    5 ++---
 6 files changed, 49 insertions(+), 40 deletions(-)

diff --git a/hw/spapr.c b/hw/spapr.c
index bfaf260..cca20f9 100644
--- a/hw/spapr.c
+++ b/hw/spapr.c
@@ -631,8 +631,7 @@ static void ppc_spapr_init(ram_addr_t ram_size,
 
     for (i = 0; i < MAX_SERIAL_PORTS; i++) {
         if (serial_hds[i]) {
-            spapr_vty_create(spapr->vio_bus, SPAPR_VTY_BASE_ADDRESS + i,
-                             serial_hds[i]);
+            spapr_vty_create(spapr->vio_bus, serial_hds[i]);
         }
     }
 
@@ -650,14 +649,14 @@ static void ppc_spapr_init(ram_addr_t ram_size,
         }
 
         if (strcmp(nd->model, "ibmveth") == 0) {
-            spapr_vlan_create(spapr->vio_bus, 0x1000 + i, nd);
+            spapr_vlan_create(spapr->vio_bus, nd);
         } else {
             pci_nic_init_nofail(&nd_table[i], nd->model, NULL);
         }
     }
 
     for (i = 0; i <= drive_get_max_bus(IF_SCSI); i++) {
-        spapr_vscsi_create(spapr->vio_bus, 0x2000 + i);
+        spapr_vscsi_create(spapr->vio_bus);
     }
 
     if (rma_size < (MIN_RMA_SLOF << 20)) {
diff --git a/hw/spapr_llan.c b/hw/spapr_llan.c
index e18d2eb..8313043 100644
--- a/hw/spapr_llan.c
+++ b/hw/spapr_llan.c
@@ -204,12 +204,11 @@ static int spapr_vlan_init(VIOsPAPRDevice *sdev)
     return 0;
 }
 
-void spapr_vlan_create(VIOsPAPRBus *bus, uint32_t reg, NICInfo *nd)
+void spapr_vlan_create(VIOsPAPRBus *bus, NICInfo *nd)
 {
     DeviceState *dev;
 
     dev = qdev_create(&bus->bus, "spapr-vlan");
-    qdev_prop_set_uint32(dev, "reg", reg);
 
     qdev_set_nic_properties(dev, nd);
 
@@ -480,7 +479,7 @@ static target_ulong h_multicast_ctrl(CPUPPCState *env, sPAPREnvironment *spapr,
 }
 
 static Property spapr_vlan_properties[] = {
-    DEFINE_SPAPR_PROPERTIES(VIOsPAPRVLANDevice, sdev, 0x1000, 0x10000000),
+    DEFINE_SPAPR_PROPERTIES(VIOsPAPRVLANDevice, sdev, 0x10000000),
     DEFINE_NIC_PROPERTIES(VIOsPAPRVLANDevice, nicconf),
     DEFINE_PROP_END_OF_LIST(),
 };
diff --git a/hw/spapr_vio.c b/hw/spapr_vio.c
index fccf48b..315ab80 100644
--- a/hw/spapr_vio.c
+++ b/hw/spapr_vio.c
@@ -620,28 +620,22 @@ static void rtas_quiesce(sPAPREnvironment *spapr, uint32_t token,
     rtas_st(rets, 0, 0);
 }
 
-static int spapr_vio_check_reg(VIOsPAPRDevice *sdev)
+static VIOsPAPRDevice *reg_conflict(VIOsPAPRDevice *dev)
 {
-    VIOsPAPRDevice *other_sdev;
+    VIOsPAPRBus *bus = DO_UPCAST(VIOsPAPRBus, bus, dev->qdev.parent_bus);
     DeviceState *qdev;
-    VIOsPAPRBus *sbus;
-
-    sbus = DO_UPCAST(VIOsPAPRBus, bus, sdev->qdev.parent_bus);
+    VIOsPAPRDevice *other;
 
     /*
-     * Check two device aren't given clashing addresses by the user (or some
-     * other mechanism). We have to open code this because we have to check
-     * for matches with devices other than us.
+     * Check for a device other than the given one which is already
+     * using the requested address. We have to open code this because
+     * the given dev might already be in the list.
      */
-    QTAILQ_FOREACH(qdev, &sbus->bus.children, sibling) {
-        other_sdev = DO_UPCAST(VIOsPAPRDevice, qdev, qdev);
+    QTAILQ_FOREACH(qdev, &bus->bus.children, sibling) {
+        other = DO_UPCAST(VIOsPAPRDevice, qdev, qdev);
 
-        if (other_sdev != sdev && other_sdev->reg == sdev->reg) {
-            fprintf(stderr, "vio: %s and %s devices conflict at address %#x\n",
-                    object_get_typename(OBJECT(sdev)),
-                    object_get_typename(OBJECT(qdev)),
-                    sdev->reg);
-            return -EEXIST;
+        if (other != dev && other->reg == dev->reg) {
+            return other;
         }
     }
 
@@ -667,11 +661,30 @@ static int spapr_vio_busdev_init(DeviceState *qdev)
     VIOsPAPRDevice *dev = (VIOsPAPRDevice *)qdev;
     VIOsPAPRDeviceClass *pc = VIO_SPAPR_DEVICE_GET_CLASS(dev);
     char *id;
-    int ret;
 
-    ret = spapr_vio_check_reg(dev);
-    if (ret) {
-        return ret;
+    if (dev->reg != -1) {
+        /*
+         * Explicitly assigned address, just verify that no-one else
+         * is using it.  other mechanism). We have to open code this
+         * rather than using spapr_vio_find_by_reg() because sdev
+         * itself is already in the list.
+         */
+        VIOsPAPRDevice *other = reg_conflict(dev);
+
+        if (other) {
+            fprintf(stderr, "vio: %s and %s devices conflict at address %#x\n",
+                    object_get_typename(OBJECT(qdev)),
+                    object_get_typename(OBJECT(&other->qdev)),
+                    dev->reg);
+            return -1;
+        }
+    } else {
+        /* Need to assign an address */
+        VIOsPAPRBus *bus = DO_UPCAST(VIOsPAPRBus, bus, dev->qdev.parent_bus);
+
+        do {
+            dev->reg = bus->next_reg++;
+        } while (reg_conflict(dev));
     }
 
     /* Don't overwrite ids assigned on the command line */
@@ -731,6 +744,7 @@ VIOsPAPRBus *spapr_vio_bus_init(void)
 
     qbus = qbus_create(&spapr_vio_bus_info, dev, "spapr-vio");
     bus = DO_UPCAST(VIOsPAPRBus, bus, qbus);
+    bus->next_reg = 0x1000;
 
     /* hcall-vio */
     spapr_register_hypercall(H_VIO_SIGNAL, h_vio_signal);
diff --git a/hw/spapr_vio.h b/hw/spapr_vio.h
index 10ab359..87816e4 100644
--- a/hw/spapr_vio.h
+++ b/hw/spapr_vio.h
@@ -32,8 +32,6 @@ enum VIOsPAPR_TCEAccess {
     SPAPR_TCE_RW = 3,
 };
 
-#define SPAPR_VTY_BASE_ADDRESS     0x30000000
-
 #define TYPE_VIO_SPAPR_DEVICE "vio-spapr-device"
 #define VIO_SPAPR_DEVICE(obj) \
      OBJECT_CHECK(VIOsPAPRDevice, (obj), TYPE_VIO_SPAPR_DEVICE)
@@ -82,13 +80,14 @@ struct VIOsPAPRDevice {
     VIOsPAPR_CRQ crq;
 };
 
-#define DEFINE_SPAPR_PROPERTIES(type, field, default_reg, default_dma_window) \
-        DEFINE_PROP_UINT32("reg", type, field.reg, default_reg), \
+#define DEFINE_SPAPR_PROPERTIES(type, field, default_dma_window)       \
+        DEFINE_PROP_UINT32("reg", type, field.reg, -1),                \
         DEFINE_PROP_UINT32("dma-window", type, field.rtce_window_size, \
                            default_dma_window)
 
 struct VIOsPAPRBus {
     BusState bus;
+    uint32_t next_reg;
     int (*init)(VIOsPAPRDevice *dev);
     int (*devnode)(VIOsPAPRDevice *dev, void *fdt, int node_off);
 };
@@ -119,9 +118,9 @@ int spapr_vio_send_crq(VIOsPAPRDevice *dev, uint8_t *crq);
 
 VIOsPAPRDevice *vty_lookup(sPAPREnvironment *spapr, target_ulong reg);
 void vty_putchars(VIOsPAPRDevice *sdev, uint8_t *buf, int len);
-void spapr_vty_create(VIOsPAPRBus *bus, uint32_t reg, CharDriverState *chardev);
-void spapr_vlan_create(VIOsPAPRBus *bus, uint32_t reg, NICInfo *nd);
-void spapr_vscsi_create(VIOsPAPRBus *bus, uint32_t reg);
+void spapr_vty_create(VIOsPAPRBus *bus, CharDriverState *chardev);
+void spapr_vlan_create(VIOsPAPRBus *bus, NICInfo *nd);
+void spapr_vscsi_create(VIOsPAPRBus *bus);
 
 VIOsPAPRDevice *spapr_vty_get_default(VIOsPAPRBus *bus);
 
diff --git a/hw/spapr_vscsi.c b/hw/spapr_vscsi.c
index 538e0b7..037867a 100644
--- a/hw/spapr_vscsi.c
+++ b/hw/spapr_vscsi.c
@@ -918,12 +918,11 @@ static int spapr_vscsi_init(VIOsPAPRDevice *dev)
     return 0;
 }
 
-void spapr_vscsi_create(VIOsPAPRBus *bus, uint32_t reg)
+void spapr_vscsi_create(VIOsPAPRBus *bus)
 {
     DeviceState *dev;
 
     dev = qdev_create(&bus->bus, "spapr-vscsi");
-    qdev_prop_set_uint32(dev, "reg", reg);
 
     qdev_init_nofail(dev);
 }
@@ -946,7 +945,7 @@ static int spapr_vscsi_devnode(VIOsPAPRDevice *dev, void *fdt, int node_off)
 }
 
 static Property spapr_vscsi_properties[] = {
-    DEFINE_SPAPR_PROPERTIES(VSCSIState, vdev, 0x2000, 0x10000000),
+    DEFINE_SPAPR_PROPERTIES(VSCSIState, vdev, 0x10000000),
     DEFINE_PROP_END_OF_LIST(),
 };
 
diff --git a/hw/spapr_vty.c b/hw/spapr_vty.c
index a30c040..c9674f3 100644
--- a/hw/spapr_vty.c
+++ b/hw/spapr_vty.c
@@ -123,18 +123,17 @@ static target_ulong h_get_term_char(CPUPPCState *env, sPAPREnvironment *spapr,
     return H_SUCCESS;
 }
 
-void spapr_vty_create(VIOsPAPRBus *bus, uint32_t reg, CharDriverState *chardev)
+void spapr_vty_create(VIOsPAPRBus *bus, CharDriverState *chardev)
 {
     DeviceState *dev;
 
     dev = qdev_create(&bus->bus, "spapr-vty");
-    qdev_prop_set_uint32(dev, "reg", reg);
     qdev_prop_set_chr(dev, "chardev", chardev);
     qdev_init_nofail(dev);
 }
 
 static Property spapr_vty_properties[] = {
-    DEFINE_SPAPR_PROPERTIES(VIOsPAPRVTYDevice, sdev, SPAPR_VTY_BASE_ADDRESS, 0),
+    DEFINE_SPAPR_PROPERTIES(VIOsPAPRVTYDevice, sdev, 0),
     DEFINE_PROP_CHR("chardev", VIOsPAPRVTYDevice, chardev),
     DEFINE_PROP_END_OF_LIST(),
 };
-- 
1.6.0.2

[Qemu-devel] [PATCH 5/8] pseries: Use the same interrupt swizzling for host bridges as p2p bridges

[Alexander Graf]

From: David Gibson <david@gibson.dropbear.id.au>

Currently the pseries PCI code uses a somewhat strange scheme of PCI irq
allocation - one per slot up to a maximum that's greater than the usual 4.
This scheme more or less worked, because we were able to tell the guest the
irq mapping in the device tree, however it's a bit odd and may break
assumptions in the future.  Worse, the array used to construct the dev
tree interrupt map was mis-sized, we got away with it only because it
happened that our SPAPR_PCI_NUM_LSI value was greater than 7.

This patch changes the pseries PCI code to use the same interrupt swizzling
scheme as is standardized for PCI to PCI bridges.  This makes for better
consistency, deals better with any devices which use multiple interrupt
pins and will make life easier in the future when we add passthrough of
what may be either a host bridge or a PCI to PCI bridge.  This won't break
existing guests, because they don't assume a particular mapping scheme for
host bridges, but just follow what we tell them in the device tree (also
updated to match, of course).  This patch also fixes the allocation of the
irq map.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/spapr_pci.c |   49 ++++++++++++++++++++++++++++---------------------
 hw/spapr_pci.h |    5 ++---
 2 files changed, 30 insertions(+), 24 deletions(-)

diff --git a/hw/spapr_pci.c b/hw/spapr_pci.c
index a564c00..25b400a 100644
--- a/hw/spapr_pci.c
+++ b/hw/spapr_pci.c
@@ -198,16 +198,20 @@ static void rtas_write_pci_config(sPAPREnvironment *spapr,
     finish_write_pci_config(spapr, 0, addr, size, val, rets);
 }
 
+static int pci_spapr_swizzle(int slot, int pin)
+{
+    return (slot + pin) % PCI_NUM_PINS;
+}
+
 static int pci_spapr_map_irq(PCIDevice *pci_dev, int irq_num)
 {
     /*
      * Here we need to convert pci_dev + irq_num to some unique value
-     * which is less than number of IRQs on the specific bus (now it
-     * is 16).  At the moment irq_num == device_id (number of the
-     * slot?)
-     * FIXME: we should swizzle in fn and irq_num
+     * which is less than number of IRQs on the specific bus (4).  We
+     * use standard PCI swizzling, that is (slot number + pin number)
+     * % 4.
      */
-    return (pci_dev->devfn >> 3) % SPAPR_PCI_NUM_LSI;
+    return pci_spapr_swizzle(PCI_SLOT(pci_dev->devfn), irq_num);
 }
 
 static void pci_spapr_set_irq(void *opaque, int irq_num, int level)
@@ -304,13 +308,13 @@ static int spapr_phb_init(SysBusDevice *s)
                            phb->busname ? phb->busname : phb->dtbusname,
                            pci_spapr_set_irq, pci_spapr_map_irq, phb,
                            &phb->memspace, &phb->iospace,
-                           PCI_DEVFN(0, 0), SPAPR_PCI_NUM_LSI);
+                           PCI_DEVFN(0, 0), PCI_NUM_PINS);
     phb->host_state.bus = bus;
 
     QLIST_INSERT_HEAD(&spapr->phbs, phb, list);
 
     /* Initialize the LSI table */
-    for (i = 0; i < SPAPR_PCI_NUM_LSI; i++) {
+    for (i = 0; i < PCI_NUM_PINS; i++) {
         qemu_irq qirq;
         uint32_t num;
 
@@ -392,8 +396,7 @@ int spapr_populate_pci_devices(sPAPRPHBState *phb,
                                uint32_t xics_phandle,
                                void *fdt)
 {
-    PCIBus *bus = phb->host_state.bus;
-    int bus_off, i;
+    int bus_off, i, j;
     char nodename[256];
     uint32_t bus_range[] = { cpu_to_be32(0), cpu_to_be32(0xff) };
     struct {
@@ -415,8 +418,8 @@ int spapr_populate_pci_devices(sPAPRPHBState *phb,
     };
     uint64_t bus_reg[] = { cpu_to_be64(phb->buid), 0 };
     uint32_t interrupt_map_mask[] = {
-        cpu_to_be32(b_ddddd(-1)|b_fff(0)), 0x0, 0x0, 0x0};
-    uint32_t interrupt_map[bus->nirq][7];
+        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];
 
     /* Start populating the FDT */
     sprintf(nodename, "pci@%" PRIx64, phb->buid);
@@ -450,19 +453,23 @@ int spapr_populate_pci_devices(sPAPRPHBState *phb,
      */
     _FDT(fdt_setprop(fdt, bus_off, "interrupt-map-mask",
                      &interrupt_map_mask, sizeof(interrupt_map_mask)));
-    for (i = 0; i < 7; i++) {
-        uint32_t *irqmap = interrupt_map[i];
-        irqmap[0] = cpu_to_be32(b_ddddd(i)|b_fff(0));
-        irqmap[1] = 0;
-        irqmap[2] = 0;
-        irqmap[3] = 0;
-        irqmap[4] = cpu_to_be32(xics_phandle);
-        irqmap[5] = cpu_to_be32(phb->lsi_table[i % SPAPR_PCI_NUM_LSI].dt_irq);
-        irqmap[6] = cpu_to_be32(0x8);
+    for (i = 0; i < PCI_SLOT_MAX; i++) {
+        for (j = 0; j < PCI_NUM_PINS; j++) {
+            uint32_t *irqmap = interrupt_map[i*PCI_NUM_PINS + j];
+            int lsi_num = pci_spapr_swizzle(i, j);
+
+            irqmap[0] = cpu_to_be32(b_ddddd(i)|b_fff(0));
+            irqmap[1] = 0;
+            irqmap[2] = 0;
+            irqmap[3] = cpu_to_be32(j+1);
+            irqmap[4] = cpu_to_be32(xics_phandle);
+            irqmap[5] = cpu_to_be32(phb->lsi_table[lsi_num].dt_irq);
+            irqmap[6] = cpu_to_be32(0x8);
+        }
     }
     /* Write interrupt map */
     _FDT(fdt_setprop(fdt, bus_off, "interrupt-map", &interrupt_map,
-                     7 * sizeof(interrupt_map[0])));
+                     sizeof(interrupt_map)));
 
     return 0;
 }
diff --git a/hw/spapr_pci.h b/hw/spapr_pci.h
index 039f85b..f54c2e8 100644
--- a/hw/spapr_pci.h
+++ b/hw/spapr_pci.h
@@ -23,11 +23,10 @@
 #if !defined(__HW_SPAPR_PCI_H__)
 #define __HW_SPAPR_PCI_H__
 
+#include "hw/pci.h"
 #include "hw/pci_host.h"
 #include "hw/xics.h"
 
-#define SPAPR_PCI_NUM_LSI   16
-
 typedef struct sPAPRPHBState {
     SysBusDevice busdev;
     PCIHostState host_state;
@@ -43,7 +42,7 @@ typedef struct sPAPRPHBState {
     struct {
         uint32_t dt_irq;
         qemu_irq qirq;
-    } lsi_table[SPAPR_PCI_NUM_LSI];
+    } lsi_table[PCI_NUM_PINS];
 
     QLIST_ENTRY(sPAPRPHBState) list;
 } sPAPRPHBState;
-- 
1.6.0.2

[Qemu-devel] [PATCH 6/8] pseries: Fix use of global CPU state

[Alexander Graf]

From: Peter Portante <peter.portante@redhat.com>

Commit ed120055c7f9b26b5707d3ceabbe5a3f06aaf937 (Implement PAPR VPA
functions for pSeries shared processor partitions) introduced the
deregister_dtl() function and typo "emv" as name of its argument.
This went unnoticed because the code in that function can access the
global variable "env" so that no build failure resulted.

Fix the argument to read "env". Resolves LP#986241.

Signed-off-by: Peter Portante <peter.portante@redhat.com>
Acked-by: Andreas Färber <afaerber@suse.de>
[agraf: fixed typo in commit message]
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/spapr_hcall.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/hw/spapr_hcall.c b/hw/spapr_hcall.c
index 634763e..94bb504 100644
--- a/hw/spapr_hcall.c
+++ b/hw/spapr_hcall.c
@@ -482,7 +482,7 @@ static target_ulong register_dtl(CPUPPCState *env, target_ulong addr)
     return H_SUCCESS;
 }
 
-static target_ulong deregister_dtl(CPUPPCState *emv, target_ulong addr)
+static target_ulong deregister_dtl(CPUPPCState *env, target_ulong addr)
 {
     env->dispatch_trace_log = 0;
     env->dtl_size = 0;
-- 
1.6.0.2

[Qemu-devel] [PATCH 7/8] ppce500_spin: Replace assert by hw_error (fixes compiler warning)

[Alexander Graf]

From: Stefan Weil <sw@weilnetz.de>

The default case in function spin_read should never be reached,
therefore the old code used assert(0) to abort QEMU.

This does not work when QEMU is compiled with macro NDEBUG defined.
In this case (and also when the compiler does not know that assert
never returns), there is a compiler warning because of the missing
return value.

Using hw_error allows an improved error message and aborts always.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
[agraf: use __func__]
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 hw/ppce500_spin.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/hw/ppce500_spin.c b/hw/ppce500_spin.c
index 95a2825..fddf219 100644
--- a/hw/ppce500_spin.c
+++ b/hw/ppce500_spin.c
@@ -179,7 +179,7 @@ static uint64_t spin_read(void *opaque, target_phys_addr_t addr, unsigned len)
     case 4:
         return ldl_p(spin_p);
     default:
-        assert(0);
+        hw_error("ppce500: unexpected %s with len = %u", __func__, len);
     }
 }
 
-- 
1.6.0.2

[Qemu-devel] [PATCH 8/8] target-ppc: Some support for dumping TLB_EMB TLBs

[Alexander Graf]

From: François Revol <revol@free.fr>

Add mmubooke_dump_mmu().

TODO: Add printing of individual flags.

Signed-off-by: François Revol <revol@free.fr>
[agraf: fix coding style]
Signed-off-by: Alexander Graf <agraf@suse.de>
---
 target-ppc/helper.c |   50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 50 insertions(+), 0 deletions(-)

diff --git a/target-ppc/helper.c b/target-ppc/helper.c
index c610ce3..e97e496 100644
--- a/target-ppc/helper.c
+++ b/target-ppc/helper.c
@@ -1466,6 +1466,53 @@ static const char *book3e_tsize_to_str[32] = {
     "1T", "2T"
 };
 
+static void mmubooke_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
+                                 CPUPPCState *env)
+{
+    ppcemb_tlb_t *entry;
+    int i;
+
+    if (kvm_enabled() && !env->kvm_sw_tlb) {
+        cpu_fprintf(f, "Cannot access KVM TLB\n");
+        return;
+    }
+
+    cpu_fprintf(f, "\nTLB:\n");
+    cpu_fprintf(f, "Effective          Physical           Size PID   Prot     "
+                "Attr\n");
+
+    entry = &env->tlb.tlbe[0];
+    for (i = 0; i < env->nb_tlb; i++, entry++) {
+        target_phys_addr_t ea, pa;
+        target_ulong mask;
+        uint64_t size = (uint64_t)entry->size;
+        char size_buf[20];
+
+        /* Check valid flag */
+        if (!(entry->prot & PAGE_VALID)) {
+            continue;
+        }
+
+        mask = ~(entry->size - 1);
+        ea = entry->EPN & mask;
+        pa = entry->RPN & mask;
+#if (TARGET_PHYS_ADDR_BITS >= 36)
+        /* Extend the physical address to 36 bits */
+        pa |= (target_phys_addr_t)(entry->RPN & 0xF) << 32;
+#endif
+        size /= 1024;
+        if (size >= 1024) {
+            snprintf(size_buf, sizeof(size_buf), "%3" PRId64 "M", size / 1024);
+        } else {
+            snprintf(size_buf, sizeof(size_buf), "%3" PRId64 "k", size);
+        }
+        cpu_fprintf(f, "0x%016" PRIx64 " 0x%016" PRIx64 " %s %-5u %08x %08x\n",
+                    (uint64_t)ea, (uint64_t)pa, size_buf, (uint32_t)entry->PID,
+                    entry->prot, entry->attr);
+    }
+
+}
+
 static void mmubooke206_dump_one_tlb(FILE *f, fprintf_function cpu_fprintf,
                                      CPUPPCState *env, int tlbn, int offset,
                                      int tlbsize)
@@ -1561,6 +1608,9 @@ static void mmubooks_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
 void dump_mmu(FILE *f, fprintf_function cpu_fprintf, CPUPPCState *env)
 {
     switch (env->mmu_model) {
+    case POWERPC_MMU_BOOKE:
+        mmubooke_dump_mmu(f, cpu_fprintf, env);
+        break;
     case POWERPC_MMU_BOOKE206:
         mmubooke206_dump_mmu(f, cpu_fprintf, env);
         break;
-- 
1.6.0.2

Re: [Qemu-devel] [PATCH 3/8] linux-user: Fix undefined HOST_LONG_SIZE on PPC hosts

[Stefan Weil]

Am 01.05.2012 21:50, schrieb Alexander Graf:
> On my PPC host, HOST_LONG_SIZE is not defined even after
> running configure. Use the normal C way of determining the
> long size instead.
>
> Signed-off-by: Alexander Graf <agraf@suse.de>
> ---
> thunk.h | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/thunk.h b/thunk.h
> index 5be8f91..87025c3 100644
> --- a/thunk.h
> +++ b/thunk.h
> @@ -113,7 +113,7 @@ static inline int thunk_type_size(const argtype 
> *type_ptr, int is_host)
> defined(HOST_PARISC) || defined(HOST_SPARC64)
> return 4;
> #elif defined(HOST_PPC)
> - return HOST_LONG_SIZE;
> + return sizeof(void *);
> #else
> return 2;
> #endif


Malc reverted his change and applied Alexander's previous patch,
so this one is no longer needed (and would conflict with
latest git master).

Regards,

Stefan W.

Re: [Qemu-devel] [PATCH 3/8] linux-user: Fix undefined HOST_LONG_SIZE on PPC hosts

[Andreas Färber]

Am 01.05.2012 22:51, schrieb Stefan Weil:
> Am 01.05.2012 21:50, schrieb Alexander Graf:
>> On my PPC host, HOST_LONG_SIZE is not defined even after
>> running configure. Use the normal C way of determining the
>> long size instead.
>>
>> Signed-off-by: Alexander Graf <agraf@suse.de>
>> ---
>> thunk.h | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/thunk.h b/thunk.h
>> index 5be8f91..87025c3 100644
>> --- a/thunk.h
>> +++ b/thunk.h
>> @@ -113,7 +113,7 @@ static inline int thunk_type_size(const argtype
>> *type_ptr, int is_host)
>> defined(HOST_PARISC) || defined(HOST_SPARC64)
>> return 4;
>> #elif defined(HOST_PPC)
>> - return HOST_LONG_SIZE;
>> + return sizeof(void *);
>> #else
>> return 2;
>> #endif
> 
> 
> Malc reverted his change and applied Alexander's previous patch,
> so this one is no longer needed (and would conflict with
> latest git master).

Since malc does seem to have committed Alex' exact patch, I would hope
that a PULL applies fine now, just like a rebase of a local branch onto
committed patch versions from master...

Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg