[Qemu-devel] [0/8] Assorted pseries updates

[Qemu-devel] [0/8] Assorted pseries updates

[David Gibson]

Now that qemu 1.0 is out there, here are a number of cleanups and
minor bugfixes for the pseries machine to queue up for post 1.0.  Many
of these fix problems with non-default device configurations exposed
by the w.i.p. port of libvirt for powerkvm.

[Qemu-devel] [PATCH 1/8] pseries: Remove hcalls callback

[David Gibson]

For forgotten historical reasons, PAPR hypercalls for specific virtual IO
devices (oh which there are quite a number) are registered via a callback
in the VIOsPAPRDeviceInfo structure.

This is kind of ugly, so this patch instead registers hypercalls from
device_init() functions for each device type.  This works just as well,
and is cleaner.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr_llan.c |   17 ++++++-----------
 hw/spapr_vio.c  |   13 -------------
 hw/spapr_vio.h  |    1 -
 hw/spapr_vty.c  |    9 ++-------
 4 files changed, 8 insertions(+), 32 deletions(-)

diff --git a/hw/spapr_llan.c b/hw/spapr_llan.c
index abe1297..45674c4 100644
--- a/hw/spapr_llan.c
+++ b/hw/spapr_llan.c
@@ -474,16 +474,6 @@ static target_ulong h_multicast_ctrl(CPUState *env, sPAPREnvironment *spapr,
     return H_SUCCESS;
 }
 
-static void vlan_hcalls(VIOsPAPRBus *bus)
-{
-    spapr_register_hypercall(H_REGISTER_LOGICAL_LAN, h_register_logical_lan);
-    spapr_register_hypercall(H_FREE_LOGICAL_LAN, h_free_logical_lan);
-    spapr_register_hypercall(H_SEND_LOGICAL_LAN, h_send_logical_lan);
-    spapr_register_hypercall(H_ADD_LOGICAL_LAN_BUFFER,
-                             h_add_logical_lan_buffer);
-    spapr_register_hypercall(H_MULTICAST_CTRL, h_multicast_ctrl);
-}
-
 static VIOsPAPRDeviceInfo spapr_vlan = {
     .init = spapr_vlan_init,
     .devnode = spapr_vlan_devnode,
@@ -491,7 +481,6 @@ static VIOsPAPRDeviceInfo spapr_vlan = {
     .dt_type = "network",
     .dt_compatible = "IBM,l-lan",
     .signal_mask = 0x1,
-    .hcalls = vlan_hcalls,
     .qdev.name = "spapr-vlan",
     .qdev.size = sizeof(VIOsPAPRVLANDevice),
     .qdev.props = (Property[]) {
@@ -504,5 +493,11 @@ static VIOsPAPRDeviceInfo spapr_vlan = {
 static void spapr_vlan_register(void)
 {
     spapr_vio_bus_register_withprop(&spapr_vlan);
+    spapr_register_hypercall(H_REGISTER_LOGICAL_LAN, h_register_logical_lan);
+    spapr_register_hypercall(H_FREE_LOGICAL_LAN, h_free_logical_lan);
+    spapr_register_hypercall(H_SEND_LOGICAL_LAN, h_send_logical_lan);
+    spapr_register_hypercall(H_ADD_LOGICAL_LAN_BUFFER,
+                             h_add_logical_lan_buffer);
+    spapr_register_hypercall(H_MULTICAST_CTRL, h_multicast_ctrl);
 }
 device_init(spapr_vlan_register);
diff --git a/hw/spapr_vio.c b/hw/spapr_vio.c
index 2dcc036..5a35541 100644
--- a/hw/spapr_vio.c
+++ b/hw/spapr_vio.c
@@ -684,7 +684,6 @@ VIOsPAPRBus *spapr_vio_bus_init(void)
     VIOsPAPRBus *bus;
     BusState *qbus;
     DeviceState *dev;
-    DeviceInfo *qinfo;
 
     /* Create bridge device */
     dev = qdev_create(NULL, "spapr-vio-bridge");
@@ -711,18 +710,6 @@ VIOsPAPRBus *spapr_vio_bus_init(void)
     spapr_rtas_register("ibm,set-tce-bypass", rtas_set_tce_bypass);
     spapr_rtas_register("quiesce", rtas_quiesce);
 
-    for (qinfo = device_info_list; qinfo; qinfo = qinfo->next) {
-        VIOsPAPRDeviceInfo *info = (VIOsPAPRDeviceInfo *)qinfo;
-
-        if (qinfo->bus_info != &spapr_vio_bus_info) {
-            continue;
-        }
-
-        if (info->hcalls) {
-            info->hcalls(bus);
-        }
-    }
-
     return bus;
 }
 
diff --git a/hw/spapr_vio.h b/hw/spapr_vio.h
index a325a5f..2430d45 100644
--- a/hw/spapr_vio.h
+++ b/hw/spapr_vio.h
@@ -75,7 +75,6 @@ typedef struct {
     const char *dt_name, *dt_type, *dt_compatible;
     target_ulong signal_mask;
     int (*init)(VIOsPAPRDevice *dev);
-    void (*hcalls)(VIOsPAPRBus *bus);
     int (*devnode)(VIOsPAPRDevice *dev, void *fdt, int node_off);
 } VIOsPAPRDeviceInfo;
 
diff --git a/hw/spapr_vty.c b/hw/spapr_vty.c
index f23cc36..e217314 100644
--- a/hw/spapr_vty.c
+++ b/hw/spapr_vty.c
@@ -135,18 +135,11 @@ void spapr_vty_create(VIOsPAPRBus *bus, uint32_t reg, CharDriverState *chardev)
     qdev_init_nofail(dev);
 }
 
-static void vty_hcalls(VIOsPAPRBus *bus)
-{
-    spapr_register_hypercall(H_PUT_TERM_CHAR, h_put_term_char);
-    spapr_register_hypercall(H_GET_TERM_CHAR, h_get_term_char);
-}
-
 static VIOsPAPRDeviceInfo spapr_vty = {
     .init = spapr_vty_init,
     .dt_name = "vty",
     .dt_type = "serial",
     .dt_compatible = "hvterm1",
-    .hcalls = vty_hcalls,
     .qdev.name = "spapr-vty",
     .qdev.size = sizeof(VIOsPAPRVTYDevice),
     .qdev.props = (Property[]) {
@@ -182,5 +175,7 @@ static VIOsPAPRDevice *vty_lookup(sPAPREnvironment *spapr, target_ulong reg)
 static void spapr_vty_register(void)
 {
     spapr_vio_bus_register_withprop(&spapr_vty);
+    spapr_register_hypercall(H_PUT_TERM_CHAR, h_put_term_char);
+    spapr_register_hypercall(H_GET_TERM_CHAR, h_get_term_char);
 }
 device_init(spapr_vty_register);
-- 
1.7.7.3

[Qemu-devel] [PATCH 2/8] monitor: add ability to dump SLB entries

[David Gibson]

From: Nishanth Aravamudan <nacc@us.ibm.com>

When run with a PPC Book3S (server) CPU Currently 'info tlb' in the
qemu monitor reports "dump_mmu: unimplemented".  However, during
bringup work, it can be quite handy to have the SLB entries, which are
available in the CPUPPCState.  This patch adds an implementation of
info tlb for book3s, which dumps the SLB.

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 target-ppc/helper.c |   32 +++++++++++++++++++++++++++-----
 1 files changed, 27 insertions(+), 5 deletions(-)

diff --git a/target-ppc/helper.c b/target-ppc/helper.c
index 137a494..29c7050 100644
--- a/target-ppc/helper.c
+++ b/target-ppc/helper.c
@@ -1545,14 +1545,36 @@ static void mmubooke206_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
     }
 }
 
+static void mmubooks_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
+                              CPUState *env)
+{
+    int i;
+    uint64_t slbe, slbv;
+
+    cpu_synchronize_state(env);
+
+    cpu_fprintf(f, "SLB\tESID\t\t\tVSID\n");
+    for (i = 0; i < env->slb_nr; i++) {
+        slbe = env->slb[i].esid;
+        slbv = env->slb[i].vsid;
+        if (slbe == 0 && slbv == 0) {
+            continue;
+        }
+        cpu_fprintf(f, "%d\t0x%016" PRIx64 "\t0x%016" PRIx64 "\n",
+                    i, slbe, slbv);
+    }
+}
+
 void dump_mmu(FILE *f, fprintf_function cpu_fprintf, CPUState *env)
 {
-    switch (env->mmu_model) {
-    case POWERPC_MMU_BOOKE206:
+    if (env->mmu_model == POWERPC_MMU_BOOKE206) {
         mmubooke206_dump_mmu(f, cpu_fprintf, env);
-        break;
-    default:
-        cpu_fprintf(f, "%s: unimplemented\n", __func__);
+    } else {
+        if ((env->mmu_model & POWERPC_MMU_64B) != 0) {
+            mmubooks_dump_mmu(f, cpu_fprintf, env);
+        } else {
+            cpu_fprintf(f, "%s: unimplemented\n", __func__);
+        }
     }
 }
 
-- 
1.7.7.3

Re: [Qemu-devel] [PATCH 2/8] monitor: add ability to dump SLB entries

[Alexander Graf]

On 13.12.2011, at 05:24, David Gibson wrote:

> From: Nishanth Aravamudan <nacc@us.ibm.com>
> 
> When run with a PPC Book3S (server) CPU Currently 'info tlb' in the
> qemu monitor reports "dump_mmu: unimplemented".  However, during
> bringup work, it can be quite handy to have the SLB entries, which are
> available in the CPUPPCState.  This patch adds an implementation of
> info tlb for book3s, which dumps the SLB.
> 
> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
> target-ppc/helper.c |   32 +++++++++++++++++++++++++++-----
> 1 files changed, 27 insertions(+), 5 deletions(-)
> 
> diff --git a/target-ppc/helper.c b/target-ppc/helper.c
> index 137a494..29c7050 100644
> --- a/target-ppc/helper.c
> +++ b/target-ppc/helper.c
> @@ -1545,14 +1545,36 @@ static void mmubooke206_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
>     }
> }
> 
> +static void mmubooks_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
> +                              CPUState *env)
> +{
> +    int i;
> +    uint64_t slbe, slbv;
> +
> +    cpu_synchronize_state(env);
> +
> +    cpu_fprintf(f, "SLB\tESID\t\t\tVSID\n");
> +    for (i = 0; i < env->slb_nr; i++) {
> +        slbe = env->slb[i].esid;
> +        slbv = env->slb[i].vsid;
> +        if (slbe == 0 && slbv == 0) {
> +            continue;
> +        }
> +        cpu_fprintf(f, "%d\t0x%016" PRIx64 "\t0x%016" PRIx64 "\n",
> +                    i, slbe, slbv);
> +    }
> +}
> +
> void dump_mmu(FILE *f, fprintf_function cpu_fprintf, CPUState *env)
> {
> -    switch (env->mmu_model) {
> -    case POWERPC_MMU_BOOKE206:
> +    if (env->mmu_model == POWERPC_MMU_BOOKE206) {
>         mmubooke206_dump_mmu(f, cpu_fprintf, env);
> -        break;
> -    default:
> -        cpu_fprintf(f, "%s: unimplemented\n", __func__);
> +    } else {
> +        if ((env->mmu_model & POWERPC_MMU_64B) != 0) {
> +            mmubooks_dump_mmu(f, cpu_fprintf, env);
> +        } else {
> +            cpu_fprintf(f, "%s: unimplemented\n", __func__);
> +        }

Didn't I comment on this one before, saying that it should stay a switch? In fact, I even applied his v2 already.


Alex

Re: [Qemu-devel] [Qemu-ppc] [PATCH 2/8] monitor: add ability to dump SLB entries

[David Gibson]

On Tue, Dec 20, 2011 at 04:28:47PM +0100, Alexander Graf wrote:
> 
> On 13.12.2011, at 05:24, David Gibson wrote:
> 
> > From: Nishanth Aravamudan <nacc@us.ibm.com>
> > 
> > When run with a PPC Book3S (server) CPU Currently 'info tlb' in the
> > qemu monitor reports "dump_mmu: unimplemented".  However, during
> > bringup work, it can be quite handy to have the SLB entries, which are
> > available in the CPUPPCState.  This patch adds an implementation of
> > info tlb for book3s, which dumps the SLB.
> > 
> > Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> > ---
> > target-ppc/helper.c |   32 +++++++++++++++++++++++++++-----
> > 1 files changed, 27 insertions(+), 5 deletions(-)
> > 
> > diff --git a/target-ppc/helper.c b/target-ppc/helper.c
> > index 137a494..29c7050 100644
> > --- a/target-ppc/helper.c
> > +++ b/target-ppc/helper.c
> > @@ -1545,14 +1545,36 @@ static void mmubooke206_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
> >     }
> > }
> > 
> > +static void mmubooks_dump_mmu(FILE *f, fprintf_function cpu_fprintf,
> > +                              CPUState *env)
> > +{
> > +    int i;
> > +    uint64_t slbe, slbv;
> > +
> > +    cpu_synchronize_state(env);
> > +
> > +    cpu_fprintf(f, "SLB\tESID\t\t\tVSID\n");
> > +    for (i = 0; i < env->slb_nr; i++) {
> > +        slbe = env->slb[i].esid;
> > +        slbv = env->slb[i].vsid;
> > +        if (slbe == 0 && slbv == 0) {
> > +            continue;
> > +        }
> > +        cpu_fprintf(f, "%d\t0x%016" PRIx64 "\t0x%016" PRIx64 "\n",
> > +                    i, slbe, slbv);
> > +    }
> > +}
> > +
> > void dump_mmu(FILE *f, fprintf_function cpu_fprintf, CPUState *env)
> > {
> > -    switch (env->mmu_model) {
> > -    case POWERPC_MMU_BOOKE206:
> > +    if (env->mmu_model == POWERPC_MMU_BOOKE206) {
> >         mmubooke206_dump_mmu(f, cpu_fprintf, env);
> > -        break;
> > -    default:
> > -        cpu_fprintf(f, "%s: unimplemented\n", __func__);
> > +    } else {
> > +        if ((env->mmu_model & POWERPC_MMU_64B) != 0) {
> > +            mmubooks_dump_mmu(f, cpu_fprintf, env);
> > +        } else {
> > +            cpu_fprintf(f, "%s: unimplemented\n", __func__);
> > +        }
> 
> Didn't I comment on this one before, saying that it should stay a
> switch? In fact, I even applied his v2 already.

Ah, yes, sorry.  I was tracking upstream, so I forgot it was in your
tree.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[Qemu-devel] [PATCH 3/8] pseries: FDT NUMA extensions to support multi-node guests

[David Gibson]

From: Bharata B Rao <bharata@linux.vnet.ibm.com>

Add NUMA specific properties to guest's device tree to boot a multi-node
guests. This patch adds the following properties:

ibm,associativity
ibm,architecture-vec-5
ibm,associativity-reference-points

With this, it becomes possible to use -numa option on pseries targets.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr.c |  112 ++++++++++++++++++++++++++++++++++++++++++++++++++++++------
 hw/spapr.h |    1 +
 2 files changed, 102 insertions(+), 11 deletions(-)

diff --git a/hw/spapr.c b/hw/spapr.c
index 2b901f1..1883270 100644
--- a/hw/spapr.c
+++ b/hw/spapr.c
@@ -97,6 +97,44 @@ qemu_irq spapr_allocate_irq(uint32_t hint, uint32_t *irq_num)
     return qirq;
 }
 
+static int spapr_set_associativity(void *fdt, sPAPREnvironment *spapr)
+{
+    int ret = 0, offset;
+    CPUState *env;
+    char cpu_model[32];
+    int smt = kvmppc_smt_threads();
+
+    assert(spapr->cpu_model);
+
+    for (env = first_cpu; env != NULL; env = env->next_cpu) {
+        uint32_t associativity[] = {cpu_to_be32(0x5),
+                                    cpu_to_be32(0x0),
+                                    cpu_to_be32(0x0),
+                                    cpu_to_be32(0x0),
+                                    cpu_to_be32(env->numa_node),
+                                    cpu_to_be32(env->cpu_index)};
+
+        if ((env->cpu_index % smt) != 0) {
+            continue;
+        }
+
+        snprintf(cpu_model, 32, "/cpus/%s@%x", spapr->cpu_model,
+                 env->cpu_index);
+
+        offset = fdt_path_offset(fdt, cpu_model);
+        if (offset < 0) {
+            return offset;
+        }
+
+        ret = fdt_setprop(fdt, offset, "ibm,associativity", associativity,
+                          sizeof(associativity));
+        if (ret < 0) {
+            return ret;
+        }
+    }
+    return ret;
+}
+
 static void *spapr_create_fdt_skel(const char *cpu_model,
                                    target_phys_addr_t rma_size,
                                    target_phys_addr_t initrd_base,
@@ -107,9 +145,7 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
 {
     void *fdt;
     CPUState *env;
-    uint64_t mem_reg_property_rma[] = { 0, cpu_to_be64(rma_size) };
-    uint64_t mem_reg_property_nonrma[] = { cpu_to_be64(rma_size),
-                                           cpu_to_be64(ram_size - rma_size) };
+    uint64_t mem_reg_property[2];
     uint32_t start_prop = cpu_to_be32(initrd_base);
     uint32_t end_prop = cpu_to_be32(initrd_base + initrd_size);
     uint32_t pft_size_prop[] = {0, cpu_to_be32(hash_shift)};
@@ -119,6 +155,13 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
     int i;
     char *modelname;
     int smt = kvmppc_smt_threads();
+    unsigned char vec5[] = {0x0, 0x0, 0x0, 0x0, 0x0, 0x80};
+    uint32_t refpoints[] = {cpu_to_be32(0x4), cpu_to_be32(0x4)};
+    uint32_t associativity[] = {cpu_to_be32(0x4), cpu_to_be32(0x0),
+                                cpu_to_be32(0x0), cpu_to_be32(0x0),
+                                cpu_to_be32(0x0)};
+    char mem_name[32];
+    target_phys_addr_t node0_size, mem_start;
 
 #define _FDT(exp) \
     do { \
@@ -146,6 +189,9 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
     /* /chosen */
     _FDT((fdt_begin_node(fdt, "chosen")));
 
+    /* Set Form1_affinity */
+    _FDT((fdt_property(fdt, "ibm,architecture-vec-5", vec5, sizeof(vec5))));
+
     _FDT((fdt_property_string(fdt, "bootargs", kernel_cmdline)));
     _FDT((fdt_property(fdt, "linux,initrd-start",
                        &start_prop, sizeof(start_prop))));
@@ -164,24 +210,54 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
     _FDT((fdt_end_node(fdt)));
 
     /* memory node(s) */
-    _FDT((fdt_begin_node(fdt, "memory@0")));
+    node0_size = (nb_numa_nodes > 1) ? node_mem[0] : ram_size;
+    if (rma_size > node0_size) {
+        rma_size = node0_size;
+    }
 
+    /* RMA */
+    mem_reg_property[0] = 0;
+    mem_reg_property[1] = cpu_to_be64(rma_size);
+    _FDT((fdt_begin_node(fdt, "memory@0")));
     _FDT((fdt_property_string(fdt, "device_type", "memory")));
-    _FDT((fdt_property(fdt, "reg", mem_reg_property_rma,
-                       sizeof(mem_reg_property_rma))));
+    _FDT((fdt_property(fdt, "reg", mem_reg_property,
+        sizeof(mem_reg_property))));
+    _FDT((fdt_property(fdt, "ibm,associativity", associativity,
+        sizeof(associativity))));
     _FDT((fdt_end_node(fdt)));
 
-    if (ram_size > rma_size) {
-        char mem_name[32];
+    /* RAM: Node 0 */
+    if (node0_size > rma_size) {
+        mem_reg_property[0] = cpu_to_be64(rma_size);
+        mem_reg_property[1] = cpu_to_be64(node0_size - rma_size);
 
-        sprintf(mem_name, "memory@%" PRIx64, (uint64_t)rma_size);
+        sprintf(mem_name, "memory@" TARGET_FMT_lx, rma_size);
         _FDT((fdt_begin_node(fdt, mem_name)));
         _FDT((fdt_property_string(fdt, "device_type", "memory")));
-        _FDT((fdt_property(fdt, "reg", mem_reg_property_nonrma,
-                           sizeof(mem_reg_property_nonrma))));
+        _FDT((fdt_property(fdt, "reg", mem_reg_property,
+                           sizeof(mem_reg_property))));
+        _FDT((fdt_property(fdt, "ibm,associativity", associativity,
+                           sizeof(associativity))));
         _FDT((fdt_end_node(fdt)));
     }
 
+    /* RAM: Node 1 and beyond */
+    mem_start = node0_size;
+    for (i = 1; i < nb_numa_nodes; i++) {
+        mem_reg_property[0] = cpu_to_be64(mem_start);
+        mem_reg_property[1] = cpu_to_be64(node_mem[i]);
+        associativity[3] = associativity[4] = cpu_to_be32(i);
+        sprintf(mem_name, "memory@" TARGET_FMT_lx, mem_start);
+        _FDT((fdt_begin_node(fdt, mem_name)));
+        _FDT((fdt_property_string(fdt, "device_type", "memory")));
+        _FDT((fdt_property(fdt, "reg", mem_reg_property,
+            sizeof(mem_reg_property))));
+        _FDT((fdt_property(fdt, "ibm,associativity", associativity,
+            sizeof(associativity))));
+        _FDT((fdt_end_node(fdt)));
+        mem_start += node_mem[i];
+    }
+
     /* cpus */
     _FDT((fdt_begin_node(fdt, "cpus")));
 
@@ -194,6 +270,9 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
         modelname[i] = toupper(modelname[i]);
     }
 
+    /* This is needed during FDT finalization */
+    spapr->cpu_model = g_strdup(modelname);
+
     for (env = first_cpu; env != NULL; env = env->next_cpu) {
         int index = env->cpu_index;
         uint32_t servers_prop[smp_threads];
@@ -280,6 +359,9 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
     _FDT((fdt_property(fdt, "ibm,hypertas-functions", hypertas_prop,
                        sizeof(hypertas_prop))));
 
+    _FDT((fdt_property(fdt, "ibm,associativity-reference-points",
+        refpoints, sizeof(refpoints))));
+
     _FDT((fdt_end_node(fdt)));
 
     /* interrupt controller */
@@ -351,6 +433,14 @@ static void spapr_finalize_fdt(sPAPREnvironment *spapr,
         fprintf(stderr, "Couldn't set up RTAS device tree properties\n");
     }
 
+    /* Advertise NUMA via ibm,associativity */
+    if (nb_numa_nodes > 1) {
+        ret = spapr_set_associativity(fdt, spapr);
+        if (ret < 0) {
+            fprintf(stderr, "Couldn't set up NUMA device tree properties\n");
+        }
+    }
+
     _FDT((fdt_pack(fdt)));
 
     cpu_physical_memory_write(fdt_addr, fdt, fdt_totalsize(fdt));
diff --git a/hw/spapr.h b/hw/spapr.h
index df88f6a..a172e9f 100644
--- a/hw/spapr.h
+++ b/hw/spapr.h
@@ -21,6 +21,7 @@ typedef struct sPAPREnvironment {
     target_ulong entry_point;
     int next_irq;
     int rtc_offset;
+    char *cpu_model;
 } sPAPREnvironment;
 
 #define H_SUCCESS         0
-- 
1.7.7.3

[Qemu-devel] [PATCH 5/8] pseries: Emit device tree nodes in reg order

[David Gibson]

Although in theory the device tree has no inherent ordering, in practice
the order of nodes in the device tree does effect the order that devices
are detected by software.

Currently the ordering is determined by the order the devices appear on
the QEMU command line. Although that does give the user control over the
ordering, it is fragile, especially when the user does not generate the
command line manually - eg. when using libvirt etc.

So order the device tree based on the reg value, ie. the address of on
the VIO bus of the devices. This gives us a sane and stable ordering.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr_vio.c |   48 +++++++++++++++++++++++++++++++++++++++++++-----
 1 files changed, 43 insertions(+), 5 deletions(-)

diff --git a/hw/spapr_vio.c b/hw/spapr_vio.c
index 5a35541..339c1c3 100644
--- a/hw/spapr_vio.c
+++ b/hw/spapr_vio.c
@@ -736,21 +736,59 @@ static void spapr_vio_register_devices(void)
 device_init(spapr_vio_register_devices)
 
 #ifdef CONFIG_FDT
+static int compare_reg(const void *p1, const void *p2)
+{
+    VIOsPAPRDevice const *dev1, *dev2;
+
+    dev1 = (VIOsPAPRDevice *)*(DeviceState **)p1;
+    dev2 = (VIOsPAPRDevice *)*(DeviceState **)p2;
+
+    if (dev1->reg < dev2->reg)
+        return -1;
+    if (dev1->reg == dev2->reg)
+        return 0;
+
+    /* dev1->reg > dev2->reg */
+    return 1;
+}
+
 int spapr_populate_vdevice(VIOsPAPRBus *bus, void *fdt)
 {
-    DeviceState *qdev;
-    int ret = 0;
+    DeviceState *qdev, **qdevs;
+    int i, num, ret = 0;
 
+    /* Count qdevs on the bus list */
+    num = 0;
     QTAILQ_FOREACH(qdev, &bus->bus.children, sibling) {
-        VIOsPAPRDevice *dev = (VIOsPAPRDevice *)qdev;
+        num++;
+    }
+
+    /* Copy out into an array of pointers */
+    qdevs = g_malloc(sizeof(qdev) * num);
+    num = 0;
+    QTAILQ_FOREACH(qdev, &bus->bus.children, sibling) {
+        qdevs[num++] = qdev;
+    }
+
+    /* Sort the array */
+    qsort(qdevs, num, sizeof(qdev), compare_reg);
+
+    /* Hack alert. Give the devices to libfdt in reverse order, we happen
+     * to know that will mean they are in forward order in the tree. */
+    for (i = num - 1; i >= 0; i--) {
+        VIOsPAPRDevice *dev = (VIOsPAPRDevice *)(qdevs[i]);
 
         ret = vio_make_devnode(dev, fdt);
 
         if (ret < 0) {
-            return ret;
+            goto out;
         }
     }
 
-    return 0;
+    ret = 0;
+out:
+    free(qdevs);
+
+    return ret;
 }
 #endif /* CONFIG_FDT */
-- 
1.7.7.3

[Qemu-devel] [PATCH 6/8] pseries: Add a routine to find a stable "default" vty and use it

[David Gibson]

In vty_lookup() we have a special case for supporting early debug in
the kernel. This accepts reg == 0 as a special case to mean "any vty".

We implement this by searching the vtys on the bus and returning the
first we find. This means that the vty we chose depends on the order
the vtys are specified on the QEMU command line - because that determines
the order of the vtys on the bus.

We'd rather the command line order was irrelevant, so instead return
the vty with the lowest reg value. This is still a guess as to what the
user really means, but it is at least stable WRT command line ordering.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr_vty.c |   46 +++++++++++++++++++++++++++++++++++++---------
 1 files changed, 37 insertions(+), 9 deletions(-)

diff --git a/hw/spapr_vty.c b/hw/spapr_vty.c
index e217314..fc8d91a 100644
--- a/hw/spapr_vty.c
+++ b/hw/spapr_vty.c
@@ -149,24 +149,52 @@ static VIOsPAPRDeviceInfo spapr_vty = {
     },
 };
 
+static VIOsPAPRDevice *spapr_vty_get_default(VIOsPAPRBus *bus)
+{
+    VIOsPAPRDevice *sdev, *selected;
+    DeviceState *iter;
+
+    /*
+     * To avoid the console bouncing around we want one VTY to be
+     * the "default". We haven't really got anything to go on, so
+     * arbitrarily choose the one with the lowest reg value.
+     */
+
+    selected = NULL;
+    QTAILQ_FOREACH(iter, &bus->bus.children, sibling) {
+        /* Only look at VTY devices */
+        if (iter->info != &spapr_vty.qdev) {
+            continue;
+        }
+
+        sdev = DO_UPCAST(VIOsPAPRDevice, qdev, iter);
+
+        /* First VTY we've found, so it is selected for now */
+        if (!selected) {
+            selected = sdev;
+            continue;
+        }
+
+        /* Choose VTY with lowest reg value */
+        if (sdev->reg < selected->reg)
+            selected = sdev;
+    }
+
+    return selected;
+}
+
 static VIOsPAPRDevice *vty_lookup(sPAPREnvironment *spapr, target_ulong reg)
 {
     VIOsPAPRDevice *sdev;
 
     sdev = spapr_vio_find_by_reg(spapr->vio_bus, reg);
     if (!sdev && reg == 0) {
-        DeviceState *qdev;
-
         /* Hack for kernel early debug, which always specifies reg==0.
-         * We search all VIO devices, and grab the first available vty
-         * device.  This attempts to mimic existing PowerVM behaviour
+         * We search all VIO devices, and grab the vty with the lowest
+         * reg.  This attempts to mimic existing PowerVM behaviour
          * (early debug does work there, despite having no vty with
          * reg==0. */
-        QTAILQ_FOREACH(qdev, &spapr->vio_bus->bus.children, sibling) {
-            if (qdev->info == &spapr_vty.qdev) {
-                return DO_UPCAST(VIOsPAPRDevice, qdev, qdev);
-            }
-        }
+        return spapr_vty_get_default(spapr->vio_bus);
     }
 
     return sdev;
-- 
1.7.7.3

[Qemu-devel] [PATCH 7/8] pseries: Populate "/chosen/linux, stdout-path" in the FDT

[David Gibson]

There is a device tree property "/chosen/linux,stdout-path" which indicates
which device should be used as stdout - ie. "the console".

Currently we don't specify anything, which means both firmware and Linux
choose something arbitrarily. Use the routine we added in the last patch
to pick a default vty and specify it as stdout.

Currently SLOF doesn't use the property, but we are hoping to update it
to do so.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr.c     |    2 ++
 hw/spapr_vio.c |   34 ++++++++++++++++++++++++++++++++++
 hw/spapr_vio.h |    3 +++
 hw/spapr_vty.c |    2 +-
 4 files changed, 40 insertions(+), 1 deletions(-)

diff --git a/hw/spapr.c b/hw/spapr.c
index 1883270..6a543de 100644
--- a/hw/spapr.c
+++ b/hw/spapr.c
@@ -441,6 +441,8 @@ static void spapr_finalize_fdt(sPAPREnvironment *spapr,
         }
     }
 
+    spapr_populate_chosen_stdout(fdt, spapr->vio_bus);
+
     _FDT((fdt_pack(fdt)));
 
     cpu_physical_memory_write(fdt_addr, fdt, fdt_totalsize(fdt));
diff --git a/hw/spapr_vio.c b/hw/spapr_vio.c
index 339c1c3..b188a71 100644
--- a/hw/spapr_vio.c
+++ b/hw/spapr_vio.c
@@ -791,4 +791,38 @@ out:
 
     return ret;
 }
+
+int spapr_populate_chosen_stdout(void *fdt, VIOsPAPRBus *bus)
+{
+    VIOsPAPRDevice *dev;
+    char *name, *path;
+    int ret, offset;
+
+    dev = spapr_vty_get_default(bus);
+    if (!dev)
+        return 0;
+
+    offset = fdt_path_offset(fdt, "/chosen");
+    if (offset < 0) {
+        return offset;
+    }
+
+    name = vio_format_dev_name(dev);
+    if (!name) {
+        return -ENOMEM;
+    }
+
+    if (asprintf(&path, "/vdevice/%s", name) < 0) {
+        path = NULL;
+        ret = -ENOMEM;
+        goto out;
+    }
+
+    ret = fdt_setprop_string(fdt, offset, "linux,stdout-path", path);
+out:
+    free(name);
+    free(path);
+
+    return ret;
+}
 #endif /* CONFIG_FDT */
diff --git a/hw/spapr_vio.h b/hw/spapr_vio.h
index 2430d45..0984d55 100644
--- a/hw/spapr_vio.h
+++ b/hw/spapr_vio.h
@@ -82,6 +82,7 @@ extern VIOsPAPRBus *spapr_vio_bus_init(void);
 extern VIOsPAPRDevice *spapr_vio_find_by_reg(VIOsPAPRBus *bus, uint32_t reg);
 extern void spapr_vio_bus_register_withprop(VIOsPAPRDeviceInfo *info);
 extern int spapr_populate_vdevice(VIOsPAPRBus *bus, void *fdt);
+extern int spapr_populate_chosen_stdout(void *fdt, VIOsPAPRBus *bus);
 
 extern int spapr_vio_signal(VIOsPAPRDevice *dev, target_ulong mode);
 
@@ -107,6 +108,8 @@ 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);
 
+VIOsPAPRDevice *spapr_vty_get_default(VIOsPAPRBus *bus);
+
 int spapr_tce_set_bypass(uint32_t unit, uint32_t enable);
 void spapr_vio_quiesce(void);
 
diff --git a/hw/spapr_vty.c b/hw/spapr_vty.c
index fc8d91a..0a8016f 100644
--- a/hw/spapr_vty.c
+++ b/hw/spapr_vty.c
@@ -149,7 +149,7 @@ static VIOsPAPRDeviceInfo spapr_vty = {
     },
 };
 
-static VIOsPAPRDevice *spapr_vty_get_default(VIOsPAPRBus *bus)
+VIOsPAPRDevice *spapr_vty_get_default(VIOsPAPRBus *bus)
 {
     VIOsPAPRDevice *sdev, *selected;
     DeviceState *iter;
-- 
1.7.7.3

[Qemu-devel] [PATCH 8/8] pseries: Check for duplicate addresses on the spapr-vio bus

[David Gibson]

From: Michael Ellerman <michael@ellerman.id.au>

Check that devices on the spapr vio bus aren't given duplicate
addresses. Currently we will not run with duplicate devices, the
fdt code will spot it, but the error reporting is not great. With
this patch we can report the error nicely in terms of the device
names given by the user.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/spapr_vio.c |   32 ++++++++++++++++++++++++++++++++
 1 files changed, 32 insertions(+), 0 deletions(-)

diff --git a/hw/spapr_vio.c b/hw/spapr_vio.c
index b188a71..0e5581c 100644
--- a/hw/spapr_vio.c
+++ b/hw/spapr_vio.c
@@ -621,11 +621,43 @@ static void rtas_quiesce(sPAPREnvironment *spapr, uint32_t token,
     rtas_st(rets, 0, 0);
 }
 
+static int spapr_vio_check_reg(VIOsPAPRDevice *sdev, VIOsPAPRDeviceInfo *info)
+{
+    VIOsPAPRDevice *other_sdev;
+    DeviceState *qdev;
+    VIOsPAPRBus *sbus;
+
+    sbus = DO_UPCAST(VIOsPAPRBus, bus, sdev->qdev.parent_bus);
+
+    /*
+     * 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.
+     */
+    QTAILQ_FOREACH(qdev, &sbus->bus.children, sibling) {
+        other_sdev = 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",
+                    info->qdev.name, other_sdev->qdev.info->name, sdev->reg);
+            return -EEXIST;
+        }
+    }
+
+    return 0;
+}
+
 static int spapr_vio_busdev_init(DeviceState *qdev, DeviceInfo *qinfo)
 {
     VIOsPAPRDeviceInfo *info = (VIOsPAPRDeviceInfo *)qinfo;
     VIOsPAPRDevice *dev = (VIOsPAPRDevice *)qdev;
     char *id;
+    int ret;
+
+    ret = spapr_vio_check_reg(dev, info);
+    if (ret) {
+        return ret;
+    }
 
     /* Don't overwrite ids assigned on the command line */
     if (!dev->qdev.id) {
-- 
1.7.7.3

Re: [Qemu-devel] [0/8] Assorted pseries updates

[Alexander Graf]

On 13.12.2011, at 05:24, David Gibson wrote:

> Now that qemu 1.0 is out there, here are a number of cleanups and
> minor bugfixes for the pseries machine to queue up for post 1.0.  Many
> of these fix problems with non-default device configurations exposed
> by the w.i.p. port of libvirt for powerkvm.
> 

Thanks, applied all to ppc-next, except:

  2) was already in
  4) is missing - did you forget to send it out? It's not in my inbox and not on patchwork either
  5) added braces - please run checkpatch before you send out patchsets!
  6) same as above


Alex

Re: [Qemu-devel] [Qemu-ppc] [0/8] Assorted pseries updates

[David Gibson]

On Tue, Dec 20, 2011 at 04:41:25PM +0100, Alexander Graf wrote:
> 
> On 13.12.2011, at 05:24, David Gibson wrote:
> 
> > Now that qemu 1.0 is out there, here are a number of cleanups and
> > minor bugfixes for the pseries machine to queue up for post 1.0.  Many
> > of these fix problems with non-default device configurations exposed
> > by the w.i.p. port of libvirt for powerkvm.
> > 
> 
> Thanks, applied all to ppc-next, except:
> 
>   2) was already in
>   4) is missing - did you forget to send it out? It's not in my
>   inbox and not on patchwork either

Odd, something must have screwed up with my mail setup somehow.  I'll
resend it along with the next batch of fixes.

>   5) added braces - please run checkpatch before you send out patchsets!
>   6) same as above
> 
> 
> Alex
> 
> 

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson