Skip to contentSkip to conversation

ppc/spapr: Implement H_RANDOM hypercall in QEMU

PATCH Applied Thomas Huth 4 participants 10 messages Sep 17, 2015 – Sep 22, 2015

lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 ppc/spapr: Implement H_RANDOM hypercall in QEMU

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

---
 hw/ppc/Makefile.objs   |   2 +-
 hw/ppc/spapr.c         |   8 +++
 hw/ppc/spapr_rng.c     | 186 +++++++++++++++++++++++++++++++++++++++++++++++++
 include/hw/ppc/spapr.h |   4 ++
 target-ppc/kvm.c       |   9 +++
 target-ppc/kvm_ppc.h   |   5 ++
 6 files changed, 213 insertions(+), 1 deletion(-)
 create mode 100644 hw/ppc/spapr_rng.c
hw/ppc/Makefile.objs +1 -1
index c8ab06e..c1ffc77 100644
@@ -3,7 +3,7 @@ obj-y += ppc.o ppc_booke.o
 # IBM pSeries (sPAPR)
 obj-$(CONFIG_PSERIES) += spapr.o spapr_vio.o spapr_events.o
 obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o
-obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o
+obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o spapr_rng.o
 ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)
 obj-y += spapr_pci_vfio.o
 endif
hw/ppc/spapr.c +8 -0
index bf0c64f..34e7d24 100644
@@ -768,6 +768,14 @@ static void spapr_finalize_fdt(sPAPRMachineState *spapr,
         exit(1);
     }
 
+    if (object_resolve_path_type("", TYPE_SPAPR_RNG, NULL)) {
+        ret = spapr_rng_populate_dt(fdt);
+        if (ret < 0) {
+            fprintf(stderr, "could not set up rng device in the fdt\n");
+            exit(1);
+        }
+    }
+
     QLIST_FOREACH(phb, &spapr->phbs, list) {
         ret = spapr_populate_pci_dt(phb, PHANDLE_XICP, fdt);
     }
hw/ppc/spapr_rng.c new file +186 -0

186 lines not on this page · raw message

include/hw/ppc/spapr.h +4 -0
index 91a61ab..4e8aa2d 100644
@@ -331,6 +331,7 @@ struct sPAPRMachineState {
 #define H_SET_MPP               0x2D0
 #define H_GET_MPP               0x2D4
 #define H_XIRR_X                0x2FC
+#define H_RANDOM                0x300
 #define H_SET_MODE              0x31C
 #define MAX_HCALL_OPCODE        H_SET_MODE
 
@@ -603,10 +604,13 @@ struct sPAPRConfigureConnectorState {
 void spapr_ccs_reset_hook(void *opaque);
 
 #define TYPE_SPAPR_RTC "spapr-rtc"
+#define TYPE_SPAPR_RNG "spapr-rng"
 
 void spapr_rtc_read(DeviceState *dev, struct tm *tm, uint32_t *ns);
 int spapr_rtc_import_offset(DeviceState *dev, int64_t legacy_offset);
 
+int spapr_rng_populate_dt(void *fdt);
+
 #define SPAPR_MEMORY_BLOCK_SIZE (1 << 28) /* 256MB */
 
 #endif /* !defined (__HW_SPAPR_H__) */
target-ppc/kvm.c +9 -0
index 110436d..42f66fe 100644
@@ -2484,3 +2484,12 @@ int kvm_arch_msi_data_to_gsi(uint32_t data)
 {
     return data & 0xffff;
 }
+
+int kvmppc_enable_hwrng(void)
+{
+    if (!kvm_enabled() || !kvm_check_extension(kvm_state, KVM_CAP_PPC_HWRNG)) {
+        return -1;
+    }
+
+    return kvmppc_enable_hcall(kvm_state, H_RANDOM);
+}
target-ppc/kvm_ppc.h +5 -0
index 4d30e27..68836b4 100644
@@ -55,6 +55,7 @@ void kvmppc_hash64_free_pteg(uint64_t token);
 void kvmppc_hash64_write_pte(CPUPPCState *env, target_ulong pte_index,
                              target_ulong pte0, target_ulong pte1);
 bool kvmppc_has_cap_fixup_hcalls(void);
+int kvmppc_enable_hwrng(void);
 
 #else
 
@@ -248,6 +249,10 @@ static inline bool kvmppc_has_cap_fixup_hcalls(void)
     abort();
 }
 
+static inline int kvmppc_enable_hwrng(void)
+{
+    return -1;
+}
 #endif
 
 #ifndef CONFIG_KVM
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 22 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

Quoted 8 lines
 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

---

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

hw/ppc/spapr.c +16 -0
index 240fab72e7af..4b92efed2ef3 100644
@@ -939,6 +939,14 @@ static int spapr_check_htab_fd(sPAPRMachineState *spapr)
     return rc;
 }
 
+static void spapr_rng_create(void)
+{
+    Object *rng = object_new(TYPE_SPAPR_RNG);
+
+    object_property_set_bool(rng, true, "use-kvm", &error_abort);
+    object_property_set_bool(rng, true, "realized", &error_abort);
+}
+
 static void ppc_spapr_reset(void)
 {
     sPAPRMachineState *spapr = SPAPR_MACHINE(qdev_get_machine());
@@ -962,6 +970,14 @@ static void ppc_spapr_reset(void)
     spapr->rtas_addr = rtas_limit - RTAS_MAX_SIZE;
     spapr->fdt_addr = spapr->rtas_addr - FDT_MAX_SIZE;
 
+    /* Create a rng device if the user did not provide it already and
+     * KVM has hwrng support.
+     */
+    if (kvmppc_hwrng_present() &&
+        !object_resolve_path_type("", TYPE_SPAPR_RNG, NULL)) {
+        spapr_rng_create();
+    }
+
     /* Load the fdt */
     spapr_finalize_fdt(spapr, spapr->fdt_addr, spapr->rtas_addr,
                        spapr->rtas_size);
hw/ppc/spapr_rng.c +1 -1
index ed43d5e04221..ee5af302bd4d 100644
@@ -114,7 +114,7 @@ static void spapr_rng_realize(DeviceState *dev, Error **errp)
     sPAPRRngState *rngstate = SPAPR_RNG(dev);
 
     if (rngstate->use_kvm) {
-        if (kvmppc_enable_hwrng() = 0) {
+        if (kvmppc_hwrng_present() && kvmppc_enable_hwrng() = 0) {
             return;
         }
         /*
target-ppc/kvm.c +5 -4
index 42f66fea23e9..008f8a26ab17 100644
@@ -2485,11 +2485,12 @@ int kvm_arch_msi_data_to_gsi(uint32_t data)
     return data & 0xffff;
 }
 
-int kvmppc_enable_hwrng(void)
+bool kvmppc_hwrng_present(void)
 {
-    if (!kvm_enabled() || !kvm_check_extension(kvm_state, KVM_CAP_PPC_HWRNG)) {
-        return -1;
-    }
+    return kvm_enabled() && kvm_check_extension(kvm_state, KVM_CAP_PPC_HWRNG);
+}
 
+int kvmppc_enable_hwrng(void)
+{
     return kvmppc_enable_hcall(kvm_state, H_RANDOM);
 }
target-ppc/kvm_ppc.h +6 -0
index 68836b401105..4b78bfe5224a 100644
@@ -55,6 +55,7 @@ void kvmppc_hash64_free_pteg(uint64_t token);
 void kvmppc_hash64_write_pte(CPUPPCState *env, target_ulong pte_index,
                              target_ulong pte0, target_ulong pte1);
 bool kvmppc_has_cap_fixup_hcalls(void);
+bool kvmppc_hwrng_present(void);
 int kvmppc_enable_hwrng(void);
 
 #else
@@ -249,6 +250,11 @@ static inline bool kvmppc_has_cap_fixup_hcalls(void)
     abort();
 }
 
+static inline bool kvmppc_hwrng_present(void)
+{
+    return false;
+}
+
 static inline int kvmppc_enable_hwrng(void)
 {
     return -1;
Quoted 299 lines
 hw/ppc/Makefile.objs   |   2 +-
 hw/ppc/spapr.c         |   8 +++
 hw/ppc/spapr_rng.c     | 186 +++++++++++++++++++++++++++++++++++++++++++++++++
 include/hw/ppc/spapr.h |   4 ++
 target-ppc/kvm.c       |   9 +++
 target-ppc/kvm_ppc.h   |   5 ++
 6 files changed, 213 insertions(+), 1 deletion(-)
 create mode 100644 hw/ppc/spapr_rng.c
hw/ppc/Makefile.objs +1 -1
index c8ab06e..c1ffc77 100644
@@ -3,7 +3,7 @@ obj-y += ppc.o ppc_booke.o
 # IBM pSeries (sPAPR)
 obj-$(CONFIG_PSERIES) += spapr.o spapr_vio.o spapr_events.o
 obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o
-obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o
+obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o spapr_rng.o
 ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)
 obj-y += spapr_pci_vfio.o
 endif
hw/ppc/spapr.c +8 -0
index bf0c64f..34e7d24 100644
@@ -768,6 +768,14 @@ static void spapr_finalize_fdt(sPAPRMachineState *spapr,
         exit(1);
     }
 
+    if (object_resolve_path_type("", TYPE_SPAPR_RNG, NULL)) {
+        ret = spapr_rng_populate_dt(fdt);
+        if (ret < 0) {
+            fprintf(stderr, "could not set up rng device in the fdt\n");
+            exit(1);
+        }
+    }
+
     QLIST_FOREACH(phb, &spapr->phbs, list) {
         ret = spapr_populate_pci_dt(phb, PHANDLE_XICP, fdt);
     }
hw/ppc/spapr_rng.c new file +186 -0
new file mode 100644
index 0000000..ed43d5e
@@ -0,0 +1,186 @@
+/*
+ * QEMU sPAPR random number generator "device" for H_RANDOM hypercall
+ *
+ * Copyright 2015 Thomas Huth, Red Hat Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/error-report.h"
+#include "sysemu/sysemu.h"
+#include "sysemu/device_tree.h"
+#include "sysemu/rng.h"
+#include "hw/ppc/spapr.h"
+#include "kvm_ppc.h"
+
+#define SPAPR_RNG(obj) \
+    OBJECT_CHECK(sPAPRRngState, (obj), TYPE_SPAPR_RNG)
+
+struct sPAPRRngState {
+    /*< private >*/
+    DeviceState ds;
+    RngBackend *backend;
+    bool use_kvm;
+};
+typedef struct sPAPRRngState sPAPRRngState;
+
+struct HRandomData {
+    QemuSemaphore sem;
+    union {
+        uint64_t v64;
+        uint8_t v8[8];
+    } val;
+    int received;
+};
+typedef struct HRandomData HRandomData;
+
+/* Callback function for the RngBackend */
+static void random_recv(void *dest, const void *src, size_t size)
+{
+    HRandomData *hrdp = dest;
+
+    if (src && size > 0) {
+        assert(size + hrdp->received <= sizeof(hrdp->val.v8));
+        memcpy(&hrdp->val.v8[hrdp->received], src, size);
+        hrdp->received += size;
+    }
+
+    qemu_sem_post(&hrdp->sem);
+}
+
+/* Handler for the H_RANDOM hypercall */
+static target_ulong h_random(PowerPCCPU *cpu, sPAPRMachineState *spapr,
+                             target_ulong opcode, target_ulong *args)
+{
+    sPAPRRngState *rngstate;
+    HRandomData hrdata;
+
+    rngstate = SPAPR_RNG(object_resolve_path_type("", TYPE_SPAPR_RNG, NULL));
+
+    if (!rngstate || !rngstate->backend) {
+        return H_HARDWARE;
+    }
+
+    qemu_sem_init(&hrdata.sem, 0);
+    hrdata.val.v64 = 0;
+    hrdata.received = 0;
+
+    qemu_mutex_unlock_iothread();
+    while (hrdata.received < 8) {
+        rng_backend_request_entropy(rngstate->backend, 8 - hrdata.received,
+                                    random_recv, &hrdata);
+        qemu_sem_wait(&hrdata.sem);
+    }
+    qemu_mutex_lock_iothread();
+
+    qemu_sem_destroy(&hrdata.sem);
+    args[0] = hrdata.val.v64;
+
+    return H_SUCCESS;
+}
+
+static void spapr_rng_instance_init(Object *obj)
+{
+    sPAPRRngState *rngstate = SPAPR_RNG(obj);
+
+    if (object_resolve_path_type("", TYPE_SPAPR_RNG, NULL) != NULL) {
+        error_report("spapr-rng can not be instantiated twice!");
+        return;
+    }
+
+    object_property_add_link(obj, "rng", TYPE_RNG_BACKEND,
+                             (Object **)&rngstate->backend,
+                             object_property_allow_set_link,
+                             OBJ_PROP_LINK_UNREF_ON_RELEASE, NULL);
+    object_property_set_description(obj, "rng",
+                                    "ID of the random number generator backend",
+                                    NULL);
+}
+
+static void spapr_rng_realize(DeviceState *dev, Error **errp)
+{
+
+    sPAPRRngState *rngstate = SPAPR_RNG(dev);
+
+    if (rngstate->use_kvm) {
+        if (kvmppc_enable_hwrng() = 0) {
+            return;
+        }
+        /*
+         * If user specified both, use-kvm and a backend, we fall back to
+         * the backend now. If not, provide an appropriate error message.
+         */
+        if (!rngstate->backend) {
+            error_setg(errp, "Could not initialize in-kernel H_RANDOM call!");
+            return;
+        }
+    }
+
+    if (rngstate->backend) {
+        spapr_register_hypercall(H_RANDOM, h_random);
+    } else {
+        error_setg(errp, "spapr-rng needs an RNG backend!");
+    }
+}
+
+int spapr_rng_populate_dt(void *fdt)
+{
+    int node;
+    int ret;
+
+    node = qemu_fdt_add_subnode(fdt, "/ibm,platform-facilities");
+    if (node <= 0) {
+        return -1;
+    }
+    ret = fdt_setprop_string(fdt, node, "device_type",
+                             "ibm,platform-facilities");
+    ret |= fdt_setprop_cell(fdt, node, "#address-cells", 0x1);
+    ret |= fdt_setprop_cell(fdt, node, "#size-cells", 0x0);
+
+    node = fdt_add_subnode(fdt, node, "ibm,random-v1");
+    if (node <= 0) {
+        return -1;
+    }
+    ret |= fdt_setprop_string(fdt, node, "compatible", "ibm,random");
+
+    return ret ? -1 : 0;
+}
+
+static Property spapr_rng_properties[] = {
+    DEFINE_PROP_BOOL("use-kvm", sPAPRRngState, use_kvm, false),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void spapr_rng_class_init(ObjectClass *oc, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(oc);
+
+    dc->realize = spapr_rng_realize;
+    set_bit(DEVICE_CATEGORY_MISC, dc->categories);
+    dc->props = spapr_rng_properties;
+}
+
+static const TypeInfo spapr_rng_info = {
+    .name          = TYPE_SPAPR_RNG,
+    .parent        = TYPE_DEVICE,
+    .instance_size = sizeof(sPAPRRngState),
+    .instance_init = spapr_rng_instance_init,
+    .class_init    = spapr_rng_class_init,
+};
+
+static void spapr_rng_register_type(void)
+{
+    type_register_static(&spapr_rng_info);
+}
+type_init(spapr_rng_register_type)
include/hw/ppc/spapr.h +4 -0
index 91a61ab..4e8aa2d 100644
@@ -331,6 +331,7 @@ struct sPAPRMachineState {
 #define H_SET_MPP               0x2D0
 #define H_GET_MPP               0x2D4
 #define H_XIRR_X                0x2FC
+#define H_RANDOM                0x300
 #define H_SET_MODE              0x31C
 #define MAX_HCALL_OPCODE        H_SET_MODE
 
@@ -603,10 +604,13 @@ struct sPAPRConfigureConnectorState {
 void spapr_ccs_reset_hook(void *opaque);
 
 #define TYPE_SPAPR_RTC "spapr-rtc"
+#define TYPE_SPAPR_RNG "spapr-rng"
 
 void spapr_rtc_read(DeviceState *dev, struct tm *tm, uint32_t *ns);
 int spapr_rtc_import_offset(DeviceState *dev, int64_t legacy_offset);
 
+int spapr_rng_populate_dt(void *fdt);
+
 #define SPAPR_MEMORY_BLOCK_SIZE (1 << 28) /* 256MB */
 
 #endif /* !defined (__HW_SPAPR_H__) */
target-ppc/kvm.c +9 -0
index 110436d..42f66fe 100644
@@ -2484,3 +2484,12 @@ int kvm_arch_msi_data_to_gsi(uint32_t data)
 {
     return data & 0xffff;
 }
+
+int kvmppc_enable_hwrng(void)
+{
+    if (!kvm_enabled() || !kvm_check_extension(kvm_state, KVM_CAP_PPC_HWRNG)) {
+        return -1;
+    }
+
+    return kvmppc_enable_hcall(kvm_state, H_RANDOM);
+}
target-ppc/kvm_ppc.h +5 -0
index 4d30e27..68836b4 100644
@@ -55,6 +55,7 @@ void kvmppc_hash64_free_pteg(uint64_t token);
 void kvmppc_hash64_write_pte(CPUPPCState *env, target_ulong pte_index,
                              target_ulong pte0, target_ulong pte1);
 bool kvmppc_has_cap_fixup_hcalls(void);
+int kvmppc_enable_hwrng(void);
 
 #else
 
@@ -248,6 +249,10 @@ static inline bool kvmppc_has_cap_fixup_hcalls(void)
     abort();
 }
 
+static inline int kvmppc_enable_hwrng(void)
+{
+    return -1;
+}
 #endif
 
 #ifndef CONFIG_KVM
#2 lore raw
Quoted 37 lines · Greg Kurz
Quoted 22 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

Quoted 8 lines
 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

---

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

Quoted 7 lines

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Signature · 4 lines
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
#3 lore raw
Quoted 46 lines · David Gibson
Quoted 37 lines · Greg Kurz
Quoted 22 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

Quoted 8 lines
 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

---

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

Quoted 7 lines

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

Quoted 7 lines

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

I thought about this question on the weekend and came to the same conclusion. I think if we want to enable this by default, it likely should rather be done at the libvirt level instead?

Thomas

#4 lore raw
Quoted 16 lines · Thomas Huth
Quoted 8 lines
Quoted 7 lines

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

Quoted 7 lines

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Quoted 4 lines

I thought about this question on the weekend and came to the same conclusion. I think if we want to enable this by default, it likely should rather be done at the libvirt level instead?

Adding hardware by default to existing machine types has been a bane to libvirt usage in the past. If upgrading from an old qemu to a new one suddenly turns on new guest-visible hardware with no change to the command line, then libvirt has a much harder time migrating that guest.
It's okay to have a new machine type turn on a feature by default, and
to have knobs so that the feature can be turned on even for older machine types, but experience has shown that any new feature MUST come with knobs and a way to learn if the feature can be turned on/off, rather than just blindly assuming that turning it on is the right thing.

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
Acked-by Tested-by
#5 lore raw
In reply to David Gibson
Quoted 43 lines · On Mon, 21 Sep 2015 12:10:00 +1000 David Gibson <[email protected]>…
Quoted 42 lines · Greg Kurz
Quoted 22 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

Quoted 8 lines
 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

---

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

Quoted 6 lines
Quoted 2 lines

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It

This could be handled with a new spapr property, say 'use-hwrng', defaulting to true.

Quoted 5 lines

also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Yeah, describing the guest hw is really essential for migration... this is best addressed at the libvirt level with a full XML description of the machine... but FWIW if we are talking about running pseries on a POWER8 or newer host, I am not aware about "hwrng-less" boards... but I am probably missing something :)

Back to Thomas' patch, it does the job and brings H_RANDOM, which is currently missing.

I could test both use-kvm and backend flavors (including migration).

Thanks.

--
Greg

#6 lore raw
Quoted 18 lines · Greg Kurz
Quoted 15 lines · On Mon, 21 Sep 2015 12:10:00 +1000 David Gibson <[email protected]>…
Quoted 14 lines · Greg Kurz
Quoted 11 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available.

...

Quoted 19 lines
Quoted 11 lines
Quoted 10 lines

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

Quoted 7 lines

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Quoted 5 lines

Yeah, describing the guest hw is really essential for migration... this is best addressed at the libvirt level with a full XML description of the machine... but FWIW if we are talking about running pseries on a POWER8 or newer host, I am not aware about "hwrng-less" boards... but I am probably missing something :)

Maybe it would be at least ok to enable it by default as long as "-nodefaults" has not been specified as command line option?

Quoted 8 lines

Back to Thomas' patch, it does the job and brings H_RANDOM, which is currently missing.

I could test both use-kvm and backend flavors (including migration).

Thanks!

Thomas

#7 lore raw
Quoted 39 lines · On Mon, 21 Sep 2015 10:26:52 +0200 Thomas Huth <[email protected]> wrote:
Quoted 18 lines · Greg Kurz
Quoted 15 lines · On Mon, 21 Sep 2015 12:10:00 +1000 David Gibson <[email protected]>…
Quoted 14 lines · Greg Kurz
Quoted 11 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available.

...

Quoted 19 lines
Quoted 18 lines
Quoted 10 lines

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Quoted 9 lines
Quoted 5 lines

Yeah, describing the guest hw is really essential for migration... this is best addressed at the libvirt level with a full XML description of the machine... but FWIW if we are talking about running pseries on a POWER8 or newer host, I am not aware about "hwrng-less" boards... but I am probably missing something :)

Maybe it would be at least ok to enable it by default as long as "-nodefaults" has not been specified as command line option?

It makes a lot of sense indeed. I guess David should take your patch as it is now and the default behavior could be a follow up.

Quoted 14 lines
Quoted 8 lines

Back to Thomas' patch, it does the job and brings H_RANDOM, which is currently missing.

I could test both use-kvm and backend flavors (including migration).

Thanks!

Thomas

#8 lore raw
Quoted 42 lines · Greg Kurz
Quoted 39 lines · On Mon, 21 Sep 2015 10:26:52 +0200 Thomas Huth <[email protected]> wrote:
Quoted 18 lines · Greg Kurz
Quoted 15 lines · On Mon, 21 Sep 2015 12:10:00 +1000 David Gibson <[email protected]>…
Quoted 14 lines · Greg Kurz
Quoted 11 lines · On Thu, 17 Sep 2015 10:49:41 +0200 Thomas Huth <[email protected]> wrote:

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available.

...

Quoted 19 lines
Quoted 18 lines
Quoted 10 lines

It is a good thing that the user can choose between in-kernel and backend, and this patch does the work.

This being said, I am not sure about the use case where a user has a hwrng capable platform and wants to run guests without any hwrng support at all is an appropriate default behavior... I guess we will find more users that want in-kernel being the default if it is available.

The patch below modifies yours to do just this: the pseudo-device is only created if hwrng is present and not already created.

I have mixed feelings about this. On the one hand, I agree that it would be nice to allow H_RANDOM support by default. On the other hand the patch below leaves no way to turn it off for testing purposes. It also adds another place where the guest hardware depends on the host configuration, which adds to the already substantial mess of ensuring that source and destination hardware configuration matches for migration.

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines

Yeah, describing the guest hw is really essential for migration... this is best addressed at the libvirt level with a full XML description of the machine... but FWIW if we are talking about running pseries on a POWER8 or newer host, I am not aware about "hwrng-less" boards... but I am probably missing something :)

Maybe it would be at least ok to enable it by default as long as "-nodefaults" has not been specified as command line option?

I like that in principle, but the -nodefaults option isn't exposed outside vl.c

Quoted 2 lines

It makes a lot of sense indeed. I guess David should take your patch as it is now and the default behavior could be a follow up.

That's the plan. I've already taken the base patch.

Signature · 4 lines
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
Applied
#9 lore raw
In reply to Thomas Huth
Quoted 22 lines · Thomas Huth

The PAPR interface defines a hypercall to pass high-quality hardware generated random numbers to guests. Recent kernels can already provide this hypercall to the guest if the right hardware random number generator is available. But in case the user wants to use another source like EGD, or QEMU is running with an older kernel, we should also have this call in QEMU, so that guests that do not support virtio-rng yet can get good random numbers, too.

This patch now adds a new pseudo-device to QEMU that either directly provides this hypercall to the guest or is able to enable the in-kernel hypercall if available. The in-kernel hypercall can be enabled with the use-kvm property, e.g.:

qemu-system-ppc64 -device spapr-rng,use-kvm=true

For handling the hypercall in QEMU instead, a "RngBackend" is required since the hypercall should provide "good" random data instead of pseudo-random (like from a "simple" library function like rand() or g_random_int()). Since there are multiple RngBackends available, the user must select an appropriate back-end via the "rng" property of the device, e.g.:

Quoted 7 lines
 qemu-system-ppc64 -object rng-random,filename=/dev/hwrng,id=gid0 \
                   -device spapr-rng,rng=gid0 ...

See http://wiki.qemu-project.org/Features-Done/VirtIORNG for other example of specifying RngBackends.

Thanks, applied to spapr-next.

Signature · 4 lines
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