* [Qemu-devel] [PATCH 01/12] piix4/ich9: do not raise SMI on ACPI enable/disable commands
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 02/12] target-i386: add support for SMBASE MSR and SMIs Paolo Bonzini
` (10 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
These commands are handled entirely by QEMU. Do not raise an SMI
when they happen, because Windows (at least 2008r2) expects these
commands to work and (depending on the value of APMC_EN at
startup) the firmware might not have installed an SMI handler.
When this happens (e.g. the kernel supports SMIs, or you are using
TCG, but you have used "-machine smm=off") RIP is moved to 0x38000
where there is no code to execute.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/acpi/piix4.c | 3 +++
hw/isa/lpc_ich9.c | 3 +++
2 files changed, 6 insertions(+)
diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
index b730ca6..8a64c60 100644
--- a/hw/acpi/piix4.c
+++ b/hw/acpi/piix4.c
@@ -112,6 +112,9 @@ static void apm_ctrl_changed(uint32_t val, void *arg)
/* ACPI specs 3.0, 4.7.2.5 */
acpi_pm1_cnt_update(&s->ar, val == ACPI_ENABLE, val == ACPI_DISABLE);
+ if (val == ACPI_ENABLE || val == ACPI_DISABLE) {
+ return;
+ }
if (d->config[0x5b] & (1 << 1)) {
if (s->smi_irq) {
diff --git a/hw/isa/lpc_ich9.c b/hw/isa/lpc_ich9.c
index 0269cfe..247158c 100644
--- a/hw/isa/lpc_ich9.c
+++ b/hw/isa/lpc_ich9.c
@@ -378,6 +378,9 @@ static void ich9_apm_ctrl_changed(uint32_t val, void *arg)
acpi_pm1_cnt_update(&lpc->pm.acpi_regs,
val == ICH9_APM_ACPI_ENABLE,
val == ICH9_APM_ACPI_DISABLE);
+ if (val == ICH9_APM_ACPI_ENABLE || val == ICH9_APM_ACPI_DISABLE) {
+ return;
+ }
/* SMI_EN = PMBASE + 30. SMI control and enable register */
if (lpc->pm.smi_en & ICH9_PMIO_SMI_EN_APMC_EN) {
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 02/12] target-i386: add support for SMBASE MSR and SMIs
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 01/12] piix4/ich9: do not raise SMI on ACPI enable/disable commands Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 03/12] kvm-all: put kvm_mem_flags to more work Paolo Bonzini
` (9 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
Apart from the MSR, the smi field of struct kvm_vcpu_events has to be
translated into the corresponding CPUX86State fields. Also,
memory transaction flags depend on SMM state, so pull it from struct
kvm_run on every exit from KVM to userspace.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target-i386/cpu.h | 1 +
target-i386/kvm.c | 94 +++++++++++++++++++++++++++++++++++++++++++++++++------
2 files changed, 86 insertions(+), 9 deletions(-)
diff --git a/target-i386/cpu.h b/target-i386/cpu.h
index df6e885..684e3a0 100644
--- a/target-i386/cpu.h
+++ b/target-i386/cpu.h
@@ -314,6 +314,7 @@
#define MSR_P6_PERFCTR0 0xc1
+#define MSR_IA32_SMBASE 0x9e
#define MSR_MTRRcap 0xfe
#define MSR_MTRRcap_VCNT 8
#define MSR_MTRRcap_FIXRANGE_SUPPORT (1 << 8)
diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index 009bf74..8d23d15 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -74,6 +74,7 @@ static bool has_msr_feature_control;
static bool has_msr_async_pf_en;
static bool has_msr_pv_eoi_en;
static bool has_msr_misc_enable;
+static bool has_msr_smbase;
static bool has_msr_bndcfgs;
static bool has_msr_kvm_steal_time;
static int lm_capable_kernel;
@@ -820,6 +821,10 @@ static int kvm_get_supported_msrs(KVMState *s)
has_msr_tsc_deadline = true;
continue;
}
+ if (kvm_msr_list->indices[i] == MSR_IA32_SMBASE) {
+ has_msr_smbase = true;
+ continue;
+ }
if (kvm_msr_list->indices[i] == MSR_IA32_MISC_ENABLE) {
has_msr_misc_enable = true;
continue;
@@ -1246,6 +1251,9 @@ static int kvm_put_msrs(X86CPU *cpu, int level)
kvm_msr_entry_set(&msrs[n++], MSR_IA32_MISC_ENABLE,
env->msr_ia32_misc_enable);
}
+ if (has_msr_smbase) {
+ kvm_msr_entry_set(&msrs[n++], MSR_IA32_SMBASE, env->smbase);
+ }
if (has_msr_bndcfgs) {
kvm_msr_entry_set(&msrs[n++], MSR_IA32_BNDCFGS, env->msr_bndcfgs);
}
@@ -1607,6 +1615,9 @@ static int kvm_get_msrs(X86CPU *cpu)
if (has_msr_misc_enable) {
msrs[n++].index = MSR_IA32_MISC_ENABLE;
}
+ if (has_msr_smbase) {
+ msrs[n++].index = MSR_IA32_SMBASE;
+ }
if (has_msr_feature_control) {
msrs[n++].index = MSR_IA32_FEATURE_CONTROL;
}
@@ -1761,6 +1772,9 @@ static int kvm_get_msrs(X86CPU *cpu)
case MSR_IA32_MISC_ENABLE:
env->msr_ia32_misc_enable = msrs[i].data;
break;
+ case MSR_IA32_SMBASE:
+ env->smbase = msrs[i].data;
+ break;
case MSR_IA32_FEATURE_CONTROL:
env->msr_ia32_feature_control = msrs[i].data;
break;
@@ -1948,6 +1962,21 @@ static int kvm_put_vcpu_events(X86CPU *cpu, int level)
events.sipi_vector = env->sipi_vector;
+ if (has_msr_smbase) {
+ events.smi.smm = !!(env->hflags & HF_SMM_MASK);
+ events.smi.smm_inside_nmi = !!(env->hflags2 & HF2_SMM_INSIDE_NMI_MASK);
+ if (kvm_irqchip_in_kernel()) {
+ events.smi.pending = cpu->interrupt_request & CPU_INTERRUPT_SMI;
+ events.smi.latched_init = cpu->interrupt_request & CPU_INTERRUPT_INIT;
+ cpu->interrupt_request &= ~(CPU_INTERRUPT_INIT | CPU_INTERRUPT_SMI);
+ } else {
+ /* Keep these in cpu->interrupt_request. */
+ events.smi.pending = 0;
+ events.smi.latched_init = 0;
+ }
+ events.flags |= KVM_VCPUEVENT_VALID_SMM;
+ }
+
events.flags = 0;
if (level >= KVM_PUT_RESET_STATE) {
events.flags |=
@@ -1967,6 +1996,7 @@ static int kvm_get_vcpu_events(X86CPU *cpu)
return 0;
}
+ memset(&events, 0, sizeof(events));
ret = kvm_vcpu_ioctl(CPU(cpu), KVM_GET_VCPU_EVENTS, &events);
if (ret < 0) {
return ret;
@@ -1988,6 +2018,29 @@ static int kvm_get_vcpu_events(X86CPU *cpu)
env->hflags2 &= ~HF2_NMI_MASK;
}
+ if (events.flags & KVM_VCPUEVENT_VALID_SMM) {
+ if (events.smi.smm) {
+ env->hflags |= HF_SMM_MASK;
+ } else {
+ env->hflags &= ~HF_SMM_MASK;
+ }
+ if (events.smi.pending) {
+ cpu_interrupt(CPU(cpu), CPU_INTERRUPT_SMI);
+ } else {
+ cpu_reset_interrupt(CPU(cpu), CPU_INTERRUPT_SMI);
+ }
+ if (events.smi.smm_inside_nmi) {
+ env->hflags2 |= HF2_SMM_INSIDE_NMI_MASK;
+ } else {
+ env->hflags2 &= ~HF2_SMM_INSIDE_NMI_MASK;
+ }
+ if (events.smi.latched_init) {
+ cpu_interrupt(CPU(cpu), CPU_INTERRUPT_INIT);
+ } else {
+ cpu_reset_interrupt(CPU(cpu), CPU_INTERRUPT_INIT);
+ }
+ }
+
env->sipi_vector = events.sipi_vector;
return 0;
@@ -2191,13 +2244,24 @@ void kvm_arch_pre_run(CPUState *cpu, struct kvm_run *run)
int ret;
/* Inject NMI */
- if (cpu->interrupt_request & CPU_INTERRUPT_NMI) {
- cpu->interrupt_request &= ~CPU_INTERRUPT_NMI;
- DPRINTF("injected NMI\n");
- ret = kvm_vcpu_ioctl(cpu, KVM_NMI);
- if (ret < 0) {
- fprintf(stderr, "KVM: injection failed, NMI lost (%s)\n",
- strerror(-ret));
+ if (cpu->interrupt_request & (CPU_INTERRUPT_NMI | CPU_INTERRUPT_SMI)) {
+ if (cpu->interrupt_request & CPU_INTERRUPT_NMI) {
+ cpu->interrupt_request &= ~CPU_INTERRUPT_NMI;
+ DPRINTF("injected NMI\n");
+ ret = kvm_vcpu_ioctl(cpu, KVM_NMI);
+ if (ret < 0) {
+ fprintf(stderr, "KVM: injection failed, NMI lost (%s)\n",
+ strerror(-ret));
+ }
+ }
+ if (cpu->interrupt_request & CPU_INTERRUPT_SMI) {
+ cpu->interrupt_request &= ~CPU_INTERRUPT_SMI;
+ DPRINTF("injected SMI\n");
+ ret = kvm_vcpu_ioctl(cpu, KVM_SMI);
+ if (ret < 0) {
+ fprintf(stderr, "KVM: injection failed, SMI lost (%s)\n",
+ strerror(-ret));
+ }
}
}
@@ -2206,7 +2270,13 @@ void kvm_arch_pre_run(CPUState *cpu, struct kvm_run *run)
* pending TPR access reports.
*/
if (cpu->interrupt_request & (CPU_INTERRUPT_INIT | CPU_INTERRUPT_TPR)) {
- cpu->exit_request = 1;
+ if ((cpu->interrupt_request & CPU_INTERRUPT_INIT) &&
+ !(env->hflags & HF_SMM_MASK)) {
+ cpu->exit_request = 1;
+ }
+ if (cpu->interrupt_request & CPU_INTERRUPT_TPR) {
+ cpu->exit_request = 1;
+ }
}
if (!kvm_irqchip_in_kernel()) {
@@ -2252,6 +2322,11 @@ MemTxAttrs kvm_arch_post_run(CPUState *cpu, struct kvm_run *run)
X86CPU *x86_cpu = X86_CPU(cpu);
CPUX86State *env = &x86_cpu->env;
+ if (run->flags & KVM_RUN_X86_SMM) {
+ env->hflags |= HF_SMM_MASK;
+ } else {
+ env->hflags &= HF_SMM_MASK;
+ }
if (run->if_flag) {
env->eflags |= IF_MASK;
} else {
@@ -2290,7 +2365,8 @@ int kvm_arch_process_async_events(CPUState *cs)
}
}
- if (cs->interrupt_request & CPU_INTERRUPT_INIT) {
+ if ((cs->interrupt_request & CPU_INTERRUPT_INIT) &&
+ !(env->hflags & HF_SMM_MASK)) {
kvm_cpu_synchronize_state(cs);
do_cpu_init(cpu);
}
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 03/12] kvm-all: put kvm_mem_flags to more work
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 01/12] piix4/ich9: do not raise SMI on ACPI enable/disable commands Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 02/12] target-i386: add support for SMBASE MSR and SMIs Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 04/12] kvm-all: remove useless typedef Paolo Bonzini
` (8 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: Andrew Jones, lersek, kraxel, mst
From: Andrew Jones <drjones@redhat.com>
Currently kvm_mem_flags just translates bools to bits, let's
make it also determine the bools first. This avoids its parameter
list growing each time we add a flag.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
kvm-all.c | 42 ++++++++++++++++++++----------------------
1 file changed, 20 insertions(+), 22 deletions(-)
diff --git a/kvm-all.c b/kvm-all.c
index 8af8e8d..20012c6 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -291,45 +291,47 @@ err:
* dirty pages logging control
*/
-static int kvm_mem_flags(KVMState *s, bool log_dirty, bool readonly)
+static int kvm_mem_flags(MemoryRegion *mr)
{
+ bool readonly = mr->readonly || memory_region_is_romd(mr);
int flags = 0;
- flags = log_dirty ? KVM_MEM_LOG_DIRTY_PAGES : 0;
+
+ if (memory_region_get_dirty_log_mask(mr) != 0) {
+ flags |= KVM_MEM_LOG_DIRTY_PAGES;
+ }
if (readonly && kvm_readonly_mem_allowed) {
flags |= KVM_MEM_READONLY;
}
return flags;
}
-static int kvm_slot_dirty_pages_log_change(KVMSlot *mem, bool log_dirty)
+static int kvm_slot_update_flags(KVMSlot *mem, MemoryRegion *mr)
{
KVMState *s = kvm_state;
- int flags, mask = KVM_MEM_LOG_DIRTY_PAGES;
int old_flags;
old_flags = mem->flags;
-
- flags = (mem->flags & ~mask) | kvm_mem_flags(s, log_dirty, false);
- mem->flags = flags;
+ mem->flags = kvm_mem_flags(mr);
/* If nothing changed effectively, no need to issue ioctl */
- if (flags == old_flags) {
+ if (mem->flags == old_flags) {
return 0;
}
return kvm_set_user_memory_region(s, mem);
}
-static int kvm_dirty_pages_log_change(hwaddr phys_addr,
- ram_addr_t size, bool log_dirty)
+static int kvm_section_update_flags(MemoryRegionSection *section)
{
KVMState *s = kvm_state;
+ hwaddr phys_addr = section->offset_within_address_space;
+ ram_addr_t size = int128_get64(section->size);
KVMSlot *mem = kvm_lookup_matching_slot(s, phys_addr, phys_addr + size);
if (mem == NULL) {
return 0;
} else {
- return kvm_slot_dirty_pages_log_change(mem, log_dirty);
+ return kvm_slot_update_flags(mem, section->mr);
}
}
@@ -343,8 +345,7 @@ static void kvm_log_start(MemoryListener *listener,
return;
}
- r = kvm_dirty_pages_log_change(section->offset_within_address_space,
- int128_get64(section->size), true);
+ r = kvm_section_update_flags(section);
if (r < 0) {
abort();
}
@@ -360,8 +361,7 @@ static void kvm_log_stop(MemoryListener *listener,
return;
}
- r = kvm_dirty_pages_log_change(section->offset_within_address_space,
- int128_get64(section->size), false);
+ r = kvm_section_update_flags(section);
if (r < 0) {
abort();
}
@@ -638,9 +638,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
KVMSlot *mem, old;
int err;
MemoryRegion *mr = section->mr;
- bool log_dirty = memory_region_get_dirty_log_mask(mr) != 0;
bool writeable = !mr->readonly && !mr->rom_device;
- bool readonly_flag = mr->readonly || memory_region_is_romd(mr);
hwaddr start_addr = section->offset_within_address_space;
ram_addr_t size = int128_get64(section->size);
void *ram = NULL;
@@ -684,7 +682,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
(ram - start_addr == mem->ram - mem->start_addr)) {
/* The new slot fits into the existing one and comes with
* identical parameters - update flags and done. */
- kvm_slot_dirty_pages_log_change(mem, log_dirty);
+ kvm_slot_update_flags(mem, mr);
return;
}
@@ -717,7 +715,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
mem->memory_size = old.memory_size;
mem->start_addr = old.start_addr;
mem->ram = old.ram;
- mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);
+ mem->flags = kvm_mem_flags(mr);
err = kvm_set_user_memory_region(s, mem);
if (err) {
@@ -738,7 +736,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
mem->memory_size = start_addr - old.start_addr;
mem->start_addr = old.start_addr;
mem->ram = old.ram;
- mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);
+ mem->flags = kvm_mem_flags(mr);
err = kvm_set_user_memory_region(s, mem);
if (err) {
@@ -762,7 +760,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
size_delta = mem->start_addr - old.start_addr;
mem->memory_size = old.memory_size - size_delta;
mem->ram = old.ram + size_delta;
- mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);
+ mem->flags = kvm_mem_flags(mr);
err = kvm_set_user_memory_region(s, mem);
if (err) {
@@ -784,7 +782,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
mem->memory_size = size;
mem->start_addr = start_addr;
mem->ram = ram;
- mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);
+ mem->flags = kvm_mem_flags(mr);
err = kvm_set_user_memory_region(s, mem);
if (err) {
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 04/12] kvm-all: remove useless typedef
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (2 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 03/12] kvm-all: put kvm_mem_flags to more work Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 05/12] kvm-all: move internal types to kvm_int.h Paolo Bonzini
` (7 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
kvm-all.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/kvm-all.c b/kvm-all.c
index 20012c6..2e13aff 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -69,8 +69,6 @@ typedef struct KVMSlot
int flags;
} KVMSlot;
-typedef struct kvm_dirty_log KVMDirtyLog;
-
struct KVMState
{
AccelState parent_obj;
@@ -393,7 +391,7 @@ static int kvm_physical_sync_dirty_bitmap(MemoryRegionSection *section)
{
KVMState *s = kvm_state;
unsigned long size, allocated_size = 0;
- KVMDirtyLog d = {};
+ struct kvm_dirty_log d = {};
KVMSlot *mem;
int ret = 0;
hwaddr start_addr = section->offset_within_address_space;
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 05/12] kvm-all: move internal types to kvm_int.h
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (3 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 04/12] kvm-all: remove useless typedef Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 06/12] kvm-all: make KVM's memory listener more generic Paolo Bonzini
` (6 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
i386 code will have to define a different KVMMemoryListener. Create
an internal header so that KVMSlot is not exposed outside.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
include/sysemu/kvm_int.h | 30 ++++++++++++++++++++++++++++++
kvm-all.c | 18 +-----------------
2 files changed, 31 insertions(+), 17 deletions(-)
create mode 100644 include/sysemu/kvm_int.h
diff --git a/include/sysemu/kvm_int.h b/include/sysemu/kvm_int.h
new file mode 100644
index 0000000..e8dcbd7
--- /dev/null
+++ b/include/sysemu/kvm_int.h
@@ -0,0 +1,30 @@
+/*
+ * Internal definitions for a target's KVM support
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#ifndef QEMU_KVM_INT_H
+#define QEMU_KVM_INT_H
+
+#include "sysemu/sysemu.h"
+#include "sysemu/accel.h"
+#include "sysemu/kvm.h"
+
+typedef struct KVMSlot
+{
+ hwaddr start_addr;
+ ram_addr_t memory_size;
+ void *ram;
+ int slot;
+ int flags;
+} KVMSlot;
+
+#define TYPE_KVM_ACCEL ACCEL_CLASS_NAME("kvm")
+
+#define KVM_STATE(obj) \
+ OBJECT_CHECK(KVMState, (obj), TYPE_KVM_ACCEL)
+
+#endif
diff --git a/kvm-all.c b/kvm-all.c
index 2e13aff..a6e92d2 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -24,13 +24,11 @@
#include "qemu/atomic.h"
#include "qemu/option.h"
#include "qemu/config-file.h"
-#include "sysemu/sysemu.h"
-#include "sysemu/accel.h"
#include "hw/hw.h"
#include "hw/pci/msi.h"
#include "hw/s390x/adapter.h"
#include "exec/gdbstub.h"
-#include "sysemu/kvm.h"
+#include "sysemu/kvm_int.h"
#include "qemu/bswap.h"
#include "exec/memory.h"
#include "exec/ram_addr.h"
@@ -60,15 +58,6 @@
#define KVM_MSI_HASHTAB_SIZE 256
-typedef struct KVMSlot
-{
- hwaddr start_addr;
- ram_addr_t memory_size;
- void *ram;
- int slot;
- int flags;
-} KVMSlot;
-
struct KVMState
{
AccelState parent_obj;
@@ -106,11 +95,6 @@ struct KVMState
#endif
};
-#define TYPE_KVM_ACCEL ACCEL_CLASS_NAME("kvm")
-
-#define KVM_STATE(obj) \
- OBJECT_CHECK(KVMState, (obj), TYPE_KVM_ACCEL)
-
KVMState *kvm_state;
bool kvm_kernel_irqchip;
bool kvm_async_interrupts_allowed;
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 06/12] kvm-all: make KVM's memory listener more generic
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (4 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 05/12] kvm-all: move internal types to kvm_int.h Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 07/12] kvm-all: add support for multiple address spaces Paolo Bonzini
` (5 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
No semantic change, but s->slots moves into a new struct
KVMMemoryListener. KVM's memory listener becomes a member of struct
KVMState, and becomes of type KVMMemoryListener.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
include/sysemu/kvm_int.h | 5 ++
kvm-all.c | 151 ++++++++++++++++++++++++++++-------------------
2 files changed, 94 insertions(+), 62 deletions(-)
diff --git a/include/sysemu/kvm_int.h b/include/sysemu/kvm_int.h
index e8dcbd7..d15f8bb 100644
--- a/include/sysemu/kvm_int.h
+++ b/include/sysemu/kvm_int.h
@@ -22,6 +22,11 @@ typedef struct KVMSlot
int flags;
} KVMSlot;
+typedef struct KVMMemoryListener {
+ MemoryListener listener;
+ KVMSlot *slots;
+} KVMMemoryListener;
+
#define TYPE_KVM_ACCEL ACCEL_CLASS_NAME("kvm")
#define KVM_STATE(obj) \
diff --git a/kvm-all.c b/kvm-all.c
index a6e92d2..ddb1848 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -62,7 +62,6 @@ struct KVMState
{
AccelState parent_obj;
- KVMSlot *slots;
int nr_slots;
int fd;
int vmfd;
@@ -93,6 +92,7 @@ struct KVMState
QTAILQ_HEAD(msi_hashtab, KVMMSIRoute) msi_hashtab[KVM_MSI_HASHTAB_SIZE];
bool direct_msi;
#endif
+ KVMMemoryListener memory_listener;
};
KVMState *kvm_state;
@@ -115,13 +115,14 @@ static const KVMCapabilityInfo kvm_required_capabilites[] = {
KVM_CAP_LAST_INFO
};
-static KVMSlot *kvm_get_free_slot(KVMState *s)
+static KVMSlot *kvm_get_free_slot(KVMMemoryListener *kml)
{
+ KVMState *s = kvm_state;
int i;
for (i = 0; i < s->nr_slots; i++) {
- if (s->slots[i].memory_size == 0) {
- return &s->slots[i];
+ if (kml->slots[i].memory_size == 0) {
+ return &kml->slots[i];
}
}
@@ -130,12 +131,14 @@ static KVMSlot *kvm_get_free_slot(KVMState *s)
bool kvm_has_free_slot(MachineState *ms)
{
- return kvm_get_free_slot(KVM_STATE(ms->accelerator));
+ KVMState *s = KVM_STATE(ms->accelerator);
+
+ return kvm_get_free_slot(&s->memory_listener);
}
-static KVMSlot *kvm_alloc_slot(KVMState *s)
+static KVMSlot *kvm_alloc_slot(KVMMemoryListener *kml)
{
- KVMSlot *slot = kvm_get_free_slot(s);
+ KVMSlot *slot = kvm_get_free_slot(kml);
if (slot) {
return slot;
@@ -145,14 +148,15 @@ static KVMSlot *kvm_alloc_slot(KVMState *s)
abort();
}
-static KVMSlot *kvm_lookup_matching_slot(KVMState *s,
+static KVMSlot *kvm_lookup_matching_slot(KVMMemoryListener *kml,
hwaddr start_addr,
hwaddr end_addr)
{
+ KVMState *s = kvm_state;
int i;
for (i = 0; i < s->nr_slots; i++) {
- KVMSlot *mem = &s->slots[i];
+ KVMSlot *mem = &kml->slots[i];
if (start_addr == mem->start_addr &&
end_addr == mem->start_addr + mem->memory_size) {
@@ -166,15 +170,16 @@ static KVMSlot *kvm_lookup_matching_slot(KVMState *s,
/*
* Find overlapping slot with lowest start address
*/
-static KVMSlot *kvm_lookup_overlapping_slot(KVMState *s,
+static KVMSlot *kvm_lookup_overlapping_slot(KVMMemoryListener *kml,
hwaddr start_addr,
hwaddr end_addr)
{
+ KVMState *s = kvm_state;
KVMSlot *found = NULL;
int i;
for (i = 0; i < s->nr_slots; i++) {
- KVMSlot *mem = &s->slots[i];
+ KVMSlot *mem = &kml->slots[i];
if (mem->memory_size == 0 ||
(found && found->start_addr < mem->start_addr)) {
@@ -193,10 +198,11 @@ static KVMSlot *kvm_lookup_overlapping_slot(KVMState *s,
int kvm_physical_memory_addr_from_host(KVMState *s, void *ram,
hwaddr *phys_addr)
{
+ KVMMemoryListener *kml = &s->memory_listener;
int i;
for (i = 0; i < s->nr_slots; i++) {
- KVMSlot *mem = &s->slots[i];
+ KVMSlot *mem = &kml->slots[i];
if (ram >= mem->ram && ram < mem->ram + mem->memory_size) {
*phys_addr = mem->start_addr + (ram - mem->ram);
@@ -207,8 +213,9 @@ int kvm_physical_memory_addr_from_host(KVMState *s, void *ram,
return 0;
}
-static int kvm_set_user_memory_region(KVMState *s, KVMSlot *slot)
+static int kvm_set_user_memory_region(KVMMemoryListener *kml, KVMSlot *slot)
{
+ KVMState *s = kvm_state;
struct kvm_userspace_memory_region mem;
mem.slot = slot->slot;
@@ -287,9 +294,9 @@ static int kvm_mem_flags(MemoryRegion *mr)
return flags;
}
-static int kvm_slot_update_flags(KVMSlot *mem, MemoryRegion *mr)
+static int kvm_slot_update_flags(KVMMemoryListener *kml, KVMSlot *mem,
+ MemoryRegion *mr)
{
- KVMState *s = kvm_state;
int old_flags;
old_flags = mem->flags;
@@ -300,20 +307,20 @@ static int kvm_slot_update_flags(KVMSlot *mem, MemoryRegion *mr)
return 0;
}
- return kvm_set_user_memory_region(s, mem);
+ return kvm_set_user_memory_region(kml, mem);
}
-static int kvm_section_update_flags(MemoryRegionSection *section)
+static int kvm_section_update_flags(KVMMemoryListener *kml,
+ MemoryRegionSection *section)
{
- KVMState *s = kvm_state;
hwaddr phys_addr = section->offset_within_address_space;
ram_addr_t size = int128_get64(section->size);
- KVMSlot *mem = kvm_lookup_matching_slot(s, phys_addr, phys_addr + size);
+ KVMSlot *mem = kvm_lookup_matching_slot(kml, phys_addr, phys_addr + size);
if (mem == NULL) {
return 0;
} else {
- return kvm_slot_update_flags(mem, section->mr);
+ return kvm_slot_update_flags(kml, mem, section->mr);
}
}
@@ -321,13 +328,14 @@ static void kvm_log_start(MemoryListener *listener,
MemoryRegionSection *section,
int old, int new)
{
+ KVMMemoryListener *kml = container_of(listener, KVMMemoryListener, listener);
int r;
if (old != 0) {
return;
}
- r = kvm_section_update_flags(section);
+ r = kvm_section_update_flags(kml, section);
if (r < 0) {
abort();
}
@@ -337,13 +345,14 @@ static void kvm_log_stop(MemoryListener *listener,
MemoryRegionSection *section,
int old, int new)
{
+ KVMMemoryListener *kml = container_of(listener, KVMMemoryListener, listener);
int r;
if (new != 0) {
return;
}
- r = kvm_section_update_flags(section);
+ r = kvm_section_update_flags(kml, section);
if (r < 0) {
abort();
}
@@ -371,7 +380,8 @@ static int kvm_get_dirty_pages_log_range(MemoryRegionSection *section,
* @start_add: start of logged region.
* @end_addr: end of logged region.
*/
-static int kvm_physical_sync_dirty_bitmap(MemoryRegionSection *section)
+static int kvm_physical_sync_dirty_bitmap(KVMMemoryListener *kml,
+ MemoryRegionSection *section)
{
KVMState *s = kvm_state;
unsigned long size, allocated_size = 0;
@@ -383,7 +393,7 @@ static int kvm_physical_sync_dirty_bitmap(MemoryRegionSection *section)
d.dirty_bitmap = NULL;
while (start_addr < end_addr) {
- mem = kvm_lookup_overlapping_slot(s, start_addr, end_addr);
+ mem = kvm_lookup_overlapping_slot(kml, start_addr, end_addr);
if (mem == NULL) {
break;
}
@@ -614,7 +624,8 @@ kvm_check_extension_list(KVMState *s, const KVMCapabilityInfo *list)
return NULL;
}
-static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
+static void kvm_set_phys_mem(KVMMemoryListener *kml,
+ MemoryRegionSection *section, bool add)
{
KVMState *s = kvm_state;
KVMSlot *mem, old;
@@ -654,7 +665,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
ram = memory_region_get_ram_ptr(mr) + section->offset_within_region + delta;
while (1) {
- mem = kvm_lookup_overlapping_slot(s, start_addr, start_addr + size);
+ mem = kvm_lookup_overlapping_slot(kml, start_addr, start_addr + size);
if (!mem) {
break;
}
@@ -664,19 +675,19 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
(ram - start_addr == mem->ram - mem->start_addr)) {
/* The new slot fits into the existing one and comes with
* identical parameters - update flags and done. */
- kvm_slot_update_flags(mem, mr);
+ kvm_slot_update_flags(kml, mem, mr);
return;
}
old = *mem;
if (mem->flags & KVM_MEM_LOG_DIRTY_PAGES) {
- kvm_physical_sync_dirty_bitmap(section);
+ kvm_physical_sync_dirty_bitmap(kml, section);
}
/* unregister the overlapping slot */
mem->memory_size = 0;
- err = kvm_set_user_memory_region(s, mem);
+ err = kvm_set_user_memory_region(kml, mem);
if (err) {
fprintf(stderr, "%s: error unregistering overlapping slot: %s\n",
__func__, strerror(-err));
@@ -693,13 +704,13 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
* - and actually require a recent KVM version. */
if (s->broken_set_mem_region &&
old.start_addr == start_addr && old.memory_size < size && add) {
- mem = kvm_alloc_slot(s);
+ mem = kvm_alloc_slot(kml);
mem->memory_size = old.memory_size;
mem->start_addr = old.start_addr;
mem->ram = old.ram;
mem->flags = kvm_mem_flags(mr);
- err = kvm_set_user_memory_region(s, mem);
+ err = kvm_set_user_memory_region(kml, mem);
if (err) {
fprintf(stderr, "%s: error updating slot: %s\n", __func__,
strerror(-err));
@@ -714,13 +725,13 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
/* register prefix slot */
if (old.start_addr < start_addr) {
- mem = kvm_alloc_slot(s);
+ mem = kvm_alloc_slot(kml);
mem->memory_size = start_addr - old.start_addr;
mem->start_addr = old.start_addr;
mem->ram = old.ram;
mem->flags = kvm_mem_flags(mr);
- err = kvm_set_user_memory_region(s, mem);
+ err = kvm_set_user_memory_region(kml, mem);
if (err) {
fprintf(stderr, "%s: error registering prefix slot: %s\n",
__func__, strerror(-err));
@@ -737,14 +748,14 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
if (old.start_addr + old.memory_size > start_addr + size) {
ram_addr_t size_delta;
- mem = kvm_alloc_slot(s);
+ mem = kvm_alloc_slot(kml);
mem->start_addr = start_addr + size;
size_delta = mem->start_addr - old.start_addr;
mem->memory_size = old.memory_size - size_delta;
mem->ram = old.ram + size_delta;
mem->flags = kvm_mem_flags(mr);
- err = kvm_set_user_memory_region(s, mem);
+ err = kvm_set_user_memory_region(kml, mem);
if (err) {
fprintf(stderr, "%s: error registering suffix slot: %s\n",
__func__, strerror(-err));
@@ -760,13 +771,13 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
if (!add) {
return;
}
- mem = kvm_alloc_slot(s);
+ mem = kvm_alloc_slot(kml);
mem->memory_size = size;
mem->start_addr = start_addr;
mem->ram = ram;
mem->flags = kvm_mem_flags(mr);
- err = kvm_set_user_memory_region(s, mem);
+ err = kvm_set_user_memory_region(kml, mem);
if (err) {
fprintf(stderr, "%s: error registering slot: %s\n", __func__,
strerror(-err));
@@ -777,23 +788,28 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
static void kvm_region_add(MemoryListener *listener,
MemoryRegionSection *section)
{
+ KVMMemoryListener *kml = container_of(listener, KVMMemoryListener, listener);
+
memory_region_ref(section->mr);
- kvm_set_phys_mem(section, true);
+ kvm_set_phys_mem(kml, section, true);
}
static void kvm_region_del(MemoryListener *listener,
MemoryRegionSection *section)
{
- kvm_set_phys_mem(section, false);
+ KVMMemoryListener *kml = container_of(listener, KVMMemoryListener, listener);
+
+ kvm_set_phys_mem(kml, section, false);
memory_region_unref(section->mr);
}
static void kvm_log_sync(MemoryListener *listener,
MemoryRegionSection *section)
{
+ KVMMemoryListener *kml = container_of(listener, KVMMemoryListener, listener);
int r;
- r = kvm_physical_sync_dirty_bitmap(section);
+ r = kvm_physical_sync_dirty_bitmap(kml, section);
if (r < 0) {
abort();
}
@@ -868,18 +884,27 @@ static void kvm_io_ioeventfd_del(MemoryListener *listener,
}
}
-static MemoryListener kvm_memory_listener = {
- .region_add = kvm_region_add,
- .region_del = kvm_region_del,
- .log_start = kvm_log_start,
- .log_stop = kvm_log_stop,
- .log_sync = kvm_log_sync,
- .eventfd_add = kvm_mem_ioeventfd_add,
- .eventfd_del = kvm_mem_ioeventfd_del,
- .coalesced_mmio_add = kvm_coalesce_mmio_region,
- .coalesced_mmio_del = kvm_uncoalesce_mmio_region,
- .priority = 10,
-};
+static void kvm_memory_listener_register(KVMState *s,
+ KVMMemoryListener *kml,
+ AddressSpace *as)
+{
+ int i;
+
+ kml->slots = g_malloc0(s->nr_slots * sizeof(KVMSlot));
+
+ for (i = 0; i < s->nr_slots; i++) {
+ kml->slots[i].slot = i;
+ }
+
+ kml->listener.region_add = kvm_region_add;
+ kml->listener.region_del = kvm_region_del;
+ kml->listener.log_start = kvm_log_start;
+ kml->listener.log_stop = kvm_log_stop;
+ kml->listener.log_sync = kvm_log_sync;
+ kml->listener.priority = 10;
+
+ memory_listener_register(&kml->listener, as);
+}
static MemoryListener kvm_io_listener = {
.eventfd_add = kvm_io_ioeventfd_add,
@@ -1387,7 +1412,7 @@ static int kvm_init(MachineState *ms)
KVMState *s;
const KVMCapabilityInfo *missing_cap;
int ret;
- int i, type = 0;
+ int type = 0;
const char *kvm_type;
s = KVM_STATE(ms->accelerator);
@@ -1436,12 +1461,6 @@ static int kvm_init(MachineState *ms)
s->nr_slots = 32;
}
- s->slots = g_malloc0(s->nr_slots * sizeof(KVMSlot));
-
- for (i = 0; i < s->nr_slots; i++) {
- s->slots[i].slot = i;
- }
-
/* check the vcpu limits */
soft_vcpus_limit = kvm_recommended_vcpus(s);
hard_vcpus_limit = kvm_max_vcpus(s);
@@ -1579,8 +1598,16 @@ static int kvm_init(MachineState *ms)
}
kvm_state = s;
- memory_listener_register(&kvm_memory_listener, &address_space_memory);
- memory_listener_register(&kvm_io_listener, &address_space_io);
+
+ s->memory_listener.listener.eventfd_add = kvm_mem_ioeventfd_add;
+ s->memory_listener.listener.eventfd_del = kvm_mem_ioeventfd_del;
+ s->memory_listener.listener.coalesced_mmio_add = kvm_coalesce_mmio_region;
+ s->memory_listener.listener.coalesced_mmio_del = kvm_uncoalesce_mmio_region;
+
+ kvm_memory_listener_register(s, &s->memory_listener,
+ &address_space_memory);
+ memory_listener_register(&kvm_io_listener,
+ &address_space_io);
s->many_ioeventfds = kvm_check_many_ioeventfds();
@@ -1596,7 +1623,7 @@ err:
if (s->fd != -1) {
close(s->fd);
}
- g_free(s->slots);
+ g_free(s->memory_listener.slots);
return ret;
}
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 07/12] kvm-all: add support for multiple address spaces
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (5 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 06/12] kvm-all: make KVM's memory listener more generic Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 08/12] kvm-all: kvm_irqchip_create is not expected to fail Paolo Bonzini
` (4 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
Make kvm_memory_listener_register public, and assign a kernel
address space id to each KVMMemoryListener.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
include/sysemu/kvm_int.h | 4 ++++
kvm-all.c | 13 ++++++-------
2 files changed, 10 insertions(+), 7 deletions(-)
diff --git a/include/sysemu/kvm_int.h b/include/sysemu/kvm_int.h
index d15f8bb..888557a 100644
--- a/include/sysemu/kvm_int.h
+++ b/include/sysemu/kvm_int.h
@@ -25,6 +25,7 @@ typedef struct KVMSlot
typedef struct KVMMemoryListener {
MemoryListener listener;
KVMSlot *slots;
+ int as_id;
} KVMMemoryListener;
#define TYPE_KVM_ACCEL ACCEL_CLASS_NAME("kvm")
@@ -32,4 +33,7 @@ typedef struct KVMMemoryListener {
#define KVM_STATE(obj) \
OBJECT_CHECK(KVMState, (obj), TYPE_KVM_ACCEL)
+void kvm_memory_listener_register(KVMState *s, KVMMemoryListener *kml,
+ AddressSpace *as, int as_id);
+
#endif
diff --git a/kvm-all.c b/kvm-all.c
index ddb1848..9d88b97 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -218,7 +218,7 @@ static int kvm_set_user_memory_region(KVMMemoryListener *kml, KVMSlot *slot)
KVMState *s = kvm_state;
struct kvm_userspace_memory_region mem;
- mem.slot = slot->slot;
+ mem.slot = slot->slot | (kml->as_id << 16);
mem.guest_phys_addr = slot->start_addr;
mem.userspace_addr = (unsigned long)slot->ram;
mem.flags = slot->flags;
@@ -420,8 +420,7 @@ static int kvm_physical_sync_dirty_bitmap(KVMMemoryListener *kml,
allocated_size = size;
memset(d.dirty_bitmap, 0, allocated_size);
- d.slot = mem->slot;
-
+ d.slot = mem->slot | (kml->as_id << 16);
if (kvm_vm_ioctl(s, KVM_GET_DIRTY_LOG, &d) == -1) {
DPRINTF("ioctl failed %d\n", errno);
ret = -1;
@@ -884,13 +883,13 @@ static void kvm_io_ioeventfd_del(MemoryListener *listener,
}
}
-static void kvm_memory_listener_register(KVMState *s,
- KVMMemoryListener *kml,
- AddressSpace *as)
+void kvm_memory_listener_register(KVMState *s, KVMMemoryListener *kml,
+ AddressSpace *as, int as_id)
{
int i;
kml->slots = g_malloc0(s->nr_slots * sizeof(KVMSlot));
+ kml->as_id = as_id;
for (i = 0; i < s->nr_slots; i++) {
kml->slots[i].slot = i;
@@ -1605,7 +1604,7 @@ static int kvm_init(MachineState *ms)
s->memory_listener.listener.coalesced_mmio_del = kvm_uncoalesce_mmio_region;
kvm_memory_listener_register(s, &s->memory_listener,
- &address_space_memory);
+ &address_space_memory, 0);
memory_listener_register(&kvm_io_listener,
&address_space_io);
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 08/12] kvm-all: kvm_irqchip_create is not expected to fail
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (6 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 07/12] kvm-all: add support for multiple address spaces Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 09/12] target-i386: register a separate KVM address space including SMRAM regions Paolo Bonzini
` (3 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
KVM_CREATE_IRQCHIP should never fail, and so should its userspace
wrapper kvm_irqchip_create. The function does not do anything
if the irqchip capability is not available, as is the case for PPC.
With this patch, kvm_arch_init can allocate memory and it will not
be leaked.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
kvm-all.c | 35 ++++++++++++++++++-----------------
1 file changed, 18 insertions(+), 17 deletions(-)
diff --git a/kvm-all.c b/kvm-all.c
index 9d88b97..d263908 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -1342,27 +1342,31 @@ int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq)
false);
}
-static int kvm_irqchip_create(MachineState *machine, KVMState *s)
+static void kvm_irqchip_create(MachineState *machine, KVMState *s)
{
int ret;
- if (!machine_kernel_irqchip_allowed(machine) ||
- (!kvm_check_extension(s, KVM_CAP_IRQCHIP) &&
- (kvm_vm_enable_cap(s, KVM_CAP_S390_IRQCHIP, 0) < 0))) {
- return 0;
+ if (kvm_check_extension(s, KVM_CAP_IRQCHIP)) {
+ ;
+ } else if (kvm_check_extension(s, KVM_CAP_S390_IRQCHIP)) {
+ ret = kvm_vm_enable_cap(s, KVM_CAP_S390_IRQCHIP, 0);
+ if (ret < 0) {
+ fprintf(stderr, "Enable kernel irqchip failed: %s\n", strerror(-ret));
+ exit(1);
+ }
+ } else {
+ return;
}
/* First probe and see if there's a arch-specific hook to create the
* in-kernel irqchip for us */
ret = kvm_arch_irqchip_create(s);
- if (ret < 0) {
- return ret;
- } else if (ret == 0) {
+ if (ret == 0) {
ret = kvm_vm_ioctl(s, KVM_CREATE_IRQCHIP);
- if (ret < 0) {
- fprintf(stderr, "Create kernel irqchip failed\n");
- return ret;
- }
+ }
+ if (ret < 0) {
+ fprintf(stderr, "Create kernel irqchip failed: %s\n", strerror(-ret));
+ exit(1);
}
kvm_kernel_irqchip = true;
@@ -1373,8 +1377,6 @@ static int kvm_irqchip_create(MachineState *machine, KVMState *s)
kvm_halt_in_kernel_allowed = true;
kvm_init_irq_routing(s);
-
- return 0;
}
/* Find number of supported CPUs using the recommended
@@ -1591,9 +1593,8 @@ static int kvm_init(MachineState *ms)
goto err;
}
- ret = kvm_irqchip_create(ms, s);
- if (ret < 0) {
- goto err;
+ if (machine_kernel_irqchip_allowed(ms)) {
+ kvm_irqchip_create(ms, s);
}
kvm_state = s;
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 09/12] target-i386: register a separate KVM address space including SMRAM regions
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (7 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 08/12] kvm-all: kvm_irqchip_create is not expected to fail Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 10/12] pc_piix: rename kvm_enabled to smm_enabled Paolo Bonzini
` (2 subsequent siblings)
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target-i386/kvm.c | 41 ++++++++++++++++++++++++++++++++++++++++-
1 file changed, 40 insertions(+), 1 deletion(-)
diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index 8d23d15..8fa24cb 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -22,7 +22,7 @@
#include "qemu-common.h"
#include "sysemu/sysemu.h"
-#include "sysemu/kvm.h"
+#include "sysemu/kvm_int.h"
#include "kvm_i386.h"
#include "cpu.h"
#include "exec/gdbstub.h"
@@ -846,6 +846,40 @@ static int kvm_get_supported_msrs(KVMState *s)
return ret;
}
+static Notifier smram_machine_done;
+static KVMMemoryListener smram_listener;
+static AddressSpace smram_address_space;
+static MemoryRegion smram_as_root;
+static MemoryRegion smram_as_mem;
+
+static void register_smram_listener(Notifier *n, void *unused)
+{
+ MemoryRegion *smram =
+ (MemoryRegion *) object_resolve_path("/machine/smram", NULL);
+
+ /* Outer container... */
+ memory_region_init(&smram_as_root, OBJECT(kvm_state), "mem-container-smram", ~0ull);
+ memory_region_set_enabled(&smram_as_root, true);
+
+ /* ... with two regions inside: normal system memory with low
+ * priority, and...
+ */
+ memory_region_init_alias(&smram_as_mem, OBJECT(kvm_state), "mem-smram",
+ get_system_memory(), 0, ~0ull);
+ memory_region_add_subregion_overlap(&smram_as_root, 0, &smram_as_mem, 0);
+ memory_region_set_enabled(&smram_as_mem, true);
+
+ if (smram) {
+ /* ... SMRAM with higher priority */
+ memory_region_add_subregion_overlap(&smram_as_root, 0, smram, 10);
+ memory_region_set_enabled(smram, true);
+ }
+
+ address_space_init(&smram_address_space, &smram_as_root, "KVM-SMRAM");
+ kvm_memory_listener_register(kvm_state, &smram_listener,
+ &smram_address_space, 1);
+}
+
int kvm_arch_init(MachineState *ms, KVMState *s)
{
uint64_t identity_base = 0xfffbc000;
@@ -904,6 +938,11 @@ int kvm_arch_init(MachineState *ms, KVMState *s)
return ret;
}
}
+
+ if (kvm_check_extension(s, KVM_CAP_X86_SMM)) {
+ smram_machine_done.notify = register_smram_listener;
+ qemu_add_machine_init_done_notifier(&smram_machine_done);
+ }
return 0;
}
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 10/12] pc_piix: rename kvm_enabled to smm_enabled
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (8 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 09/12] target-i386: register a separate KVM address space including SMRAM regions Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 11/12] ich9: add smm_enabled field and arguments Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 12/12] pc: add SMM property Paolo Bonzini
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
We will enable SMM even if KVM is in use. Rename the field and
arguments.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/acpi/piix4.c | 10 +++++-----
hw/i386/pc_piix.c | 2 +-
include/hw/i386/pc.h | 2 +-
3 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
index 8a64c60..3ba0d07 100644
--- a/hw/acpi/piix4.c
+++ b/hw/acpi/piix4.c
@@ -72,7 +72,7 @@ typedef struct PIIX4PMState {
qemu_irq irq;
qemu_irq smi_irq;
- int kvm_enabled;
+ int smm_enabled;
Notifier machine_ready;
Notifier powerdown_notifier;
@@ -324,7 +324,7 @@ static void piix4_reset(void *opaque)
pci_conf[0x40] = 0x01; /* PM io base read only bit */
pci_conf[0x80] = 0;
- if (s->kvm_enabled) {
+ if (!s->smm_enabled) {
/* Mark SMM as already inited (until KVM supports SMM). */
pci_conf[0x5B] = 0x02;
}
@@ -455,7 +455,7 @@ static void piix4_pm_realize(PCIDevice *dev, Error **errp)
/* APM */
apm_init(dev, &s->apm, apm_ctrl_changed, s);
- if (s->kvm_enabled) {
+ if (!s->smm_enabled) {
/* Mark SMM as already inited to prevent SMM from running. KVM does not
* support SMM mode. */
pci_conf[0x5B] = 0x02;
@@ -506,7 +506,7 @@ Object *piix4_pm_find(void)
I2CBus *piix4_pm_init(PCIBus *bus, int devfn, uint32_t smb_io_base,
qemu_irq sci_irq, qemu_irq smi_irq,
- int kvm_enabled, DeviceState **piix4_pm)
+ int smm_enabled, DeviceState **piix4_pm)
{
DeviceState *dev;
PIIX4PMState *s;
@@ -520,7 +520,7 @@ I2CBus *piix4_pm_init(PCIBus *bus, int devfn, uint32_t smb_io_base,
s = PIIX4_PM(dev);
s->irq = sci_irq;
s->smi_irq = smi_irq;
- s->kvm_enabled = kvm_enabled;
+ s->smm_enabled = smm_enabled;
if (xen_enabled()) {
s->use_acpi_pci_hotplug = false;
}
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 0688f0c..4f2deba 100644
--- a/hw/i386/pc_piix.c
+++ b/hw/i386/pc_piix.c
@@ -287,7 +287,7 @@ static void pc_init1(MachineState *machine)
/* TODO: Populate SPD eeprom data. */
smbus = piix4_pm_init(pci_bus, piix3_devfn + 3, 0xb100,
gsi[9], *smi_irq,
- kvm_enabled(), &piix4_pm);
+ !kvm_enabled(), &piix4_pm);
smbus_eeprom_init(smbus, 8, NULL, 0);
object_property_add_link(OBJECT(machine), PC_MACHINE_ACPI_DEVICE_PROP,
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index 8beece0..8fa42e0 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -217,7 +217,7 @@ void ioapic_init_gsi(GSIState *gsi_state, const char *parent_name);
I2CBus *piix4_pm_init(PCIBus *bus, int devfn, uint32_t smb_io_base,
qemu_irq sci_irq, qemu_irq smi_irq,
- int kvm_enabled, DeviceState **piix4_pm);
+ int smm_enabled, DeviceState **piix4_pm);
void piix4_smbus_register_device(SMBusDevice *dev, uint8_t addr);
/* hpet.c */
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 11/12] ich9: add smm_enabled field and arguments
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (9 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 10/12] pc_piix: rename kvm_enabled to smm_enabled Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
2015-06-05 14:31 ` [Qemu-devel] [PATCH 12/12] pc: add SMM property Paolo Bonzini
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
Q35's ACPI device is hard-coding SMM availability to KVM. Place the
logic where the board is created instead, so that it will be possible
to override it.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/acpi/ich9.c | 5 +++--
hw/i386/pc_q35.c | 2 +-
hw/isa/lpc_ich9.c | 4 ++--
include/hw/acpi/ich9.h | 3 ++-
include/hw/i386/ich9.h | 2 +-
5 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/hw/acpi/ich9.c b/hw/acpi/ich9.c
index 25bc023..07dbc44 100644
--- a/hw/acpi/ich9.c
+++ b/hw/acpi/ich9.c
@@ -194,7 +194,7 @@ static void pm_reset(void *opaque)
acpi_pm_tmr_reset(&pm->acpi_regs);
acpi_gpe_reset(&pm->acpi_regs);
- if (kvm_enabled()) {
+ if (!pm->smm_enabled) {
/* Mark SMM as already inited to prevent SMM from running. KVM does not
* support SMM mode. */
pm->smi_en |= ICH9_PMIO_SMI_EN_APMC_EN;
@@ -211,7 +211,7 @@ static void pm_powerdown_req(Notifier *n, void *opaque)
acpi_pm1_evt_power_down(&pm->acpi_regs);
}
-void ich9_pm_init(PCIDevice *lpc_pci, ICH9LPCPMRegs *pm,
+void ich9_pm_init(PCIDevice *lpc_pci, ICH9LPCPMRegs *pm, bool smm_enabled,
qemu_irq sci_irq)
{
memory_region_init(&pm->io, OBJECT(lpc_pci), "ich9-pm", ICH9_PMIO_SIZE);
@@ -233,6 +233,7 @@ void ich9_pm_init(PCIDevice *lpc_pci, ICH9LPCPMRegs *pm,
"acpi-smi", 8);
memory_region_add_subregion(&pm->io, ICH9_PMIO_SMI_EN, &pm->io_smi);
+ pm->smm_enabled = smm_enabled;
pm->irq = sci_irq;
qemu_register_reset(pm_reset, pm);
pm->powerdown_notifier.notify = pm_powerdown_req;
diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
index 66220b3..27c4b19 100644
--- a/hw/i386/pc_q35.c
+++ b/hw/i386/pc_q35.c
@@ -254,7 +254,7 @@ static void pc_q35_init(MachineState *machine)
(pc_machine->vmport != ON_OFF_AUTO_ON), 0xff0104);
/* connect pm stuff to lpc */
- ich9_lpc_pm_init(lpc);
+ ich9_lpc_pm_init(lpc, !kvm_enabled());
/* ahci and SATA device, for q35 1 ahci controller is built-in */
ahci = pci_create_simple_multifunction(host_bus,
diff --git a/hw/isa/lpc_ich9.c b/hw/isa/lpc_ich9.c
index 247158c..1de013d 100644
--- a/hw/isa/lpc_ich9.c
+++ b/hw/isa/lpc_ich9.c
@@ -357,13 +357,13 @@ static void ich9_set_sci(void *opaque, int irq_num, int level)
}
}
-void ich9_lpc_pm_init(PCIDevice *lpc_pci)
+void ich9_lpc_pm_init(PCIDevice *lpc_pci, bool smm_enabled)
{
ICH9LPCState *lpc = ICH9_LPC_DEVICE(lpc_pci);
qemu_irq *sci_irq;
sci_irq = qemu_allocate_irqs(ich9_set_sci, lpc, 1);
- ich9_pm_init(lpc_pci, &lpc->pm, sci_irq[0]);
+ ich9_pm_init(lpc_pci, &lpc->pm, smm_enabled, sci_irq[0]);
ich9_lpc_reset(&lpc->d.qdev);
}
diff --git a/include/hw/acpi/ich9.h b/include/hw/acpi/ich9.h
index 77cc65c..ac24bbe 100644
--- a/include/hw/acpi/ich9.h
+++ b/include/hw/acpi/ich9.h
@@ -54,10 +54,11 @@ typedef struct ICH9LPCPMRegs {
uint8_t disable_s3;
uint8_t disable_s4;
uint8_t s4_val;
+ uint8_t smm_enabled;
} ICH9LPCPMRegs;
void ich9_pm_init(PCIDevice *lpc_pci, ICH9LPCPMRegs *pm,
- qemu_irq sci_irq);
+ bool smm_enabled, qemu_irq sci_irq);
void ich9_pm_iospace_update(ICH9LPCPMRegs *pm, uint32_t pm_io_base);
extern const VMStateDescription vmstate_ich9_pm;
diff --git a/include/hw/i386/ich9.h b/include/hw/i386/ich9.h
index a2cc15c..b317a48 100644
--- a/include/hw/i386/ich9.h
+++ b/include/hw/i386/ich9.h
@@ -17,7 +17,7 @@
void ich9_lpc_set_irq(void *opaque, int irq_num, int level);
int ich9_lpc_map_irq(PCIDevice *pci_dev, int intx);
PCIINTxRoute ich9_route_intx_pin_to_irq(void *opaque, int pirq_pin);
-void ich9_lpc_pm_init(PCIDevice *pci_lpc);
+void ich9_lpc_pm_init(PCIDevice *pci_lpc, bool smm_enabled);
I2CBus *ich9_smb_init(PCIBus *bus, int devfn, uint32_t smb_io_base);
#define ICH9_CC_SIZE (16 * 1024) /* 16KB */
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Qemu-devel] [PATCH 12/12] pc: add SMM property
2015-06-05 14:31 [Qemu-devel] [PATCH 00/12] SMM part 2: KVM and -machine enablement Paolo Bonzini
` (10 preceding siblings ...)
2015-06-05 14:31 ` [Qemu-devel] [PATCH 11/12] ich9: add smm_enabled field and arguments Paolo Bonzini
@ 2015-06-05 14:31 ` Paolo Bonzini
11 siblings, 0 replies; 14+ messages in thread
From: Paolo Bonzini @ 2015-06-05 14:31 UTC (permalink / raw)
To: qemu-devel; +Cc: lersek, kraxel, mst
The property can take values on, off or auto. The default is "off"
for KVM and pre-2.4 machines, otherwise "auto" (which makes it
available on TCG or on new-enough kernels).
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/i386/pc.c | 51 +++++++++++++++++++++++++++++++++++++++++++++++++++
hw/i386/pc_piix.c | 7 ++++++-
hw/i386/pc_q35.c | 6 +++++-
include/hw/i386/pc.h | 3 +++
4 files changed, 65 insertions(+), 2 deletions(-)
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index 43bc505..4b3c31a 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1801,6 +1801,48 @@ static void pc_machine_set_vmport(Object *obj, Visitor *v, void *opaque,
visit_type_OnOffAuto(v, &pcms->vmport, name, errp);
}
+bool pc_machine_is_smm_enabled(PCMachineState *pcms)
+{
+ bool smm_available = false;
+
+ if (pcms->smm == ON_OFF_AUTO_OFF) {
+ return false;
+ }
+
+ if (tcg_enabled() || qtest_enabled()) {
+ smm_available = true;
+ } else if (kvm_enabled()) {
+ smm_available = kvm_check_extension(kvm_state, KVM_CAP_X86_SMM);
+ }
+
+ if (smm_available) {
+ return true;
+ }
+
+ if (pcms->smm == ON_OFF_AUTO_ON) {
+ error_report("System Management Mode not supported by this hypervisor.");
+ exit(1);
+ }
+ return false;
+}
+
+static void pc_machine_get_smm(Object *obj, Visitor *v, void *opaque,
+ const char *name, Error **errp)
+{
+ PCMachineState *pcms = PC_MACHINE(obj);
+ OnOffAuto smm = pcms->smm;
+
+ visit_type_OnOffAuto(v, &smm, name, errp);
+}
+
+static void pc_machine_set_smm(Object *obj, Visitor *v, void *opaque,
+ const char *name, Error **errp)
+{
+ PCMachineState *pcms = PC_MACHINE(obj);
+
+ visit_type_OnOffAuto(v, &pcms->smm, name, errp);
+}
+
static bool pc_machine_get_aligned_dimm(Object *obj, Error **errp)
{
PCMachineState *pcms = PC_MACHINE(obj);
@@ -1825,6 +1867,15 @@ static void pc_machine_initfn(Object *obj)
"Maximum ram below the 4G boundary (32bit boundary)",
NULL);
+ pcms->smm = ON_OFF_AUTO_AUTO;
+ object_property_add(obj, PC_MACHINE_SMM, "OnOffAuto",
+ pc_machine_get_smm,
+ pc_machine_set_smm,
+ NULL, NULL, NULL);
+ object_property_set_description(obj, PC_MACHINE_SMM,
+ "Enable SMM (pc & q35)",
+ NULL);
+
pcms->vmport = ON_OFF_AUTO_AUTO;
object_property_add(obj, PC_MACHINE_VMPORT, "OnOffAuto",
pc_machine_get_vmport,
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 4f2deba..e17a797 100644
--- a/hw/i386/pc_piix.c
+++ b/hw/i386/pc_piix.c
@@ -287,7 +287,8 @@ static void pc_init1(MachineState *machine)
/* TODO: Populate SPD eeprom data. */
smbus = piix4_pm_init(pci_bus, piix3_devfn + 3, 0xb100,
gsi[9], *smi_irq,
- !kvm_enabled(), &piix4_pm);
+ pc_machine_is_smm_enabled(pc_machine),
+ &piix4_pm);
smbus_eeprom_init(smbus, 8, NULL, 0);
object_property_add_link(OBJECT(machine), PC_MACHINE_ACPI_DEVICE_PROP,
@@ -306,6 +307,10 @@ static void pc_init1(MachineState *machine)
static void pc_compat_2_3(MachineState *machine)
{
+ PCMachineState *pcms = PC_MACHINE(machine);
+ if (kvm_enabled()) {
+ pcms->smm = ON_OFF_AUTO_OFF;
+ }
}
static void pc_compat_2_2(MachineState *machine)
diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
index 27c4b19..51dcc12 100644
--- a/hw/i386/pc_q35.c
+++ b/hw/i386/pc_q35.c
@@ -254,7 +254,7 @@ static void pc_q35_init(MachineState *machine)
(pc_machine->vmport != ON_OFF_AUTO_ON), 0xff0104);
/* connect pm stuff to lpc */
- ich9_lpc_pm_init(lpc, !kvm_enabled());
+ ich9_lpc_pm_init(lpc, pc_machine_is_smm_enabled(pc_machine));
/* ahci and SATA device, for q35 1 ahci controller is built-in */
ahci = pci_create_simple_multifunction(host_bus,
@@ -291,6 +291,10 @@ static void pc_q35_init(MachineState *machine)
static void pc_compat_2_3(MachineState *machine)
{
+ PCMachineState *pcms = PC_MACHINE(machine);
+ if (kvm_enabled()) {
+ pcms->smm = ON_OFF_AUTO_OFF;
+ }
}
static void pc_compat_2_2(MachineState *machine)
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index 8fa42e0..ef95535 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -40,6 +40,7 @@ struct PCMachineState {
uint64_t max_ram_below_4g;
OnOffAuto vmport;
+ OnOffAuto smm;
bool enforce_aligned_dimm;
};
@@ -47,6 +48,7 @@ struct PCMachineState {
#define PC_MACHINE_MEMHP_REGION_SIZE "hotplug-memory-region-size"
#define PC_MACHINE_MAX_RAM_BELOW_4G "max-ram-below-4g"
#define PC_MACHINE_VMPORT "vmport"
+#define PC_MACHINE_SMM "smm"
#define PC_MACHINE_ENFORCE_ALIGNED_DIMM "enforce-aligned-dimm"
/**
@@ -158,6 +160,7 @@ void i8042_setup_a20_line(ISADevice *dev, qemu_irq *a20_out);
/* pc.c */
extern int fd_bootchk;
+bool pc_machine_is_smm_enabled(PCMachineState *pcms);
void pc_register_ferr_irq(qemu_irq irq);
void pc_acpi_smi_interrupt(void *opaque, int irq, int level);
--
2.4.1
^ permalink raw reply related [flat|nested] 14+ messages in thread