* [Qemu-devel] [QEMU PATCH v10 0/3] migration: migrate QTAILQ
@ 2016-10-31 19:53 Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo Jianjun Duan
` (2 more replies)
0 siblings, 3 replies; 18+ messages in thread
From: Jianjun Duan @ 2016-10-31 19:53 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, duanj, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, quintela, amit.shah, mreitz, kwolf, rth,
aurelien, leon.alrae, blauwirbel, mark.cave-ayland, mdroth,
dgilbert
Hi all,
I addressed some review comments. Comments are welcome.
v10: - Fixed a typo.
Previous versions are:
v9: - No more hard encoding of QTAILQ layout information
(link: http://lists.nongnu.org/archive/html/qemu-ppc/2016-10/msg01042.html)
v8: - Fixed a style issue.
(link: http://lists.nongnu.org/archive/html/qemu-ppc/2016-10/msg00874.html)
v7: - Fixed merge errors.
- Simplified macro definitions related to pointer arithmetic based QTAILQ access.
- Added test case for QTAILQ migration in tests/test-vmstate.c.
(link: http://lists.nongnu.org/archive/html/qemu-ppc/2016-10/msg00711.html)
v6: - Split from Power specific patches.
- Dropped VMS_LINKED flag.
- Rebased to master.
- Added comments to clarify about put/get in VMStateInfo.
(link: http://lists.nongnu.org/archive/html/qemu-ppc/2016-10/msg00336.html)
v5: - Rebased to David's ppc-for-2.8.
(link: https://lists.nongnu.org/archive/html/qemu-devel/2016-10/msg00270.html)
v4: - Introduce a way to set customized instance_id in SaveStateEntry. Use it
to set instance_id for DRC using its unique index to address David
Gibson's concern.
- Rename VMS_CSTM to VMS_LINKED based on Paolo Bonzini's suggestions.
- Clean up qjson stuff in put_qtailq.
- Add trace for put_qtailq and get_qtailq based on David Gilbert's
suggestion.
- Based on David's ppc-for-2.7.
(link: https://lists.nongnu.org/archive/html/qemu-devel/2016-06/msg07720.html)
v3: - Simplify overall design followng discussion with Paolo. No longer need
metadata to migrate QTAILQ.
- Extend VMStateInfo instead of adding similar fields to VMStateField.
- Clean up macros in qemu/queue.h.
(link: https://lists.nongnu.org/archive/html/qemu-devel/2016-05/msg05695.html)
v2: - Introduce a general approach to migrate QTAILQ in qemu/queue.h.
- Migrate signalled field in the DRC state.
- Put the newly added migrating fields in subsections so that backward
migration is not broken.
- Set detach_cb field right after migration so that a migrated hot-unplug
event could finish its course.
(link: https://lists.nongnu.org/archive/html/qemu-devel/2016-05/msg04188.html)
v1: - Inital version.
(link: https://lists.nongnu.org/archive/html/qemu-devel/2016-04/msg02601.html)
Jianjun Duan (3):
migration: extend VMStateInfo
migration: migrate QTAILQ
tests/migration: Add test for QTAILQ migration
hw/display/virtio-gpu.c | 6 +-
hw/intc/s390_flic_kvm.c | 6 +-
hw/net/vmxnet3.c | 18 +++--
hw/nvram/eeprom93xx.c | 6 +-
hw/nvram/fw_cfg.c | 6 +-
hw/pci/msix.c | 6 +-
hw/pci/pci.c | 12 ++-
hw/pci/shpc.c | 5 +-
hw/scsi/scsi-bus.c | 6 +-
hw/timer/twl92230.c | 6 +-
hw/usb/redirect.c | 18 +++--
hw/virtio/virtio-pci.c | 6 +-
hw/virtio/virtio.c | 12 ++-
include/migration/vmstate.h | 35 ++++++++-
include/qemu/queue.h | 61 ++++++++++++++++
migration/savevm.c | 5 +-
migration/trace-events | 4 +
migration/vmstate.c | 173 ++++++++++++++++++++++++++++++++++----------
target-alpha/machine.c | 5 +-
target-arm/machine.c | 12 ++-
target-i386/machine.c | 21 ++++--
target-mips/machine.c | 10 ++-
target-ppc/machine.c | 10 ++-
target-sparc/machine.c | 5 +-
tests/test-vmstate.c | 160 ++++++++++++++++++++++++++++++++++++++++
25 files changed, 510 insertions(+), 104 deletions(-)
--
1.9.1
^ permalink raw reply [flat|nested] 18+ messages in thread
* [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo
2016-10-31 19:53 [Qemu-devel] [QEMU PATCH v10 0/3] migration: migrate QTAILQ Jianjun Duan
@ 2016-10-31 19:53 ` Jianjun Duan
2016-11-02 10:40 ` Juan Quintela
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration Jianjun Duan
2 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-10-31 19:53 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, duanj, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, quintela, amit.shah, mreitz, kwolf, rth,
aurelien, leon.alrae, blauwirbel, mark.cave-ayland, mdroth,
dgilbert
Current migration code cannot handle some data structures such as
QTAILQ in qemu/queue.h. Here we extend the signatures of put/get
in VMStateInfo so that customized handling is supported.
Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
---
hw/display/virtio-gpu.c | 6 ++-
hw/intc/s390_flic_kvm.c | 6 ++-
hw/net/vmxnet3.c | 18 +++++---
hw/nvram/eeprom93xx.c | 6 ++-
hw/nvram/fw_cfg.c | 6 ++-
hw/pci/msix.c | 6 ++-
hw/pci/pci.c | 12 +++--
hw/pci/shpc.c | 5 ++-
hw/scsi/scsi-bus.c | 6 ++-
hw/timer/twl92230.c | 6 ++-
hw/usb/redirect.c | 18 +++++---
hw/virtio/virtio-pci.c | 6 ++-
hw/virtio/virtio.c | 12 +++--
include/migration/vmstate.h | 15 +++++--
migration/savevm.c | 5 ++-
migration/vmstate.c | 106 ++++++++++++++++++++++++++++----------------
target-alpha/machine.c | 5 ++-
target-arm/machine.c | 12 +++--
target-i386/machine.c | 21 ++++++---
target-mips/machine.c | 10 +++--
target-ppc/machine.c | 10 +++--
target-sparc/machine.c | 5 ++-
22 files changed, 198 insertions(+), 104 deletions(-)
diff --git a/hw/display/virtio-gpu.c b/hw/display/virtio-gpu.c
index fa6fd0e..2a21150 100644
--- a/hw/display/virtio-gpu.c
+++ b/hw/display/virtio-gpu.c
@@ -987,7 +987,8 @@ static const VMStateDescription vmstate_virtio_gpu_scanouts = {
},
};
-static void virtio_gpu_save(QEMUFile *f, void *opaque, size_t size)
+static void virtio_gpu_save(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
VirtIOGPU *g = opaque;
struct virtio_gpu_simple_resource *res;
@@ -1014,7 +1015,8 @@ static void virtio_gpu_save(QEMUFile *f, void *opaque, size_t size)
vmstate_save_state(f, &vmstate_virtio_gpu_scanouts, g, NULL);
}
-static int virtio_gpu_load(QEMUFile *f, void *opaque, size_t size)
+static int virtio_gpu_load(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
VirtIOGPU *g = opaque;
struct virtio_gpu_simple_resource *res;
diff --git a/hw/intc/s390_flic_kvm.c b/hw/intc/s390_flic_kvm.c
index 21ac2e2..a80a812 100644
--- a/hw/intc/s390_flic_kvm.c
+++ b/hw/intc/s390_flic_kvm.c
@@ -286,7 +286,8 @@ static void kvm_s390_release_adapter_routes(S390FLICState *fs,
* increase until buffer is sufficient or maxium size is
* reached
*/
-static void kvm_flic_save(QEMUFile *f, void *opaque, size_t size)
+static void kvm_flic_save(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
KVMS390FLICState *flic = opaque;
int len = FLIC_SAVE_INITIAL_SIZE;
@@ -331,7 +332,8 @@ static void kvm_flic_save(QEMUFile *f, void *opaque, size_t size)
* Note: Do nothing when no interrupts where stored
* in QEMUFile
*/
-static int kvm_flic_load(QEMUFile *f, void *opaque, size_t size)
+static int kvm_flic_load(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
uint64_t len = 0;
uint64_t count = 0;
diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
index 90f6943..943a960 100644
--- a/hw/net/vmxnet3.c
+++ b/hw/net/vmxnet3.c
@@ -2450,7 +2450,8 @@ static void vmxnet3_put_tx_stats_to_file(QEMUFile *f,
qemu_put_be64(f, tx_stat->pktsTxDiscard);
}
-static int vmxnet3_get_txq_descr(QEMUFile *f, void *pv, size_t size)
+static int vmxnet3_get_txq_descr(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
Vmxnet3TxqDescr *r = pv;
@@ -2464,7 +2465,8 @@ static int vmxnet3_get_txq_descr(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void vmxnet3_put_txq_descr(QEMUFile *f, void *pv, size_t size)
+static void vmxnet3_put_txq_descr(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
Vmxnet3TxqDescr *r = pv;
@@ -2511,7 +2513,8 @@ static void vmxnet3_put_rx_stats_to_file(QEMUFile *f,
qemu_put_be64(f, rx_stat->pktsRxError);
}
-static int vmxnet3_get_rxq_descr(QEMUFile *f, void *pv, size_t size)
+static int vmxnet3_get_rxq_descr(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
Vmxnet3RxqDescr *r = pv;
int i;
@@ -2529,7 +2532,8 @@ static int vmxnet3_get_rxq_descr(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void vmxnet3_put_rxq_descr(QEMUFile *f, void *pv, size_t size)
+static void vmxnet3_put_rxq_descr(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
Vmxnet3RxqDescr *r = pv;
int i;
@@ -2574,7 +2578,8 @@ static const VMStateInfo rxq_descr_info = {
.put = vmxnet3_put_rxq_descr
};
-static int vmxnet3_get_int_state(QEMUFile *f, void *pv, size_t size)
+static int vmxnet3_get_int_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
Vmxnet3IntState *r = pv;
@@ -2585,7 +2590,8 @@ static int vmxnet3_get_int_state(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void vmxnet3_put_int_state(QEMUFile *f, void *pv, size_t size)
+static void vmxnet3_put_int_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
Vmxnet3IntState *r = pv;
diff --git a/hw/nvram/eeprom93xx.c b/hw/nvram/eeprom93xx.c
index 2c16fc2..76d5f41 100644
--- a/hw/nvram/eeprom93xx.c
+++ b/hw/nvram/eeprom93xx.c
@@ -94,14 +94,16 @@ struct _eeprom_t {
This is a Big hack, but it is how the old state did it.
*/
-static int get_uint16_from_uint8(QEMUFile *f, void *pv, size_t size)
+static int get_uint16_from_uint8(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint16_t *v = pv;
*v = qemu_get_ubyte(f);
return 0;
}
-static void put_unused(QEMUFile *f, void *pv, size_t size)
+static void put_unused(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
fprintf(stderr, "uint16_from_uint8 is used only for backwards compatibility.\n");
fprintf(stderr, "Never should be used to write a new state.\n");
diff --git a/hw/nvram/fw_cfg.c b/hw/nvram/fw_cfg.c
index 92aa563..a8a4a7a 100644
--- a/hw/nvram/fw_cfg.c
+++ b/hw/nvram/fw_cfg.c
@@ -524,14 +524,16 @@ static void fw_cfg_reset(DeviceState *d)
Or we broke compatibility in the state, or we can't use struct tm
*/
-static int get_uint32_as_uint16(QEMUFile *f, void *pv, size_t size)
+static int get_uint32_as_uint16(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint32_t *v = pv;
*v = qemu_get_be16(f);
return 0;
}
-static void put_unused(QEMUFile *f, void *pv, size_t size)
+static void put_unused(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
fprintf(stderr, "uint32_as_uint16 is only used for backward compatibility.\n");
fprintf(stderr, "This functions shouldn't be called.\n");
diff --git a/hw/pci/msix.c b/hw/pci/msix.c
index 0ec1cb1..69e7a50 100644
--- a/hw/pci/msix.c
+++ b/hw/pci/msix.c
@@ -587,12 +587,14 @@ void msix_unset_vector_notifiers(PCIDevice *dev)
dev->msix_vector_poll_notifier = NULL;
}
-static void put_msix_state(QEMUFile *f, void *pv, size_t size)
+static void put_msix_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
msix_save(pv, f);
}
-static int get_msix_state(QEMUFile *f, void *pv, size_t size)
+static int get_msix_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
msix_load(pv, f);
return 0;
diff --git a/hw/pci/pci.c b/hw/pci/pci.c
index 24fae16..08c4547 100644
--- a/hw/pci/pci.c
+++ b/hw/pci/pci.c
@@ -445,7 +445,8 @@ int pci_bus_numa_node(PCIBus *bus)
return PCI_BUS_GET_CLASS(bus)->numa_node(bus);
}
-static int get_pci_config_device(QEMUFile *f, void *pv, size_t size)
+static int get_pci_config_device(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
PCIDevice *s = container_of(pv, PCIDevice, config);
PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(s);
@@ -484,7 +485,8 @@ static int get_pci_config_device(QEMUFile *f, void *pv, size_t size)
}
/* just put buffer */
-static void put_pci_config_device(QEMUFile *f, void *pv, size_t size)
+static void put_pci_config_device(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
const uint8_t **v = pv;
assert(size == pci_config_size(container_of(pv, PCIDevice, config)));
@@ -497,7 +499,8 @@ static VMStateInfo vmstate_info_pci_config = {
.put = put_pci_config_device,
};
-static int get_pci_irq_state(QEMUFile *f, void *pv, size_t size)
+static int get_pci_irq_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
PCIDevice *s = container_of(pv, PCIDevice, irq_state);
uint32_t irq_state[PCI_NUM_PINS];
@@ -518,7 +521,8 @@ static int get_pci_irq_state(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_pci_irq_state(QEMUFile *f, void *pv, size_t size)
+static void put_pci_irq_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
int i;
PCIDevice *s = container_of(pv, PCIDevice, irq_state);
diff --git a/hw/pci/shpc.c b/hw/pci/shpc.c
index 3dcd472..9f82aa6 100644
--- a/hw/pci/shpc.c
+++ b/hw/pci/shpc.c
@@ -695,13 +695,14 @@ void shpc_cap_write_config(PCIDevice *d, uint32_t addr, uint32_t val, int l)
shpc_cap_update_dword(d);
}
-static void shpc_save(QEMUFile *f, void *pv, size_t size)
+static void shpc_save(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
PCIDevice *d = container_of(pv, PCIDevice, shpc);
qemu_put_buffer(f, d->shpc->config, SHPC_SIZEOF(d));
}
-static int shpc_load(QEMUFile *f, void *pv, size_t size)
+static int shpc_load(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
PCIDevice *d = container_of(pv, PCIDevice, shpc);
int ret = qemu_get_buffer(f, d->shpc->config, SHPC_SIZEOF(d));
diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
index 297216d..f40c10b 100644
--- a/hw/scsi/scsi-bus.c
+++ b/hw/scsi/scsi-bus.c
@@ -1945,7 +1945,8 @@ SCSIDevice *scsi_device_find(SCSIBus *bus, int channel, int id, int lun)
/* SCSI request list. For simplicity, pv points to the whole device */
-static void put_scsi_requests(QEMUFile *f, void *pv, size_t size)
+static void put_scsi_requests(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
SCSIDevice *s = pv;
SCSIBus *bus = DO_UPCAST(SCSIBus, qbus, s->qdev.parent_bus);
@@ -1970,7 +1971,8 @@ static void put_scsi_requests(QEMUFile *f, void *pv, size_t size)
qemu_put_sbyte(f, 0);
}
-static int get_scsi_requests(QEMUFile *f, void *pv, size_t size)
+static int get_scsi_requests(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
SCSIDevice *s = pv;
SCSIBus *bus = DO_UPCAST(SCSIBus, qbus, s->qdev.parent_bus);
diff --git a/hw/timer/twl92230.c b/hw/timer/twl92230.c
index 7ba4e9a..95eb7f3 100644
--- a/hw/timer/twl92230.c
+++ b/hw/timer/twl92230.c
@@ -747,14 +747,16 @@ static int menelaus_rx(I2CSlave *i2c)
Or we broke compatibility in the state, or we can't use struct tm
*/
-static int get_int32_as_uint16(QEMUFile *f, void *pv, size_t size)
+static int get_int32_as_uint16(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
int *v = pv;
*v = qemu_get_be16(f);
return 0;
}
-static void put_int32_as_uint16(QEMUFile *f, void *pv, size_t size)
+static void put_int32_as_uint16(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
int *v = pv;
qemu_put_be16(f, *v);
diff --git a/hw/usb/redirect.c b/hw/usb/redirect.c
index d4ca026..7f28c3e 100644
--- a/hw/usb/redirect.c
+++ b/hw/usb/redirect.c
@@ -2158,7 +2158,8 @@ static int usbredir_post_load(void *priv, int version_id)
}
/* For usbredirparser migration */
-static void usbredir_put_parser(QEMUFile *f, void *priv, size_t unused)
+static void usbredir_put_parser(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field, QJSON *vmdesc)
{
USBRedirDevice *dev = priv;
uint8_t *data;
@@ -2178,7 +2179,8 @@ static void usbredir_put_parser(QEMUFile *f, void *priv, size_t unused)
free(data);
}
-static int usbredir_get_parser(QEMUFile *f, void *priv, size_t unused)
+static int usbredir_get_parser(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field)
{
USBRedirDevice *dev = priv;
uint8_t *data;
@@ -2221,7 +2223,8 @@ static const VMStateInfo usbredir_parser_vmstate_info = {
/* For buffered packets (iso/irq) queue migration */
-static void usbredir_put_bufpq(QEMUFile *f, void *priv, size_t unused)
+static void usbredir_put_bufpq(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field, QJSON *vmdesc)
{
struct endp_data *endp = priv;
USBRedirDevice *dev = endp->dev;
@@ -2241,7 +2244,8 @@ static void usbredir_put_bufpq(QEMUFile *f, void *priv, size_t unused)
assert(i == endp->bufpq_size);
}
-static int usbredir_get_bufpq(QEMUFile *f, void *priv, size_t unused)
+static int usbredir_get_bufpq(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field)
{
struct endp_data *endp = priv;
USBRedirDevice *dev = endp->dev;
@@ -2344,7 +2348,8 @@ static const VMStateDescription usbredir_ep_vmstate = {
/* For PacketIdQueue migration */
-static void usbredir_put_packet_id_q(QEMUFile *f, void *priv, size_t unused)
+static void usbredir_put_packet_id_q(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field, QJSON *vmdesc)
{
struct PacketIdQueue *q = priv;
USBRedirDevice *dev = q->dev;
@@ -2360,7 +2365,8 @@ static void usbredir_put_packet_id_q(QEMUFile *f, void *priv, size_t unused)
assert(remain == 0);
}
-static int usbredir_get_packet_id_q(QEMUFile *f, void *priv, size_t unused)
+static int usbredir_get_packet_id_q(QEMUFile *f, void *priv, size_t unused,
+ VMStateField *field)
{
struct PacketIdQueue *q = priv;
USBRedirDevice *dev = q->dev;
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index 06831de..4bd12f0 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -108,7 +108,8 @@ static bool virtio_pci_has_extra_state(DeviceState *d)
return proxy->flags & VIRTIO_PCI_FLAG_MIGRATE_EXTRA;
}
-static int get_virtio_pci_modern_state(QEMUFile *f, void *pv, size_t size)
+static int get_virtio_pci_modern_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
VirtIOPCIProxy *proxy = pv;
int i;
@@ -137,7 +138,8 @@ static void virtio_pci_save_modern_queue_state(VirtIOPCIQueue *vq,
qemu_put_be32(f, vq->used[1]);
}
-static void put_virtio_pci_modern_state(QEMUFile *f, void *pv, size_t size)
+static void put_virtio_pci_modern_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
VirtIOPCIProxy *proxy = pv;
int i;
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index d48d1a9..c799c5c 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -1490,7 +1490,8 @@ static const VMStateDescription vmstate_virtio_ringsize = {
}
};
-static int get_extra_state(QEMUFile *f, void *pv, size_t size)
+static int get_extra_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
VirtIODevice *vdev = pv;
BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
@@ -1503,7 +1504,8 @@ static int get_extra_state(QEMUFile *f, void *pv, size_t size)
}
}
-static void put_extra_state(QEMUFile *f, void *pv, size_t size)
+static void put_extra_state(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
VirtIODevice *vdev = pv;
BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
@@ -1640,13 +1642,15 @@ void virtio_save(VirtIODevice *vdev, QEMUFile *f)
}
/* A wrapper for use as a VMState .put function */
-static void virtio_device_put(QEMUFile *f, void *opaque, size_t size)
+static void virtio_device_put(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
virtio_save(VIRTIO_DEVICE(opaque), f);
}
/* A wrapper for use as a VMState .get function */
-static int virtio_device_get(QEMUFile *f, void *opaque, size_t size)
+static int virtio_device_get(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
VirtIODevice *vdev = VIRTIO_DEVICE(opaque);
DeviceClass *dc = DEVICE_CLASS(VIRTIO_DEVICE_GET_CLASS(vdev));
diff --git a/include/migration/vmstate.h b/include/migration/vmstate.h
index 1638ee5..d0e37b5 100644
--- a/include/migration/vmstate.h
+++ b/include/migration/vmstate.h
@@ -81,11 +81,18 @@ void unregister_savevm(DeviceState *dev, const char *idstr, void *opaque);
typedef struct VMStateInfo VMStateInfo;
typedef struct VMStateDescription VMStateDescription;
+typedef struct VMStateField VMStateField;
+/* VMStateInfo allows customized migration of objects that don't fit in
+ * any category in VMStateFlags. Additional information can be passed
+ * into get and put in terms of field and vmdesc parameters.
+ * For primitive data types such as integer, field and vmdesc parameters
+ * should be ignored inside get/put. */
struct VMStateInfo {
const char *name;
- int (*get)(QEMUFile *f, void *pv, size_t size);
- void (*put)(QEMUFile *f, void *pv, size_t size);
+ int (*get)(QEMUFile *f, void *pv, size_t size, VMStateField *field);
+ void (*put)(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc);
};
enum VMStateFlags {
@@ -186,7 +193,7 @@ enum VMStateFlags {
VMS_MULTIPLY_ELEMENTS = 0x4000,
};
-typedef struct {
+struct VMStateField {
const char *name;
size_t offset;
size_t size;
@@ -199,7 +206,7 @@ typedef struct {
const VMStateDescription *vmsd;
int version_id;
bool (*field_exists)(void *opaque, int version_id);
-} VMStateField;
+};
struct VMStateDescription {
const char *name;
diff --git a/migration/savevm.c b/migration/savevm.c
index 33a2911..12b7f8d 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -220,14 +220,15 @@ void timer_get(QEMUFile *f, QEMUTimer *ts)
* Not in vmstate.c to not add qemu-timer.c as dependency to vmstate.c
*/
-static int get_timer(QEMUFile *f, void *pv, size_t size)
+static int get_timer(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
QEMUTimer *v = pv;
timer_get(f, v);
return 0;
}
-static void put_timer(QEMUFile *f, void *pv, size_t size)
+static void put_timer(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
QEMUTimer *v = pv;
timer_put(f, v);
diff --git a/migration/vmstate.c b/migration/vmstate.c
index fc29acf..d188afa 100644
--- a/migration/vmstate.c
+++ b/migration/vmstate.c
@@ -6,6 +6,7 @@
#include "qemu/bitops.h"
#include "qemu/error-report.h"
#include "trace.h"
+#include "migration/qjson.h"
static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
void *opaque, QJSON *vmdesc);
@@ -83,6 +84,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
trace_vmstate_load_state(vmsd->name, version_id);
if (version_id > vmsd->version_id) {
+ error_report("%s %s", vmsd->name, "too new");
trace_vmstate_load_state_end(vmsd->name, "too new", -EINVAL);
return -EINVAL;
}
@@ -93,6 +95,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
trace_vmstate_load_state_end(vmsd->name, "old path", ret);
return ret;
}
+ error_report("%s %s", vmsd->name, "too old");
trace_vmstate_load_state_end(vmsd->name, "too old", -EINVAL);
return -EINVAL;
}
@@ -122,8 +125,10 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
ret = vmstate_load_state(f, field->vmsd, addr,
field->vmsd->version_id);
} else {
- ret = field->info->get(f, addr, size);
-
+ /* field is always passed in. But it should be ignored by
+ * get when not needed. It is only needed in cases* of
+ * customized handling, such as migrating QTAILQ. */
+ ret = field->info->get(f, addr, size, field);
}
if (ret >= 0) {
ret = qemu_file_get_error(f);
@@ -328,7 +333,11 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd,
if (field->flags & VMS_STRUCT) {
vmstate_save_state(f, field->vmsd, addr, vmdesc_loop);
} else {
- field->info->put(f, addr, size);
+ /* field and vmdesc_loop are always passed in. But they
+ * should be ignored by put when not needed. They are
+ * only needed in cases f customized handling, such as
+ * migrating QTAILQ. */
+ field->info->put(f, addr, size, field, vmdesc_loop);
}
written_bytes = qemu_ftell_fast(f) - old_offset;
@@ -461,14 +470,15 @@ static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
/* bool */
-static int get_bool(QEMUFile *f, void *pv, size_t size)
+static int get_bool(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
bool *v = pv;
*v = qemu_get_byte(f);
return 0;
}
-static void put_bool(QEMUFile *f, void *pv, size_t size)
+static void put_bool(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
bool *v = pv;
qemu_put_byte(f, *v);
@@ -482,14 +492,15 @@ const VMStateInfo vmstate_info_bool = {
/* 8 bit int */
-static int get_int8(QEMUFile *f, void *pv, size_t size)
+static int get_int8(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int8_t *v = pv;
qemu_get_s8s(f, v);
return 0;
}
-static void put_int8(QEMUFile *f, void *pv, size_t size)
+static void put_int8(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
int8_t *v = pv;
qemu_put_s8s(f, v);
@@ -503,14 +514,15 @@ const VMStateInfo vmstate_info_int8 = {
/* 16 bit int */
-static int get_int16(QEMUFile *f, void *pv, size_t size)
+static int get_int16(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int16_t *v = pv;
qemu_get_sbe16s(f, v);
return 0;
}
-static void put_int16(QEMUFile *f, void *pv, size_t size)
+static void put_int16(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
int16_t *v = pv;
qemu_put_sbe16s(f, v);
@@ -524,14 +536,15 @@ const VMStateInfo vmstate_info_int16 = {
/* 32 bit int */
-static int get_int32(QEMUFile *f, void *pv, size_t size)
+static int get_int32(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int32_t *v = pv;
qemu_get_sbe32s(f, v);
return 0;
}
-static void put_int32(QEMUFile *f, void *pv, size_t size)
+static void put_int32(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
int32_t *v = pv;
qemu_put_sbe32s(f, v);
@@ -546,7 +559,8 @@ const VMStateInfo vmstate_info_int32 = {
/* 32 bit int. See that the received value is the same than the one
in the field */
-static int get_int32_equal(QEMUFile *f, void *pv, size_t size)
+static int get_int32_equal(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
int32_t *v = pv;
int32_t v2;
@@ -568,7 +582,7 @@ const VMStateInfo vmstate_info_int32_equal = {
* and less than or equal to the one in the field.
*/
-static int get_int32_le(QEMUFile *f, void *pv, size_t size)
+static int get_int32_le(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int32_t *cur = pv;
int32_t loaded;
@@ -589,14 +603,15 @@ const VMStateInfo vmstate_info_int32_le = {
/* 64 bit int */
-static int get_int64(QEMUFile *f, void *pv, size_t size)
+static int get_int64(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int64_t *v = pv;
qemu_get_sbe64s(f, v);
return 0;
}
-static void put_int64(QEMUFile *f, void *pv, size_t size)
+static void put_int64(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
int64_t *v = pv;
qemu_put_sbe64s(f, v);
@@ -610,14 +625,15 @@ const VMStateInfo vmstate_info_int64 = {
/* 8 bit unsigned int */
-static int get_uint8(QEMUFile *f, void *pv, size_t size)
+static int get_uint8(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
uint8_t *v = pv;
qemu_get_8s(f, v);
return 0;
}
-static void put_uint8(QEMUFile *f, void *pv, size_t size)
+static void put_uint8(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint8_t *v = pv;
qemu_put_8s(f, v);
@@ -631,14 +647,15 @@ const VMStateInfo vmstate_info_uint8 = {
/* 16 bit unsigned int */
-static int get_uint16(QEMUFile *f, void *pv, size_t size)
+static int get_uint16(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
uint16_t *v = pv;
qemu_get_be16s(f, v);
return 0;
}
-static void put_uint16(QEMUFile *f, void *pv, size_t size)
+static void put_uint16(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint16_t *v = pv;
qemu_put_be16s(f, v);
@@ -652,14 +669,15 @@ const VMStateInfo vmstate_info_uint16 = {
/* 32 bit unsigned int */
-static int get_uint32(QEMUFile *f, void *pv, size_t size)
+static int get_uint32(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
uint32_t *v = pv;
qemu_get_be32s(f, v);
return 0;
}
-static void put_uint32(QEMUFile *f, void *pv, size_t size)
+static void put_uint32(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint32_t *v = pv;
qemu_put_be32s(f, v);
@@ -674,7 +692,8 @@ const VMStateInfo vmstate_info_uint32 = {
/* 32 bit uint. See that the received value is the same than the one
in the field */
-static int get_uint32_equal(QEMUFile *f, void *pv, size_t size)
+static int get_uint32_equal(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint32_t *v = pv;
uint32_t v2;
@@ -694,14 +713,15 @@ const VMStateInfo vmstate_info_uint32_equal = {
/* 64 bit unsigned int */
-static int get_uint64(QEMUFile *f, void *pv, size_t size)
+static int get_uint64(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
uint64_t *v = pv;
qemu_get_be64s(f, v);
return 0;
}
-static void put_uint64(QEMUFile *f, void *pv, size_t size)
+static void put_uint64(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint64_t *v = pv;
qemu_put_be64s(f, v);
@@ -716,7 +736,8 @@ const VMStateInfo vmstate_info_uint64 = {
/* 64 bit unsigned int. See that the received value is the same than the one
in the field */
-static int get_uint64_equal(QEMUFile *f, void *pv, size_t size)
+static int get_uint64_equal(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint64_t *v = pv;
uint64_t v2;
@@ -737,7 +758,8 @@ const VMStateInfo vmstate_info_uint64_equal = {
/* 8 bit int. See that the received value is the same than the one
in the field */
-static int get_uint8_equal(QEMUFile *f, void *pv, size_t size)
+static int get_uint8_equal(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint8_t *v = pv;
uint8_t v2;
@@ -758,7 +780,8 @@ const VMStateInfo vmstate_info_uint8_equal = {
/* 16 bit unsigned int int. See that the received value is the same than the one
in the field */
-static int get_uint16_equal(QEMUFile *f, void *pv, size_t size)
+static int get_uint16_equal(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint16_t *v = pv;
uint16_t v2;
@@ -778,7 +801,8 @@ const VMStateInfo vmstate_info_uint16_equal = {
/* floating point */
-static int get_float64(QEMUFile *f, void *pv, size_t size)
+static int get_float64(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
float64 *v = pv;
@@ -786,7 +810,8 @@ static int get_float64(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_float64(QEMUFile *f, void *pv, size_t size)
+static void put_float64(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint64_t *v = pv;
@@ -801,7 +826,8 @@ const VMStateInfo vmstate_info_float64 = {
/* CPU_DoubleU type */
-static int get_cpudouble(QEMUFile *f, void *pv, size_t size)
+static int get_cpudouble(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
CPU_DoubleU *v = pv;
qemu_get_be32s(f, &v->l.upper);
@@ -809,7 +835,8 @@ static int get_cpudouble(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_cpudouble(QEMUFile *f, void *pv, size_t size)
+static void put_cpudouble(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
CPU_DoubleU *v = pv;
qemu_put_be32s(f, &v->l.upper);
@@ -824,14 +851,16 @@ const VMStateInfo vmstate_info_cpudouble = {
/* uint8_t buffers */
-static int get_buffer(QEMUFile *f, void *pv, size_t size)
+static int get_buffer(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint8_t *v = pv;
qemu_get_buffer(f, v, size);
return 0;
}
-static void put_buffer(QEMUFile *f, void *pv, size_t size)
+static void put_buffer(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
uint8_t *v = pv;
qemu_put_buffer(f, v, size);
@@ -846,7 +875,8 @@ const VMStateInfo vmstate_info_buffer = {
/* unused buffers: space that was used for some fields that are
not useful anymore */
-static int get_unused_buffer(QEMUFile *f, void *pv, size_t size)
+static int get_unused_buffer(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint8_t buf[1024];
int block_len;
@@ -859,7 +889,8 @@ static int get_unused_buffer(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_unused_buffer(QEMUFile *f, void *pv, size_t size)
+static void put_unused_buffer(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
static const uint8_t buf[1024];
int block_len;
@@ -884,7 +915,7 @@ const VMStateInfo vmstate_info_unused_buffer = {
*/
/* This is the number of 64 bit words sent over the wire */
#define BITS_TO_U64S(nr) DIV_ROUND_UP(nr, 64)
-static int get_bitmap(QEMUFile *f, void *pv, size_t size)
+static int get_bitmap(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
unsigned long *bmp = pv;
int i, idx = 0;
@@ -898,7 +929,8 @@ static int get_bitmap(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_bitmap(QEMUFile *f, void *pv, size_t size)
+static void put_bitmap(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
unsigned long *bmp = pv;
int i, idx = 0;
diff --git a/target-alpha/machine.c b/target-alpha/machine.c
index 710b783..48e3278 100644
--- a/target-alpha/machine.c
+++ b/target-alpha/machine.c
@@ -5,14 +5,15 @@
#include "hw/boards.h"
#include "migration/cpu.h"
-static int get_fpcr(QEMUFile *f, void *opaque, size_t size)
+static int get_fpcr(QEMUFile *f, void *opaque, size_t size, VMStateField *field)
{
CPUAlphaState *env = opaque;
cpu_alpha_store_fpcr(env, qemu_get_be64(f));
return 0;
}
-static void put_fpcr(QEMUFile *f, void *opaque, size_t size)
+static void put_fpcr(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
CPUAlphaState *env = opaque;
qemu_put_be64(f, cpu_alpha_load_fpcr(env));
diff --git a/target-arm/machine.c b/target-arm/machine.c
index d90943b..96ff2da 100644
--- a/target-arm/machine.c
+++ b/target-arm/machine.c
@@ -17,7 +17,8 @@ static bool vfp_needed(void *opaque)
return arm_feature(env, ARM_FEATURE_VFP);
}
-static int get_fpscr(QEMUFile *f, void *opaque, size_t size)
+static int get_fpscr(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
ARMCPU *cpu = opaque;
CPUARMState *env = &cpu->env;
@@ -27,7 +28,8 @@ static int get_fpscr(QEMUFile *f, void *opaque, size_t size)
return 0;
}
-static void put_fpscr(QEMUFile *f, void *opaque, size_t size)
+static void put_fpscr(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
ARMCPU *cpu = opaque;
CPUARMState *env = &cpu->env;
@@ -163,7 +165,8 @@ static const VMStateDescription vmstate_pmsav7 = {
}
};
-static int get_cpsr(QEMUFile *f, void *opaque, size_t size)
+static int get_cpsr(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
ARMCPU *cpu = opaque;
CPUARMState *env = &cpu->env;
@@ -180,7 +183,8 @@ static int get_cpsr(QEMUFile *f, void *opaque, size_t size)
return 0;
}
-static void put_cpsr(QEMUFile *f, void *opaque, size_t size)
+static void put_cpsr(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
ARMCPU *cpu = opaque;
CPUARMState *env = &cpu->env;
diff --git a/target-i386/machine.c b/target-i386/machine.c
index 71c0e4d..1df19e2 100644
--- a/target-i386/machine.c
+++ b/target-i386/machine.c
@@ -139,7 +139,8 @@ static const VMStateDescription vmstate_mtrr_var = {
#define VMSTATE_MTRR_VARS(_field, _state, _n, _v) \
VMSTATE_STRUCT_ARRAY(_field, _state, _n, _v, vmstate_mtrr_var, MTRRVar)
-static void put_fpreg_error(QEMUFile *f, void *opaque, size_t size)
+static void put_fpreg_error(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
fprintf(stderr, "call put_fpreg() with invalid arguments\n");
exit(0);
@@ -167,7 +168,8 @@ static void fp64_to_fp80(union x86_longdouble *p, uint64_t temp)
p->exp = e;
}
-static int get_fpreg(QEMUFile *f, void *opaque, size_t size)
+static int get_fpreg(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
FPReg *fp_reg = opaque;
uint64_t mant;
@@ -179,7 +181,8 @@ static int get_fpreg(QEMUFile *f, void *opaque, size_t size)
return 0;
}
-static void put_fpreg(QEMUFile *f, void *opaque, size_t size)
+static void put_fpreg(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
FPReg *fp_reg = opaque;
uint64_t mant;
@@ -197,7 +200,8 @@ static const VMStateInfo vmstate_fpreg = {
.put = put_fpreg,
};
-static int get_fpreg_1_mmx(QEMUFile *f, void *opaque, size_t size)
+static int get_fpreg_1_mmx(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
union x86_longdouble *p = opaque;
uint64_t mant;
@@ -214,7 +218,8 @@ static const VMStateInfo vmstate_fpreg_1_mmx = {
.put = put_fpreg_error,
};
-static int get_fpreg_1_no_mmx(QEMUFile *f, void *opaque, size_t size)
+static int get_fpreg_1_no_mmx(QEMUFile *f, void *opaque, size_t size,
+ VMStateField *field)
{
union x86_longdouble *p = opaque;
uint64_t mant;
@@ -276,14 +281,16 @@ static bool less_than_7(void *opaque, int version_id)
return version_id < 7;
}
-static int get_uint64_as_uint32(QEMUFile *f, void *pv, size_t size)
+static int get_uint64_as_uint32(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field)
{
uint64_t *v = pv;
*v = qemu_get_be32(f);
return 0;
}
-static void put_uint64_as_uint32(QEMUFile *f, void *pv, size_t size)
+static void put_uint64_as_uint32(QEMUFile *f, void *pv, size_t size,
+ VMStateField *field, QJSON *vmdesc)
{
uint64_t *v = pv;
qemu_put_be32(f, *v);
diff --git a/target-mips/machine.c b/target-mips/machine.c
index a27f2f1..179084c 100644
--- a/target-mips/machine.c
+++ b/target-mips/machine.c
@@ -20,7 +20,7 @@ static int cpu_post_load(void *opaque, int version_id)
/* FPU state */
-static int get_fpr(QEMUFile *f, void *pv, size_t size)
+static int get_fpr(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
int i;
fpr_t *v = pv;
@@ -31,7 +31,8 @@ static int get_fpr(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_fpr(QEMUFile *f, void *pv, size_t size)
+static void put_fpr(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
int i;
fpr_t *v = pv;
@@ -125,7 +126,7 @@ const VMStateDescription vmstate_mvp = {
/* TLB state */
-static int get_tlb(QEMUFile *f, void *pv, size_t size)
+static int get_tlb(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
r4k_tlb_t *v = pv;
uint16_t flags;
@@ -152,7 +153,8 @@ static int get_tlb(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_tlb(QEMUFile *f, void *pv, size_t size)
+static void put_tlb(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
r4k_tlb_t *v = pv;
diff --git a/target-ppc/machine.c b/target-ppc/machine.c
index 4820f22..0e1822c 100644
--- a/target-ppc/machine.c
+++ b/target-ppc/machine.c
@@ -106,7 +106,7 @@ static int cpu_load_old(QEMUFile *f, void *opaque, int version_id)
return 0;
}
-static int get_avr(QEMUFile *f, void *pv, size_t size)
+static int get_avr(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
ppc_avr_t *v = pv;
@@ -116,7 +116,8 @@ static int get_avr(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_avr(QEMUFile *f, void *pv, size_t size)
+static void put_avr(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
ppc_avr_t *v = pv;
@@ -324,7 +325,7 @@ static const VMStateDescription vmstate_sr = {
};
#ifdef TARGET_PPC64
-static int get_slbe(QEMUFile *f, void *pv, size_t size)
+static int get_slbe(QEMUFile *f, void *pv, size_t size, VMStateField *field)
{
ppc_slb_t *v = pv;
@@ -334,7 +335,8 @@ static int get_slbe(QEMUFile *f, void *pv, size_t size)
return 0;
}
-static void put_slbe(QEMUFile *f, void *pv, size_t size)
+static void put_slbe(QEMUFile *f, void *pv, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
ppc_slb_t *v = pv;
diff --git a/target-sparc/machine.c b/target-sparc/machine.c
index 59c92f7..3194e03 100644
--- a/target-sparc/machine.c
+++ b/target-sparc/machine.c
@@ -59,7 +59,7 @@ static const VMStateDescription vmstate_tlb_entry = {
};
#endif
-static int get_psr(QEMUFile *f, void *opaque, size_t size)
+static int get_psr(QEMUFile *f, void *opaque, size_t size, VMStateField *field)
{
SPARCCPU *cpu = opaque;
CPUSPARCState *env = &cpu->env;
@@ -72,7 +72,8 @@ static int get_psr(QEMUFile *f, void *opaque, size_t size)
return 0;
}
-static void put_psr(QEMUFile *f, void *opaque, size_t size)
+static void put_psr(QEMUFile *f, void *opaque, size_t size, VMStateField *field,
+ QJSON *vmdesc)
{
SPARCCPU *cpu = opaque;
CPUSPARCState *env = &cpu->env;
--
1.9.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-10-31 19:53 [Qemu-devel] [QEMU PATCH v10 0/3] migration: migrate QTAILQ Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo Jianjun Duan
@ 2016-10-31 19:53 ` Jianjun Duan
2016-11-02 10:45 ` Juan Quintela
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration Jianjun Duan
2 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-10-31 19:53 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, duanj, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, quintela, amit.shah, mreitz, kwolf, rth,
aurelien, leon.alrae, blauwirbel, mark.cave-ayland, mdroth,
dgilbert
Currently we cannot directly transfer a QTAILQ instance because of the
limitation in the migration code. Here we introduce an approach to
transfer such structures. We created VMStateInfo vmstate_info_qtailq
for QTAILQ. Similar VMStateInfo can be created for other data structures
such as list.
This approach will be used to transfer pending_events and ccs_list in spapr
state.
We also create some macros in qemu/queue.h to access a QTAILQ using pointer
arithmetic. This ensures that we do not depend on the implementation
details about QTAILQ in the migration code.
Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
---
include/migration/vmstate.h | 20 ++++++++++++++
include/qemu/queue.h | 61 +++++++++++++++++++++++++++++++++++++++++
migration/trace-events | 4 +++
migration/vmstate.c | 67 +++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 152 insertions(+)
diff --git a/include/migration/vmstate.h b/include/migration/vmstate.h
index d0e37b5..318a6f1 100644
--- a/include/migration/vmstate.h
+++ b/include/migration/vmstate.h
@@ -251,6 +251,7 @@ extern const VMStateInfo vmstate_info_timer;
extern const VMStateInfo vmstate_info_buffer;
extern const VMStateInfo vmstate_info_unused_buffer;
extern const VMStateInfo vmstate_info_bitmap;
+extern const VMStateInfo vmstate_info_qtailq;
#define type_check_2darray(t1,t2,n,m) ((t1(*)[n][m])0 - (t2*)0)
#define type_check_array(t1,t2,n) ((t1(*)[n])0 - (t2*)0)
@@ -662,6 +663,25 @@ extern const VMStateInfo vmstate_info_bitmap;
.offset = offsetof(_state, _field), \
}
+/* For QTAILQ that need customized handling.
+ * Target QTAILQ needs be properly initialized.
+ * _type: type of QTAILQ element
+ * _next: name of QTAILQ entry field in QTAILQ element
+ * _vmsd: VMSD for QTAILQ element
+ * size: size of QTAILQ element
+ * start: offset of QTAILQ entry in QTAILQ element
+ */
+#define VMSTATE_QTAILQ_V(_field, _state, _version, _vmsd, _type, _next) \
+{ \
+ .name = (stringify(_field)), \
+ .version_id = (_version), \
+ .vmsd = &(_vmsd), \
+ .size = sizeof(_type), \
+ .info = &vmstate_info_qtailq, \
+ .offset = offsetof(_state, _field), \
+ .start = offsetof(_type, _next), \
+}
+
/* _f : field name
_f_n : num of elements field_name
_n : num of elements
diff --git a/include/qemu/queue.h b/include/qemu/queue.h
index 342073f..aae23f2 100644
--- a/include/qemu/queue.h
+++ b/include/qemu/queue.h
@@ -438,4 +438,65 @@ struct { \
#define QTAILQ_PREV(elm, headname, field) \
(*(((struct headname *)((elm)->field.tqe_prev))->tqh_last))
+#define field_at_offset(base, offset, type) \
+ ((type) (((char *) (base)) + (offset)))
+
+typedef struct DUMMY_Q_ENTRY DUMMY_Q_ENTRY;
+typedef struct DUMMY_Q DUMMY_Q;
+
+struct DUMMY_Q_ENTRY {
+ QTAILQ_ENTRY(DUMMY_Q_ENTRY) next;
+};
+
+struct DUMMY_Q {
+ QTAILQ_HEAD(DUMMY_Q_HEAD, DUMMY_Q_ENTRY) head;
+};
+
+#define dummy_q ((DUMMY_Q *) 0)
+#define dummy_qh ((typeof(dummy_q->head) *) 0)
+#define dummy_qe ((DUMMY_Q_ENTRY *) 0)
+
+/*
+ * Offsets of layout of a tail queue head.
+ */
+#define QTAILQ_FIRST_OFFSET ((size_t) &(QTAILQ_FIRST(dummy_qh)))
+#define QTAILQ_LAST_OFFSET (offsetof(typeof(dummy_q->head), tqh_last))
+/*
+ * Raw access of elements of a tail queue
+ */
+#define QTAILQ_RAW_FIRST(head) \
+ (*field_at_offset(head, QTAILQ_FIRST_OFFSET, void **))
+#define QTAILQ_RAW_LAST(head) \
+ (*field_at_offset(head, QTAILQ_LAST_OFFSET, void ***))
+
+/*
+ * Offsets of layout of a tail queue element.
+ */
+#define QTAILQ_NEXT_OFFSET ((size_t) (&QTAILQ_NEXT(dummy_qe, next)))
+#define QTAILQ_PREV_OFFSET (offsetof(typeof(dummy_qe->next), tqe_prev))
+
+/*
+ * Raw access of elements of a tail entry
+ */
+#define QTAILQ_RAW_NEXT(elm, entry) \
+ (*field_at_offset(elm, entry + QTAILQ_NEXT_OFFSET, void **))
+#define QTAILQ_RAW_PREV(elm, entry) \
+ (*field_at_offset(elm, entry + QTAILQ_PREV_OFFSET, void ***))
+/*
+ * Tail queue tranversal using pointer arithmetic.
+ */
+#define QTAILQ_RAW_FOREACH(elm, head, entry) \
+ for ((elm) = QTAILQ_RAW_FIRST(head); \
+ (elm); \
+ (elm) = QTAILQ_RAW_NEXT(elm, entry))
+/*
+ * Tail queue insertion using pointer arithmetic.
+ */
+#define QTAILQ_RAW_INSERT_TAIL(head, elm, entry) do { \
+ QTAILQ_RAW_NEXT(elm, entry) = NULL; \
+ QTAILQ_RAW_PREV(elm, entry) = QTAILQ_RAW_LAST(head); \
+ *QTAILQ_RAW_LAST(head) = (elm); \
+ QTAILQ_RAW_LAST(head) = &QTAILQ_RAW_NEXT(elm, entry); \
+} while (/*CONSTCOND*/0)
+
#endif /* QEMU_SYS_QUEUE_H */
diff --git a/migration/trace-events b/migration/trace-events
index dfee75a..9a6ec59 100644
--- a/migration/trace-events
+++ b/migration/trace-events
@@ -52,6 +52,10 @@ vmstate_n_elems(const char *name, int n_elems) "%s: %d"
vmstate_subsection_load(const char *parent) "%s"
vmstate_subsection_load_bad(const char *parent, const char *sub, const char *sub2) "%s: %s/%s"
vmstate_subsection_load_good(const char *parent) "%s"
+get_qtailq(const char *name, int version_id) "%s v%d"
+get_qtailq_end(const char *name, const char *reason, int val) "%s %s/%d"
+put_qtailq(const char *name, int version_id) "%s v%d"
+put_qtailq_end(const char *name, const char *reason) "%s %s"
# migration/qemu-file.c
qemu_file_fclose(void) ""
diff --git a/migration/vmstate.c b/migration/vmstate.c
index d188afa..70e0472 100644
--- a/migration/vmstate.c
+++ b/migration/vmstate.c
@@ -5,6 +5,7 @@
#include "migration/vmstate.h"
#include "qemu/bitops.h"
#include "qemu/error-report.h"
+#include "qemu/queue.h"
#include "trace.h"
#include "migration/qjson.h"
@@ -948,3 +949,69 @@ const VMStateInfo vmstate_info_bitmap = {
.get = get_bitmap,
.put = put_bitmap,
};
+
+/* get for QTAILQ
+ * meta data about the QTAILQ is encoded in a VMStateField structure
+ */
+static int get_qtailq(QEMUFile *f, void *pv, size_t unused_size,
+ VMStateField *field)
+{
+ int ret = 0;
+ const VMStateDescription *vmsd = field->vmsd;
+ /* size of a QTAILQ element */
+ size_t size = field->size;
+ /* offset of the QTAILQ entry in a QTAILQ element */
+ size_t entry_offset = field->start;
+ int version_id = field->version_id;
+ void *elm;
+
+ trace_get_qtailq(vmsd->name, version_id);
+ if (version_id > vmsd->version_id) {
+ error_report("%s %s", vmsd->name, "too new");
+ trace_get_qtailq_end(vmsd->name, "too new", -EINVAL);
+
+ return -EINVAL;
+ }
+ if (version_id < vmsd->minimum_version_id) {
+ error_report("%s %s", vmsd->name, "too old");
+ trace_get_qtailq_end(vmsd->name, "too old", -EINVAL);
+ return -EINVAL;
+ }
+
+ while (qemu_get_byte(f)) {
+ elm = g_malloc(size);
+ ret = vmstate_load_state(f, vmsd, elm, version_id);
+ if (ret) {
+ return ret;
+ }
+ QTAILQ_RAW_INSERT_TAIL(pv, elm, entry_offset);
+ }
+
+ trace_get_qtailq_end(vmsd->name, "end", ret);
+ return ret;
+}
+
+/* put for QTAILQ */
+static void put_qtailq(QEMUFile *f, void *pv, size_t unused_size,
+ VMStateField *field, QJSON *vmdesc)
+{
+ const VMStateDescription *vmsd = field->vmsd;
+ /* offset of the QTAILQ entry in a QTAILQ element*/
+ size_t entry_offset = field->start;
+ void *elm;
+
+ trace_put_qtailq(vmsd->name, vmsd->version_id);
+
+ QTAILQ_RAW_FOREACH(elm, pv, entry_offset) {
+ qemu_put_byte(f, true);
+ vmstate_save_state(f, vmsd, elm, vmdesc);
+ }
+ qemu_put_byte(f, false);
+
+ trace_put_qtailq_end(vmsd->name, "end");
+}
+const VMStateInfo vmstate_info_qtailq = {
+ .name = "qtailq",
+ .get = get_qtailq,
+ .put = put_qtailq,
+};
--
1.9.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-10-31 19:53 [Qemu-devel] [QEMU PATCH v10 0/3] migration: migrate QTAILQ Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ Jianjun Duan
@ 2016-10-31 19:53 ` Jianjun Duan
2016-11-02 10:47 ` Juan Quintela
2 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-10-31 19:53 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, duanj, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, quintela, amit.shah, mreitz, kwolf, rth,
aurelien, leon.alrae, blauwirbel, mark.cave-ayland, mdroth,
dgilbert
Add a test for QTAILQ migration to tests/test-vmstate.c.
Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
---
tests/test-vmstate.c | 160 +++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 160 insertions(+)
diff --git a/tests/test-vmstate.c b/tests/test-vmstate.c
index d8da26f..a992408 100644
--- a/tests/test-vmstate.c
+++ b/tests/test-vmstate.c
@@ -475,6 +475,164 @@ static void test_load_skip(void)
qemu_fclose(loading);
}
+
+/* test QTAILQ migration */
+typedef struct TestQtailqElement TestQtailqElement;
+
+struct TestQtailqElement {
+ bool b;
+ uint8_t u8;
+ QTAILQ_ENTRY(TestQtailqElement) next;
+};
+
+typedef struct TestQtailq {
+ int16_t i16;
+ QTAILQ_HEAD(TestQtailqHead, TestQtailqElement) q;
+ int32_t i32;
+} TestQtailq;
+
+static const VMStateDescription vmstate_q_element = {
+ .name = "test/queue-element",
+ .version_id = 1,
+ .minimum_version_id = 1,
+ .fields = (VMStateField[]) {
+ VMSTATE_BOOL(b, TestQtailqElement),
+ VMSTATE_UINT8(u8, TestQtailqElement),
+ VMSTATE_END_OF_LIST()
+ },
+};
+
+static const VMStateDescription vmstate_q = {
+ .name = "test/queue",
+ .version_id = 1,
+ .minimum_version_id = 1,
+ .fields = (VMStateField[]) {
+ VMSTATE_INT16(i16, TestQtailq),
+ VMSTATE_QTAILQ_V(q, TestQtailq, 1, vmstate_q_element, TestQtailqElement,
+ next),
+ VMSTATE_INT32(i32, TestQtailq),
+ VMSTATE_END_OF_LIST()
+ }
+};
+
+static void test_save_q(void)
+{
+ TestQtailq obj_q = {
+ .i16 = -512,
+ .i32 = 70000,
+ };
+
+ TestQtailqElement obj_qe1 = {
+ .b = true,
+ .u8 = 130,
+ };
+
+ TestQtailqElement obj_qe2 = {
+ .b = false,
+ .u8 = 65,
+ };
+
+ uint8_t wire_q[] = {
+ /* i16 */ 0xfe, 0x0,
+ /* start of element 0 of q */ 0x01,
+ /* .b */ 0x01,
+ /* .u8 */ 0x82,
+ /* start of element 1 of q */ 0x01,
+ /* b */ 0x00,
+ /* u8 */ 0x41,
+ /* end of q */ 0x00,
+ /* i32 */ 0x00, 0x01, 0x11, 0x70,
+ QEMU_VM_EOF, /* just to ensure we won't get EOF reported prematurely */
+ };
+
+ QTAILQ_INIT(&obj_q.q);
+ QTAILQ_INSERT_TAIL(&obj_q.q, &obj_qe1, next);
+ QTAILQ_INSERT_TAIL(&obj_q.q, &obj_qe2, next);
+
+ save_vmstate(&vmstate_q, &obj_q);
+ compare_vmstate(wire_q, sizeof(wire_q));
+}
+
+static void test_load_q(void)
+{
+ TestQtailq obj_q = {
+ .i16 = -512,
+ .i32 = 70000,
+ };
+
+ TestQtailqElement obj_qe1 = {
+ .b = true,
+ .u8 = 130,
+ };
+
+ TestQtailqElement obj_qe2 = {
+ .b = false,
+ .u8 = 65,
+ };
+
+ uint8_t wire_q[] = {
+ /* i16 */ 0xfe, 0x0,
+ /* start of element 0 of q */ 0x01,
+ /* .b */ 0x01,
+ /* .u8 */ 0x82,
+ /* start of element 1 of q */ 0x01,
+ /* b */ 0x00,
+ /* u8 */ 0x41,
+ /* end of q */ 0x00,
+ /* i32 */ 0x00, 0x01, 0x11, 0x70,
+ };
+
+ QTAILQ_INIT(&obj_q.q);
+ QTAILQ_INSERT_TAIL(&obj_q.q, &obj_qe1, next);
+ QTAILQ_INSERT_TAIL(&obj_q.q, &obj_qe2, next);
+
+ QEMUFile *fsave = open_test_file(true);
+
+ qemu_put_buffer(fsave, wire_q, sizeof(wire_q));
+ qemu_put_byte(fsave, QEMU_VM_EOF);
+ g_assert(!qemu_file_get_error(fsave));
+ qemu_fclose(fsave);
+
+ QEMUFile *fload = open_test_file(false);
+ TestQtailq tgt;
+
+ QTAILQ_INIT(&tgt.q);
+ vmstate_load_state(fload, &vmstate_q, &tgt, 1);
+ char eof = qemu_get_byte(fload);
+ g_assert(!qemu_file_get_error(fload));
+ g_assert_cmpint(tgt.i16, ==, obj_q.i16);
+ g_assert_cmpint(tgt.i32, ==, obj_q.i32);
+ g_assert_cmpint(eof, ==, QEMU_VM_EOF);
+
+ TestQtailqElement *qele_from = QTAILQ_FIRST(&obj_q.q);
+ TestQtailqElement *qlast_from = QTAILQ_LAST(&obj_q.q, TestQtailqHead);
+ TestQtailqElement *qele_to = QTAILQ_FIRST(&tgt.q);
+ TestQtailqElement *qlast_to = QTAILQ_LAST(&tgt.q, TestQtailqHead);
+
+ while (1) {
+ g_assert_cmpint(qele_to->b, ==, qele_from->b);
+ g_assert_cmpint(qele_to->u8, ==, qele_from->u8);
+ if ((qele_from == qlast_from) || (qele_to == qlast_to)) {
+ break;
+ }
+ qele_from = QTAILQ_NEXT(qele_from, next);
+ qele_to = QTAILQ_NEXT(qele_to, next);
+ }
+
+ g_assert_cmpint((uint64_t) qele_from, ==, (uint64_t) qlast_from);
+ g_assert_cmpint((uint64_t) qele_to, ==, (uint64_t) qlast_to);
+
+ /* clean up */
+ TestQtailqElement *qele;
+ while (!QTAILQ_EMPTY(&tgt.q)) {
+ qele = QTAILQ_LAST(&tgt.q, TestQtailqHead);
+ QTAILQ_REMOVE(&tgt.q, qele, next);
+ free(qele);
+ qele = NULL;
+ }
+ qemu_fclose(fload);
+}
+
int main(int argc, char **argv)
{
temp_fd = mkstemp(temp_file);
@@ -489,6 +647,8 @@ int main(int argc, char **argv)
g_test_add_func("/vmstate/field_exists/load/skip", test_load_skip);
g_test_add_func("/vmstate/field_exists/save/noskip", test_save_noskip);
g_test_add_func("/vmstate/field_exists/save/skip", test_save_skip);
+ g_test_add_func("/vmstate/qtailq/save/saveq", test_save_q);
+ g_test_add_func("/vmstate/qtailq/load/loadq", test_load_q);
g_test_run();
close(temp_fd);
--
1.9.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo Jianjun Duan
@ 2016-11-02 10:40 ` Juan Quintela
2016-11-02 16:54 ` Jianjun Duan
0 siblings, 1 reply; 18+ messages in thread
From: Juan Quintela @ 2016-11-02 10:40 UTC (permalink / raw)
To: Jianjun Duan
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
> Current migration code cannot handle some data structures such as
> QTAILQ in qemu/queue.h. Here we extend the signatures of put/get
> in VMStateInfo so that customized handling is supported.
>
> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
> +/* VMStateInfo allows customized migration of objects that don't fit in
> + * any category in VMStateFlags. Additional information can be passed
> + * into get and put in terms of field and vmdesc parameters.
> + * For primitive data types such as integer, field and vmdesc parameters
> + * should be ignored inside get/put. */
> struct VMStateInfo {
> const char *name;
> - int (*get)(QEMUFile *f, void *pv, size_t size);
> - void (*put)(QEMUFile *f, void *pv, size_t size);
> + int (*get)(QEMUFile *f, void *pv, size_t size, VMStateField *field);
> + void (*put)(QEMUFile *f, void *pv, size_t size, VMStateField *field,
> + QJSON *vmdesc);
If we are changing put type, I would like to add an error value to it.
> static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
> void *opaque, QJSON *vmdesc);
> @@ -83,6 +84,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>
> trace_vmstate_load_state(vmsd->name, version_id);
> if (version_id > vmsd->version_id) {
> + error_report("%s %s", vmsd->name, "too new");
> trace_vmstate_load_state_end(vmsd->name, "too new", -EINVAL);
> return -EINVAL;
> }
> @@ -93,6 +95,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
> trace_vmstate_load_state_end(vmsd->name, "old path", ret);
> return ret;
> }
> + error_report("%s %s", vmsd->name, "too old");
> trace_vmstate_load_state_end(vmsd->name, "too old", -EINVAL);
> return -EINVAL;
> }
This two are good, but don't belong to this patch, please sent separately.
> @@ -122,8 +125,10 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
> ret = vmstate_load_state(f, field->vmsd, addr,
> field->vmsd->version_id);
> } else {
> - ret = field->info->get(f, addr, size);
> -
> + /* field is always passed in. But it should be ignored by
> + * get when not needed. It is only needed in cases* of
> + * customized handling, such as migrating QTAILQ. */
> + ret = field->info->get(f, addr, size, field);
I would not put this comment here, better on the header when the field
is declared? Same for the next one.
> }
> if (ret >= 0) {
> ret = qemu_file_get_error(f);
> @@ -328,7 +333,11 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd,
> if (field->flags & VMS_STRUCT) {
> vmstate_save_state(f, field->vmsd, addr, vmdesc_loop);
> } else {
> - field->info->put(f, addr, size);
> + /* field and vmdesc_loop are always passed in. But they
> + * should be ignored by put when not needed. They are
> + * only needed in cases f customized handling, such as
> + * migrating QTAILQ. */
> + field->info->put(f, addr, size, field, vmdesc_loop);
> }
Rest of patch is ok.
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ Jianjun Duan
@ 2016-11-02 10:45 ` Juan Quintela
2016-11-02 16:38 ` Paolo Bonzini
2016-11-02 17:05 ` Jianjun Duan
0 siblings, 2 replies; 18+ messages in thread
From: Juan Quintela @ 2016-11-02 10:45 UTC (permalink / raw)
To: Jianjun Duan
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
> Currently we cannot directly transfer a QTAILQ instance because of the
> limitation in the migration code. Here we introduce an approach to
> transfer such structures. We created VMStateInfo vmstate_info_qtailq
> for QTAILQ. Similar VMStateInfo can be created for other data structures
> such as list.
>
> This approach will be used to transfer pending_events and ccs_list in spapr
> state.
>
> We also create some macros in qemu/queue.h to access a QTAILQ using pointer
> arithmetic. This ensures that we do not depend on the implementation
> details about QTAILQ in the migration code.
>
> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
> +
> + trace_get_qtailq(vmsd->name, version_id);
> + if (version_id > vmsd->version_id) {
> + error_report("%s %s", vmsd->name, "too new");
> + trace_get_qtailq_end(vmsd->name, "too new", -EINVAL);
> +
> + return -EINVAL;
> + }
> + if (version_id < vmsd->minimum_version_id) {
> + error_report("%s %s", vmsd->name, "too old");
> + trace_get_qtailq_end(vmsd->name, "too old", -EINVAL);
> + return -EINVAL;
> + }
> +
> + while (qemu_get_byte(f)) {
> + elm = g_malloc(size);
I think this is not generic enough. We really need to allocate a new
element, and then fill it with default values.
virtio list code use it in this way.
Thanks, Juan.
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration Jianjun Duan
@ 2016-11-02 10:47 ` Juan Quintela
2016-11-03 12:22 ` Halil Pasic
0 siblings, 1 reply; 18+ messages in thread
From: Juan Quintela @ 2016-11-02 10:47 UTC (permalink / raw)
To: Jianjun Duan
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
> Add a test for QTAILQ migration to tests/test-vmstate.c.
>
> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-11-02 10:45 ` Juan Quintela
@ 2016-11-02 16:38 ` Paolo Bonzini
2016-11-03 11:14 ` Halil Pasic
2016-11-02 17:05 ` Jianjun Duan
1 sibling, 1 reply; 18+ messages in thread
From: Paolo Bonzini @ 2016-11-02 16:38 UTC (permalink / raw)
To: quintela, Jianjun Duan
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien, leon.alrae,
blauwirbel, mark.cave-ayland, mdroth, dgilbert
On 02/11/2016 11:45, Juan Quintela wrote:
>> > + while (qemu_get_byte(f)) {
>> > + elm = g_malloc(size);
> I think this is not generic enough. We really need to allocate a new
> element, and then fill it with default values.
>
> virtio list code use it in this way.
One thing at a time? The allocation and "looping" are different
mechanisms. We can customize the allocation later.
Paolo
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo
2016-11-02 10:40 ` Juan Quintela
@ 2016-11-02 16:54 ` Jianjun Duan
2016-11-03 10:18 ` Juan Quintela
0 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-11-02 16:54 UTC (permalink / raw)
To: quintela
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
On 11/02/2016 03:40 AM, Juan Quintela wrote:
> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>> Current migration code cannot handle some data structures such as
>> QTAILQ in qemu/queue.h. Here we extend the signatures of put/get
>> in VMStateInfo so that customized handling is supported.
>>
>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>
>> +/* VMStateInfo allows customized migration of objects that don't fit in
>> + * any category in VMStateFlags. Additional information can be passed
>> + * into get and put in terms of field and vmdesc parameters.
>> + * For primitive data types such as integer, field and vmdesc parameters
>> + * should be ignored inside get/put. */
>> struct VMStateInfo {
>> const char *name;
>> - int (*get)(QEMUFile *f, void *pv, size_t size);
>> - void (*put)(QEMUFile *f, void *pv, size_t size);
>> + int (*get)(QEMUFile *f, void *pv, size_t size, VMStateField *field);
>> + void (*put)(QEMUFile *f, void *pv, size_t size, VMStateField *field,
>> + QJSON *vmdesc);
>
> If we are changing put type, I would like to add an error value to it.
>
so we change the return type to int?
>> static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
>> void *opaque, QJSON *vmdesc);
>> @@ -83,6 +84,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>>
>> trace_vmstate_load_state(vmsd->name, version_id);
>> if (version_id > vmsd->version_id) {
>> + error_report("%s %s", vmsd->name, "too new");
>> trace_vmstate_load_state_end(vmsd->name, "too new", -EINVAL);
>> return -EINVAL;
>> }
>> @@ -93,6 +95,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>> trace_vmstate_load_state_end(vmsd->name, "old path", ret);
>> return ret;
>> }
>> + error_report("%s %s", vmsd->name, "too old");
>> trace_vmstate_load_state_end(vmsd->name, "too old", -EINVAL);
>> return -EINVAL;
>> }
>
> This two are good, but don't belong to this patch, please sent separately.
>
OK.
>> @@ -122,8 +125,10 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>> ret = vmstate_load_state(f, field->vmsd, addr,
>> field->vmsd->version_id);
>> } else {
>> - ret = field->info->get(f, addr, size);
>> -
>> + /* field is always passed in. But it should be ignored by
>> + * get when not needed. It is only needed in cases* of
>> + * customized handling, such as migrating QTAILQ. */
>> + ret = field->info->get(f, addr, size, field);
>
> I would not put this comment here, better on the header when the field
> is declared? Same for the next one.
>
OK.
Thanks,
Jianjun
>> }
>> if (ret >= 0) {
>> ret = qemu_file_get_error(f);
>> @@ -328,7 +333,11 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd,
>> if (field->flags & VMS_STRUCT) {
>> vmstate_save_state(f, field->vmsd, addr, vmdesc_loop);
>> } else {
>> - field->info->put(f, addr, size);
>> + /* field and vmdesc_loop are always passed in. But they
>> + * should be ignored by put when not needed. They are
>> + * only needed in cases f customized handling, such as
>> + * migrating QTAILQ. */
>> + field->info->put(f, addr, size, field, vmdesc_loop);
>> }
>
>
> Rest of patch is ok.
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-11-02 10:45 ` Juan Quintela
2016-11-02 16:38 ` Paolo Bonzini
@ 2016-11-02 17:05 ` Jianjun Duan
1 sibling, 0 replies; 18+ messages in thread
From: Jianjun Duan @ 2016-11-02 17:05 UTC (permalink / raw)
To: quintela
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
On 11/02/2016 03:45 AM, Juan Quintela wrote:
> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>> Currently we cannot directly transfer a QTAILQ instance because of the
>> limitation in the migration code. Here we introduce an approach to
>> transfer such structures. We created VMStateInfo vmstate_info_qtailq
>> for QTAILQ. Similar VMStateInfo can be created for other data structures
>> such as list.
>>
>> This approach will be used to transfer pending_events and ccs_list in spapr
>> state.
>>
>> We also create some macros in qemu/queue.h to access a QTAILQ using pointer
>> arithmetic. This ensures that we do not depend on the implementation
>> details about QTAILQ in the migration code.
>>
>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>
>
>> +
>> + trace_get_qtailq(vmsd->name, version_id);
>> + if (version_id > vmsd->version_id) {
>> + error_report("%s %s", vmsd->name, "too new");
>> + trace_get_qtailq_end(vmsd->name, "too new", -EINVAL);
>> +
>> + return -EINVAL;
>> + }
>> + if (version_id < vmsd->minimum_version_id) {
>> + error_report("%s %s", vmsd->name, "too old");
>> + trace_get_qtailq_end(vmsd->name, "too old", -EINVAL);
>> + return -EINVAL;
>> + }
>> +
>> + while (qemu_get_byte(f)) {
>> + elm = g_malloc(size);
>
> I think this is not generic enough. We really need to allocate a new
> element, and then fill it with default values.
>
> virtio list code use it in this way.
>
>
To do that we need probably to expand VMStateDescription and/or
VMStateField so that a default value can be supplied. Or we can always
fill the untouched fields in post_load.
Thanks,
Jianjun
> Thanks, Juan.
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo
2016-11-02 16:54 ` Jianjun Duan
@ 2016-11-03 10:18 ` Juan Quintela
0 siblings, 0 replies; 18+ messages in thread
From: Juan Quintela @ 2016-11-03 10:18 UTC (permalink / raw)
To: Jianjun Duan
Cc: qemu-devel, qemu-ppc, dmitry, peter.maydell, kraxel, mst, david,
pbonzini, veroniabahaa, amit.shah, mreitz, kwolf, rth, aurelien,
leon.alrae, blauwirbel, mark.cave-ayland, mdroth, dgilbert
Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
> On 11/02/2016 03:40 AM, Juan Quintela wrote:
>> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>>> Current migration code cannot handle some data structures such as
>>> QTAILQ in qemu/queue.h. Here we extend the signatures of put/get
>>> in VMStateInfo so that customized handling is supported.
>>>
>>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>>
>>> +/* VMStateInfo allows customized migration of objects that don't fit in
>>> + * any category in VMStateFlags. Additional information can be passed
>>> + * into get and put in terms of field and vmdesc parameters.
>>> + * For primitive data types such as integer, field and vmdesc parameters
>>> + * should be ignored inside get/put. */
>>> struct VMStateInfo {
>>> const char *name;
>>> - int (*get)(QEMUFile *f, void *pv, size_t size);
>>> - void (*put)(QEMUFile *f, void *pv, size_t size);
>>> + int (*get)(QEMUFile *f, void *pv, size_t size, VMStateField *field);
>>> + void (*put)(QEMUFile *f, void *pv, size_t size, VMStateField *field,
>>> + QJSON *vmdesc);
>>
>> If we are changing put type, I would like to add an error value to it.
>>
> so we change the return type to int?
Yeap. Right now it is impossible to return errors. If we change the
prototype, better to just add the return of the error, and that
everything return 0 for now.
Thanks, Juan.
>>> static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
>>> void *opaque, QJSON *vmdesc);
>>> @@ -83,6 +84,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>>>
>>> trace_vmstate_load_state(vmsd->name, version_id);
>>> if (version_id > vmsd->version_id) {
>>> + error_report("%s %s", vmsd->name, "too new");
>>> trace_vmstate_load_state_end(vmsd->name, "too new", -EINVAL);
>>> return -EINVAL;
>>> }
>>> @@ -93,6 +95,7 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>>> trace_vmstate_load_state_end(vmsd->name, "old path", ret);
>>> return ret;
>>> }
>>> + error_report("%s %s", vmsd->name, "too old");
>>> trace_vmstate_load_state_end(vmsd->name, "too old", -EINVAL);
>>> return -EINVAL;
>>> }
>>
>> This two are good, but don't belong to this patch, please sent separately.
>>
> OK.
>>> @@ -122,8 +125,10 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
>>> ret = vmstate_load_state(f, field->vmsd, addr,
>>> field->vmsd->version_id);
>>> } else {
>>> - ret = field->info->get(f, addr, size);
>>> -
>>> + /* field is always passed in. But it should be ignored by
>>> + * get when not needed. It is only needed in cases* of
>>> + * customized handling, such as migrating QTAILQ. */
>>> + ret = field->info->get(f, addr, size, field);
>>
>> I would not put this comment here, better on the header when the field
>> is declared? Same for the next one.
>>
> OK.
>
> Thanks,
> Jianjun
>>> }
>>> if (ret >= 0) {
>>> ret = qemu_file_get_error(f);
>>> @@ -328,7 +333,11 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd,
>>> if (field->flags & VMS_STRUCT) {
>>> vmstate_save_state(f, field->vmsd, addr, vmdesc_loop);
>>> } else {
>>> - field->info->put(f, addr, size);
>>> + /* field and vmdesc_loop are always passed in. But they
>>> + * should be ignored by put when not needed. They are
>>> + * only needed in cases f customized handling, such as
>>> + * migrating QTAILQ. */
>>> + field->info->put(f, addr, size, field, vmdesc_loop);
>>> }
>>
>>
>> Rest of patch is ok.
>>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-11-02 16:38 ` Paolo Bonzini
@ 2016-11-03 11:14 ` Halil Pasic
2016-11-03 11:32 ` Halil Pasic
0 siblings, 1 reply; 18+ messages in thread
From: Halil Pasic @ 2016-11-03 11:14 UTC (permalink / raw)
To: Paolo Bonzini, quintela, Jianjun Duan
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, dmitry, kwolf, rth, leon.alrae, aurelien, david
On 11/02/2016 05:38 PM, Paolo Bonzini wrote:
>
>
> On 02/11/2016 11:45, Juan Quintela wrote:
>>>> + while (qemu_get_byte(f)) {
>>>> + elm = g_malloc(size);
>> I think this is not generic enough. We really need to allocate a new
>> element, and then fill it with default values.
>>
>> virtio list code use it in this way.
>
> One thing at a time? The allocation and "looping" are different
> mechanisms. We can customize the allocation later.
>
> Paolo
>
+1
And while at it how about s/g_malloc/gmalloc0? This would possibly make
something like mismatch in .start between source and target easier to
detect, along with the 'untouched' fields mentioned by Jianjun.
(Although I'm not sure how untouched fields and the virtio list use is
connected. In my understanding the problem pointed out by Juan has
something to do with the address of a element/node having significance
beyond the list.)
Halil
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ
2016-11-03 11:14 ` Halil Pasic
@ 2016-11-03 11:32 ` Halil Pasic
0 siblings, 0 replies; 18+ messages in thread
From: Halil Pasic @ 2016-11-03 11:32 UTC (permalink / raw)
To: Paolo Bonzini, quintela, Jianjun Duan
Cc: veroniabahaa, peter.maydell, mst, qemu-devel, mark.cave-ayland,
dgilbert, mdroth, blauwirbel, dmitry, qemu-ppc, kraxel, david,
amit.shah, kwolf, mreitz, leon.alrae, aurelien, rth
On 11/03/2016 12:14 PM, Halil Pasic wrote:
>
>
> On 11/02/2016 05:38 PM, Paolo Bonzini wrote:
>>
>>
>> On 02/11/2016 11:45, Juan Quintela wrote:
>>>>> + while (qemu_get_byte(f)) {
>>>>> + elm = g_malloc(size);
>>> I think this is not generic enough. We really need to allocate a new
>>> element, and then fill it with default values.
>>>
>>> virtio list code use it in this way.
>>
>> One thing at a time? The allocation and "looping" are different
>> mechanisms. We can customize the allocation later.
>>
>> Paolo
>>
>
[..]
>
> And while at it how about s/g_malloc/gmalloc0? This would possibly make
> something like mismatch in .start between source and target easier to
> detect
Sorry guys, I just realized this was stupid of me. Please ignore.
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-11-02 10:47 ` Juan Quintela
@ 2016-11-03 12:22 ` Halil Pasic
2016-11-03 16:47 ` Jianjun Duan
0 siblings, 1 reply; 18+ messages in thread
From: Halil Pasic @ 2016-11-03 12:22 UTC (permalink / raw)
To: quintela, Jianjun Duan
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, kwolf, dmitry, pbonzini, rth, leon.alrae,
aurelien, david
On 11/02/2016 11:47 AM, Juan Quintela wrote:
> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>> Add a test for QTAILQ migration to tests/test-vmstate.c.
>>
>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>
> Reviewed-by: Juan Quintela <quintela@redhat.com>
>
Empty QTAILQ seems to be broken. Have written a small
test to prove my point. It May even make sense to have such
a test in the test-suite (some prettyfication might be
necessary though).
Halil
-----------------------8<-------------------------------------
>From 4323c308c5f56ed86eb0a5bb6027bca4617ecc8c Mon Sep 17 00:00:00 2001
From: Halil Pasic <pasic@linux.vnet.ibm.com>
Date: Thu, 3 Nov 2016 13:07:05 +0100
Subject: [PATCH] tests/test-vmstate.c: add test empty qtailq
Add test for empty qtailq.
Signed-off-by: Halil Pasic <pasic@linux.vnet.ibm.com>
--
For fixup.
---
tests/test-vmstate.c | 28 ++++++++++++++++++++++++++++
1 files changed, 28 insertions(+), 0 deletions(-)
diff --git a/tests/test-vmstate.c b/tests/test-vmstate.c
index a992408..789f07a 100644
--- a/tests/test-vmstate.c
+++ b/tests/test-vmstate.c
@@ -126,6 +126,14 @@ static int load_vmstate_one(const VMStateDescription *desc, void *obj,
return ret;
}
+static void load_vmstate_one_obj(const VMStateDescription *vmsd, void *obj,
+ int version_id)
+{
+ QEMUFile *fload = open_test_file(false);
+
+ SUCCESS(vmstate_load_state(fload, vmsd, obj, version_id));
+ qemu_fclose(fload);
+}
static int load_vmstate(const VMStateDescription *desc,
void *obj, void *obj_clone,
@@ -633,6 +641,25 @@ static void test_load_q(void)
qemu_fclose(fload);
}
+
+static void test_sl_empty_q(void)
+{
+ TestQtailq obj_q = {
+ .i16 = -512,
+ .i32 = 70000,
+ };
+ TestQtailq tgt = {.q = {.tqh_first = (void *)1}};
+
+ QTAILQ_INIT(&obj_q.q);
+
+ save_vmstate(&vmstate_q, &obj_q);
+ load_vmstate_one_obj(&vmstate_q, &tgt, 1);
+ g_assert_cmpint(tgt.i16, ==, obj_q.i16);
+ g_assert_cmpint(tgt.i32, ==, obj_q.i32);
+ g_assert_cmpint(QTAILQ_EMPTY(&(tgt.q)),!=, false);
+
+}
+
int main(int argc, char **argv)
{
temp_fd = mkstemp(temp_file);
@@ -649,6 +676,7 @@ int main(int argc, char **argv)
g_test_add_func("/vmstate/field_exists/save/skip", test_save_skip);
g_test_add_func("/vmstate/qtailq/save/saveq", test_save_q);
g_test_add_func("/vmstate/qtailq/load/loadq", test_load_q);
+ g_test_add_func("/vmstate/qtailq/empty", test_sl_empty_q);
g_test_run();
close(temp_fd);
--
1.7.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-11-03 12:22 ` Halil Pasic
@ 2016-11-03 16:47 ` Jianjun Duan
2016-11-03 17:17 ` Halil Pasic
0 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-11-03 16:47 UTC (permalink / raw)
To: Halil Pasic, quintela
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, kwolf, dmitry, pbonzini, rth, leon.alrae,
aurelien, david
On 11/03/2016 05:22 AM, Halil Pasic wrote:
>
>
> On 11/02/2016 11:47 AM, Juan Quintela wrote:
>> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>>> Add a test for QTAILQ migration to tests/test-vmstate.c.
>>>
>>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>>
>> Reviewed-by: Juan Quintela <quintela@redhat.com>
>>
>
> Empty QTAILQ seems to be broken. Have written a small
> test to prove my point. It May even make sense to have such
> a test in the test-suite (some prettyfication might be
> necessary though).
>
It is working as intended.
The current design is to append the qtailq from source to the
corresponding one on target. It works well for the task in hard
such as migrating ccs_list and pending_events for DRC objects.
I suspect in most cases the qtailqs on target are empty. If not,
appending to them is a good choice. Clearing them is tricky since
each queue probably require a specialized routine to clean. If they
are not empty there are must be good reasons for that.
Thanks,
Jianjun
> Halil
>
> -----------------------8<-------------------------------------
>
> From 4323c308c5f56ed86eb0a5bb6027bca4617ecc8c Mon Sep 17 00:00:00 2001
> From: Halil Pasic <pasic@linux.vnet.ibm.com>
> Date: Thu, 3 Nov 2016 13:07:05 +0100
> Subject: [PATCH] tests/test-vmstate.c: add test empty qtailq
>
> Add test for empty qtailq.
>
> Signed-off-by: Halil Pasic <pasic@linux.vnet.ibm.com>
>
> --
>
> For fixup.
> ---
> tests/test-vmstate.c | 28 ++++++++++++++++++++++++++++
> 1 files changed, 28 insertions(+), 0 deletions(-)
>
> diff --git a/tests/test-vmstate.c b/tests/test-vmstate.c
> index a992408..789f07a 100644
> --- a/tests/test-vmstate.c
> +++ b/tests/test-vmstate.c
> @@ -126,6 +126,14 @@ static int load_vmstate_one(const VMStateDescription *desc, void *obj,
> return ret;
> }
>
> +static void load_vmstate_one_obj(const VMStateDescription *vmsd, void *obj,
> + int version_id)
> +{
> + QEMUFile *fload = open_test_file(false);
> +
> + SUCCESS(vmstate_load_state(fload, vmsd, obj, version_id));
> + qemu_fclose(fload);
> +}
>
> static int load_vmstate(const VMStateDescription *desc,
> void *obj, void *obj_clone,
> @@ -633,6 +641,25 @@ static void test_load_q(void)
> qemu_fclose(fload);
> }
>
> +
> +static void test_sl_empty_q(void)
> +{
> + TestQtailq obj_q = {
> + .i16 = -512,
> + .i32 = 70000,
> + };
> + TestQtailq tgt = {.q = {.tqh_first = (void *)1}};
> +
> + QTAILQ_INIT(&obj_q.q);
> +
> + save_vmstate(&vmstate_q, &obj_q);
> + load_vmstate_one_obj(&vmstate_q, &tgt, 1);
> + g_assert_cmpint(tgt.i16, ==, obj_q.i16);
> + g_assert_cmpint(tgt.i32, ==, obj_q.i32);
> + g_assert_cmpint(QTAILQ_EMPTY(&(tgt.q)),!=, false);
> +
> +}
> +
> int main(int argc, char **argv)
> {
> temp_fd = mkstemp(temp_file);
> @@ -649,6 +676,7 @@ int main(int argc, char **argv)
> g_test_add_func("/vmstate/field_exists/save/skip", test_save_skip);
> g_test_add_func("/vmstate/qtailq/save/saveq", test_save_q);
> g_test_add_func("/vmstate/qtailq/load/loadq", test_load_q);
> + g_test_add_func("/vmstate/qtailq/empty", test_sl_empty_q);
> g_test_run();
>
> close(temp_fd);
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-11-03 16:47 ` Jianjun Duan
@ 2016-11-03 17:17 ` Halil Pasic
2016-11-03 18:40 ` Jianjun Duan
0 siblings, 1 reply; 18+ messages in thread
From: Halil Pasic @ 2016-11-03 17:17 UTC (permalink / raw)
To: Jianjun Duan, quintela
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, kwolf, dmitry, pbonzini, rth, leon.alrae,
aurelien, david
On 11/03/2016 05:47 PM, Jianjun Duan wrote:
>
> On 11/03/2016 05:22 AM, Halil Pasic wrote:
>> >
>> >
>> > On 11/02/2016 11:47 AM, Juan Quintela wrote:
>>> >> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>>>> >>> Add a test for QTAILQ migration to tests/test-vmstate.c.
>>>> >>>
>>>> >>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>>> >>
>>> >> Reviewed-by: Juan Quintela <quintela@redhat.com>
>>> >>
>> >
>> > Empty QTAILQ seems to be broken. Have written a small
>> > test to prove my point. It May even make sense to have such
>> > a test in the test-suite (some prettyfication might be
>> > necessary though).
>> >
> It is working as intended.
>
My train of thought was that the object holding the queue might
be dynamically allocated by the migration code or otherwise
uninitialized. I was unaware these scenarios are prohibited.
> The current design is to append the qtailq from source to the
> corresponding one on target.
I do not see this documented. I'm used to vmstate_load overwriting
values and following pointers, so IMHO it is not obvious that
qtailq load does append.
> It works well for the task in hard
> such as migrating ccs_list and pending_events for DRC objects.
>
Because target head is always properly initialized to empty queue?
> I suspect in most cases the qtailqs on target are empty.
If I think about migration having no queues populated with
elements on a target site sounds very reasonable since IFAIU
the target should not do any work which would populate these
data structures.
> If not,
> appending to them is a good choice. Clearing them is tricky since
> each queue probably require a specialized routine to clean. If they
> are not empty there are must be good reasons for that.
Have you some code or a scenario in mind where this is legit? I
mean creating a mix of the state(?) we found at the target and
the state captured at the source does not sound right. I would
argue that the target should not have any state which is subject
to migration.
You are right a non-empty queue is trouble, and frankly I never
considered it as a valid scenario.
Sorry if I'm bothering you with nonsense.
Greetings,
Halil
>
> Thanks,
> Jianjun
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-11-03 17:17 ` Halil Pasic
@ 2016-11-03 18:40 ` Jianjun Duan
2016-11-03 18:51 ` Halil Pasic
0 siblings, 1 reply; 18+ messages in thread
From: Jianjun Duan @ 2016-11-03 18:40 UTC (permalink / raw)
To: Halil Pasic, quintela
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, kwolf, dmitry, pbonzini, rth, leon.alrae,
aurelien, david
On 11/03/2016 10:17 AM, Halil Pasic wrote:
>
>
> On 11/03/2016 05:47 PM, Jianjun Duan wrote:
>>
>> On 11/03/2016 05:22 AM, Halil Pasic wrote:
>>>>
>>>>
>>>> On 11/02/2016 11:47 AM, Juan Quintela wrote:
>>>>>> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>>>>>>>> Add a test for QTAILQ migration to tests/test-vmstate.c.
>>>>>>>>
>>>>>>>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>>>>>>
>>>>>> Reviewed-by: Juan Quintela <quintela@redhat.com>
>>>>>>
>>>>
>>>> Empty QTAILQ seems to be broken. Have written a small
>>>> test to prove my point. It May even make sense to have such
>>>> a test in the test-suite (some prettyfication might be
>>>> necessary though).
>>>>
>> It is working as intended.
>>
>
> My train of thought was that the object holding the queue might
> be dynamically allocated by the migration code or otherwise
> uninitialized. I was unaware these scenarios are prohibited.
>
>
This is a valid point. To get this covered vmstate_load_state needs to
be revised so that at any moment of recursion we know if the field is in
a dynamic created structure. If yes the structures which need
initialization such as QTAILQ can be initialized.
I would leave this until the need is there. In current device migration
code I imagine such scenarios would be rare if they should appear at
all. Because all the devices (even the hotplugged ones) are already
initialized on target. So a QTAILQ in such context should already be
initialized. Otherwise it should be fixed.
>> The current design is to append the qtailq from source to the
>> corresponding one on target.
>
> I do not see this documented. I'm used to vmstate_load overwriting
> values and following pointers, so IMHO it is not obvious that
> qtailq load does append.
>
I will document this.
>> It works well for the task in hard
>> such as migrating ccs_list and pending_events for DRC objects.
>>
>
> Because target head is always properly initialized to empty queue?
>
They may not be empty. But they should be initialized.
>> I suspect in most cases the qtailqs on target are empty.
>
> If I think about migration having no queues populated with
> elements on a target site sounds very reasonable since IFAIU
> the target should not do any work which would populate these
> data structures.
>
See above.
>
>
>> If not,
>> appending to them is a good choice. Clearing them is tricky since
>> each queue probably require a specialized routine to clean. If they
>> are not empty there are must be good reasons for that.
>
> Have you some code or a scenario in mind where this is legit? I
> mean creating a mix of the state(?) we found at the target and
> the state captured at the source does not sound right. I would
> argue that the target should not have any state which is subject
> to migration.
>
> You are right a non-empty queue is trouble, and frankly I never
> considered it as a valid scenario.
>
It may not be a mix of state. It really depends on how the overall state
of the devices is designed. If there are dependence between different
elements of the state, then difference in these elements may break some
consistency. One example is that migrating the length of the qtailq but
appending the content to a non-empty qtailq. In such a case the length
should not be migrated. It should be calculated on target instead.
It should be treated case by case.
Thanks,
Jianjun
> Sorry if I'm bothering you with nonsense.
>
> Greetings,
> Halil
>
>>
>> Thanks,
>> Jianjun
>>
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration
2016-11-03 18:40 ` Jianjun Duan
@ 2016-11-03 18:51 ` Halil Pasic
0 siblings, 0 replies; 18+ messages in thread
From: Halil Pasic @ 2016-11-03 18:51 UTC (permalink / raw)
To: Jianjun Duan, quintela
Cc: veroniabahaa, peter.maydell, dgilbert, mdroth, mst,
mark.cave-ayland, qemu-devel, mreitz, blauwirbel, amit.shah,
qemu-ppc, kraxel, kwolf, dmitry, pbonzini, rth, leon.alrae,
aurelien, david
On 11/03/2016 07:40 PM, Jianjun Duan wrote:
>
>
> On 11/03/2016 10:17 AM, Halil Pasic wrote:
>>
>>
>> On 11/03/2016 05:47 PM, Jianjun Duan wrote:
>>>
>>> On 11/03/2016 05:22 AM, Halil Pasic wrote:
>>>>>
>>>>>
>>>>> On 11/02/2016 11:47 AM, Juan Quintela wrote:
>>>>>>> Jianjun Duan <duanj@linux.vnet.ibm.com> wrote:
>>>>>>>>> Add a test for QTAILQ migration to tests/test-vmstate.c.
>>>>>>>>>
>>>>>>>>> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
>>>>>>>
>>>>>>> Reviewed-by: Juan Quintela <quintela@redhat.com>
>>>>>>>
>>>>>
>>>>> Empty QTAILQ seems to be broken. Have written a small
>>>>> test to prove my point. It May even make sense to have such
>>>>> a test in the test-suite (some prettyfication might be
>>>>> necessary though).
>>>>>
>>> It is working as intended.
>>>
>>
>> My train of thought was that the object holding the queue might
>> be dynamically allocated by the migration code or otherwise
>> uninitialized. I was unaware these scenarios are prohibited.
>>
>>
> This is a valid point. To get this covered vmstate_load_state needs to
> be revised so that at any moment of recursion we know if the field is in
> a dynamic created structure. If yes the structures which need
> initialization such as QTAILQ can be initialized.
>
Or you just zero out the head in VMStateInfo.get right away and to not care
what was there. Of course this is only if loading to non-empty
lists is invalid. This is why I cared describing why I think
it is (invalid).
> I would leave this until the need is there. In current device migration
> code I imagine such scenarios would be rare if they should appear at
> all. Because all the devices (even the hotplugged ones) are already
> initialized on target. So a QTAILQ in such context should already be
> initialized. Otherwise it should be fixed.
>
I agree that this type of a solution is an overkill for something
nobody needs at the moment.
>>> The current design is to append the qtailq from source to the
>>> corresponding one on target.
>>
>> I do not see this documented. I'm used to vmstate_load overwriting
>> values and following pointers, so IMHO it is not obvious that
>> qtailq load does append.
>>
>
> I will document this.
>
I'm fine with this option too but I think I would slightly prefer
the solution described above. Maybe some of the more experienced
guys (think Paolo, Juan, Dave) will come up with some guidance.
>>> It works well for the task in hard
>>> such as migrating ccs_list and pending_events for DRC objects.
>>>
>>
>> Because target head is always properly initialized to empty queue?
>>
> They may not be empty. But they should be initialized.
>>> I suspect in most cases the qtailqs on target are empty.
>>
>> If I think about migration having no queues populated with
>> elements on a target site sounds very reasonable since IFAIU
>> the target should not do any work which would populate these
>> data structures.
>>
> See above.
>>
>
>>
>>> If not,
>>> appending to them is a good choice. Clearing them is tricky since
>>> each queue probably require a specialized routine to clean. If they
>>> are not empty there are must be good reasons for that.
>>
>> Have you some code or a scenario in mind where this is legit? I
>> mean creating a mix of the state(?) we found at the target and
>> the state captured at the source does not sound right. I would
>> argue that the target should not have any state which is subject
>> to migration.
>>
>> You are right a non-empty queue is trouble, and frankly I never
>> considered it as a valid scenario.
>>
> It may not be a mix of state. It really depends on how the overall state
> of the devices is designed. If there are dependence between different
> elements of the state, then difference in these elements may break some
> consistency. One example is that migrating the length of the qtailq but
> appending the content to a non-empty qtailq. In such a case the length
> should not be migrated. It should be calculated on target instead.
>
> It should be treated case by case.
>
I did not get your argument here, but I think we can both live
without me understanding this.
Halil
^ permalink raw reply [flat|nested] 18+ messages in thread
end of thread, other threads:[~2016-11-03 18:51 UTC | newest]
Thread overview: 18+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-10-31 19:53 [Qemu-devel] [QEMU PATCH v10 0/3] migration: migrate QTAILQ Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 1/3] migration: extend VMStateInfo Jianjun Duan
2016-11-02 10:40 ` Juan Quintela
2016-11-02 16:54 ` Jianjun Duan
2016-11-03 10:18 ` Juan Quintela
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 2/3] migration: migrate QTAILQ Jianjun Duan
2016-11-02 10:45 ` Juan Quintela
2016-11-02 16:38 ` Paolo Bonzini
2016-11-03 11:14 ` Halil Pasic
2016-11-03 11:32 ` Halil Pasic
2016-11-02 17:05 ` Jianjun Duan
2016-10-31 19:53 ` [Qemu-devel] [QEMU PATCH v10 3/3] tests/migration: Add test for QTAILQ migration Jianjun Duan
2016-11-02 10:47 ` Juan Quintela
2016-11-03 12:22 ` Halil Pasic
2016-11-03 16:47 ` Jianjun Duan
2016-11-03 17:17 ` Halil Pasic
2016-11-03 18:40 ` Jianjun Duan
2016-11-03 18:51 ` Halil Pasic
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).