* [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices
@ 2025-12-15 17:18 Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 1/7] hw/ppc: Add VMSTATE information for LPC model Caleb Schlossin
` (7 more replies)
0 siblings, 8 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
Addressing comments from V1 review:
Updates in V2:
- Added new patch set for PnvPsi support as it fits with the rest
- Added vmstate support for Power8 and Power9 for LPC
- Fixed pnv_core.c commit message
Tested:
passed make check
Thanks,
Caleb
Michael Kowal (2):
hw/ppc: Add VMSTATE information for LPC model
hw/ppc: Add VMSTATE information to PnvPsi
Caleb Schlossin (2):
hw/ppc: Add pnv_spi vmstate support
hw/ppc: Add pnv_i2c vmstate support
Angelo Jaramillo (3):
hw/ppc: pnv_adu.c added vmstate support
hw/ppc: pnv_core.c add vmstate support
hw/ppc: pnv_chiptod.c add vmstate support
hw/ppc/pnv_adu.c | 12 ++++++++++
hw/ppc/pnv_chiptod.c | 38 +++++++++++++++++++++++++++++
hw/ppc/pnv_core.c | 22 +++++++++++++++++
hw/ppc/pnv_i2c.c | 11 +++++++++
hw/ppc/pnv_lpc.c | 41 ++++++++++++++++++++++++++++++++
hw/ppc/pnv_psi.c | 46 ++++++++++++++++++++++++++++++++++--
hw/ssi/pnv_spi.c | 27 +++++++++++++++++++++
include/hw/ppc/pnv_chiptod.h | 2 ++
8 files changed, 197 insertions(+), 2 deletions(-)
--
2.47.3
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH v2 1/7] hw/ppc: Add VMSTATE information for LPC model
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 2/7] hw/ppc: Add pnv_spi vmstate support Caleb Schlossin
` (6 subsequent siblings)
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
The PNV LPC model needs snapshot/migration support. Added a VMSTATE
descriptor to save model data and an associated post_load() method.
Snapshot support added for Power8, Power9, and Power10.
Signed-off-by: Michael Kowal <kowal@linux.ibm.com>
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_lpc.c | 41 +++++++++++++++++++++++++++++++++++++++++
1 file changed, 41 insertions(+)
diff --git a/hw/ppc/pnv_lpc.c b/hw/ppc/pnv_lpc.c
index f6beba0917..e52a062181 100644
--- a/hw/ppc/pnv_lpc.c
+++ b/hw/ppc/pnv_lpc.c
@@ -30,6 +30,7 @@
#include "hw/ppc/pnv_lpc.h"
#include "hw/ppc/pnv_xscom.h"
#include "hw/ppc/fdt.h"
+#include "migration/vmstate.h"
#include <libfdt.h>
@@ -696,6 +697,43 @@ static const MemoryRegionOps opb_master_ops = {
},
};
+static int vmstate_pnv_lpc_post_load(void *opaque, int version_id)
+{
+ PnvLpcController *lpc = PNV_LPC(opaque);
+
+ memory_region_set_alias_offset(&lpc->opb_isa_fw,
+ lpc->lpc_hc_fw_seg_idsel * LPC_FW_OPB_SIZE);
+ pnv_lpc_eval_serirq_routes(lpc);
+
+ pnv_lpc_eval_irqs(lpc);
+ return 0;
+}
+
+static const VMStateDescription vmstate_pnv_lpc = {
+ .name = TYPE_PNV_LPC,
+ .version_id = 1,
+ .minimum_version_id = 1,
+ .post_load = vmstate_pnv_lpc_post_load,
+ .fields = (const VMStateField[]) {
+ VMSTATE_UINT64(eccb_stat_reg, PnvLpcController),
+ VMSTATE_UINT32(eccb_data_reg, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_route0, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_route1, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_stat, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_mask, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_pol, PnvLpcController),
+ VMSTATE_UINT32(opb_irq_input, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_irq_inputs, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_fw_seg_idsel, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_irqser_ctrl, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_irqmask, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_irqstat, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_error_addr, PnvLpcController),
+ VMSTATE_UINT32(lpc_hc_fw_rd_acc_size, PnvLpcController),
+ VMSTATE_END_OF_LIST()
+ }
+};
+
static void pnv_lpc_power8_realize(DeviceState *dev, Error **errp)
{
PnvLpcController *lpc = PNV_LPC(dev);
@@ -721,6 +759,7 @@ static void pnv_lpc_power8_class_init(ObjectClass *klass, const void *data)
PnvLpcClass *plc = PNV_LPC_CLASS(klass);
dc->desc = "PowerNV LPC Controller POWER8";
+ dc->vmsd = &vmstate_pnv_lpc;
xdc->dt_xscom = pnv_lpc_dt_xscom;
@@ -766,6 +805,7 @@ static void pnv_lpc_power9_class_init(ObjectClass *klass, const void *data)
PnvLpcClass *plc = PNV_LPC_CLASS(klass);
dc->desc = "PowerNV LPC Controller POWER9";
+ dc->vmsd = &vmstate_pnv_lpc;
device_class_set_parent_realize(dc, pnv_lpc_power9_realize,
&plc->parent_realize);
@@ -782,6 +822,7 @@ static void pnv_lpc_power10_class_init(ObjectClass *klass, const void *data)
DeviceClass *dc = DEVICE_CLASS(klass);
dc->desc = "PowerNV LPC Controller POWER10";
+ dc->vmsd = &vmstate_pnv_lpc;
}
static const TypeInfo pnv_lpc_power10_info = {
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 2/7] hw/ppc: Add pnv_spi vmstate support
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 1/7] hw/ppc: Add VMSTATE information for LPC model Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 3/7] hw/ppc: Add pnv_i2c " Caleb Schlossin
` (5 subsequent siblings)
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
- Add support for needed PnvSpi structure variables
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ssi/pnv_spi.c | 27 +++++++++++++++++++++++++++
1 file changed, 27 insertions(+)
diff --git a/hw/ssi/pnv_spi.c b/hw/ssi/pnv_spi.c
index f40e8836b9..389a2cca6b 100644
--- a/hw/ssi/pnv_spi.c
+++ b/hw/ssi/pnv_spi.c
@@ -13,6 +13,7 @@
#include "hw/ssi/pnv_spi.h"
#include "hw/ssi/pnv_spi_regs.h"
#include "hw/ssi/ssi.h"
+#include "migration/vmstate.h"
#include <libfdt.h>
#include "hw/irq.h"
#include "trace.h"
@@ -1199,6 +1200,31 @@ static int pnv_spi_dt_xscom(PnvXScomInterface *dev, void *fdt,
return 0;
}
+static const VMStateDescription pnv_spi_vmstate = {
+ .name = TYPE_PNV_SPI,
+ .version_id = 1,
+ .fields = (const VMStateField[]) {
+ VMSTATE_UINT8(fail_count, PnvSpi),
+ VMSTATE_UINT8(transfer_len, PnvSpi),
+ VMSTATE_UINT8(responder_select, PnvSpi),
+ VMSTATE_BOOL(shift_n1_done, PnvSpi),
+ VMSTATE_UINT8(loop_counter_1, PnvSpi),
+ VMSTATE_UINT8(loop_counter_2, PnvSpi),
+ VMSTATE_UINT8(N1_bits, PnvSpi),
+ VMSTATE_UINT8(N2_bits, PnvSpi),
+ VMSTATE_UINT8(N1_bytes, PnvSpi),
+ VMSTATE_UINT8(N2_bytes, PnvSpi),
+ VMSTATE_UINT8(N1_tx, PnvSpi),
+ VMSTATE_UINT8(N2_tx, PnvSpi),
+ VMSTATE_UINT8(N1_rx, PnvSpi),
+ VMSTATE_UINT8(N2_rx, PnvSpi),
+ VMSTATE_UINT64_ARRAY(regs, PnvSpi, PNV_SPI_REGS),
+ VMSTATE_UINT8_ARRAY(seq_op, PnvSpi, PNV_SPI_REG_SIZE),
+ VMSTATE_UINT64(status, PnvSpi),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_spi_class_init(ObjectClass *klass, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
@@ -1209,6 +1235,7 @@ static void pnv_spi_class_init(ObjectClass *klass, const void *data)
dc->desc = "PowerNV SPI";
dc->realize = pnv_spi_realize;
device_class_set_legacy_reset(dc, do_reset);
+ dc->vmsd = &pnv_spi_vmstate;
device_class_set_props(dc, pnv_spi_properties);
}
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 3/7] hw/ppc: Add pnv_i2c vmstate support
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 1/7] hw/ppc: Add VMSTATE information for LPC model Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 2/7] hw/ppc: Add pnv_spi vmstate support Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 4/7] hw/ppc: pnv_adu.c added " Caleb Schlossin
` (4 subsequent siblings)
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
- Add vmstate support for i2c registers
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_i2c.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/hw/ppc/pnv_i2c.c b/hw/ppc/pnv_i2c.c
index 60de479491..1018078228 100644
--- a/hw/ppc/pnv_i2c.c
+++ b/hw/ppc/pnv_i2c.c
@@ -19,6 +19,7 @@
#include "hw/ppc/pnv_i2c.h"
#include "hw/ppc/pnv_xscom.h"
#include "hw/ppc/fdt.h"
+#include "migration/vmstate.h"
#include <libfdt.h>
@@ -549,6 +550,15 @@ static const Property pnv_i2c_properties[] = {
DEFINE_PROP_UINT32("num-busses", PnvI2C, num_busses, 1),
};
+static const VMStateDescription pnv_i2c_vmstate = {
+ .name = TYPE_PNV_I2C,
+ .version_id = 1,
+ .fields = (const VMStateField[]) {
+ VMSTATE_UINT64_ARRAY(regs, PnvI2C, PNV_I2C_REGS),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_i2c_class_init(ObjectClass *klass, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
@@ -561,6 +571,7 @@ static void pnv_i2c_class_init(ObjectClass *klass, const void *data)
dc->desc = "PowerNV I2C";
dc->realize = pnv_i2c_realize;
+ dc->vmsd = &pnv_i2c_vmstate;
device_class_set_props(dc, pnv_i2c_properties);
}
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 4/7] hw/ppc: pnv_adu.c added vmstate support
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
` (2 preceding siblings ...)
2025-12-15 17:18 ` [PATCH v2 3/7] hw/ppc: Add pnv_i2c " Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 5/7] hw/ppc: pnv_core.c add " Caleb Schlossin
` (3 subsequent siblings)
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
- Added vmstate support for ADU model
Signed-off-by: Angelo Jaramillo <angelo.jaramillo@linux.ibm.com>
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_adu.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/hw/ppc/pnv_adu.c b/hw/ppc/pnv_adu.c
index 005fbda475..bd2a9e233a 100644
--- a/hw/ppc/pnv_adu.c
+++ b/hw/ppc/pnv_adu.c
@@ -23,6 +23,7 @@
#include "hw/ppc/pnv_chip.h"
#include "hw/ppc/pnv_lpc.h"
#include "hw/ppc/pnv_xscom.h"
+#include "migration/vmstate.h"
#include "trace.h"
#define ADU_LPC_BASE_REG 0x40
@@ -189,6 +190,16 @@ static const Property pnv_adu_properties[] = {
DEFINE_PROP_LINK("lpc", PnvADU, lpc, TYPE_PNV_LPC, PnvLpcController *),
};
+static const VMStateDescription pnv_adu_vmstate = {
+ .name = TYPE_PNV_ADU,
+ .version_id = 1,
+ .fields = (const VMStateField[]) {
+ VMSTATE_UINT64(lpc_cmd_reg, PnvADU),
+ VMSTATE_UINT64(lpc_data_reg, PnvADU),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_adu_class_init(ObjectClass *klass, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
@@ -197,6 +208,7 @@ static void pnv_adu_class_init(ObjectClass *klass, const void *data)
dc->desc = "PowerNV ADU";
device_class_set_props(dc, pnv_adu_properties);
dc->user_creatable = false;
+ dc->vmsd = &pnv_adu_vmstate;
}
static const TypeInfo pnv_adu_type_info = {
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 5/7] hw/ppc: pnv_core.c add vmstate support
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
` (3 preceding siblings ...)
2025-12-15 17:18 ` [PATCH v2 4/7] hw/ppc: pnv_adu.c added " Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 6/7] hw/ppc: pnv_chiptod.c " Caleb Schlossin
` (2 subsequent siblings)
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
- Add vmstate support PnvCore and PnvQuad capturing scratch
registers and special wakeup registers
Signed-off-by: Angelo Jaramillo <angelo.jaramillo@linux.ibm.com>
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_core.c | 22 ++++++++++++++++++++++
1 file changed, 22 insertions(+)
diff --git a/hw/ppc/pnv_core.c b/hw/ppc/pnv_core.c
index fb2dfc7ba2..03b64f0013 100644
--- a/hw/ppc/pnv_core.c
+++ b/hw/ppc/pnv_core.c
@@ -31,6 +31,7 @@
#include "hw/ppc/xics.h"
#include "hw/qdev-properties.h"
#include "helper_regs.h"
+#include "migration/vmstate.h"
static const char *pnv_core_cpu_typename(PnvCore *pc)
{
@@ -478,6 +479,15 @@ static void pnv_core_power11_class_init(ObjectClass *oc, const void *data)
pnv_core_power10_class_init(oc, data);
}
+static const VMStateDescription pnv_core_vmstate = {
+ .name = TYPE_PNV_CORE,
+ .version_id = 1,
+ .fields = (const VMStateField[]) {
+ VMSTATE_UINT64_ARRAY(scratch, PnvCore, 8),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_core_class_init(ObjectClass *oc, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(oc);
@@ -486,6 +496,7 @@ static void pnv_core_class_init(ObjectClass *oc, const void *data)
dc->unrealize = pnv_core_unrealize;
device_class_set_props(dc, pnv_core_properties);
dc->user_creatable = false;
+ dc->vmsd = &pnv_core_vmstate;
}
#define DEFINE_PNV_CORE_TYPE(family, cpu_model) \
@@ -737,12 +748,23 @@ static void pnv_quad_power11_class_init(ObjectClass *oc, const void *data)
pnv_quad_power10_class_init(oc, data);
}
+static const VMStateDescription pnv_quad_vmstate = {
+ .name = TYPE_PNV_QUAD,
+ .version_id = 1,
+ .fields = (const VMStateField[]) {
+ VMSTATE_BOOL(special_wakeup_done, PnvQuad),
+ VMSTATE_BOOL_ARRAY(special_wakeup, PnvQuad, 4),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_quad_class_init(ObjectClass *oc, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(oc);
device_class_set_props(dc, pnv_quad_properties);
dc->user_creatable = false;
+ dc->vmsd = &pnv_quad_vmstate;
}
static const TypeInfo pnv_quad_infos[] = {
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 6/7] hw/ppc: pnv_chiptod.c add vmstate support
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
` (4 preceding siblings ...)
2025-12-15 17:18 ` [PATCH v2 5/7] hw/ppc: pnv_core.c add " Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi Caleb Schlossin
2025-12-16 13:02 ` [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Aditya Gupta
7 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
- Added pre_save and post_load methods to handle slave_pc_target and tod_state
Signed-off-by: Angelo Jaramillo <angelo.jaramillo@linux.ibm.com>
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_chiptod.c | 38 ++++++++++++++++++++++++++++++++++++
include/hw/ppc/pnv_chiptod.h | 2 ++
2 files changed, 40 insertions(+)
diff --git a/hw/ppc/pnv_chiptod.c b/hw/ppc/pnv_chiptod.c
index f887a18cde..9dc5942ca0 100644
--- a/hw/ppc/pnv_chiptod.c
+++ b/hw/ppc/pnv_chiptod.c
@@ -37,6 +37,7 @@
#include "hw/ppc/pnv_core.h"
#include "hw/ppc/pnv_xscom.h"
#include "hw/ppc/pnv_chiptod.h"
+#include "migration/vmstate.h"
#include "trace.h"
#include <libfdt.h>
@@ -341,6 +342,8 @@ static void pnv_chiptod_xscom_write(void *opaque, hwaddr addr,
" TOD_TX_TTYPE_CTRL_REG val 0x%" PRIx64
" invalid slave address\n", val);
}
+ /* Write slave_pc_target to a uint64_t variable for vmstate support. */
+ chiptod->tx_ttype_ctrl = val;
break;
case TOD_ERROR_REG:
chiptod->tod_error &= ~val;
@@ -613,6 +616,40 @@ static void pnv_chiptod_unrealize(DeviceState *dev)
qemu_unregister_reset(pnv_chiptod_reset, chiptod);
}
+static int vmstate_pnv_chiptod_pre_save(void *opaque)
+{
+ PnvChipTOD *chiptod = PNV_CHIPTOD(opaque);
+ chiptod->tod_state_val = (uint8_t)chiptod->tod_state;
+ return 0;
+}
+
+static int vmstate_pnv_chiptod_post_load(void *opaque)
+{
+ PnvChipTOD *chiptod = PNV_CHIPTOD(opaque);
+ if (chiptod->tx_ttype_ctrl != 0) {
+ pnv_chiptod_xscom_write(chiptod, TOD_TX_TTYPE_CTRL_REG << 3,
+ chiptod->tx_ttype_ctrl, 8);
+ }
+ chiptod->tod_state = (enum tod_state)chiptod->tod_state_val;
+ return 0;
+}
+
+static const VMStateDescription pnv_chiptod_vmstate = {
+ .name = TYPE_PNV_CHIPTOD,
+ .version_id = 1,
+ .pre_save = vmstate_pnv_chiptod_pre_save,
+ .pre_load = vmstate_pnv_chiptod_post_load,
+ .fields = (const VMStateField[]) {
+ VMSTATE_BOOL(primary, PnvChipTOD),
+ VMSTATE_BOOL(secondary, PnvChipTOD),
+ VMSTATE_UINT64(tod_error, PnvChipTOD),
+ VMSTATE_UINT64(pss_mss_ctrl_reg, PnvChipTOD),
+ VMSTATE_UINT64(tx_ttype_ctrl, PnvChipTOD),
+ VMSTATE_UINT8(tod_state_val, PnvChipTOD),
+ VMSTATE_END_OF_LIST(),
+ },
+};
+
static void pnv_chiptod_class_init(ObjectClass *klass, const void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
@@ -621,6 +658,7 @@ static void pnv_chiptod_class_init(ObjectClass *klass, const void *data)
dc->unrealize = pnv_chiptod_unrealize;
dc->desc = "PowerNV ChipTOD Controller";
dc->user_creatable = false;
+ dc->vmsd = &pnv_chiptod_vmstate;
}
static const TypeInfo pnv_chiptod_type_info = {
diff --git a/include/hw/ppc/pnv_chiptod.h b/include/hw/ppc/pnv_chiptod.h
index 466b06560a..3e5e3b02b2 100644
--- a/include/hw/ppc/pnv_chiptod.h
+++ b/include/hw/ppc/pnv_chiptod.h
@@ -41,6 +41,8 @@ struct PnvChipTOD {
uint64_t tod_error;
uint64_t pss_mss_ctrl_reg;
PnvCore *slave_pc_target;
+ uint64_t tx_ttype_ctrl;
+ uint8_t tod_state_val;
};
struct PnvChipTODClass {
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
` (5 preceding siblings ...)
2025-12-15 17:18 ` [PATCH v2 6/7] hw/ppc: pnv_chiptod.c " Caleb Schlossin
@ 2025-12-15 17:18 ` Caleb Schlossin
2025-12-16 12:58 ` Aditya Gupta
2025-12-16 13:02 ` [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Aditya Gupta
7 siblings, 1 reply; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-15 17:18 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-ppc, npiggin, adityag, milesg, alistair, kowal, chalapathi.v,
calebs, angeloj
PnvPsi needs to be able to save/load snapshots. Add VMSTATE information
to the device class and a post_load() method to restore dynamic data items and
memory region mappings.
Signed-off-by: Michael Kowal <kowal@linux.ibm.com>
Signed-off-by: Caleb Schlossin <calebs@linux.ibm.com>
---
hw/ppc/pnv_psi.c | 46 ++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 44 insertions(+), 2 deletions(-)
diff --git a/hw/ppc/pnv_psi.c b/hw/ppc/pnv_psi.c
index 5d947d8b52..88d5f1d45d 100644
--- a/hw/ppc/pnv_psi.c
+++ b/hw/ppc/pnv_psi.c
@@ -25,6 +25,7 @@
#include "qemu/module.h"
#include "system/reset.h"
#include "qapi/error.h"
+#include "migration/vmstate.h"
#include "hw/ppc/fdt.h"
@@ -35,6 +36,8 @@
#include <libfdt.h>
+#undef PSI_DEBUG
+
#define PSIHB_XSCOM_FIR_RW 0x00
#define PSIHB_XSCOM_FIR_AND 0x01
#define PSIHB_XSCOM_FIR_OR 0x02
@@ -130,12 +133,11 @@ static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
{
PnvPsiClass *ppc = PNV_PSI_GET_CLASS(psi);
MemoryRegion *sysmem = get_system_memory();
- uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
psi->regs[PSIHB_XSCOM_BAR] = bar & (ppc->bar_mask | PSIHB_BAR_EN);
/* Update MR, always remove it first */
- if (old & PSIHB_BAR_EN) {
+ if (memory_region_is_mapped(&psi->regs_mr)) {
memory_region_del_subregion(sysmem, &psi->regs_mr);
}
@@ -975,6 +977,40 @@ static void pnv_psi_register_types(void)
type_init(pnv_psi_register_types);
+#ifdef PSI_DEBUG
+static void psi_regs_pic_print_info(uint64_t *regs, uint32_t nr_regs,
+ GString *buf) {
+ uint i, prev_idx = -1;
+ uint64_t reg1, prev_reg1 = -1;
+ uint64_t reg2, prev_reg2 = -1;
+ uint64_t reg3, prev_reg3 = -1;
+ uint64_t reg4, prev_reg4 = -1;
+ for (i = 0; i < nr_regs; i = i + 4) {
+ /* Don't print if values do not change, but print last*/
+ reg1 = regs[i];
+ reg2 = regs[i + 1];
+ reg3 = regs[i + 2];
+ reg4 = regs[i + 3];
+ if (reg1 == prev_reg1 && reg2 == prev_reg2 &&
+ reg3 == prev_reg3 && reg4 == prev_reg4 &&
+ i < (nr_regs - 4)) {
+ if (i == (prev_idx + 4)) {
+ g_string_append_printf(buf, " . . .\n");
+ }
+ continue;
+ }
+
+ g_string_append_printf(buf, " [%03X] 0x%016lX %016lX %016lX %016lX\n",
+ i, reg1, reg2, reg3, reg4);
+ prev_idx = i;
+ prev_reg1 = reg1;
+ prev_reg2 = reg2;
+ prev_reg3 = reg3;
+ prev_reg4 = reg4;
+ }
+}
+#endif
+
void pnv_psi_pic_print_info(Pnv9Psi *psi9, GString *buf)
{
PnvPsi *psi = PNV_PSI(psi9);
@@ -985,4 +1021,10 @@ void pnv_psi_pic_print_info(Pnv9Psi *psi9, GString *buf)
g_string_append_printf(buf, "PSIHB Source %08x .. %08x\n",
offset, offset + psi9->source.nr_irqs - 1);
xive_source_pic_print_info(&psi9->source, offset, buf);
+#ifdef PSI_DEBUG
+ /* Print PSI registers */
+ g_string_append_printf(buf, "\nPSI Regs[0x0..%X]\n",
+ PSIHB_XSCOM_MAX);
+ psi_regs_pic_print_info(psi->regs, PSIHB_XSCOM_MAX, buf);
+#endif
}
--
2.47.3
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi
2025-12-15 17:18 ` [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi Caleb Schlossin
@ 2025-12-16 12:58 ` Aditya Gupta
2025-12-16 13:35 ` Caleb Schlossin
0 siblings, 1 reply; 13+ messages in thread
From: Aditya Gupta @ 2025-12-16 12:58 UTC (permalink / raw)
To: Caleb Schlossin
Cc: qemu-devel, qemu-ppc, npiggin, milesg, alistair, kowal,
chalapathi.v, angeloj
Hello Caleb,
On 25/12/15 11:18AM, Caleb Schlossin wrote:
> <...snip...>
>
> --- a/hw/ppc/pnv_psi.c
> +++ b/hw/ppc/pnv_psi.c
> @@ -25,6 +25,7 @@
> #include "qemu/module.h"
> #include "system/reset.h"
> #include "qapi/error.h"
> +#include "migration/vmstate.h"
>
>
> #include "hw/ppc/fdt.h"
> @@ -35,6 +36,8 @@
>
> #include <libfdt.h>
>
> +#undef PSI_DEBUG
> +
Is this intended or got left over from debugging ?
> #define PSIHB_XSCOM_FIR_RW 0x00
> #define PSIHB_XSCOM_FIR_AND 0x01
> #define PSIHB_XSCOM_FIR_OR 0x02
> @@ -130,12 +133,11 @@ static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
> {
> PnvPsiClass *ppc = PNV_PSI_GET_CLASS(psi);
> MemoryRegion *sysmem = get_system_memory();
> - uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
>
> psi->regs[PSIHB_XSCOM_BAR] = bar & (ppc->bar_mask | PSIHB_BAR_EN);
>
> /* Update MR, always remove it first */
> - if (old & PSIHB_BAR_EN) {
> + if (memory_region_is_mapped(&psi->regs_mr)) {
> memory_region_del_subregion(sysmem, &psi->regs_mr);
> }
>
> @@ -975,6 +977,40 @@ static void pnv_psi_register_types(void)
>
> type_init(pnv_psi_register_types);
>
> +#ifdef PSI_DEBUG
> +static void psi_regs_pic_print_info(uint64_t *regs, uint32_t nr_regs,
> + GString *buf) {
> + uint i, prev_idx = -1;
> + uint64_t reg1, prev_reg1 = -1;
> + uint64_t reg2, prev_reg2 = -1;
> + uint64_t reg3, prev_reg3 = -1;
> + uint64_t reg4, prev_reg4 = -1;
Very minor nitpick, 2 spaces in the declaration between type and name.
checkpatch doesn't point it out, so it's okay with me.
Looks good to me overall. Please just see if the #undef was intentional,
if so:
Reviewed-by: Aditya Gupta <adityag@linux.ibm.com>
Thanks,
- Aditya G
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
` (6 preceding siblings ...)
2025-12-15 17:18 ` [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi Caleb Schlossin
@ 2025-12-16 13:02 ` Aditya Gupta
7 siblings, 0 replies; 13+ messages in thread
From: Aditya Gupta @ 2025-12-16 13:02 UTC (permalink / raw)
To: Caleb Schlossin
Cc: qemu-devel, qemu-ppc, npiggin, milesg, alistair, kowal,
chalapathi.v, angeloj
On 25/12/15 11:18AM, Caleb Schlossin wrote:
> Addressing comments from V1 review:
>
> Updates in V2:
> - Added new patch set for PnvPsi support as it fits with the rest
> - Added vmstate support for Power8 and Power9 for LPC
> - Fixed pnv_core.c commit message
>
> Tested:
> passed make check
Looks good to me.
Had one comment in PSI about #undef. With that sorted:
For the series:
Reviewed-by: Aditya Gupta <adityag@linux.ibm.com>
Thanks,
- Aditya G
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi
2025-12-16 12:58 ` Aditya Gupta
@ 2025-12-16 13:35 ` Caleb Schlossin
2025-12-16 14:36 ` Aditya Gupta
0 siblings, 1 reply; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-16 13:35 UTC (permalink / raw)
To: Aditya Gupta
Cc: qemu-devel, qemu-ppc, npiggin, milesg, alistair, kowal,
chalapathi.v, angeloj
On 12/16/25 6:58 AM, Aditya Gupta wrote:
> Hello Caleb,
>
> On 25/12/15 11:18AM, Caleb Schlossin wrote:
>> <...snip...>
>>
>> --- a/hw/ppc/pnv_psi.c
>> +++ b/hw/ppc/pnv_psi.c
>> @@ -25,6 +25,7 @@
>> #include "qemu/module.h"
>> #include "system/reset.h"
>> #include "qapi/error.h"
>> +#include "migration/vmstate.h"
>>
>>
>> #include "hw/ppc/fdt.h"
>> @@ -35,6 +36,8 @@
>>
>> #include <libfdt.h>
>>
>> +#undef PSI_DEBUG
>> +
>
> Is this intended or got left over from debugging ?
This was indented to aid future debug, if needed.
>
>> #define PSIHB_XSCOM_FIR_RW 0x00
>> #define PSIHB_XSCOM_FIR_AND 0x01
>> #define PSIHB_XSCOM_FIR_OR 0x02
>> @@ -130,12 +133,11 @@ static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
>> {
>> PnvPsiClass *ppc = PNV_PSI_GET_CLASS(psi);
>> MemoryRegion *sysmem = get_system_memory();
>> - uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
>>
>> psi->regs[PSIHB_XSCOM_BAR] = bar & (ppc->bar_mask | PSIHB_BAR_EN);
>>
>> /* Update MR, always remove it first */
>> - if (old & PSIHB_BAR_EN) {
>> + if (memory_region_is_mapped(&psi->regs_mr)) {
>> memory_region_del_subregion(sysmem, &psi->regs_mr);
>> }
>>
>> @@ -975,6 +977,40 @@ static void pnv_psi_register_types(void)
>>
>> type_init(pnv_psi_register_types);
>>
>> +#ifdef PSI_DEBUG
>> +static void psi_regs_pic_print_info(uint64_t *regs, uint32_t nr_regs,
>> + GString *buf) {
>> + uint i, prev_idx = -1;
>> + uint64_t reg1, prev_reg1 = -1;
>> + uint64_t reg2, prev_reg2 = -1;
>> + uint64_t reg3, prev_reg3 = -1;
>> + uint64_t reg4, prev_reg4 = -1;
>
> Very minor nitpick, 2 spaces in the declaration between type and name.
> checkpatch doesn't point it out, so it's okay with me.
>
> Looks good to me overall. Please just see if the #undef was intentional,
> if so:
>
> Reviewed-by: Aditya Gupta <adityag@linux.ibm.com>
>
> Thanks,
> - Aditya G
>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi
2025-12-16 13:35 ` Caleb Schlossin
@ 2025-12-16 14:36 ` Aditya Gupta
2025-12-16 14:56 ` Caleb Schlossin
0 siblings, 1 reply; 13+ messages in thread
From: Aditya Gupta @ 2025-12-16 14:36 UTC (permalink / raw)
To: Caleb Schlossin
Cc: qemu-devel, qemu-ppc, npiggin, milesg, alistair, kowal,
chalapathi.v, angeloj
On 25/12/16 07:35AM, Caleb Schlossin wrote:
>
>
> On 12/16/25 6:58 AM, Aditya Gupta wrote:
> > Hello Caleb,
> >
> > On 25/12/15 11:18AM, Caleb Schlossin wrote:
> >> <...snip...>
> >>
> >> --- a/hw/ppc/pnv_psi.c
> >> +++ b/hw/ppc/pnv_psi.c
> >> @@ -25,6 +25,7 @@
> >> #include "qemu/module.h"
> >> #include "system/reset.h"
> >> #include "qapi/error.h"
> >> +#include "migration/vmstate.h"
> >>
> >>
> >> #include "hw/ppc/fdt.h"
> >> @@ -35,6 +36,8 @@
> >>
> >> #include <libfdt.h>
> >>
> >> +#undef PSI_DEBUG
> >> +
> >
> > Is this intended or got left over from debugging ?
>
> This was indented to aid future debug, if needed.
>
If so, i believe this should be removed.
With this we just unconditionally noped out the #ifdef PSI_DEBUG
portions in this file, even if the user explicitly compiles with PSI_DEBUG
What do you say ?
- Aditya G
> >
> >> #define PSIHB_XSCOM_FIR_RW 0x00
> >> #define PSIHB_XSCOM_FIR_AND 0x01
> >> #define PSIHB_XSCOM_FIR_OR 0x02
> >> @@ -130,12 +133,11 @@ static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
> >> {
> >> PnvPsiClass *ppc = PNV_PSI_GET_CLASS(psi);
> >> MemoryRegion *sysmem = get_system_memory();
> >> - uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
> >>
> >> psi->regs[PSIHB_XSCOM_BAR] = bar & (ppc->bar_mask | PSIHB_BAR_EN);
> >>
> >> /* Update MR, always remove it first */
> >> - if (old & PSIHB_BAR_EN) {
> >> + if (memory_region_is_mapped(&psi->regs_mr)) {
> >> memory_region_del_subregion(sysmem, &psi->regs_mr);
> >> }
> >>
> >> @@ -975,6 +977,40 @@ static void pnv_psi_register_types(void)
> >>
> >> type_init(pnv_psi_register_types);
> >>
> >> +#ifdef PSI_DEBUG
> >> +static void psi_regs_pic_print_info(uint64_t *regs, uint32_t nr_regs,
> >> + GString *buf) {
> >> + uint i, prev_idx = -1;
> >> + uint64_t reg1, prev_reg1 = -1;
> >> + uint64_t reg2, prev_reg2 = -1;
> >> + uint64_t reg3, prev_reg3 = -1;
> >> + uint64_t reg4, prev_reg4 = -1;
> >
> > Very minor nitpick, 2 spaces in the declaration between type and name.
> > checkpatch doesn't point it out, so it's okay with me.
> >
> > Looks good to me overall. Please just see if the #undef was intentional,
> > if so:
> >
> > Reviewed-by: Aditya Gupta <adityag@linux.ibm.com>
> >
> > Thanks,
> > - Aditya G
> >
>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi
2025-12-16 14:36 ` Aditya Gupta
@ 2025-12-16 14:56 ` Caleb Schlossin
0 siblings, 0 replies; 13+ messages in thread
From: Caleb Schlossin @ 2025-12-16 14:56 UTC (permalink / raw)
To: Aditya Gupta
Cc: qemu-devel, qemu-ppc, npiggin, milesg, alistair, kowal,
chalapathi.v, angeloj
On 12/16/25 8:36 AM, Aditya Gupta wrote:
> On 25/12/16 07:35AM, Caleb Schlossin wrote:
>>
>>
>> On 12/16/25 6:58 AM, Aditya Gupta wrote:
>>> Hello Caleb,
>>>
>>> On 25/12/15 11:18AM, Caleb Schlossin wrote:
>>>> <...snip...>
>>>>
>>>> --- a/hw/ppc/pnv_psi.c
>>>> +++ b/hw/ppc/pnv_psi.c
>>>> @@ -25,6 +25,7 @@
>>>> #include "qemu/module.h"
>>>> #include "system/reset.h"
>>>> #include "qapi/error.h"
>>>> +#include "migration/vmstate.h"
>>>>
>>>>
>>>> #include "hw/ppc/fdt.h"
>>>> @@ -35,6 +36,8 @@
>>>>
>>>> #include <libfdt.h>
>>>>
>>>> +#undef PSI_DEBUG
>>>> +
>>>
>>> Is this intended or got left over from debugging ?
>>
>> This was indented to aid future debug, if needed.
>>
>
> If so, i believe this should be removed.
>
> With this we just unconditionally noped out the #ifdef PSI_DEBUG
> portions in this file, even if the user explicitly compiles with PSI_DEBUG
>
> What do you say ?
>
> - Aditya G
Sounds good, I'll remove it and send up a new set of patches.
>
>>>
>>>> #define PSIHB_XSCOM_FIR_RW 0x00
>>>> #define PSIHB_XSCOM_FIR_AND 0x01
>>>> #define PSIHB_XSCOM_FIR_OR 0x02
>>>> @@ -130,12 +133,11 @@ static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
>>>> {
>>>> PnvPsiClass *ppc = PNV_PSI_GET_CLASS(psi);
>>>> MemoryRegion *sysmem = get_system_memory();
>>>> - uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
>>>>
>>>> psi->regs[PSIHB_XSCOM_BAR] = bar & (ppc->bar_mask | PSIHB_BAR_EN);
>>>>
>>>> /* Update MR, always remove it first */
>>>> - if (old & PSIHB_BAR_EN) {
>>>> + if (memory_region_is_mapped(&psi->regs_mr)) {
>>>> memory_region_del_subregion(sysmem, &psi->regs_mr);
>>>> }
>>>>
>>>> @@ -975,6 +977,40 @@ static void pnv_psi_register_types(void)
>>>>
>>>> type_init(pnv_psi_register_types);
>>>>
>>>> +#ifdef PSI_DEBUG
>>>> +static void psi_regs_pic_print_info(uint64_t *regs, uint32_t nr_regs,
>>>> + GString *buf) {
>>>> + uint i, prev_idx = -1;
>>>> + uint64_t reg1, prev_reg1 = -1;
>>>> + uint64_t reg2, prev_reg2 = -1;
>>>> + uint64_t reg3, prev_reg3 = -1;
>>>> + uint64_t reg4, prev_reg4 = -1;
>>>
>>> Very minor nitpick, 2 spaces in the declaration between type and name.
>>> checkpatch doesn't point it out, so it's okay with me.
>>>
>>> Looks good to me overall. Please just see if the #undef was intentional,
>>> if so:
>>>
>>> Reviewed-by: Aditya Gupta <adityag@linux.ibm.com>
>>>
>>> Thanks,
>>> - Aditya G
>>>
>>
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2025-12-16 14:57 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-12-15 17:18 [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 1/7] hw/ppc: Add VMSTATE information for LPC model Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 2/7] hw/ppc: Add pnv_spi vmstate support Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 3/7] hw/ppc: Add pnv_i2c " Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 4/7] hw/ppc: pnv_adu.c added " Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 5/7] hw/ppc: pnv_core.c add " Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 6/7] hw/ppc: pnv_chiptod.c " Caleb Schlossin
2025-12-15 17:18 ` [PATCH v2 7/7] hw/ppc: Add VMSTATE information to PnvPsi Caleb Schlossin
2025-12-16 12:58 ` Aditya Gupta
2025-12-16 13:35 ` Caleb Schlossin
2025-12-16 14:36 ` Aditya Gupta
2025-12-16 14:56 ` Caleb Schlossin
2025-12-16 13:02 ` [PATCH v2 0/7] hw/ppc: Snapshot support for several ppc devices Aditya Gupta
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).