vmstate: Fix info field of VMSTATE_MACADDR
PATCH 1/1 vmstate: Fix info field of VMSTATE_MACADDR
Signed-off-by Jan Kiszka <[email protected]>
---
This may partly or fully explain the e1000 issue Pierre reported first. However, there are more bugs in the current migration code, e.g. in qemu_fseek when used with streams, it still doesn't work properly here. Need a break...
hw/hw.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
hw/hw.h +1 -1
@@ -546,7 +546,7 @@ extern const VMStateDescription vmstate_i2c_slave;#define VMSTATE_MACADDR (_field, _state) { \ .name = (stringify(_field)), \ .size = sizeof(MACAddr), \ - .info = &vmstate_info_uint8, \ + .info = &vmstate_info_buffer, \ .flags = VMS_BUFFER, \ .offset = vmstate_offset_macaddr(_state, _field), \ }
Quoted 7 lines · Jan Kiszka
Signed-off-by Jan Kiszka <[email protected]>---
This may partly or fully explain the e1000 issue Pierre reported first. However, there are more bugs in the current migration code, e.g. in
qemu_fseekwhen used with streams, it still doesn't work properly here. Need a break...
You arrived at the same conclussion than me :(
Quoted 8 lines
hw/hw.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)hw/hw.h +0
-0
hw/hw.h extern const VMStateDescription vmstate_i2c_slave; · line 552
Quoted 8 lines
hw/hw.h +1
-1@@ -546,7 +546,7 @@ extern const VMStateDescription vmstate_i2c_slave;#define VMSTATE_MACADDR (_field, _state) { \ .name = (stringify(_field)), \ .size = sizeof(MACAddr), \ - .info = &vmstate_info_uint8, \ + .info = &vmstate_info_buffer, \ .flags = VMS_BUFFER, \ .offset = vmstate_offset_macaddr(_state, _field), \ }
Good catch.
Later, Juan.