On 2025/10/28 21:28, Michael Tokarev wrote:
On 10/27/25 04:05, Akihiko Odaki wrote:
s points to the MemoryRegion itself. dev points to DS1225Y, the real
owner.
Signed-off-by: Akihiko Odaki <[email protected]>
---
hw/nvram/ds1225y.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/nvram/ds1225y.c b/hw/nvram/ds1225y.c
index dbfd0d2e5369..0945e36652e9 100644
--- a/hw/nvram/ds1225y.c
+++ b/hw/nvram/ds1225y.c
@@ -126,7 +126,7 @@ static void nvram_sysbus_realize(DeviceState *dev,
Error **errp)
s->contents = g_malloc0(s->chip_size);
- memory_region_init_io(&s->iomem, OBJECT(s), &nvram_ops, s,
+ memory_region_init_io(&s->iomem, OBJECT(dev), &nvram_ops, s,
"nvram", s->chip_size);
sysbus_init_mmio(SYS_BUS_DEVICE(dev), &s->iomem);
What's the impact of this issue?
Should it go to the stable series perhaps?
There should be no perceivable impact. It causes memory leak if the
device is hotpluggable, but it is apparently not the case.
The same goes for "[PATCH] hw/riscv/riscv-iommu: Fix MemoryRegion owner":
https://lore.kernel.org/qemu-devel/[email protected]/
Regards,
Akihiko Odaki
