Hi all,
I am sorry that this method to fix this bug was not perfect.
I will find another way to do this.
Plz ignore this patch.
Regards,
Zhu
On Fri, 2015-01-16 at 13:45 +0800, Zhu Guihua wrote:
> If hot add 100MiB memory like this:
> (monitor) object_add memory-backend-ram,id=ram0,size=100M
> (monitor) device_add pc-dimm,id=d0,memdev=ram0
>
> The hotplug operation will faile, and the guest will print error message:
> Section-unaligned hotplug range: start 0x100000000, size 0x6400000
> acpi PNP0C80:00: add_memory failed
> acpi PNP0C80:00: acpi_memory_enable_device() error
>
> Then I found that, according to the func check_hotplug_memory_range() in linux
> kernel, backend memory size must be multiple of 128MiB, not 2MiB.
>
> So this patch checks whether backend memory size is multiple of 128MiB.
>
> Signed-off-by: Zhu Guihua <zhugh.f...@cn.fujitsu.com>
> ---
> hw/mem/pc-dimm.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/hw/mem/pc-dimm.c b/hw/mem/pc-dimm.c
> index d431834..3b8a015 100644
> --- a/hw/mem/pc-dimm.c
> +++ b/hw/mem/pc-dimm.c
> @@ -161,9 +161,9 @@ uint64_t pc_dimm_get_free_addr(uint64_t
> address_space_start,
> goto out;
> }
>
> - if (QEMU_ALIGN_UP(size, align) != size) {
> + if (QEMU_ALIGN_UP(size, align * 64) != size) {
> error_setg(errp, "backend memory size must be multiple of 0x%"
> - PRIx64, align);
> + PRIx64, align * 64);
> goto out;
> }
>
