Merge pull request kata-containers#1888 from yuanzhe-liu0/qemu_align

qemu: align before memory hotplug on arm64
yubo-yue · May 24, 2021 · c09d8fc · c09d8fc
2 parents c078628 bc36b7b
commit c09d8fc
Showing 1 changed file with 41 additions and 0 deletions.
diff --git a/tools/packaging/qemu/patches/5.1.x/0013-qemu-align-before-memory-hotplug-on-arm64.patch b/tools/packaging/qemu/patches/5.1.x/0013-qemu-align-before-memory-hotplug-on-arm64.patch
@@ -0,0  1,41 @@
 From 0a153bdd2c31932c8357970c8f71f710dbfc7d11 Mon Sep 17 00:00:00 2001
 From: Yuanzhe Liu <[email protected]>
 Date: Wed, 19 May 2021 14:17:03  0000
 Subject: [PATCH] qemu: align before memory hotplug on arm64
 
 When hotplug memory on arm64 in kata, kernel will shout:
 
 [ 0.396551] Block size [0x40000000] unaligned hotplug range: start 0xc8000000, size 0x40000000
 [ 0.396556] acpi PNP0C80:01: add_memory failed
 [ 0.396834] acpi PNP0C80:01: acpi_memory_enable_device() error
 [ 0.396948] acpi PNP0C80:01: Enumeration failure
 
 It means that kernel will check if the memory range to be hotplugged
 align with 1G before plug the memory. So we should twist the qemu to
 make sure the memory range align with 1G to pass the kernel check.
 
 Signed-off-by: Yuanzhe Liu <[email protected]>
 ---
  hw/arm/virt.c | 5     -
  1 file changed, 4 insertions( ), 1 deletion(-)
 
 diff --git a/hw/arm/virt.c b/hw/arm/virt.c
 index 09eb6c06e4..513f191dfe 100644
 --- a/hw/arm/virt.c
     b/hw/arm/virt.c
 @@ -2217,8  2217,11 @@ static void virt_memory_plug(HotplugHandler *hotplug_dev,
      MachineState *ms = MACHINE(hotplug_dev);
      bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
      Error *local_err = NULL;
      PCDIMMDevice *dimm = PC_DIMM(dev);
 
 -    pc_dimm_plug(PC_DIMM(dev), MACHINE(vms), &local_err);
      dimm->addr = QEMU_ALIGN_UP(dimm->addr, 0x40000000);
  
      pc_dimm_plug(dimm, MACHINE(vms), &local_err);
      if (local_err) {
          goto out;
      }
 -- 
 2.20.1