On 2/11/2019 10:48 PM, Leif Lindholm wrote:
*bangs head on desk*
That question I just asked as a reply to ("Silicon/Hisilicon/D06: Set TA as Node 0 for TA boot") was meant to be a comment on this patch.
So - was this change one that was meant to happen together with edk2-platforms "Silicon/Hisilicon/D06: Set TA as Node 0 for TA boot"?
Yes, it is better to happen together with that patch.
What is the user visible behaviour that this change addresses?
Numa info of kernel boot is different: Use old MemoryInit.efi: [ 0.000000] ACPI: SRAT: Node 1 PXM 0 [mem 0x2080000000-0x23ffffffff] [ 0.000000] ACPI: SRAT: Node 1 PXM 0 [mem 0x00000000-0x7fffffff] [ 0.000000] ACPI: SRAT: Node 3 PXM 2 [mem 0xa2000000000-0xa23ffffffff] Use new MemoryInit.efi: [ 0.000000] ACPI: SRAT: Node 1 PXM 1 [mem 0x2080000000-0x23ffffffff] [ 0.000000] ACPI: SRAT: Node 1 PXM 1 [mem 0x00000000-0x7fffffff] [ 0.000000] ACPI: SRAT: Node 3 PXM 3 [mem 0xa2000000000-0xa23ffffffff]o
Thanks.
/ Leif
On Fri, Feb 01, 2019 at 10:25:06PM +0800, Ming Huang wrote:
Numa informations are acquired from HOB that build from memory initialization module. Correct numa informations to match booting from TA(Totem A or super cpu cluster A).
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Ming Huang ming.huang@linaro.org
Platform/Hisilicon/D06/MemoryInitPei/MemoryInit.efi | Bin 297696 -> 358656 bytes 1 file changed, 0 insertions(+), 0 deletions(-)
diff --git a/Platform/Hisilicon/D06/MemoryInitPei/MemoryInit.efi b/Platform/Hisilicon/D06/MemoryInitPei/MemoryInit.efi index 5fba353..fea1475 100644 Binary files a/Platform/Hisilicon/D06/MemoryInitPei/MemoryInit.efi and b/Platform/Hisilicon/D06/MemoryInitPei/MemoryInit.efi differ -- 2.9.5