Re: [edk2] [PATCH] MdeModulepkg VarCheckLib: Return NULL when no property set

2015-10-18 Thread Zeng, Star
On 2015/10/19 10:48, Zeng, Star wrote: On 2015/10/18 14:48, Jordan Justen wrote: This patch appears to have caused OVMF to stop booting. I hit an assert now in BDS after attempting to set the BootOrder variable fails. An ideas of how to fix this? There is bug in this patch, it was thoughtless

Re: [edk2] SMM core problems

2015-10-18 Thread Zeng, Star
On 2015/10/16 16:04, Dimitri wrote: Hi, Fist of all: I am DimitRi. UDK2014/UEFI 2.4 does not have EFI_PROPERTIES_TABLE feature. Anyway we have a bug by default. Collegue from Apple confirmed that os x boot loader move EfiRuntime memory to different physical address. I did not find in UEFI spec

Re: [edk2] [PATCH] MdeModulePkg VarCheckLib: R18611 was thoughtless for property set

2015-10-18 Thread Yao, Jiewen
Looks good. Reviewed by: jiewen@intel.com Would you please add some comments in the code to indicate why we need skip check in Set() and add check in Get()? -Original Message- From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Behalf Of Star Zeng Sent: Monday, October 19,

Re: [edk2] [PATCH] MdeModulepkg VarCheckLib: Return NULL when no property set

2015-10-18 Thread Zeng, Star
On 2015/10/18 14:48, Jordan Justen wrote: This patch appears to have caused OVMF to stop booting. I hit an assert now in BDS after attempting to set the BootOrder variable fails. An ideas of how to fix this? There is bug in this patch, it was thoughtless for property set. I have resent patch to

[edk2] [PATCH] MdeModulePkg VarCheckLib: R18611 was thoughtless for property set

2015-10-18 Thread Star Zeng
R18611 only updated the logic to return correct property when no property set to variable with wildcard name. But VariablePropertySet needs the pointer to property data for set. So roll back the change in VariablePropertyGetWithWildcardName at R18611, and check the property revision first in Variab

[edk2] [PATCH] MdeModulePkg VarCheckLib: R18611 was thoughtless for property set

2015-10-18 Thread Star Zeng
R18611 only updated the logic to return correct property when no property set to variable with wildcard name. But VariablePropertySet needs the pointer to property data for set. So roll back the change in VariablePropertyGetWithWildcardName at R18611, and check the property revision first in Variab

Re: [edk2] [PATCH v3 44/52] OvmfPkg: QemuFlashFvbServicesRuntimeDxe: no dual addressing needed

2015-10-18 Thread Andrew Fish
> On Oct 18, 2015, at 12:29 AM, Jordan Justen wrote: > > I can't find a fault with your argument, but something tells me it > might be good to get some input from Mike or Andrew (or others on the > list) in the form of a history lesson to know why the two modes might > have been supported. > T

Re: [edk2] [PATCH v3 26/52] OvmfPkg: SmmCpuFeaturesLib: customize state save map format

2015-10-18 Thread Jordan Justen
On 2015-10-14 15:26:22, Laszlo Ersek wrote: > From: Paolo Bonzini > > This adjusts the previously introduced state save map access functions, to > account for QEMU and KVM's 64-bit state save map following the AMD spec > rather than the Intel one. Shouldn't this layout match the processor being

Re: [edk2] [PATCH v3 44/52] OvmfPkg: QemuFlashFvbServicesRuntimeDxe: no dual addressing needed

2015-10-18 Thread Jordan Justen
I can't find a fault with your argument, but something tells me it might be good to get some input from Mike or Andrew (or others on the list) in the form of a history lesson to know why the two modes might have been supported. I've always seen it done this way, but likely this has been due to a l