Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-24 Thread Dong, Guo 

Thanks all for these comments. I have sent an updated patch, please help review.

Thanks,
Guo

From: Mudusuru, Giri P
Sent: Sunday, October 23, 2016 6:53 AM
To: Yao, Jiewen ; Dong, Guo ; 
edk2-devel@lists.01.org
Cc: Ma, Maurice ; Yarlagadda, Satya P 

Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Agree with Jiewen's solution. Default should be 0 in IntelFsp2WrapperPkg and 
CorebootPayloadPkg can set BIT16 to skip the API.

Thanks,
-Giri

From: Yao, Jiewen
Sent: Friday, October 21, 2016 10:15 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

My thinking is to define mask for ALL APIs, not 3.

We have 5 APIs, and 3 sub API for notify phase. How about this :
# BIT0~15 is for function:
#   BIT0 - Mask TempRamInit
#   BIT1 - Mask MemoryInit
#   BIT2 - Mask TempRamExit
#   BIT3 - Mask SiliconInit
#   BIT4 - Mask NotifyPhase
# BIT16~32 is for sub-function:
#   BIT16 - Mask NotifyPhase (AfterPciEnumeration)
#   BIT17 - Mask NotifyPhase (ReadyToBoot)
#   BIT18 - Mask NotifyPhase (EndOfFirmware)
# Any undefined BITs are reserved for future use
PcdFspApiSkipMask|UINT32

You can use BIT16 now. We may use BIT0+BIT2 later. And any other combination in 
the future.

From: Dong, Guo
Sent: Saturday, October 22, 2016 7:52 AM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


I am OK with bit mask PCD.
Now only 3 APIs in this driver. Do we really need UINT64?

Thanks,
Guo

From: Yao, Jiewen
Sent: Friday, October 21, 2016 4:23 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Understood. I agree that is a valid usage model.

So how about we define a bit mask for every function? Such as 
"PcdFspApiSkipMask|UINT64" - I am open on any naming proposal.

In the future we may need skip some other functions such as InitTempRam or 
TempRamExit.

Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 1:11 PM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Jiewen,

For Coreboot boot loader, PCI device enumeration is done in coreboot, so it is 
nature Coreboot sends PciEnumerationComplete to FSP. UEFI payload would send 
FSP ReadyToBoot and EndOfFirmware notification.

If there is other boot loader that sends ReadyToBoot and EndOfFirmware in 
different firmware component, we may add PCDs for them later.

Thanks,
Guo

From: Yao, Jiewen
Sent: Monday, October 17, 2016 9:10 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>; Yao, 
Jiewen >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-23 Thread Mudusuru, Giri P 
Agree with Jiewen's solution. Default should be 0 in IntelFsp2WrapperPkg and 
CorebootPayloadPkg can set BIT16 to skip the API.

Thanks,
-Giri

From: Yao, Jiewen
Sent: Friday, October 21, 2016 10:15 PM
To: Dong, Guo ; Mudusuru, Giri P 
; edk2-devel@lists.01.org
Cc: Ma, Maurice ; Yarlagadda, Satya P 

Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

My thinking is to define mask for ALL APIs, not 3.

We have 5 APIs, and 3 sub API for notify phase. How about this :
# BIT0~15 is for function:
#   BIT0 - Mask TempRamInit
#   BIT1 - Mask MemoryInit
#   BIT2 - Mask TempRamExit
#   BIT3 - Mask SiliconInit
#   BIT4 - Mask NotifyPhase
# BIT16~32 is for sub-function:
#   BIT16 - Mask NotifyPhase (AfterPciEnumeration)
#   BIT17 - Mask NotifyPhase (ReadyToBoot)
#   BIT18 - Mask NotifyPhase (EndOfFirmware)
# Any undefined BITs are reserved for future use
PcdFspApiSkipMask|UINT32

You can use BIT16 now. We may use BIT0+BIT2 later. And any other combination in 
the future.

From: Dong, Guo
Sent: Saturday, October 22, 2016 7:52 AM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


I am OK with bit mask PCD.
Now only 3 APIs in this driver. Do we really need UINT64?

Thanks,
Guo

From: Yao, Jiewen
Sent: Friday, October 21, 2016 4:23 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Understood. I agree that is a valid usage model.

So how about we define a bit mask for every function? Such as 
"PcdFspApiSkipMask|UINT64" - I am open on any naming proposal.

In the future we may need skip some other functions such as InitTempRam or 
TempRamExit.

Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 1:11 PM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Jiewen,

For Coreboot boot loader, PCI device enumeration is done in coreboot, so it is 
nature Coreboot sends PciEnumerationComplete to FSP. UEFI payload would send 
FSP ReadyToBoot and EndOfFirmware notification.

If there is other boot loader that sends ReadyToBoot and EndOfFirmware in 
different firmware component, we may add PCDs for them later.

Thanks,
Guo

From: Yao, Jiewen
Sent: Monday, October 17, 2016 9:10 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>; Yao, 
Jiewen >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 12:04 PM
To: Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Giri,

Thank you for the comment.

In our POC implementation, Coreboot will

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-21 Thread Yao, Jiewen 
My thinking is to define mask for ALL APIs, not 3.

We have 5 APIs, and 3 sub API for notify phase. How about this :
# BIT0~15 is for function:
#   BIT0 - Mask TempRamInit
#   BIT1 - Mask MemoryInit
#   BIT2 - Mask TempRamExit
#   BIT3 - Mask SiliconInit
#   BIT4 - Mask NotifyPhase
# BIT16~32 is for sub-function:
#   BIT16 - Mask NotifyPhase (AfterPciEnumeration)
#   BIT17 - Mask NotifyPhase (ReadyToBoot)
#   BIT18 - Mask NotifyPhase (EndOfFirmware)
# Any undefined BITs are reserved for future use
PcdFspApiSkipMask|UINT32

You can use BIT16 now. We may use BIT0+BIT2 later. And any other combination in 
the future.

From: Dong, Guo
Sent: Saturday, October 22, 2016 7:52 AM
To: Yao, Jiewen ; Mudusuru, Giri P 
; edk2-devel@lists.01.org
Cc: Ma, Maurice ; Yarlagadda, Satya P 

Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


I am OK with bit mask PCD.
Now only 3 APIs in this driver. Do we really need UINT64?

Thanks,
Guo

From: Yao, Jiewen
Sent: Friday, October 21, 2016 4:23 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Understood. I agree that is a valid usage model.

So how about we define a bit mask for every function? Such as 
"PcdFspApiSkipMask|UINT64" - I am open on any naming proposal.

In the future we may need skip some other functions such as InitTempRam or 
TempRamExit.

Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 1:11 PM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Jiewen,

For Coreboot boot loader, PCI device enumeration is done in coreboot, so it is 
nature Coreboot sends PciEnumerationComplete to FSP. UEFI payload would send 
FSP ReadyToBoot and EndOfFirmware notification.

If there is other boot loader that sends ReadyToBoot and EndOfFirmware in 
different firmware component, we may add PCDs for them later.

Thanks,
Guo

From: Yao, Jiewen
Sent: Monday, October 17, 2016 9:10 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>; Yao, 
Jiewen >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 12:04 PM
To: Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Giri,

Thank you for the comment.

In our POC implementation, Coreboot will notify FSP EnumerationComplete.  And 
in UEFI payload, the other 2 FSP notifications will be called from 
FspWrapperNotifyDxe.

Thanks,
Guo

-Original Message-
From: Mudusuru, Giri P
Sent: Monday, October 17, 2016 8:32 PM
To: Dong, Guo >; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-21 Thread Dong, Guo 

I am OK with bit mask PCD.
Now only 3 APIs in this driver. Do we really need UINT64?

Thanks,
Guo

From: Yao, Jiewen
Sent: Friday, October 21, 2016 4:23 PM
To: Dong, Guo ; Mudusuru, Giri P 
; edk2-devel@lists.01.org
Cc: Ma, Maurice ; Yarlagadda, Satya P 

Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Understood. I agree that is a valid usage model.

So how about we define a bit mask for every function? Such as 
"PcdFspApiSkipMask|UINT64" - I am open on any naming proposal.

In the future we may need skip some other functions such as InitTempRam or 
TempRamExit.

Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 1:11 PM
To: Yao, Jiewen >; Mudusuru, 
Giri P >; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Jiewen,

For Coreboot boot loader, PCI device enumeration is done in coreboot, so it is 
nature Coreboot sends PciEnumerationComplete to FSP. UEFI payload would send 
FSP ReadyToBoot and EndOfFirmware notification.

If there is other boot loader that sends ReadyToBoot and EndOfFirmware in 
different firmware component, we may add PCDs for them later.

Thanks,
Guo

From: Yao, Jiewen
Sent: Monday, October 17, 2016 9:10 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>; Yao, 
Jiewen >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 12:04 PM
To: Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Giri,

Thank you for the comment.

In our POC implementation, Coreboot will notify FSP EnumerationComplete.  And 
in UEFI payload, the other 2 FSP notifications will be called from 
FspWrapperNotifyDxe.

Thanks,
Guo

-Original Message-
From: Mudusuru, Giri P
Sent: Monday, October 17, 2016 8:32 PM
To: Dong, Guo >; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hello Guo,
FSP wrapper and Coreboot are bootloaders consuming FSP binary.

FSP must get called Post PCI bus enumeration to do the required silicon 
initialization. Why do we need a PCD to skip it?

Thanks,
-Giri

> -Original Message-
> From: Dong, Guo
> Sent: Monday, October 17, 2016 3:53 PM
> To: edk2-devel@lists.01.org
> Cc: Yao, Jiewen >; 
> Mudusuru, Giri P
> >; Dong, Guo 
> >
> Subject: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if
> signaling first event.
>
> PciEnumerationComplete might be signaled to FSP in Coreboot. Add a PCD
> PcdNotifyPciEnumerationComplete so FspWrapperNotifyDxe driver could be
> used in this case.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: gdong1 >
> ---
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.c  | 18 
> ++
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf|  3 ++-
>

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-21 Thread Yao, Jiewen 
Understood. I agree that is a valid usage model.

So how about we define a bit mask for every function? Such as 
"PcdFspApiSkipMask|UINT64" - I am open on any naming proposal.

In the future we may need skip some other functions such as InitTempRam or 
TempRamExit.

Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 1:11 PM
To: Yao, Jiewen ; Mudusuru, Giri P 
; edk2-devel@lists.01.org
Cc: Ma, Maurice ; Yarlagadda, Satya P 

Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Jiewen,

For Coreboot boot loader, PCI device enumeration is done in coreboot, so it is 
nature Coreboot sends PciEnumerationComplete to FSP. UEFI payload would send 
FSP ReadyToBoot and EndOfFirmware notification.

If there is other boot loader that sends ReadyToBoot and EndOfFirmware in 
different firmware component, we may add PCDs for them later.

Thanks,
Guo

From: Yao, Jiewen
Sent: Monday, October 17, 2016 9:10 PM
To: Dong, Guo >; Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Ma, Maurice >; 
Yarlagadda, Satya P 
>; Yao, 
Jiewen >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 12:04 PM
To: Mudusuru, Giri P 
>; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Giri,

Thank you for the comment.

In our POC implementation, Coreboot will notify FSP EnumerationComplete.  And 
in UEFI payload, the other 2 FSP notifications will be called from 
FspWrapperNotifyDxe.

Thanks,
Guo

-Original Message-
From: Mudusuru, Giri P
Sent: Monday, October 17, 2016 8:32 PM
To: Dong, Guo >; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hello Guo,
FSP wrapper and Coreboot are bootloaders consuming FSP binary.

FSP must get called Post PCI bus enumeration to do the required silicon 
initialization. Why do we need a PCD to skip it?

Thanks,
-Giri

> -Original Message-
> From: Dong, Guo
> Sent: Monday, October 17, 2016 3:53 PM
> To: edk2-devel@lists.01.org
> Cc: Yao, Jiewen >; 
> Mudusuru, Giri P
> >; Dong, Guo 
> >
> Subject: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if
> signaling first event.
>
> PciEnumerationComplete might be signaled to FSP in Coreboot. Add a PCD
> PcdNotifyPciEnumerationComplete so FspWrapperNotifyDxe driver could be
> used in this case.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: gdong1 >
> ---
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.c  | 18 
> ++
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf|  3 ++-
>  IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec|  5 +
>  3 files changed, 17 insertions(+), 9 deletions(-)
>
> diff --git
> a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> index 0797f44..45eae4b 100644
> --- a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> +++ b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> @@ -246,14 +246,16 @@ FspWrapperNotifyDxeEntryPoint (
>  return EFI_SUCCESS;
>}
>
> -  ProtocolNotifyEvent = EfiCreateProtocolNotifyEvent (
> -

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-17 Thread Yao, Jiewen 
Hi
I am a little concern on this as well.

If today, there is a POC require skipping PciEnumerationComplete.
There might be the other POC tomorrow to skip ReadyToBoot
Then another POC to skip EventExitBootServices.

We will be exhausted to add POC to skip each separately.

Can this POC handle all 3 event together? Or not handle 3 event completely?


Thank you
Yao Jiewen

From: Dong, Guo
Sent: Tuesday, October 18, 2016 12:04 PM
To: Mudusuru, Giri P ; edk2-devel@lists.01.org
Cc: Yao, Jiewen ; Ma, Maurice ; 
Yarlagadda, Satya P 
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.


Hi Giri,

Thank you for the comment.

In our POC implementation, Coreboot will notify FSP EnumerationComplete.  And 
in UEFI payload, the other 2 FSP notifications will be called from 
FspWrapperNotifyDxe.

Thanks,
Guo

-Original Message-
From: Mudusuru, Giri P
Sent: Monday, October 17, 2016 8:32 PM
To: Dong, Guo >; 
edk2-devel@lists.01.org
Cc: Yao, Jiewen >; Ma, 
Maurice >; Yarlagadda, Satya 
P >
Subject: RE: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if 
signaling first event.

Hello Guo,
FSP wrapper and Coreboot are bootloaders consuming FSP binary.

FSP must get called Post PCI bus enumeration to do the required silicon 
initialization. Why do we need a PCD to skip it?

Thanks,
-Giri

> -Original Message-
> From: Dong, Guo
> Sent: Monday, October 17, 2016 3:53 PM
> To: edk2-devel@lists.01.org
> Cc: Yao, Jiewen >; 
> Mudusuru, Giri P
> >; Dong, Guo 
> >
> Subject: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if
> signaling first event.
>
> PciEnumerationComplete might be signaled to FSP in Coreboot. Add a PCD
> PcdNotifyPciEnumerationComplete so FspWrapperNotifyDxe driver could be
> used in this case.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: gdong1 >
> ---
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.c  | 18 
> ++
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf|  3 ++-
>  IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec|  5 +
>  3 files changed, 17 insertions(+), 9 deletions(-)
>
> diff --git
> a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> index 0797f44..45eae4b 100644
> --- a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> +++ b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> @@ -246,14 +246,16 @@ FspWrapperNotifyDxeEntryPoint (
>  return EFI_SUCCESS;
>}
>
> -  ProtocolNotifyEvent = EfiCreateProtocolNotifyEvent (
> -  ,
> -  TPL_CALLBACK,
> -  OnPciEnumerationComplete,
> -  NULL,
> -  
> -  );
> -  ASSERT (ProtocolNotifyEvent != NULL);
> +  if (PcdGetBool(PcdNotifyPciEnumerationComplete)) {
> +ProtocolNotifyEvent = EfiCreateProtocolNotifyEvent (
> +,
> +TPL_CALLBACK,
> +OnPciEnumerationComplete,
> +NULL,
> +
> +);
> +ASSERT (ProtocolNotifyEvent != NULL);  }
>
>Status = EfiCreateEventReadyToBootEx (
>   TPL_CALLBACK,
> diff --git
> a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> index f851f68..f28356d 100644
> --- a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> +++ b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> @@ -61,7 +61,8 @@
>gFspHobGuid   ## CONSUMES ## HOB
>
>  [Pcd]
> -  gIntelFsp2WrapperTokenSpaceGuid.PcdFspsBaseAddress  ## CONSUMES
> +  gIntelFsp2WrapperTokenSpaceGuid.PcdFspsBaseAddress   ##
> CONSUMES
> +  gIntelFsp2WrapperTokenSpaceGuid.PcdNotifyPciEnumerationComplete  ##
> CONSUMES
>
>  [Depex]
>TRUE
> diff --git a/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> b/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> index d9d2d80..52ed9d8 100644
> --- a/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> +++ b/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> @@ -74,6 +74,11 @@
>## This is the base address of FSP-T/M/S
>
>

Re: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling first event.

2016-10-17 Thread Mudusuru, Giri P 
Hello Guo,
FSP wrapper and Coreboot are bootloaders consuming FSP binary. 

FSP must get called Post PCI bus enumeration to do the required silicon 
initialization. Why do we need a PCD to skip it?

Thanks,
-Giri

> -Original Message-
> From: Dong, Guo
> Sent: Monday, October 17, 2016 3:53 PM
> To: edk2-devel@lists.01.org
> Cc: Yao, Jiewen ; Mudusuru, Giri P
> ; Dong, Guo 
> Subject: [edk2] [PATCH] IntelFsp2WrapperPkg: Add a PCD to control if signaling
> first event.
> 
> PciEnumerationComplete might be signaled to FSP in Coreboot. Add a
> PCD PcdNotifyPciEnumerationComplete so FspWrapperNotifyDxe driver
> could be used in this case.
> 
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: gdong1 
> ---
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.c  | 18 
> ++
>  .../FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf|  3 ++-
>  IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec|  5 +
>  3 files changed, 17 insertions(+), 9 deletions(-)
> 
> diff --git a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> index 0797f44..45eae4b 100644
> --- a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> +++ b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.c
> @@ -246,14 +246,16 @@ FspWrapperNotifyDxeEntryPoint (
>  return EFI_SUCCESS;
>}
> 
> -  ProtocolNotifyEvent = EfiCreateProtocolNotifyEvent (
> -  ,
> -  TPL_CALLBACK,
> -  OnPciEnumerationComplete,
> -  NULL,
> -  
> -  );
> -  ASSERT (ProtocolNotifyEvent != NULL);
> +  if (PcdGetBool(PcdNotifyPciEnumerationComplete)) {
> +ProtocolNotifyEvent = EfiCreateProtocolNotifyEvent (
> +,
> +TPL_CALLBACK,
> +OnPciEnumerationComplete,
> +NULL,
> +
> +);
> +ASSERT (ProtocolNotifyEvent != NULL);
> +  }
> 
>Status = EfiCreateEventReadyToBootEx (
>   TPL_CALLBACK,
> diff --git
> a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> index f851f68..f28356d 100644
> --- a/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> +++ b/IntelFsp2WrapperPkg/FspWrapperNotifyDxe/FspWrapperNotifyDxe.inf
> @@ -61,7 +61,8 @@
>gFspHobGuid   ## CONSUMES ## HOB
> 
>  [Pcd]
> -  gIntelFsp2WrapperTokenSpaceGuid.PcdFspsBaseAddress  ## CONSUMES
> +  gIntelFsp2WrapperTokenSpaceGuid.PcdFspsBaseAddress   ##
> CONSUMES
> +  gIntelFsp2WrapperTokenSpaceGuid.PcdNotifyPciEnumerationComplete  ##
> CONSUMES
> 
>  [Depex]
>TRUE
> diff --git a/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> b/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> index d9d2d80..52ed9d8 100644
> --- a/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> +++ b/IntelFsp2WrapperPkg/IntelFsp2WrapperPkg.dec
> @@ -74,6 +74,11 @@
>## This is the base address of FSP-T/M/S
> 
> gIntelFsp2WrapperTokenSpaceGuid.PcdFsptBaseAddress|0x|UINT32|
> 0x0300
> 
> gIntelFsp2WrapperTokenSpaceGuid.PcdFspmBaseAddress|0x|UINT32
> |0x0301
> +  ## Indicates if the FSP wrapper will notify FSP
> PciEnumerationComplete.
> +  #   TRUE  - FSP wrapper will notify PciEnumerationComplete to FSP.
> +  #   FALSE - FSP wrapper will NOT notify PciEnumerationComplete to FSP.
> +  # @Prompt Notify FSP PciEnumerationComplete.
> +
> gIntelFsp2WrapperTokenSpaceGuid.PcdNotifyPciEnumerationComplete|TRUE|
> BOOLEAN|0x4009
> 
>  [PcdsFixedAtBuild, PcdsPatchableInModule,PcdsDynamic,PcdsDynamicEx]
> 
> gIntelFsp2WrapperTokenSpaceGuid.PcdFspsBaseAddress|0x|UINT32|
> 0x1001
> --
> 2.7.0.windows.1

___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel