Re: [edk2-devel] [PATCH v2 5/5] UefiCpuPkg/MpInitLib: Add support for multiple HOBs to SaveCpuMpData()

["Laszlo Ersek" <lersek@redhat.com>]

On 2/20/24 18:49, Gerd Hoffmann wrote:
> Add support for splitting Hand-Off data into multiple HOBs.  This is
> required for VMs with thousands of CPUs.  The actual CPU count per HOB
> is much smaller (128) for better test coverage.

(1) The mention of the count 128 now seems stale for the code.

> 
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---
>  UefiCpuPkg/Library/MpInitLib/PeiMpLib.c | 44 +++++++++++++++----------
>  1 file changed, 27 insertions(+), 17 deletions(-)
> 
> diff --git a/UefiCpuPkg/Library/MpInitLib/PeiMpLib.c b/UefiCpuPkg/Library/MpInitLib/PeiMpLib.c
> index f80e00edcff3..8a916a218016 100644
> --- a/UefiCpuPkg/Library/MpInitLib/PeiMpLib.c
> +++ b/UefiCpuPkg/Library/MpInitLib/PeiMpLib.c
> @@ -126,35 +126,45 @@ SaveCpuMpData (
>    IN CPU_MP_DATA  *CpuMpData
>    )
>  {
> +  UINT32           MaxCpusPerHob, CpusInHob;
>    UINT64           Data64;
> -  UINTN            Index;
> +  UINT32           Index, HobBase;
>    CPU_INFO_IN_HOB  *CpuInfoInHob;
>    MP_HAND_OFF      *MpHandOff;
>    UINTN            MpHandOffSize;
>  
> +  MaxCpusPerHob = (MAX_UINT16 - sizeof (EFI_HOB_GUID_TYPE) - sizeof (MP_HAND_OFF)) / sizeof (PROCESSOR_HAND_OFF);

(2) MAX_UINT16 should be 0xFFF8 instead; see subthread.

> +
>    //
>    // When APs are in a state that can be waken up by a store operation to a memory address,
>    // report the MP_HAND_OFF data for DXE to use.
>    //
> -  CpuInfoInHob  = (CPU_INFO_IN_HOB *)(UINTN)CpuMpData->CpuInfoInHob;
> -  MpHandOffSize = sizeof (MP_HAND_OFF) + sizeof (PROCESSOR_HAND_OFF) * CpuMpData->CpuCount;
> -  MpHandOff     = (MP_HAND_OFF *)BuildGuidHob (&mMpHandOffGuid, MpHandOffSize);
> -  ASSERT (MpHandOff != NULL);
> -  ZeroMem (MpHandOff, MpHandOffSize);
> -  MpHandOff->ProcessorIndex = 0;
> +  CpuInfoInHob = (CPU_INFO_IN_HOB *)(UINTN)CpuMpData->CpuInfoInHob;
>  
> -  MpHandOff->CpuCount = CpuMpData->CpuCount;
> -  if (CpuMpData->ApLoopMode != ApInHltLoop) {
> -    MpHandOff->StartupSignalValue    = MP_HAND_OFF_SIGNAL;
> -    MpHandOff->WaitLoopExecutionMode = sizeof (VOID *);
> -  }
> +  for (Index = 0; Index < CpuMpData->CpuCount; Index++) {
> +    if (Index % MaxCpusPerHob == 0) {
> +      HobBase   = Index;
> +      CpusInHob = MIN (CpuMpData->CpuCount - HobBase, MaxCpusPerHob);
>  
> -  for (Index = 0; Index < MpHandOff->CpuCount; Index++) {
> -    MpHandOff->Info[Index].ApicId = CpuInfoInHob[Index].ApicId;
> -    MpHandOff->Info[Index].Health = CpuInfoInHob[Index].Health;
> +      MpHandOffSize = sizeof (MP_HAND_OFF) + sizeof (PROCESSOR_HAND_OFF) * CpusInHob;
> +      MpHandOff     = (MP_HAND_OFF *)BuildGuidHob (&mMpHandOffGuid, MpHandOffSize);
> +      ASSERT (MpHandOff != NULL);
> +      ZeroMem (MpHandOff, MpHandOffSize);
> +
> +      MpHandOff->ProcessorIndex = HobBase;
> +      MpHandOff->CpuCount       = CpusInHob;
> +
> +      if (CpuMpData->ApLoopMode != ApInHltLoop) {
> +        MpHandOff->StartupSignalValue    = MP_HAND_OFF_SIGNAL;
> +        MpHandOff->WaitLoopExecutionMode = sizeof (VOID *);
> +      }
> +    }
> +
> +    MpHandOff->Info[Index-HobBase].ApicId = CpuInfoInHob[Index].ApicId;
> +    MpHandOff->Info[Index-HobBase].Health = CpuInfoInHob[Index].Health;
>      if (CpuMpData->ApLoopMode != ApInHltLoop) {
> -      MpHandOff->Info[Index].StartupSignalAddress    = (UINT64)(UINTN)CpuMpData->CpuData[Index].StartupApSignal;
> -      MpHandOff->Info[Index].StartupProcedureAddress = (UINT64)(UINTN)&CpuMpData->CpuData[Index].ApFunction;
> +      MpHandOff->Info[Index-HobBase].StartupSignalAddress    = (UINT64)(UINTN)CpuMpData->CpuData[Index].StartupApSignal;
> +      MpHandOff->Info[Index-HobBase].StartupProcedureAddress = (UINT64)(UINTN)&CpuMpData->CpuData[Index].ApFunction;
>      }
>    }
>  

(3) The conversion looks good otherwise, but I dislike that
StartupSignalValue and WaitLoopExecutionMode get uselessly replicated
over all HOBs. Again, it *increases* technical debt. It's fine to ignore
existent debt (if you can), but adding to it is not right.

Can you file a new TianoCore BZ, for moving these fields to separate
dynamic PCDs, or to a new singleton GUID HOB (containing both fields)?
If you assign that BZ to yourself, and reference it in patches #4 and
#5, I'll be happy to R-b version 3 of the series. (Which is something
I'd like to do.)

Version 3 may be mergeable regardless, of course, if Ray accepts it.

Thanks,
Laszlo



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#115770): https://edk2.groups.io/g/devel/message/115770
Mute This Topic: https://groups.io/mt/104472313/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-