All (I tried to add potential key stakeholders in To list, but may miss someone): I am not sure how widely the EFI_PEI_VECTOR_HANDOFF_INFO_PPI is used today.
This is an optional PPI that may be produced by SEC. If present, it provides a description of the interrupt and/or exception vectors that were established in the SEC Phase and need to persist into PEI and DXE. This PPI is an array of entries that is terminated by an entry whose Attribute is set to EFI_VECTOR_HANDOFF_LAST_ENTRY. If Attribute is set to EFI_VECTOR_HANDOFF_DO_NOT_HOOK, then the associated handler for VectorNumber must be preserved in PEI and DXE. If Attribute is set to EFI_VECTOR_HANDOFF_HOOK_BEFORE, then VectorNumber may be used in PEI and DXE, but new handlers must be invoked prior to when the existing handler is called. If Attribute is set to EFI_VECTOR_HANDOFF_HOOK_AFTER, then the associated VectorNumber may be used in PEI and DXE, but new handlers must be called after the existing handler is called According to above content I copied from PI spec, this PPI is a way that SEC tells later phase whether to honor each exception/interrupt (might be setup by SEC). I know below two facts: 1. X86/SourceLevelDebugPkg. It uses this PPI to set exceptions #1, #3, #32, #33 to "DO_NOT_HOOK" and all other exceptions to "HOOK_BEFORE". 2. ARM: The CpuExceptionLib ignores this PPI. To simplify the code logic, I'd like to check in the community for the impact of removing the support of "HOOK_AFTER" in X86 version CpuExceptionLib. Thanks, Ray -=-=-=-=-=-=-=-=-=-=-=- Groups.io Links: You receive all messages sent to this group. View/Reply Online (#88677): https://edk2.groups.io/g/devel/message/88677 Mute This Topic: https://groups.io/mt/90387418/21656 Group Owner: devel+ow...@edk2.groups.io Unsubscribe: https://edk2.groups.io/g/devel/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
