19 July 2018

Attendees

• Ilias Apalodimas (Linaro)
• Rob Herring (Linaro)
• Grant Likely (Arm)
• Bill Mills (TI)
• Peter Robinson (Red Hat)
• Tom Rini (Konsulko)
• David Rusling (Linaro)
• Kangkang Shen
• Daniel Thompson (Linaro)

Agenda

1. Issue review
2. SetVariable() behaviour revisited
3. Other business

Notes

Review distribution list

• Q: Who should we share this document with?
• Linaro membership (Daniel, engineering update?)
• 96Boards membership (a.k.a. steering committee) (Grant)
• ARM SiPs
• U-Boot list (Tom)
• TF-A List (Grant)
• Risc-V community (just send message to ebbr list)
• Arm-linux-kernel (Grant)
• FreeBSD folks (Grant)
• Microsoft (Dong)

David has issues with building document on Debian stable

Need to investigate and make sure documented correctly

TravisCI is using old Ubuntu,. And it work

SetVariable

• Expect GetVariable to always work (SuSE currently makes inferences about SetVariable() from GetVariable() not working, but Alex as relaxed about changing this
• Need to take proposal to UEFI forum to try and get wider standardization to solve this
• Microsoft reluctant to permit copy from cache to storage at reset because they prefer to use PSCI interface to UEFI (proved more reliable in practice)
  • UEFI could delegate variable storage to PSCI ;-)
• Expect SetVariable() to work during boot services
  • u-boot requires work to ensure this requirement can be met before v1.0
  • This expectation is **not **agreed by all (system designed intentionally to avoid variable setting in order to fix a boot flow? robust discussion about whether such a design benefits from being labelled EBBR compliant)