Reducing fragmentation in OpenBMC
yulei.sh at bytedance.com
Fri May 15 20:21:02 AEST 2020
On Fri, May 15, 2020 at 5:04 PM Jeremy Kerr <jk at ozlabs.org> wrote:
> So - any thoughts on how we can improve this? Comments / disagreements
> / questions always welcome.
I am all in for making the effort to upstream the separated projects.
E.g. one candidate step could be to upstream the static flash layout
and the code update in Intel-BMC/openbmc, which uses a different
layout and different update method than the upstream.
It seems to uses a single "big" chip to support A/B side of BMC
firmware, and at runtime, the rofs is mounted in ram so it could
safely update the rofs at runtime. All of these are good features that
are not implemented in OpenBMC upstream.
If one of the Intel developers could provide some initial patches, I
could test and work on shepherding those upstream.
More information about the openbmc