Progress (short week, 3 days) * UM-2 [QEMU upstream maintainership] + more code review, notably the Apple Silicon hvf support, which is nearly ready to go in * QEMU-406 [QEMU support for MVE (M-profile Vector Extension; Helium)] + Sent out v2 of the "optimized code gen for MVE" patchset; this now covers all the insns that have an easy optimized version. + Fixed a bug where we weren't correctly setting up FPSCR.LTPSIZE when using QEMU's user-mode-only emulator + Wrote some code to add support for the (not yet finalized) gdbstub XML that tells GDB that the guest CPU has MVE. This causes a GDB with the MVE handling to crash, so one or the other of us has got something wrong :-)
KVM Forum was this week, as a 2-day virtual conference. I felt the programme was comparatively a bit small this year, but there were some interesting talks. Also a BoF session on whether/how we should consider adding Rust code to QEMU: I am pushing for (a) a clearer medium-to-long-term vision of where we would be going and why we'd be doing this and (b) more design-sketch type work of "what would XYZ in rust look like", which would hopefully both (a) make the benefit/lack thereof a bit more clear and (b) demonstrate that there are enough people enthusiastic enough about the prospect to make it a success...
-- PMM