Virtualization Team's Weekly Status Meeting
This was a Google Hangout based monthly meeting.
- Round table:
- what have you done since the last meeting?
- what are you working on at the moment?
- where do you need some help / assistance?
- Open floor
- JIRA Updates
- almost done, completing this week
- no work not on JIRA ticket is allowed
- no card - poke Chris/Serge to create one
- or don’t work on that
- Also added to Linaro Meetings (private)
- Request an invitation from Christoffer to participate
- Minutes are posted to kvm-arm ML (e.g. this one)
- Xen on Calxeda: some progress, but still not in Dom0
- Xen UART is now accepting input, but not propagating to Dom0
- trying Stefano’s SWIOTLB patches that avoid 1:1 workaround
crashed my file system
- Boot protocol:
- implemented prototype support in u-boot and Xen
- formulated device tree binding proposal and sent out
- I had a look at VIRT-121 (cpu=host) and VIRT-122 (cpu!=pcpu) blueprints.
- We will need to wait a bit for VIRT-122 RFC patches because specs are not open from ARM and APM. I will still keep working in this direction anyways
- I have RFC patches for VIRT-121 ready. I will send them out with some more testing.
- Spend quite some time debugging Xen ARM64 on X-Gene Storm SoC so far Xen ARM64 bringup going quite well.
- KVM maintainer duties, pull request with minor updates only for 3.12
- Huge pages and save/restore did not make it for 3.12, please someone help review the code so we can upstream it for 3.13
- Loads of JIRA work, see separate discussion
- Last week
- Sent the v2 of the device tree editing (VIRT-76)
- Add support to non-contiguous CPU ID (VIRT-75)
- Trying to boot Linux 3.11 as dom0 on the Arndale. Doesn't work because the 1:1 mapping is not correct (memory to high)
- This week
- Off thursday. Working remotely friday
- Add support for dom0 initramfs (blueprint not yet creating)
- Update documentation
- LAVA woes running ubuntu 13.04 for use with local arndale
- got test working in cambridge lab instead:
- cleaned up, sent patch, trying to address comments
- (from both LNG and KVM side)
- aarch64 tcg host support patches to review from R.H.
- libvixl integration basically done (functional), but disassembly output quality is not good. Many missing instructions, dependency on scons and c++, limitations on 64bit LP64 for the host. Patches available, but I am arguing on the list for the objdump approach suggested by R.H.
- PSCI interface: studied the SMC calling conventions and the PSCI specification, ready to go
- virtio emergency console: not looked at it yet.
- local restrictions on data interchange and internet access create issues: change is underway however, ETA for internet “green zone” is 2 weeks. This won’t solve all problems but will mitigate a lot if it works the way I hope.
- probably need to create blueprint for aarch64 host work
- Anup requested link to Peter’s mach-virt work:
LEG/Engineering/Virtualization/TeamMeetingReports/2013-09-03 (last modified 2013-09-17 19:42:53)