I just wanted to finalize the testing, which are completed now
successfully on all targets, with a simple PCI device assignment on x86.
Enabling the QEMU setup for this was easy, but it revealed an actually
known issue when handing over PCI devices: MSI/-X vectors aren't masked,
and the new owner may stumble over invalid MSI-X vectors when only
initializing a subset.
Fortunately, there were already patches for that in wip/ivshmem2. So
this is the port of them to next plus the changes to the configs to
enable that test setup.
Hopefully really the last round...
Jan
Jan Kiszka (7):
core: ivshmem: Print addition message earlier
core: pci: Move device->cell updates into add/remove functions
core: pci: Implement device reset on cell (re-)start or ownership
changes
README: Update QEMU command line for Intel on older host kernels
configs: Reference README.md from qemu config for command lines
configs: Switch QEMU example to e1000e network adapter
configs: Allow to pass the QEMU e1000e device to non-root Linux
README.md | 4 +-
configs/linux-x86-demo.c | 98 +++++++++++++++++++++++++-
configs/qemu-vm.c | 123 ++++++++++++++++++++++++---------
hypervisor/control.c | 1 +
hypervisor/include/jailhouse/ivshmem.h | 2 +
hypervisor/include/jailhouse/pci.h | 3 +
hypervisor/ivshmem.c | 58 +++++++++++-----
hypervisor/pci.c | 73 +++++++++++++++----
8 files changed, 292 insertions(+), 70 deletions(-)
--
2.1.4
--
You received this message because you are subscribed to the Google Groups
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
