Hello,
Here are a couple of fixes for PCI hotplug issues for machines running
under the POWER hypervisor using hash MMU and the XIVE interrupt mode.
Commit 1ca3dec2b2df ("powerpc/xive: Prevent page fault issues in the
machine crash handler") forced the mapping of the XIVE ESB page and
this is now blocking the removal of a passthrough IO adapter because
the PCI isolation fails with "valid outstanding translations". Under
KVM, the ESB pages for the adapter interrupts are un-mapped from the
guest by the hypervisor in the KVM XIVE native device. This is is now
redundant but it's harmless.
Last is a fix to disable the XIVE debugfs file when XIVE is disabled.
Thanks,
C.
Cédric Le Goater (3):
powerpc/xive: Clear the page tables for the ESB IO mapping
powerpc/pci: unmap legacy INTx interrupts of passthrough IO adapters
powerpc/xive: Do not expose a debugfs file when XIVE is disabled
arch/powerpc/kernel/pci-hotplug.c | 2 ++
arch/powerpc/sysdev/xive/common.c | 10 ++++++++--
2 files changed, 10 insertions(+), 2 deletions(-)
--
2.25.4
