[PATCH 0/4] kvm-s390: revised version of kvm-s390 guest memory handling - v6

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



From: Christian Ehrhardt <ehrhardt@xxxxxxxxxx>

*updates in v6*
- drop the unrelated fix to kvm_host.h from the series to clear it thematically
- ensure the wait_on_bit waiter is notified
- move the reset of requests to kvm_vcpu_release to drop them early

Note: I beg a pardon for the change rate of these series in the last three days.
I wanted it off my todo pile and repeatedly found issues over night. I think it
is now in a state that might not be perfect, but its an self-contained step and
an improvement to what we had before.
I ensure you this time I wait for feedback or acceptance *g*.

*updates in v5*
- ensure dropping all vcpu requests while freeing a vcpu

*updates in v4*
- ensure kick allocations (might_sleep) are out of atomic context
- update vcpu->cpu in kvm-s390 arch handler for load/put
- remove a redundant declaration in kvm_host.h related to load/put

*updates in v3*
- replaced "kick to level" behaviour by "kick to low level and bail out"
- updates on running vcpus can now be handled without need to rerun the vcpu
- kvm_arch_set_memory_region waits until the update is consumed by the vcpu
- kickout only scheduled vcpus (wait might hang forever on non-scheduled vcpus)
- moved some code between patches 1 & 2 thematically

Note: further unification of make_all_cpu_request and the kick mechanism is
planned, but it might be good to split it from this step towards commonality.

*updates in v2*
added optimization to patch 3/3 to skip (addtional) kickout of vcpu's that had
the request already set.

This patch series results from our discussions about handling memslots and vcpu
mmu reloads. It streamlines kvm-s390 a bit by using slots_lock, vcpu-request
(KVM_REQ_MMU_RELOAD) and a kick mechanism to ensure vcpus come out of guest
context to catch the update.

I tested the reworked code a while with multiple smp guests and some extra
code that periodically injects kicks and/or mmu reload requests, but I'm happy
about every additional review feedback.

Patches included:
Subject: [PATCH 1/4] kvm-s390: infrastructure to kick vcpus out of guest state - v3
Subject: [PATCH 2/4] kvm-s390: fix interruption caused by signal - v2
Subject: [PATCH 3/4] kvm-s390: update vcpu->cpu
Subject: [PATCH 4/4] kvm-s390: streamline memslot handling - v6

Overall-Diffstat:
 arch/s390/include/asm/kvm_host.h |    9 ++---
 arch/s390/kvm/gaccess.h          |   23 ++++++------
 arch/s390/kvm/intercept.c        |   20 +++++++----
 arch/s390/kvm/kvm-s390.c         |   70 ++++++++++++++++++++-------------------
 arch/s390/kvm/kvm-s390.h         |   53 ++++++++++++++++++++++++++++-
 arch/s390/kvm/sigp.c             |   63 ++++++++++++++++++++++++-----------
 virt/kvm/kvm_main.c              |    4 ++
 7 files changed, 166 insertions(+), 76 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux