BMC/Intel-BMC/linux.git/virt/kvm/kvm_main.c, branch dev

sched, preempt_notifier: separate notifier registration from static_key inc/dec

2015-07-03T16:55:00+00:00

Commit 1cde2930e154 ("sched/preempt: Add static_key() to preempt_notifiers") had two problems. First, the preempt-notifier API needs to sleep with the addition of the static_key, we do however need to hold off preemption while modifying the preempt notifier list, otherwise a preemption could observe an inconsistent list state. KVM correctly registers and unregisters preempt notifiers with preemption disabled, so the sleep caused dmesg splats. Second, KVM registers and unregisters preemption notifiers very often (in vcpu_load/vcpu_put). With a single uniprocessor guest the static key would move between 0 and 1 continuously, hitting the slow path on every userspace exit. To fix this, wrap the static_key inc/dec in a new API, and call it from KVM. Fixes: 1cde2930e154 ("sched/preempt: Add static_key() to preempt_notifiers") Reported-by: Pontus Fuchs Reported-by: Takashi Iwai Tested-by: Takashi Iwai Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Paolo Bonzini

KVM: implement multiple address spaces

2015-06-05T15:26:35+00:00

Only two ioctls have to be modified; the address space id is placed in the higher 16 bits of their slot id argument. As of this patch, no architecture defines more than one address space; x86 will be the first. Reviewed-by: Radim Krčmář Signed-off-by: Paolo Bonzini

KVM: add vcpu-specific functions to read/write/translate GFNs

2015-06-05T15:26:34+00:00

We need to hide SMRAM from guests not running in SMM. Therefore, all uses of kvm_read_guest* and kvm_write_guest* must be changed to use different address spaces, depending on whether the VCPU is in system management mode. We need to introduce a new family of functions for this purpose. For now, the VCPU-based functions have the same behavior as the existing per-VM ones, they just accept a different type for the first argument. Later however they will be changed to use one of many "struct kvm_memslots" stored in struct kvm, through an architecture hook. VM-based functions will unconditionally use the first memslots pointer. Whenever possible, this patch introduces slot-based functions with an __ prefix, with two wrappers for generic and vcpu-based actions. The exceptions are kvm_read_guest and kvm_write_guest, which are copied into the new functions kvm_vcpu_read_guest and kvm_vcpu_write_guest. Reviewed-by: Radim Krčmář Signed-off-by: Paolo Bonzini

KVM: remove unused argument from mark_page_dirty_in_slot

2015-05-28T08:43:37+00:00

Signed-off-by: Paolo Bonzini

KVM: remove __gfn_to_pfn

2015-05-28T08:43:34+00:00

Most of the function that wrap it can be rewritten without it, except for gfn_to_pfn_prot. Just inline it into gfn_to_pfn_prot, and rewrite the other function on top of gfn_to_pfn_memslot*. Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini

KVM: pass kvm_memory_slot to gfn_to_page_many_atomic

2015-05-28T08:43:33+00:00

The memory slot is already available from gfn_to_memslot_dirty_bitmap. Isn't it a shame to look it up again? Plus, it makes gfn_to_page_many_atomic agnostic of multiple VCPU address spaces. Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini

KVM: add "new" argument to kvm_arch_commit_memory_region

2015-05-28T08:42:58+00:00

This lets the function access the new memory slot without going through kvm_memslots and id_to_memslot. It will simplify the code when more than one address space will be supported. Unfortunately, the "const"ness of the new argument must be casted away in two places. Fixing KVM to accept const struct kvm_memory_slot pointers would require modifications in pretty much all architectures, and is left for later. Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini

KVM: add memslots argument to kvm_arch_memslots_updated

2015-05-26T10:40:17+00:00

Prepare for the case of multiple address spaces. Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini

KVM: const-ify uses of struct kvm_userspace_memory_region

2015-05-26T10:40:13+00:00

Architecture-specific helpers are not supposed to muck with struct kvm_userspace_memory_region contents. Add const to enforce this. In order to eliminate the only write in __kvm_set_memory_region, the cleaning of deleted slots is pulled up from update_memslots to __kvm_set_memory_region. Reviewed-by: Takuya Yoshikawa Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini

KVM: use kvm_memslots whenever possible

2015-05-26T10:40:08+00:00

kvm_memslots provides lockdep checking. Use it consistently instead of explicit dereferencing of kvm->memslots. Reviewed-by: Radim Krcmar Signed-off-by: Paolo Bonzini