FEX-Emu/linux
1
0
Fork 0
mirror of https://github.com/FEX-Emu/linux.git synced 2024-12-21 00:42:16 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
697d3899dc
linux/arch/powerpc/kvm/book3s_hv_rm_mmu.c

680 lines
18 KiB
C
Raw Normal View History

KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License, version 2, as
* published by the Free Software Foundation.
*
* Copyright 2010-2011 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
*/
#include <linux/types.h>
#include <linux/string.h>
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/hugetlb.h>
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
#include <linux/module.h>
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
#include <asm/tlbflush.h>
#include <asm/kvm_ppc.h>
#include <asm/kvm_book3s.h>
#include <asm/mmu-hash64.h>
#include <asm/hvcall.h>
#include <asm/synch.h>
#include <asm/ppc-opcode.h>
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
/*
* Since this file is built in even if KVM is a module, we need
* a local copy of this function for the case where kvm_main.c is
* modular.
*/
static struct kvm_memory_slot *builtin_gfn_to_memslot(struct kvm *kvm,
gfn_t gfn)
{
struct kvm_memslots *slots;
struct kvm_memory_slot *memslot;
slots = kvm_memslots(kvm);
kvm_for_each_memslot(memslot, slots)
if (gfn >= memslot->base_gfn &&
gfn < memslot->base_gfn + memslot->npages)
return memslot;
return NULL;
}
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
/* Translate address of a vmalloc'd thing to a linear map address */
static void *real_vmalloc_addr(void *x)
{
unsigned long addr = (unsigned long) x;
pte_t *p;
p = find_linux_pte(swapper_pg_dir, addr);
if (!p || !pte_present(*p))
return NULL;
/* assume we don't have huge pages in vmalloc space... */
addr = (pte_pfn(*p) << PAGE_SHIFT) | (addr & ~PAGE_MASK);
return __va(addr);
}
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
/*
* Add this HPTE into the chain for the real page.
* Must be called with the chain locked; it unlocks the chain.
*/
static void kvmppc_add_revmap_chain(struct kvm *kvm, struct revmap_entry *rev,
unsigned long *rmap, long pte_index, int realmode)
{
struct revmap_entry *head, *tail;
unsigned long i;
if (*rmap & KVMPPC_RMAP_PRESENT) {
i = *rmap & KVMPPC_RMAP_INDEX;
head = &kvm->arch.revmap[i];
if (realmode)
head = real_vmalloc_addr(head);
tail = &kvm->arch.revmap[head->back];
if (realmode)
tail = real_vmalloc_addr(tail);
rev->forw = i;
rev->back = head->back;
tail->forw = pte_index;
head->back = pte_index;
} else {
rev->forw = rev->back = pte_index;
i = pte_index;
}
smp_wmb();
*rmap = i | KVMPPC_RMAP_REFERENCED | KVMPPC_RMAP_PRESENT; /* unlock */
}
/* Remove this HPTE from the chain for a real page */
static void remove_revmap_chain(struct kvm *kvm, long pte_index,
unsigned long hpte_v)
{
struct revmap_entry *rev, *next, *prev;
unsigned long gfn, ptel, head;
struct kvm_memory_slot *memslot;
unsigned long *rmap;
rev = real_vmalloc_addr(&kvm->arch.revmap[pte_index]);
ptel = rev->guest_rpte;
gfn = hpte_rpn(ptel, hpte_page_size(hpte_v, ptel));
memslot = builtin_gfn_to_memslot(kvm, gfn);
if (!memslot || (memslot->flags & KVM_MEMSLOT_INVALID))
return;
rmap = real_vmalloc_addr(&memslot->rmap[gfn - memslot->base_gfn]);
lock_rmap(rmap);
head = *rmap & KVMPPC_RMAP_INDEX;
next = real_vmalloc_addr(&kvm->arch.revmap[rev->forw]);
prev = real_vmalloc_addr(&kvm->arch.revmap[rev->back]);
next->back = rev->back;
prev->forw = rev->forw;
if (head == pte_index) {
head = rev->forw;
if (head == pte_index)
*rmap &= ~(KVMPPC_RMAP_PRESENT | KVMPPC_RMAP_INDEX);
else
*rmap = (*rmap & ~KVMPPC_RMAP_INDEX) | head;
}
unlock_rmap(rmap);
}
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
long kvmppc_h_enter(struct kvm_vcpu *vcpu, unsigned long flags,
long pte_index, unsigned long pteh, unsigned long ptel)
{
struct kvm *kvm = vcpu->kvm;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
unsigned long i, pa, gpa, gfn, psize;
unsigned long slot_fn;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
unsigned long *hpte;
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
struct revmap_entry *rev;
unsigned long g_ptel = ptel;
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
struct kvm_memory_slot *memslot;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
unsigned long *physp, pte_size;
KVM: PPC: Allow I/O mappings in memory slots This provides for the case where userspace maps an I/O device into the address range of a memory slot using a VM_PFNMAP mapping. In that case, we work out the pfn from vma->vm_pgoff, and record the cache enable bits from vma->vm_page_prot in two low-order bits in the slot_phys array entries. Then, in kvmppc_h_enter() we check that the cache bits in the HPTE that the guest wants to insert match the cache bits in the slot_phys array entry. However, we do allow the guest to create what it thinks is a non-cacheable or write-through mapping to memory that is actually cacheable, so that we can use normal system memory as part of an emulated device later on. In that case the actual HPTE we insert is a cacheable HPTE. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:32:27 +00:00
unsigned long is_io;
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
unsigned long *rmap;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
bool realmode = vcpu->arch.vcore->vcore_state == VCORE_RUNNING;
psize = hpte_page_size(pteh, ptel);
if (!psize)
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_PARAMETER;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
pteh &= ~(HPTE_V_HVLOCK | HPTE_V_ABSENT | HPTE_V_VALID);
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
/* Find the memslot (if any) for this address */
gpa = (ptel & HPTE_R_RPN) & ~(psize - 1);
gfn = gpa >> PAGE_SHIFT;
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
memslot = builtin_gfn_to_memslot(kvm, gfn);
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
pa = 0;
rmap = NULL;
if (!(memslot && !(memslot->flags & KVM_MEMSLOT_INVALID))) {
/* PPC970 can't do emulated MMIO */
if (!cpu_has_feature(CPU_FTR_ARCH_206))
return H_PARAMETER;
/* Emulated MMIO - mark this with key=31 */
pteh |= HPTE_V_ABSENT;
ptel |= HPTE_R_KEY_HI | HPTE_R_KEY_LO;
goto do_insert;
}
KVM: PPC: Allow use of small pages to back Book3S HV guests This relaxes the requirement that the guest memory be provided as 16MB huge pages, allowing it to be provided as normal memory, i.e. in pages of PAGE_SIZE bytes (4k or 64k). To allow this, we index the kvm->arch.slot_phys[] arrays with a small page index, even if huge pages are being used, and use the low-order 5 bits of each entry to store the order of the enclosing page with respect to normal pages, i.e. log_2(enclosing_page_size / PAGE_SIZE). Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:41 +00:00
/* Check if the requested page fits entirely in the memslot. */
if (!slot_is_aligned(memslot, psize))
return H_PARAMETER;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
slot_fn = gfn - memslot->base_gfn;
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
rmap = &memslot->rmap[slot_fn];
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
physp = kvm->arch.slot_phys[memslot->id];
if (!physp)
return H_PARAMETER;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
physp += slot_fn;
if (realmode)
physp = real_vmalloc_addr(physp);
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
pa = *physp;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
if (!pa)
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
return H_TOO_HARD;
KVM: PPC: Allow I/O mappings in memory slots This provides for the case where userspace maps an I/O device into the address range of a memory slot using a VM_PFNMAP mapping. In that case, we work out the pfn from vma->vm_pgoff, and record the cache enable bits from vma->vm_page_prot in two low-order bits in the slot_phys array entries. Then, in kvmppc_h_enter() we check that the cache bits in the HPTE that the guest wants to insert match the cache bits in the slot_phys array entry. However, we do allow the guest to create what it thinks is a non-cacheable or write-through mapping to memory that is actually cacheable, so that we can use normal system memory as part of an emulated device later on. In that case the actual HPTE we insert is a cacheable HPTE. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:32:27 +00:00
is_io = pa & (HPTE_R_I | HPTE_R_W);
KVM: PPC: Allow use of small pages to back Book3S HV guests This relaxes the requirement that the guest memory be provided as 16MB huge pages, allowing it to be provided as normal memory, i.e. in pages of PAGE_SIZE bytes (4k or 64k). To allow this, we index the kvm->arch.slot_phys[] arrays with a small page index, even if huge pages are being used, and use the low-order 5 bits of each entry to store the order of the enclosing page with respect to normal pages, i.e. log_2(enclosing_page_size / PAGE_SIZE). Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:41 +00:00
pte_size = PAGE_SIZE << (pa & KVMPPC_PAGE_ORDER_MASK);
KVM: PPC: Keep page physical addresses in per-slot arrays This allocates an array for each memory slot that is added to store the physical addresses of the pages in the slot. This array is vmalloc'd and accessed in kvmppc_h_enter using real_vmalloc_addr(). This allows us to remove the ram_pginfo field from the kvm_arch struct, and removes the 64GB guest RAM limit that we had. We use the low-order bits of the array entries to store a flag indicating that we have done get_page on the corresponding page, and therefore need to call put_page when we are finished with the page. Currently this is set for all pages except those in our special RMO regions. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:28:21 +00:00
pa &= PAGE_MASK;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
if (pte_size < psize)
return H_PARAMETER;
if (pa && pte_size > psize)
pa |= gpa & (pte_size - 1);
ptel &= ~(HPTE_R_PP0 - psize);
ptel |= pa;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
pteh |= HPTE_V_VALID;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
/* Check WIMG */
KVM: PPC: Allow I/O mappings in memory slots This provides for the case where userspace maps an I/O device into the address range of a memory slot using a VM_PFNMAP mapping. In that case, we work out the pfn from vma->vm_pgoff, and record the cache enable bits from vma->vm_page_prot in two low-order bits in the slot_phys array entries. Then, in kvmppc_h_enter() we check that the cache bits in the HPTE that the guest wants to insert match the cache bits in the slot_phys array entry. However, we do allow the guest to create what it thinks is a non-cacheable or write-through mapping to memory that is actually cacheable, so that we can use normal system memory as part of an emulated device later on. In that case the actual HPTE we insert is a cacheable HPTE. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:32:27 +00:00
if (!hpte_cache_flags_ok(ptel, is_io)) {
if (is_io)
return H_PARAMETER;
/*
* Allow guest to map emulated device memory as
* uncacheable, but actually make it cacheable.
*/
ptel &= ~(HPTE_R_W|HPTE_R_I|HPTE_R_G);
ptel |= HPTE_R_M;
}
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
do_insert:
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (pte_index >= HPT_NPTE)
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_PARAMETER;
if (likely((flags & H_EXACT) == 0)) {
pte_index &= ~7UL;
hpte = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
for (i = 0; i < 8; ++i) {
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
if ((*hpte & HPTE_V_VALID) == 0 &&
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
try_lock_hpte(hpte, HPTE_V_HVLOCK | HPTE_V_VALID |
HPTE_V_ABSENT))
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
break;
hpte += 2;
}
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
if (i == 8) {
/*
* Since try_lock_hpte doesn't retry (not even stdcx.
* failures), it could be that there is a free slot
* but we transiently failed to lock it. Try again,
* actually locking each slot and checking it.
*/
hpte -= 16;
for (i = 0; i < 8; ++i) {
while (!try_lock_hpte(hpte, HPTE_V_HVLOCK))
cpu_relax();
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (!(*hpte & (HPTE_V_VALID | HPTE_V_ABSENT)))
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
break;
*hpte &= ~HPTE_V_HVLOCK;
hpte += 2;
}
if (i == 8)
return H_PTEG_FULL;
}
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
pte_index += i;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
} else {
hpte = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (!try_lock_hpte(hpte, HPTE_V_HVLOCK | HPTE_V_VALID |
HPTE_V_ABSENT)) {
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
/* Lock the slot and check again */
while (!try_lock_hpte(hpte, HPTE_V_HVLOCK))
cpu_relax();
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (*hpte & (HPTE_V_VALID | HPTE_V_ABSENT)) {
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
*hpte &= ~HPTE_V_HVLOCK;
return H_PTEG_FULL;
}
}
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
}
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
/* Save away the guest's idea of the second HPTE dword */
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
rev = &kvm->arch.revmap[pte_index];
if (realmode)
rev = real_vmalloc_addr(rev);
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (rev)
rev->guest_rpte = g_ptel;
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
/* Link HPTE into reverse-map chain */
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (pteh & HPTE_V_VALID) {
if (realmode)
rmap = real_vmalloc_addr(rmap);
lock_rmap(rmap);
kvmppc_add_revmap_chain(kvm, rev, rmap, pte_index, realmode);
}
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
hpte[1] = ptel;
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
/* Write the first HPTE dword, unlocking the HPTE and making it valid */
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
eieio();
hpte[0] = pteh;
asm volatile("ptesync" : : : "memory");
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
vcpu->arch.gpr[4] = pte_index;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_SUCCESS;
}
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:31:00 +00:00
EXPORT_SYMBOL_GPL(kvmppc_h_enter);
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
#define LOCK_TOKEN (*(u32 *)(&get_paca()->lock_token))
static inline int try_lock_tlbie(unsigned int *lock)
{
unsigned int tmp, old;
unsigned int token = LOCK_TOKEN;
asm volatile("1:lwarx %1,0,%2\n"
" cmpwi cr0,%1,0\n"
" bne 2f\n"
" stwcx. %3,0,%2\n"
" bne- 1b\n"
" isync\n"
"2:"
: "=&r" (tmp), "=&r" (old)
: "r" (lock), "r" (token)
: "cc", "memory");
return old == 0;
}
long kvmppc_h_remove(struct kvm_vcpu *vcpu, unsigned long flags,
unsigned long pte_index, unsigned long avpn,
unsigned long va)
{
struct kvm *kvm = vcpu->kvm;
unsigned long *hpte;
unsigned long v, r, rb;
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (pte_index >= HPT_NPTE)
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_PARAMETER;
hpte = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
while (!try_lock_hpte(hpte, HPTE_V_HVLOCK))
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
cpu_relax();
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if ((hpte[0] & (HPTE_V_ABSENT | HPTE_V_VALID)) == 0 ||
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
((flags & H_AVPN) && (hpte[0] & ~0x7fUL) != avpn) ||
((flags & H_ANDCOND) && (hpte[0] & avpn) != 0)) {
hpte[0] &= ~HPTE_V_HVLOCK;
return H_NOT_FOUND;
}
if (atomic_read(&kvm->online_vcpus) == 1)
flags |= H_LOCAL;
vcpu->arch.gpr[4] = v = hpte[0] & ~HPTE_V_HVLOCK;
vcpu->arch.gpr[5] = r = hpte[1];
rb = compute_tlbie_rb(v, r, pte_index);
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (v & HPTE_V_VALID)
remove_revmap_chain(kvm, pte_index, v);
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
smp_wmb();
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
hpte[0] = 0;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (!(v & HPTE_V_VALID))
return H_SUCCESS;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
if (!(flags & H_LOCAL)) {
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
while (!try_lock_tlbie(&kvm->arch.tlbie_lock))
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
cpu_relax();
asm volatile("ptesync" : : : "memory");
asm volatile(PPC_TLBIE(%1,%0)"; eieio; tlbsync"
: : "r" (rb), "r" (kvm->arch.lpid));
asm volatile("ptesync" : : : "memory");
kvm->arch.tlbie_lock = 0;
} else {
asm volatile("ptesync" : : : "memory");
asm volatile("tlbiel %0" : : "r" (rb));
asm volatile("ptesync" : : : "memory");
}
return H_SUCCESS;
}
long kvmppc_h_bulk_remove(struct kvm_vcpu *vcpu)
{
struct kvm *kvm = vcpu->kvm;
unsigned long *args = &vcpu->arch.gpr[4];
unsigned long *hp, tlbrb[4];
long int i, found;
long int n_inval = 0;
unsigned long flags, req, pte_index;
long int local = 0;
long int ret = H_SUCCESS;
if (atomic_read(&kvm->online_vcpus) == 1)
local = 1;
for (i = 0; i < 4; ++i) {
pte_index = args[i * 2];
flags = pte_index >> 56;
pte_index &= ((1ul << 56) - 1);
req = flags >> 6;
flags &= 3;
if (req == 3)
break;
if (req != 1 || flags == 3 ||
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
pte_index >= HPT_NPTE) {
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
/* parameter error */
args[i * 2] = ((0xa0 | flags) << 56) + pte_index;
ret = H_PARAMETER;
break;
}
hp = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
while (!try_lock_hpte(hp, HPTE_V_HVLOCK))
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
cpu_relax();
found = 0;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (hp[0] & (HPTE_V_ABSENT | HPTE_V_VALID)) {
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
switch (flags & 3) {
case 0: /* absolute */
found = 1;
break;
case 1: /* andcond */
if (!(hp[0] & args[i * 2 + 1]))
found = 1;
break;
case 2: /* AVPN */
if ((hp[0] & ~0x7fUL) == args[i * 2 + 1])
found = 1;
break;
}
}
if (!found) {
hp[0] &= ~HPTE_V_HVLOCK;
args[i * 2] = ((0x90 | flags) << 56) + pte_index;
continue;
}
/* insert R and C bits from PTE */
flags |= (hp[1] >> 5) & 0x0c;
args[i * 2] = ((0x80 | flags) << 56) + pte_index;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (hp[0] & HPTE_V_VALID) {
tlbrb[n_inval++] = compute_tlbie_rb(hp[0], hp[1], pte_index);
remove_revmap_chain(kvm, pte_index, hp[0]);
}
KVM: PPC: Maintain a doubly-linked list of guest HPTEs for each gfn This expands the reverse mapping array to contain two links for each HPTE which are used to link together HPTEs that correspond to the same guest logical page. Each circular list of HPTEs is pointed to by the rmap array entry for the guest logical page, pointed to by the relevant memslot. Links are 32-bit HPT entry indexes rather than full 64-bit pointers, to save space. We use 3 of the remaining 32 bits in the rmap array entries as a lock bit, a referenced bit and a present bit (the present bit is needed since HPTE index 0 is valid). The bit lock for the rmap chain nests inside the HPTE lock bit. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:33:07 +00:00
smp_wmb();
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
hp[0] = 0;
}
if (n_inval == 0)
return ret;
if (!local) {
while(!try_lock_tlbie(&kvm->arch.tlbie_lock))
cpu_relax();
asm volatile("ptesync" : : : "memory");
for (i = 0; i < n_inval; ++i)
asm volatile(PPC_TLBIE(%1,%0)
: : "r" (tlbrb[i]), "r" (kvm->arch.lpid));
asm volatile("eieio; tlbsync; ptesync" : : : "memory");
kvm->arch.tlbie_lock = 0;
} else {
asm volatile("ptesync" : : : "memory");
for (i = 0; i < n_inval; ++i)
asm volatile("tlbiel %0" : : "r" (tlbrb[i]));
asm volatile("ptesync" : : : "memory");
}
return ret;
}
long kvmppc_h_protect(struct kvm_vcpu *vcpu, unsigned long flags,
unsigned long pte_index, unsigned long avpn,
unsigned long va)
{
struct kvm *kvm = vcpu->kvm;
unsigned long *hpte;
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
struct revmap_entry *rev;
unsigned long v, r, rb, mask, bits;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (pte_index >= HPT_NPTE)
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_PARAMETER;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
hpte = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Make the H_ENTER hcall more reliable At present, our implementation of H_ENTER only makes one try at locking each slot that it looks at, and doesn't even retry the ldarx/stdcx. atomic update sequence that it uses to attempt to lock the slot. Thus it can return the H_PTEG_FULL error unnecessarily, particularly when the H_EXACT flag is set, meaning that the caller wants a specific PTEG slot. This improves the situation by making a second pass when no free HPTE slot is found, where we spin until we succeed in locking each slot in turn and then check whether it is full while we hold the lock. If the second pass fails, then we return H_PTEG_FULL. This also moves lock_hpte to a header file (since later commits in this series will need to use it from other source files) and renames it to try_lock_hpte, which is a somewhat less misleading name. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:30:16 +00:00
while (!try_lock_hpte(hpte, HPTE_V_HVLOCK))
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
cpu_relax();
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if ((hpte[0] & (HPTE_V_ABSENT | HPTE_V_VALID)) == 0 ||
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
((flags & H_AVPN) && (hpte[0] & ~0x7fUL) != avpn)) {
hpte[0] &= ~HPTE_V_HVLOCK;
return H_NOT_FOUND;
}
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
if (atomic_read(&kvm->online_vcpus) == 1)
flags |= H_LOCAL;
v = hpte[0];
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
bits = (flags << 55) & HPTE_R_PP0;
bits |= (flags << 48) & HPTE_R_KEY_HI;
bits |= flags & (HPTE_R_PP | HPTE_R_N | HPTE_R_KEY_LO);
/* Update guest view of 2nd HPTE dword */
mask = HPTE_R_PP0 | HPTE_R_PP | HPTE_R_N |
HPTE_R_KEY_HI | HPTE_R_KEY_LO;
rev = real_vmalloc_addr(&kvm->arch.revmap[pte_index]);
if (rev) {
r = (rev->guest_rpte & ~mask) | bits;
rev->guest_rpte = r;
}
r = (hpte[1] & ~mask) | bits;
/* Update HPTE */
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (v & HPTE_V_VALID) {
rb = compute_tlbie_rb(v, r, pte_index);
hpte[0] = v & ~HPTE_V_VALID;
if (!(flags & H_LOCAL)) {
while(!try_lock_tlbie(&kvm->arch.tlbie_lock))
cpu_relax();
asm volatile("ptesync" : : : "memory");
asm volatile(PPC_TLBIE(%1,%0)"; eieio; tlbsync"
: : "r" (rb), "r" (kvm->arch.lpid));
asm volatile("ptesync" : : : "memory");
kvm->arch.tlbie_lock = 0;
} else {
asm volatile("ptesync" : : : "memory");
asm volatile("tlbiel %0" : : "r" (rb));
asm volatile("ptesync" : : : "memory");
}
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
}
hpte[1] = r;
eieio();
hpte[0] = v & ~HPTE_V_HVLOCK;
asm volatile("ptesync" : : : "memory");
return H_SUCCESS;
}
long kvmppc_h_read(struct kvm_vcpu *vcpu, unsigned long flags,
unsigned long pte_index)
{
struct kvm *kvm = vcpu->kvm;
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
unsigned long *hpte, v, r;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
int i, n = 1;
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
struct revmap_entry *rev = NULL;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (pte_index >= HPT_NPTE)
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
return H_PARAMETER;
if (flags & H_READ_4) {
pte_index &= ~3;
n = 4;
}
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (flags & H_R_XLATE)
rev = real_vmalloc_addr(&kvm->arch.revmap[pte_index]);
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
for (i = 0; i < n; ++i, ++pte_index) {
hpte = (unsigned long *)(kvm->arch.hpt_virt + (pte_index << 4));
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
v = hpte[0] & ~HPTE_V_HVLOCK;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
r = hpte[1];
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
if (v & HPTE_V_ABSENT) {
v &= ~HPTE_V_ABSENT;
v |= HPTE_V_VALID;
}
if (v & HPTE_V_VALID) {
KVM: PPC: Keep a record of HV guest view of hashed page table entries This adds an array that parallels the guest hashed page table (HPT), that is, it has one entry per HPTE, used to store the guest's view of the second doubleword of the corresponding HPTE. The first doubleword in the HPTE is the same as the guest's idea of it, so we don't need to store a copy, but the second doubleword in the HPTE has the real page number rather than the guest's logical page number. This allows us to remove the back_translate() and reverse_xlate() functions. This "reverse mapping" array is vmalloc'd, meaning that to access it in real mode we have to walk the kernel's page tables explicitly. That is done by the new real_vmalloc_addr() function. (In fact this returns an address in the linear mapping, so the result is usable both in real mode and in virtual mode.) There are also some minor cleanups here: moving the definitions of HPT_ORDER etc. to a header file and defining HPT_NPTE for HPT_NPTEG << 3. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:27:39 +00:00
if (rev)
r = rev[i].guest_rpte;
else
r = hpte[1] | HPTE_R_RPN;
}
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
vcpu->arch.gpr[4 + i * 2] = v;
KVM: PPC: Handle some PAPR hcalls in the kernel This adds the infrastructure for handling PAPR hcalls in the kernel, either early in the guest exit path while we are still in real mode, or later once the MMU has been turned back on and we are in the full kernel context. The advantage of handling hcalls in real mode if possible is that we avoid two partition switches -- and this will become more important when we support SMT4 guests, since a partition switch means we have to pull all of the threads in the core out of the guest. The disadvantage is that we can only access the kernel linear mapping, not anything vmalloced or ioremapped, since the MMU is off. This also adds code to handle the following hcalls in real mode: H_ENTER Add an HPTE to the hashed page table H_REMOVE Remove an HPTE from the hashed page table H_READ Read HPTEs from the hashed page table H_PROTECT Change the protection bits in an HPTE H_BULK_REMOVE Remove up to 4 HPTEs from the hashed page table H_SET_DABR Set the data address breakpoint register Plus code to handle the following hcalls in the kernel: H_CEDE Idle the vcpu until an interrupt or H_PROD hcall arrives H_PROD Wake up a ceded vcpu H_REGISTER_VPA Register a virtual processor area (VPA) The code that runs in real mode has to be in the base kernel, not in the module, if KVM is compiled as a module. The real-mode code can only access the kernel linear mapping, not vmalloc or ioremap space. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 00:22:05 +00:00
vcpu->arch.gpr[5 + i * 2] = r;
}
return H_SUCCESS;
}
KVM: PPC: Implement MMIO emulation support for Book3S HV guests This provides the low-level support for MMIO emulation in Book3S HV guests. When the guest tries to map a page which is not covered by any memslot, that page is taken to be an MMIO emulation page. Instead of inserting a valid HPTE, we insert an HPTE that has the valid bit clear but another hypervisor software-use bit set, which we call HPTE_V_ABSENT, to indicate that this is an absent page. An absent page is treated much like a valid page as far as guest hcalls (H_ENTER, H_REMOVE, H_READ etc.) are concerned, except of course that an absent HPTE doesn't need to be invalidated with tlbie since it was never valid as far as the hardware is concerned. When the guest accesses a page for which there is an absent HPTE, it will take a hypervisor data storage interrupt (HDSI) since we now set the VPM1 bit in the LPCR. Our HDSI handler for HPTE-not-present faults looks up the hash table and if it finds an absent HPTE mapping the requested virtual address, will switch to kernel mode and handle the fault in kvmppc_book3s_hv_page_fault(), which at present just calls kvmppc_hv_emulate_mmio() to set up the MMIO emulation. This is based on an earlier patch by Benjamin Herrenschmidt, but since heavily reworked. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 12:36:37 +00:00
static int slb_base_page_shift[4] = {
24, /* 16M */
16, /* 64k */
34, /* 16G */
20, /* 1M, unsupported */
};
long kvmppc_hv_find_lock_hpte(struct kvm *kvm, gva_t eaddr, unsigned long slb_v,
unsigned long valid)
{
unsigned int i;
unsigned int pshift;
unsigned long somask;
unsigned long vsid, hash;
unsigned long avpn;
unsigned long *hpte;
unsigned long mask, val;
unsigned long v, r;
/* Get page shift, work out hash and AVPN etc. */
mask = SLB_VSID_B | HPTE_V_AVPN | HPTE_V_SECONDARY;
val = 0;
pshift = 12;
if (slb_v & SLB_VSID_L) {
mask |= HPTE_V_LARGE;
val |= HPTE_V_LARGE;
pshift = slb_base_page_shift[(slb_v & SLB_VSID_LP) >> 4];
}
if (slb_v & SLB_VSID_B_1T) {
somask = (1UL << 40) - 1;
vsid = (slb_v & ~SLB_VSID_B) >> SLB_VSID_SHIFT_1T;
vsid ^= vsid << 25;
} else {
somask = (1UL << 28) - 1;
vsid = (slb_v & ~SLB_VSID_B) >> SLB_VSID_SHIFT;
}
hash = (vsid ^ ((eaddr & somask) >> pshift)) & HPT_HASH_MASK;
avpn = slb_v & ~(somask >> 16); /* also includes B */
avpn |= (eaddr & somask) >> 16;
if (pshift >= 24)
avpn &= ~((1UL << (pshift - 16)) - 1);
else
avpn &= ~0x7fUL;
val |= avpn;
for (;;) {
hpte = (unsigned long *)(kvm->arch.hpt_virt + (hash << 7));
for (i = 0; i < 16; i += 2) {
/* Read the PTE racily */
v = hpte[i] & ~HPTE_V_HVLOCK;
/* Check valid/absent, hash, segment size and AVPN */
if (!(v & valid) || (v & mask) != val)
continue;
/* Lock the PTE and read it under the lock */
while (!try_lock_hpte(&hpte[i], HPTE_V_HVLOCK))
cpu_relax();
v = hpte[i] & ~HPTE_V_HVLOCK;
r = hpte[i+1];
/*
* Check the HPTE again, including large page size
* Since we don't currently allow any MPSS (mixed
* page-size segment) page sizes, it is sufficient
* to check against the actual page size.
*/
if ((v & valid) && (v & mask) == val &&
hpte_page_size(v, r) == (1ul << pshift))
/* Return with the HPTE still locked */
return (hash << 3) + (i >> 1);
/* Unlock and move on */
hpte[i] = v;
}
if (val & HPTE_V_SECONDARY)
break;
val |= HPTE_V_SECONDARY;
hash = hash ^ HPT_HASH_MASK;
}
return -1;
}
EXPORT_SYMBOL(kvmppc_hv_find_lock_hpte);
/*
* Called in real mode to check whether an HPTE not found fault
* is due to accessing an emulated MMIO page.
* Returns a possibly modified status (DSISR) value if not
* (i.e. pass the interrupt to the guest),
* -1 to pass the fault up to host kernel mode code, -2 to do that
* and also load the instruction word,
* or 0 if we should make the guest retry the access.
*/
long kvmppc_hpte_hv_fault(struct kvm_vcpu *vcpu, unsigned long addr,
unsigned long slb_v, unsigned int status)
{
struct kvm *kvm = vcpu->kvm;
long int index;
unsigned long v, r, gr;
unsigned long *hpte;
unsigned long valid;
struct revmap_entry *rev;
unsigned long pp, key;
valid = HPTE_V_VALID | HPTE_V_ABSENT;
index = kvmppc_hv_find_lock_hpte(kvm, addr, slb_v, valid);
if (index < 0)
return status; /* there really was no HPTE */
hpte = (unsigned long *)(kvm->arch.hpt_virt + (index << 4));
v = hpte[0] & ~HPTE_V_HVLOCK;
r = hpte[1];
rev = real_vmalloc_addr(&kvm->arch.revmap[index]);
gr = rev->guest_rpte;
/* Unlock the HPTE */
asm volatile("lwsync" : : : "memory");
hpte[0] = v;
/* If the HPTE is valid by now, retry the instruction */
if (v & HPTE_V_VALID)
return 0;
/* Check access permissions to the page */
pp = gr & (HPTE_R_PP0 | HPTE_R_PP);
key = (vcpu->arch.shregs.msr & MSR_PR) ? SLB_VSID_KP : SLB_VSID_KS;
if (status & DSISR_ISSTORE) {
/* check write permission */
if (!hpte_write_permission(pp, slb_v & key))
goto protfault;
} else {
if (!hpte_read_permission(pp, slb_v & key))
goto protfault;
}
/* Check storage key, if applicable */
if (vcpu->arch.shregs.msr & MSR_DR) {
unsigned int perm = hpte_get_skey_perm(gr, vcpu->arch.amr);
if (status & DSISR_ISSTORE)
perm >>= 1;
if (perm & 1)
return (status & ~DSISR_NOHPTE) | DSISR_KEYFAULT;
}
/* Save HPTE info for virtual-mode handler */
vcpu->arch.pgfault_addr = addr;
vcpu->arch.pgfault_index = index;
vcpu->arch.pgfault_hpte[0] = v;
vcpu->arch.pgfault_hpte[1] = r;
if (vcpu->arch.shregs.msr & MSR_IR)
return -2; /* MMIO emulation - load instr word */
return -1; /* send fault up to host kernel mode */
protfault:
return (status & ~DSISR_NOHPTE) | DSISR_PROTFAULT;
}
Reference in New Issue Copy Permalink