openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/arch/powerpc/kvm/book3s_32_mmu.c

431 lines
10 KiB
C
Raw Normal View History

Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
/*
* 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.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* Copyright SUSE Linux Products GmbH 2009
*
* Authors: Alexander Graf <agraf@suse.de>
*/
#include <linux/types.h>
#include <linux/string.h>
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/highmem.h>
#include <asm/tlbflush.h>
#include <asm/kvm_ppc.h>
#include <asm/kvm_book3s.h>
/* #define DEBUG_MMU */
/* #define DEBUG_MMU_PTE */
/* #define DEBUG_MMU_PTE_IP 0xfff14c40 */
#ifdef DEBUG_MMU
#define dprintk(X...) printk(KERN_INFO X)
#else
#define dprintk(X...) do { } while(0)
#endif
KVM: PPC: Fix typo in book3s_32 debug code There's a typo in the debug ifdef of the book3s_32 mmu emulation. While trying to debug something I stumbled across that and wanted to save anyone after me (or myself later) from having to debug that again. So let's fix the ifdef. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-02-19 18:00:36 +08:00
#ifdef DEBUG_MMU_PTE
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
#define dprintk_pte(X...) printk(KERN_INFO X)
#else
#define dprintk_pte(X...) do { } while(0)
#endif
#define PTEG_FLAG_ACCESSED 0x00000100
#define PTEG_FLAG_DIRTY 0x00000080
KVM: PPC: Add Book3S compatibility code Some code we had so far required defines and had code that was completely Book3S_64 specific. Since we now opened book3s.c to Book3S_32 too, we need to take care of these pieces. So let's add some minor code where it makes sense to not go the Book3S_64 code paths and add compat defines on others. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-16 06:11:53 +08:00
#ifndef SID_SHIFT
#define SID_SHIFT 28
#endif
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
static inline bool check_debug_ip(struct kvm_vcpu *vcpu)
{
#ifdef DEBUG_MMU_PTE_IP
return vcpu->arch.pc == DEBUG_MMU_PTE_IP;
#else
return true;
#endif
}
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
static inline u32 sr_vsid(u32 sr_raw)
{
return sr_raw & 0x0fffffff;
}
static inline bool sr_valid(u32 sr_raw)
{
return (sr_raw & 0x80000000) ? false : true;
}
static inline bool sr_ks(u32 sr_raw)
{
return (sr_raw & 0x40000000) ? true: false;
}
static inline bool sr_kp(u32 sr_raw)
{
return (sr_raw & 0x20000000) ? true: false;
}
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
static int kvmppc_mmu_book3s_32_xlate_bat(struct kvm_vcpu *vcpu, gva_t eaddr,
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
struct kvmppc_pte *pte, bool data,
bool iswrite);
KVM: PPC: Convert u64 -> ulong There are some pieces in the code that I overlooked that still use u64s instead of longs. This slows down 32 bit hosts unnecessarily, so let's just move them to ulong. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:46 +08:00
static int kvmppc_mmu_book3s_32_esid_to_vsid(struct kvm_vcpu *vcpu, ulong esid,
KVM: PPC: Book3S_32 guest MMU fixes This patch makes the VSID of mapped pages always reflecting all special cases we have, like split mode. It also changes the tlbie mask to 0x0ffff000 according to the spec. The mask we used before was incorrect. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-03-25 04:48:20 +08:00
u64 *vsid);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
KVM: PPC: Put segment registers in shared page Now that the actual mtsr doesn't do anything anymore, we can move the sr contents over to the shared page, so a guest can directly read and write its sr contents from guest context. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 08:29:27 +08:00
static u32 find_sr(struct kvm_vcpu *vcpu, gva_t eaddr)
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
{
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
return kvmppc_get_sr(vcpu, (eaddr >> 28) & 0xf);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
}
static u64 kvmppc_mmu_book3s_32_ea_to_vp(struct kvm_vcpu *vcpu, gva_t eaddr,
bool data)
{
KVM: PPC: Book3S_32 guest MMU fixes This patch makes the VSID of mapped pages always reflecting all special cases we have, like split mode. It also changes the tlbie mask to 0x0ffff000 according to the spec. The mask we used before was incorrect. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-03-25 04:48:20 +08:00
u64 vsid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
struct kvmppc_pte pte;
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
if (!kvmppc_mmu_book3s_32_xlate_bat(vcpu, eaddr, &pte, data, false))
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
return pte.vpage;
KVM: PPC: Book3S_32 guest MMU fixes This patch makes the VSID of mapped pages always reflecting all special cases we have, like split mode. It also changes the tlbie mask to 0x0ffff000 according to the spec. The mask we used before was incorrect. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-03-25 04:48:20 +08:00
kvmppc_mmu_book3s_32_esid_to_vsid(vcpu, eaddr >> SID_SHIFT, &vsid);
return (((u64)eaddr >> 12) & 0xffff) | (vsid << 16);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
}
static void kvmppc_mmu_book3s_32_reset_msr(struct kvm_vcpu *vcpu)
{
kvmppc_set_msr(vcpu, 0);
}
KVM: PPC: Book3S PR: Allocate kvm_vcpu structs from kvm_vcpu_cache This makes PR KVM allocate its kvm_vcpu structs from the kvm_vcpu_cache rather than having them embedded in the kvmppc_vcpu_book3s struct, which is allocated with vzalloc. The reason is to reduce the differences between PR and HV KVM in order to make is easier to have them coexist in one kernel binary. With this, the kvm_vcpu struct has a pointer to the kvmppc_vcpu_book3s struct. The pointer to the kvmppc_book3s_shadow_vcpu struct has moved from the kvmppc_vcpu_book3s struct to the kvm_vcpu struct, and is only present for 32-bit, since it is only used for 32-bit. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: squash in compile fix from Aneesh] Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:49 +08:00
static hva_t kvmppc_mmu_book3s_32_get_pteg(struct kvm_vcpu *vcpu,
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
u32 sre, gva_t eaddr,
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
bool primary)
{
KVM: PPC: Book3S PR: Allocate kvm_vcpu structs from kvm_vcpu_cache This makes PR KVM allocate its kvm_vcpu structs from the kvm_vcpu_cache rather than having them embedded in the kvmppc_vcpu_book3s struct, which is allocated with vzalloc. The reason is to reduce the differences between PR and HV KVM in order to make is easier to have them coexist in one kernel binary. With this, the kvm_vcpu struct has a pointer to the kvmppc_vcpu_book3s struct. The pointer to the kvmppc_book3s_shadow_vcpu struct has moved from the kvmppc_vcpu_book3s struct to the kvm_vcpu struct, and is only present for 32-bit, since it is only used for 32-bit. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: squash in compile fix from Aneesh] Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:49 +08:00
struct kvmppc_vcpu_book3s *vcpu_book3s = to_book3s(vcpu);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
u32 page, hash, pteg, htabmask;
hva_t r;
page = (eaddr & 0x0FFFFFFF) >> 12;
htabmask = ((vcpu_book3s->sdr1 & 0x1FF) << 16) | 0xFFC0;
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
hash = ((sr_vsid(sre) ^ page) << 6);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
if (!primary)
hash = ~hash;
hash &= htabmask;
pteg = (vcpu_book3s->sdr1 & 0xffff0000) | hash;
dprintk("MMU: pc=0x%lx eaddr=0x%lx sdr1=0x%llx pteg=0x%x vsid=0x%x\n",
KVM: PPC: Book3S PR: Do dcbz32 patching with big endian instructions When the host CPU we're running on doesn't support dcbz32 itself, but the guest wants to have dcbz only clear 32 bytes of data, we loop through every executable mapped page to search for dcbz instructions and patch them with a special privileged instruction that we emulate as dcbz32. The only guests that want to see dcbz act as 32byte are book3s_32 guests, so we don't have to worry about little endian instruction ordering. So let's just always search for big endian dcbz instructions, also when we're on a little endian host. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:52:01 +08:00
kvmppc_get_pc(vcpu), eaddr, vcpu_book3s->sdr1, pteg,
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
sr_vsid(sre));
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
KVM: PPC: Book3S PR: Allocate kvm_vcpu structs from kvm_vcpu_cache This makes PR KVM allocate its kvm_vcpu structs from the kvm_vcpu_cache rather than having them embedded in the kvmppc_vcpu_book3s struct, which is allocated with vzalloc. The reason is to reduce the differences between PR and HV KVM in order to make is easier to have them coexist in one kernel binary. With this, the kvm_vcpu struct has a pointer to the kvmppc_vcpu_book3s struct. The pointer to the kvmppc_book3s_shadow_vcpu struct has moved from the kvmppc_vcpu_book3s struct to the kvm_vcpu struct, and is only present for 32-bit, since it is only used for 32-bit. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: squash in compile fix from Aneesh] Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:49 +08:00
r = gfn_to_hva(vcpu->kvm, pteg >> PAGE_SHIFT);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
if (kvm_is_error_hva(r))
return r;
return r | (pteg & ~PAGE_MASK);
}
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
static u32 kvmppc_mmu_book3s_32_get_ptem(u32 sre, gva_t eaddr, bool primary)
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
{
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
return ((eaddr & 0x0fffffff) >> 22) | (sr_vsid(sre) << 7) |
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
(primary ? 0 : 0x40) | 0x80000000;
}
static int kvmppc_mmu_book3s_32_xlate_bat(struct kvm_vcpu *vcpu, gva_t eaddr,
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
struct kvmppc_pte *pte, bool data,
bool iswrite)
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
{
struct kvmppc_vcpu_book3s *vcpu_book3s = to_book3s(vcpu);
struct kvmppc_bat *bat;
int i;
for (i = 0; i < 8; i++) {
if (data)
bat = &vcpu_book3s->dbat[i];
else
bat = &vcpu_book3s->ibat[i];
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
if (kvmppc_get_msr(vcpu) & MSR_PR) {
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
if (!bat->vp)
continue;
} else {
if (!bat->vs)
continue;
}
if (check_debug_ip(vcpu))
{
dprintk_pte("%cBAT %02d: 0x%lx - 0x%x (0x%x)\n",
data ? 'd' : 'i', i, eaddr, bat->bepi,
bat->bepi_mask);
}
if ((eaddr & bat->bepi_mask) == bat->bepi) {
KVM: PPC: Book3S_32 guest MMU fixes This patch makes the VSID of mapped pages always reflecting all special cases we have, like split mode. It also changes the tlbie mask to 0x0ffff000 according to the spec. The mask we used before was incorrect. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-03-25 04:48:20 +08:00
u64 vsid;
kvmppc_mmu_book3s_32_esid_to_vsid(vcpu,
eaddr >> SID_SHIFT, &vsid);
vsid <<= 16;
pte->vpage = (((u64)eaddr >> 12) & 0xffff) | vsid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
pte->raddr = bat->brpn | (eaddr & ~bat->bepi_mask);
pte->may_read = bat->pp;
pte->may_write = bat->pp > 1;
pte->may_execute = true;
if (!pte->may_read) {
printk(KERN_INFO "BAT is not readable!\n");
continue;
}
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
if (iswrite && !pte->may_write) {
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
dprintk_pte("BAT is read-only!\n");
continue;
}
return 0;
}
}
return -ENOENT;
}
static int kvmppc_mmu_book3s_32_xlate_pte(struct kvm_vcpu *vcpu, gva_t eaddr,
struct kvmppc_pte *pte, bool data,
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
bool iswrite, bool primary)
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
{
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
u32 sre;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
hva_t ptegp;
u32 pteg[16];
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
u32 pte0, pte1;
KVM: PPC: Convert u64 -> ulong There are some pieces in the code that I overlooked that still use u64s instead of longs. This slows down 32 bit hosts unnecessarily, so let's just move them to ulong. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:46 +08:00
u32 ptem = 0;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
int i;
int found = 0;
KVM: PPC: Put segment registers in shared page Now that the actual mtsr doesn't do anything anymore, we can move the sr contents over to the shared page, so a guest can directly read and write its sr contents from guest context. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 08:29:27 +08:00
sre = find_sr(vcpu, eaddr);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
dprintk_pte("SR 0x%lx: vsid=0x%x, raw=0x%x\n", eaddr >> 28,
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
sr_vsid(sre), sre);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
pte->vpage = kvmppc_mmu_book3s_32_ea_to_vp(vcpu, eaddr, data);
KVM: PPC: Book3S PR: Allocate kvm_vcpu structs from kvm_vcpu_cache This makes PR KVM allocate its kvm_vcpu structs from the kvm_vcpu_cache rather than having them embedded in the kvmppc_vcpu_book3s struct, which is allocated with vzalloc. The reason is to reduce the differences between PR and HV KVM in order to make is easier to have them coexist in one kernel binary. With this, the kvm_vcpu struct has a pointer to the kvmppc_vcpu_book3s struct. The pointer to the kvmppc_book3s_shadow_vcpu struct has moved from the kvmppc_vcpu_book3s struct to the kvm_vcpu struct, and is only present for 32-bit, since it is only used for 32-bit. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: squash in compile fix from Aneesh] Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:49 +08:00
ptegp = kvmppc_mmu_book3s_32_get_pteg(vcpu, sre, eaddr, primary);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
if (kvm_is_error_hva(ptegp)) {
printk(KERN_INFO "KVM: Invalid PTEG!\n");
goto no_page_found;
}
ptem = kvmppc_mmu_book3s_32_get_ptem(sre, eaddr, primary);
if(copy_from_user(pteg, (void __user *)ptegp, sizeof(pteg))) {
printk(KERN_ERR "KVM: Can't copy data from 0x%lx!\n", ptegp);
goto no_page_found;
}
for (i=0; i<16; i+=2) {
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
pte0 = be32_to_cpu(pteg[i]);
pte1 = be32_to_cpu(pteg[i + 1]);
if (ptem == pte0) {
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
u8 pp;
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
pte->raddr = (pte1 & ~(0xFFFULL)) | (eaddr & 0xFFF);
pp = pte1 & 3;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
if ((sr_kp(sre) && (kvmppc_get_msr(vcpu) & MSR_PR)) ||
(sr_ks(sre) && !(kvmppc_get_msr(vcpu) & MSR_PR)))
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
pp |= 4;
pte->may_write = false;
pte->may_read = false;
pte->may_execute = true;
switch (pp) {
case 0:
case 1:
case 2:
case 6:
pte->may_write = true;
case 3:
case 5:
case 7:
pte->may_read = true;
break;
}
dprintk_pte("MMU: Found PTE -> %x %x - %x\n",
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
pte0, pte1, pp);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
found = 1;
break;
}
}
/* Update PTE C and A bits, so the guest's swapper knows we used the
page */
if (found) {
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
u32 pte_r = pte1;
KVM: PPC: Book3S: PR: Fix C/R bit setting Commit 9308ab8e2d made C/R HTAB updates go byte-wise into the target HTAB. However, it didn't update the guest's copy of the HTAB, but instead the host local copy of it. Write to the guest's HTAB instead. Signed-off-by: Alexander Graf <agraf@suse.de> CC: Paul Mackerras <paulus@samba.org> Acked-by: Paul Mackerras <paulus@samba.org>
2014-04-24 18:48:19 +08:00
char __user *addr = (char __user *) (ptegp + (i+1) * sizeof(u32));
KVM: PPC: Book3S PR: Make HPT accesses and updates SMP-safe This adds a per-VM mutex to provide mutual exclusion between vcpus for accesses to and updates of the guest hashed page table (HPT). This also makes the code use single-byte writes to the HPT entry when updating of the reference (R) and change (C) bits. The reason for doing this, rather than writing back the whole HPTE, is that on non-PAPR virtual machines, the guest OS might be writing to the HPTE concurrently, and writing back the whole HPTE might conflict with that. Also, real hardware does single-byte writes to update R and C. The new mutex is taken in kvmppc_mmu_book3s_64_xlate() when reading the HPT and updating R and/or C, and in the PAPR HPT update hcalls (H_ENTER, H_REMOVE, etc.). Having the mutex means that we don't need to use a hypervisor lock bit in the HPT update hcalls, and we don't need to be careful about the order in which the bytes of the HPTE are updated by those hcalls. The other change here is to make emulated TLB invalidations (tlbie) effective across all vcpus. To do this we call kvmppc_mmu_pte_vflush for all vcpus in kvmppc_ppc_book3s_64_tlbie(). For 32-bit, this makes the setting of the accessed and dirty bits use single-byte writes, and makes tlbie invalidate shadow HPTEs for all vcpus. With this, PR KVM can successfully run SMP guests. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:48 +08:00
/*
* Use single-byte writes to update the HPTE, to
* conform to what real hardware does.
*/
if (pte->may_read && !(pte_r & PTEG_FLAG_ACCESSED)) {
pte_r |= PTEG_FLAG_ACCESSED;
put_user(pte_r >> 8, addr + 2);
}
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
if (iswrite && pte->may_write && !(pte_r & PTEG_FLAG_DIRTY)) {
KVM: PPC: Book3S PR: Make HPT accesses and updates SMP-safe This adds a per-VM mutex to provide mutual exclusion between vcpus for accesses to and updates of the guest hashed page table (HPT). This also makes the code use single-byte writes to the HPT entry when updating of the reference (R) and change (C) bits. The reason for doing this, rather than writing back the whole HPTE, is that on non-PAPR virtual machines, the guest OS might be writing to the HPTE concurrently, and writing back the whole HPTE might conflict with that. Also, real hardware does single-byte writes to update R and C. The new mutex is taken in kvmppc_mmu_book3s_64_xlate() when reading the HPT and updating R and/or C, and in the PAPR HPT update hcalls (H_ENTER, H_REMOVE, etc.). Having the mutex means that we don't need to use a hypervisor lock bit in the HPT update hcalls, and we don't need to be careful about the order in which the bytes of the HPTE are updated by those hcalls. The other change here is to make emulated TLB invalidations (tlbie) effective across all vcpus. To do this we call kvmppc_mmu_pte_vflush for all vcpus in kvmppc_ppc_book3s_64_tlbie(). For 32-bit, this makes the setting of the accessed and dirty bits use single-byte writes, and makes tlbie invalidate shadow HPTEs for all vcpus. With this, PR KVM can successfully run SMP guests. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:48 +08:00
pte_r |= PTEG_FLAG_DIRTY;
put_user(pte_r, addr + 3);
}
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
if (!pte->may_read || (iswrite && !pte->may_write))
return -EPERM;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
return 0;
}
no_page_found:
if (check_debug_ip(vcpu)) {
dprintk_pte("KVM MMU: No PTE found (sdr1=0x%llx ptegp=0x%lx)\n",
to_book3s(vcpu)->sdr1, ptegp);
for (i=0; i<16; i+=2) {
KVM: PPC: Book3S_32 MMU debug compile fixes Due to previous changes, the Book3S_32 guest MMU code didn't compile properly when enabling debugging. This patch repairs the broken code paths, making it possible to define DEBUG_MMU and friends again. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-07-29 21:04:16 +08:00
dprintk_pte(" %02d: 0x%x - 0x%x (0x%x)\n",
KVM: PPC: Book3S_32: PR: Access HTAB in big endian The HTAB is always big endian. We access the guest's HTAB using copy_from/to_user, but don't yet take care of the fact that we might be running on an LE host. Wrap all accesses to the guest HTAB with big endian accessors. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 18:51:44 +08:00
i, be32_to_cpu(pteg[i]),
be32_to_cpu(pteg[i+1]), ptem);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
}
}
return -ENOENT;
}
static int kvmppc_mmu_book3s_32_xlate(struct kvm_vcpu *vcpu, gva_t eaddr,
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
struct kvmppc_pte *pte, bool data,
bool iswrite)
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
{
int r;
KVM: PPC: Magic Page Book3s support We need to override EA as well as PA lookups for the magic page. When the guest tells us to project it, the magic page overrides any guest mappings. In order to reflect that, we need to hook into all the MMU layers of KVM to force map the magic page if necessary. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-07-29 20:47:54 +08:00
ulong mp_ea = vcpu->arch.magic_page_ea;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
pte->eaddr = eaddr;
KVM: PPC: Book3S PR: Use 64k host pages where possible Currently, PR KVM uses 4k pages for the host-side mappings of guest memory, regardless of the host page size. When the host page size is 64kB, we might as well use 64k host page mappings for guest mappings of 64kB and larger pages and for guest real-mode mappings. However, the magic page has to remain a 4k page. To implement this, we first add another flag bit to the guest VSID values we use, to indicate that this segment is one where host pages should be mapped using 64k pages. For segments with this bit set we set the bits in the shadow SLB entry to indicate a 64k base page size. When faulting in host HPTEs for this segment, we make them 64k HPTEs instead of 4k. We record the pagesize in struct hpte_cache for use when invalidating the HPTE. For now we restrict the segment containing the magic page (if any) to 4k pages. It should be possible to lift this restriction in future by ensuring that the magic 4k page is appropriately positioned within a host 64k page. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:45 +08:00
pte->page_size = MMU_PAGE_4K;
KVM: PPC: Magic Page Book3s support We need to override EA as well as PA lookups for the magic page. When the guest tells us to project it, the magic page overrides any guest mappings. In order to reflect that, we need to hook into all the MMU layers of KVM to force map the magic page if necessary. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-07-29 20:47:54 +08:00
/* Magic page override */
if (unlikely(mp_ea) &&
unlikely((eaddr & ~0xfffULL) == (mp_ea & ~0xfffULL)) &&
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
!(kvmppc_get_msr(vcpu) & MSR_PR)) {
KVM: PPC: Magic Page Book3s support We need to override EA as well as PA lookups for the magic page. When the guest tells us to project it, the magic page overrides any guest mappings. In order to reflect that, we need to hook into all the MMU layers of KVM to force map the magic page if necessary. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-07-29 20:47:54 +08:00
pte->vpage = kvmppc_mmu_book3s_32_ea_to_vp(vcpu, eaddr, data);
pte->raddr = vcpu->arch.magic_page_pa | (pte->raddr & 0xfff);
pte->raddr &= KVM_PAM;
pte->may_execute = true;
pte->may_read = true;
pte->may_write = true;
return 0;
}
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
r = kvmppc_mmu_book3s_32_xlate_bat(vcpu, eaddr, pte, data, iswrite);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
if (r < 0)
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
r = kvmppc_mmu_book3s_32_xlate_pte(vcpu, eaddr, pte,
data, iswrite, true);
KVM: PPC: Book3S: Stop PTE lookup on write errors When a page lookup failed because we're not allowed to write to the page, we should not overwrite that value with another lookup on the second PTEG which will return "page not found". Instead, we should just tell the caller that we had a permission problem. This fixes Mac OS X guests looping endlessly in page lookup code for me. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-11 01:22:03 +08:00
if (r == -ENOENT)
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:51 +08:00
r = kvmppc_mmu_book3s_32_xlate_pte(vcpu, eaddr, pte,
data, iswrite, false);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
return r;
}
static u32 kvmppc_mmu_book3s_32_mfsrin(struct kvm_vcpu *vcpu, u32 srnum)
{
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
return kvmppc_get_sr(vcpu, srnum);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
}
static void kvmppc_mmu_book3s_32_mtsrin(struct kvm_vcpu *vcpu, u32 srnum,
ulong value)
{
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
kvmppc_set_sr(vcpu, srnum, value);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
kvmppc_mmu_map_segment(vcpu, srnum << SID_SHIFT);
}
static void kvmppc_mmu_book3s_32_tlbie(struct kvm_vcpu *vcpu, ulong ea, bool large)
{
KVM: PPC: Book3S PR: Make HPT accesses and updates SMP-safe This adds a per-VM mutex to provide mutual exclusion between vcpus for accesses to and updates of the guest hashed page table (HPT). This also makes the code use single-byte writes to the HPT entry when updating of the reference (R) and change (C) bits. The reason for doing this, rather than writing back the whole HPTE, is that on non-PAPR virtual machines, the guest OS might be writing to the HPTE concurrently, and writing back the whole HPTE might conflict with that. Also, real hardware does single-byte writes to update R and C. The new mutex is taken in kvmppc_mmu_book3s_64_xlate() when reading the HPT and updating R and/or C, and in the PAPR HPT update hcalls (H_ENTER, H_REMOVE, etc.). Having the mutex means that we don't need to use a hypervisor lock bit in the HPT update hcalls, and we don't need to be careful about the order in which the bytes of the HPTE are updated by those hcalls. The other change here is to make emulated TLB invalidations (tlbie) effective across all vcpus. To do this we call kvmppc_mmu_pte_vflush for all vcpus in kvmppc_ppc_book3s_64_tlbie(). For 32-bit, this makes the setting of the accessed and dirty bits use single-byte writes, and makes tlbie invalidate shadow HPTEs for all vcpus. With this, PR KVM can successfully run SMP guests. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 12:52:48 +08:00
int i;
struct kvm_vcpu *v;
/* flush this VA on all cpus */
kvm_for_each_vcpu(i, v, vcpu->kvm)
kvmppc_mmu_pte_flush(v, ea, 0x0FFFF000);
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
}
KVM: PPC: Convert u64 -> ulong There are some pieces in the code that I overlooked that still use u64s instead of longs. This slows down 32 bit hosts unnecessarily, so let's just move them to ulong. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:46 +08:00
static int kvmppc_mmu_book3s_32_esid_to_vsid(struct kvm_vcpu *vcpu, ulong esid,
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
u64 *vsid)
{
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
ulong ea = esid << SID_SHIFT;
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
u32 sr;
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
u64 gvsid = esid;
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
u64 msr = kvmppc_get_msr(vcpu);
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
if (msr & (MSR_DR|MSR_IR)) {
KVM: PPC: Put segment registers in shared page Now that the actual mtsr doesn't do anything anymore, we can move the sr contents over to the shared page, so a guest can directly read and write its sr contents from guest context. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 08:29:27 +08:00
sr = find_sr(vcpu, ea);
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
if (sr_valid(sr))
gvsid = sr_vsid(sr);
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
}
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
/* In case we only have one of MSR_IR or MSR_DR set, let's put
that in the real-mode context (and hope RM doesn't access
high memory) */
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
switch (msr & (MSR_DR|MSR_IR)) {
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
case 0:
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
*vsid = VSID_REAL | esid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
break;
case MSR_IR:
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
*vsid = VSID_REAL_IR | gvsid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
break;
case MSR_DR:
KVM: PPC: Improve split mode When in split mode, instruction relocation and data relocation are not equal. So far we implemented this mode by reserving a special pseudo-VSID for the two cases and flushing all PTEs when going into split mode, which is slow. Unfortunately 32bit Linux and Mac OS X use split mode extensively. So to not slow down things too much, I came up with a different idea: Mark the split mode with a bit in the VSID and then treat it like any other segment. This means we can just flush the shadow segment cache, but keep the PTEs intact. I verified that this works with ppc32 Linux and Mac OS X 10.4 guests and does speed them up. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-04-20 08:49:48 +08:00
*vsid = VSID_REAL_DR | gvsid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
break;
case MSR_DR|MSR_IR:
KVM: PPC: Interpret SR registers on demand Right now we're examining the contents of Book3s_32's segment registers when the register is written and put the interpreted contents into a struct. There are two reasons this is bad. For starters, the struct has worse real-time performance, as it occupies more ram. But the more important part is that with segment registers being interpreted from their raw values, we can put them in the shared page, allowing guests to mess with them directly. This patch makes the internal representation of SRs be u32s. Signed-off-by: Alexander Graf <agraf@suse.de>
2010-08-03 07:06:11 +08:00
if (sr_valid(sr))
*vsid = sr_vsid(sr);
KVM: PPC: Make BAT only guest segments work When a guest sets its SR entry to invalid, we may still find a corresponding entry in a BAT. So we need to make sure we're not faulting on invalid SR entries, but instead just claim them to be BAT resolved. This resolves breakage experienced when using libogc based guests. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-06-21 21:24:55 +08:00
else
*vsid = VSID_BAT | gvsid;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
break;
default:
BUG();
}
KVM: PPC: Make shared struct aka magic page guest endian The shared (magic) page is a data structure that contains often used supervisor privileged SPRs accessible via memory to the user to reduce the number of exits we have to take to read/write them. When we actually share this structure with the guest we have to maintain it in guest endianness, because some of the patch tricks only work with native endian load/store operations. Since we only share the structure with either host or guest in little endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv. For booke, the shared struct stays big endian. For book3s_64 hv we maintain the struct in host native endian, since it never gets shared with the guest. For book3s_64 pr we introduce a variable that tells us which endianness the shared struct is in and route every access to it through helper inline functions that evaluate this variable. Signed-off-by: Alexander Graf <agraf@suse.de>
2014-04-24 19:46:24 +08:00
if (msr & MSR_PR)
KVM: PPC: Book3S_32 guest MMU fixes This patch makes the VSID of mapped pages always reflecting all special cases we have, like split mode. It also changes the tlbie mask to 0x0ffff000 according to the spec. The mask we used before was incorrect. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2010-03-25 04:48:20 +08:00
*vsid |= VSID_PR;
Add book3s_32 guest MMU This patch adds an implementation for a G3/G4 MMU, so we can run G3 and G4 guests in KVM on Book3s_64. Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-10-30 13:47:13 +08:00
return 0;
}
static bool kvmppc_mmu_book3s_32_is_dcbz32(struct kvm_vcpu *vcpu)
{
return true;
}
void kvmppc_mmu_book3s_32_init(struct kvm_vcpu *vcpu)
{
struct kvmppc_mmu *mmu = &vcpu->arch.mmu;
mmu->mtsrin = kvmppc_mmu_book3s_32_mtsrin;
mmu->mfsrin = kvmppc_mmu_book3s_32_mfsrin;
mmu->xlate = kvmppc_mmu_book3s_32_xlate;
mmu->reset_msr = kvmppc_mmu_book3s_32_reset_msr;
mmu->tlbie = kvmppc_mmu_book3s_32_tlbie;
mmu->esid_to_vsid = kvmppc_mmu_book3s_32_esid_to_vsid;
mmu->ea_to_vp = kvmppc_mmu_book3s_32_ea_to_vp;
mmu->is_dcbz32 = kvmppc_mmu_book3s_32_is_dcbz32;
mmu->slbmte = NULL;
mmu->slbmfee = NULL;
mmu->slbmfev = NULL;
mmu->slbie = NULL;
mmu->slbia = NULL;
}