OpenCloudOS-Kernel/arch/powerpc/kvm/book3s_64_vio_hv.c

571 lines
14 KiB
C
Raw Normal View History

/*
* 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 2010 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
* Copyright 2011 David Gibson, IBM Corporation <dwg@au1.ibm.com>
* Copyright 2016 Alexey Kardashevskiy, IBM Corporation <aik@au1.ibm.com>
*/
#include <linux/types.h>
#include <linux/string.h>
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/highmem.h>
#include <linux/gfp.h>
#include <linux/slab.h>
#include <linux/hugetlb.h>
#include <linux/list.h>
#include <asm/tlbflush.h>
#include <asm/kvm_ppc.h>
#include <asm/kvm_book3s.h>
#include <asm/book3s/64/mmu-hash.h>
#include <asm/mmu_context.h>
#include <asm/hvcall.h>
#include <asm/synch.h>
#include <asm/ppc-opcode.h>
#include <asm/kvm_host.h>
#include <asm/udbg.h>
#include <asm/iommu.h>
#include <asm/tce.h>
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
#ifdef CONFIG_BUG
#define WARN_ON_ONCE_RM(condition) ({ \
static bool __section(.data.unlikely) __warned; \
int __ret_warn_once = !!(condition); \
\
if (unlikely(__ret_warn_once && !__warned)) { \
__warned = true; \
pr_err("WARN_ON_ONCE_RM: (%s) at %s:%u\n", \
__stringify(condition), \
__func__, __LINE__); \
dump_stack(); \
} \
unlikely(__ret_warn_once); \
})
#else
#define WARN_ON_ONCE_RM(condition) ({ \
int __ret_warn_on = !!(condition); \
unlikely(__ret_warn_on); \
})
#endif
#define TCES_PER_PAGE (PAGE_SIZE / sizeof(u64))
/*
* Finds a TCE table descriptor by LIOBN.
*
* WARNING: This will be called in real or virtual mode on HV KVM and virtual
* mode on PR KVM
*/
struct kvmppc_spapr_tce_table *kvmppc_find_table(struct kvm *kvm,
unsigned long liobn)
{
struct kvmppc_spapr_tce_table *stt;
list_for_each_entry_lockless(stt, &kvm->arch.spapr_tce_tables, list)
if (stt->liobn == liobn)
return stt;
return NULL;
}
EXPORT_SYMBOL_GPL(kvmppc_find_table);
/*
* Validates TCE address.
* At the moment flags and page mask are validated.
* As the host kernel does not access those addresses (just puts them
* to the table and user space is supposed to process them), we can skip
* checking other things (such as TCE is a guest RAM address or the page
* was actually allocated).
*
* WARNING: This will be called in real-mode on HV KVM and virtual
* mode on PR KVM
*/
long kvmppc_tce_validate(struct kvmppc_spapr_tce_table *stt, unsigned long tce)
{
unsigned long gpa = tce & ~(TCE_PCI_READ | TCE_PCI_WRITE);
enum dma_data_direction dir = iommu_tce_direction(tce);
/* Allow userspace to poison TCE table */
if (dir == DMA_NONE)
return H_SUCCESS;
if (iommu_tce_check_gpa(stt->page_shift, gpa))
return H_PARAMETER;
return H_SUCCESS;
}
EXPORT_SYMBOL_GPL(kvmppc_tce_validate);
/* Note on the use of page_address() in real mode,
*
* It is safe to use page_address() in real mode on ppc64 because
* page_address() is always defined as lowmem_page_address()
* which returns __va(PFN_PHYS(page_to_pfn(page))) which is arithmetic
* operation and does not access page struct.
*
* Theoretically page_address() could be defined different
* but either WANT_PAGE_VIRTUAL or HASHED_PAGE_VIRTUAL
* would have to be enabled.
* WANT_PAGE_VIRTUAL is never enabled on ppc32/ppc64,
* HASHED_PAGE_VIRTUAL could be enabled for ppc32 only and only
* if CONFIG_HIGHMEM is defined. As CONFIG_SPARSEMEM_VMEMMAP
* is not expected to be enabled on ppc32, page_address()
* is safe for ppc32 as well.
*
* WARNING: This will be called in real-mode on HV KVM and virtual
* mode on PR KVM
*/
static u64 *kvmppc_page_address(struct page *page)
{
#if defined(HASHED_PAGE_VIRTUAL) || defined(WANT_PAGE_VIRTUAL)
#error TODO: fix to avoid page_address() here
#endif
return (u64 *) page_address(page);
}
/*
* Handles TCE requests for emulated devices.
* Puts guest TCE values to the table and expects user space to convert them.
* Called in both real and virtual modes.
* Cannot fail so kvmppc_tce_validate must be called before it.
*
* WARNING: This will be called in real-mode on HV KVM and virtual
* mode on PR KVM
*/
void kvmppc_tce_put(struct kvmppc_spapr_tce_table *stt,
unsigned long idx, unsigned long tce)
{
struct page *page;
u64 *tbl;
idx -= stt->offset;
page = stt->pages[idx / TCES_PER_PAGE];
tbl = kvmppc_page_address(page);
tbl[idx % TCES_PER_PAGE] = tce;
}
EXPORT_SYMBOL_GPL(kvmppc_tce_put);
long kvmppc_gpa_to_ua(struct kvm *kvm, unsigned long gpa,
unsigned long *ua, unsigned long **prmap)
{
unsigned long gfn = gpa >> PAGE_SHIFT;
struct kvm_memory_slot *memslot;
memslot = search_memslots(kvm_memslots(kvm), gfn);
if (!memslot)
return -EINVAL;
*ua = __gfn_to_hva_memslot(memslot, gfn) |
(gpa & ~(PAGE_MASK | TCE_PCI_READ | TCE_PCI_WRITE));
#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
if (prmap)
*prmap = &memslot->arch.rmap[gfn - memslot->base_gfn];
#endif
return 0;
}
EXPORT_SYMBOL_GPL(kvmppc_gpa_to_ua);
#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
static void kvmppc_rm_clear_tce(struct iommu_table *tbl, unsigned long entry)
{
unsigned long hpa = 0;
enum dma_data_direction dir = DMA_NONE;
iommu_tce_xchg_rm(tbl, entry, &hpa, &dir);
}
static long kvmppc_rm_tce_iommu_mapped_dec(struct kvm *kvm,
struct iommu_table *tbl, unsigned long entry)
{
struct mm_iommu_table_group_mem_t *mem = NULL;
const unsigned long pgsize = 1ULL << tbl->it_page_shift;
unsigned long *pua = IOMMU_TABLE_USERSPACE_ENTRY(tbl, entry);
if (!pua)
/* it_userspace allocation might be delayed */
return H_TOO_HARD;
pua = (void *) vmalloc_to_phys(pua);
if (WARN_ON_ONCE_RM(!pua))
return H_HARDWARE;
mem = mm_iommu_lookup_rm(kvm->mm, *pua, pgsize);
if (!mem)
return H_TOO_HARD;
mm_iommu_mapped_dec(mem);
*pua = 0;
return H_SUCCESS;
}
static long kvmppc_rm_tce_iommu_unmap(struct kvm *kvm,
struct iommu_table *tbl, unsigned long entry)
{
enum dma_data_direction dir = DMA_NONE;
unsigned long hpa = 0;
long ret;
if (iommu_tce_xchg_rm(tbl, entry, &hpa, &dir))
/*
* real mode xchg can fail if struct page crosses
* a page boundary
*/
return H_TOO_HARD;
if (dir == DMA_NONE)
return H_SUCCESS;
ret = kvmppc_rm_tce_iommu_mapped_dec(kvm, tbl, entry);
if (ret)
iommu_tce_xchg_rm(tbl, entry, &hpa, &dir);
return ret;
}
static long kvmppc_rm_tce_iommu_map(struct kvm *kvm, struct iommu_table *tbl,
unsigned long entry, unsigned long ua,
enum dma_data_direction dir)
{
long ret;
unsigned long hpa = 0;
unsigned long *pua = IOMMU_TABLE_USERSPACE_ENTRY(tbl, entry);
struct mm_iommu_table_group_mem_t *mem;
if (!pua)
/* it_userspace allocation might be delayed */
return H_TOO_HARD;
mem = mm_iommu_lookup_rm(kvm->mm, ua, 1ULL << tbl->it_page_shift);
if (!mem)
return H_TOO_HARD;
if (WARN_ON_ONCE_RM(mm_iommu_ua_to_hpa_rm(mem, ua, &hpa)))
return H_HARDWARE;
pua = (void *) vmalloc_to_phys(pua);
if (WARN_ON_ONCE_RM(!pua))
return H_HARDWARE;
if (WARN_ON_ONCE_RM(mm_iommu_mapped_inc(mem)))
return H_CLOSED;
ret = iommu_tce_xchg_rm(tbl, entry, &hpa, &dir);
if (ret) {
mm_iommu_mapped_dec(mem);
/*
* real mode xchg can fail if struct page crosses
* a page boundary
*/
return H_TOO_HARD;
}
if (dir != DMA_NONE)
kvmppc_rm_tce_iommu_mapped_dec(kvm, tbl, entry);
*pua = ua;
return 0;
}
long kvmppc_rm_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn,
unsigned long ioba, unsigned long tce)
{
struct kvmppc_spapr_tce_table *stt;
long ret;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
struct kvmppc_spapr_tce_iommu_table *stit;
unsigned long entry, ua = 0;
enum dma_data_direction dir;
/* udbg_printf("H_PUT_TCE(): liobn=0x%lx ioba=0x%lx, tce=0x%lx\n", */
/* liobn, ioba, tce); */
KVM: PPC: Book3S HV: Add radix checks in real-mode hypercall handlers POWER9 running a radix guest will take some hypervisor interrupts without going to real mode (turning off the MMU). This means that early hypercall handlers may now be called in virtual mode. Most of the handlers work just fine in both modes, but there are some that can crash the host if called in virtual mode, notably the TCE (IOMMU) hypercalls H_PUT_TCE, H_STUFF_TCE and H_PUT_TCE_INDIRECT. These already have both a real-mode and a virtual-mode version, so we arrange for the real-mode version to return H_TOO_HARD for radix guests, which will result in the virtual-mode version being called. The other hypercall which is sensitive to the MMU mode is H_RANDOM. It doesn't have a virtual-mode version, so this adds code to enable it to be called in either mode. An alternative solution was considered which would refuse to call any of the early hypercall handlers when doing a virtual-mode exit from a radix guest. However, the XICS-on-XIVE code depends on the XICS hypercalls being handled early even for virtual-mode exits, because the handlers need to be called before the XIVE vCPU state has been pulled off the hardware. Therefore that solution would have become quite invasive and complicated, and was rejected in favour of the simpler, though less elegant, solution presented here. Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Tested-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-05-10 14:39:41 +08:00
/* For radix, we might be in virtual mode, so punt */
if (kvm_is_radix(vcpu->kvm))
return H_TOO_HARD;
stt = kvmppc_find_table(vcpu->kvm, liobn);
if (!stt)
return H_TOO_HARD;
ret = kvmppc_ioba_validate(stt, ioba, 1);
if (ret != H_SUCCESS)
return ret;
ret = kvmppc_tce_validate(stt, tce);
if (ret != H_SUCCESS)
return ret;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
dir = iommu_tce_direction(tce);
if ((dir != DMA_NONE) && kvmppc_gpa_to_ua(vcpu->kvm,
tce & ~(TCE_PCI_READ | TCE_PCI_WRITE), &ua, NULL))
return H_PARAMETER;
entry = ioba >> stt->page_shift;
list_for_each_entry_lockless(stit, &stt->iommu_tables, next) {
if (dir == DMA_NONE)
ret = kvmppc_rm_tce_iommu_unmap(vcpu->kvm,
stit->tbl, entry);
else
ret = kvmppc_rm_tce_iommu_map(vcpu->kvm,
stit->tbl, entry, ua, dir);
if (ret == H_SUCCESS)
continue;
if (ret == H_TOO_HARD)
return ret;
WARN_ON_ONCE_RM(1);
kvmppc_rm_clear_tce(stit->tbl, entry);
}
kvmppc_tce_put(stt, entry, tce);
return H_SUCCESS;
}
static long kvmppc_rm_ua_to_hpa(struct kvm_vcpu *vcpu,
unsigned long ua, unsigned long *phpa)
{
pte_t *ptep, pte;
unsigned shift = 0;
ptep = __find_linux_pte_or_hugepte(vcpu->arch.pgdir, ua, NULL, &shift);
if (!ptep || !pte_present(*ptep))
return -ENXIO;
pte = *ptep;
if (!shift)
shift = PAGE_SHIFT;
/* Avoid handling anything potentially complicated in realmode */
if (shift > PAGE_SHIFT)
return -EAGAIN;
if (!pte_young(pte))
return -EAGAIN;
*phpa = (pte_pfn(pte) << PAGE_SHIFT) | (ua & ((1ULL << shift) - 1)) |
(ua & ~PAGE_MASK);
return 0;
}
long kvmppc_rm_h_put_tce_indirect(struct kvm_vcpu *vcpu,
unsigned long liobn, unsigned long ioba,
unsigned long tce_list, unsigned long npages)
{
struct kvmppc_spapr_tce_table *stt;
long i, ret = H_SUCCESS;
unsigned long tces, entry, ua = 0;
unsigned long *rmap = NULL;
bool prereg = false;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
struct kvmppc_spapr_tce_iommu_table *stit;
KVM: PPC: Book3S HV: Add radix checks in real-mode hypercall handlers POWER9 running a radix guest will take some hypervisor interrupts without going to real mode (turning off the MMU). This means that early hypercall handlers may now be called in virtual mode. Most of the handlers work just fine in both modes, but there are some that can crash the host if called in virtual mode, notably the TCE (IOMMU) hypercalls H_PUT_TCE, H_STUFF_TCE and H_PUT_TCE_INDIRECT. These already have both a real-mode and a virtual-mode version, so we arrange for the real-mode version to return H_TOO_HARD for radix guests, which will result in the virtual-mode version being called. The other hypercall which is sensitive to the MMU mode is H_RANDOM. It doesn't have a virtual-mode version, so this adds code to enable it to be called in either mode. An alternative solution was considered which would refuse to call any of the early hypercall handlers when doing a virtual-mode exit from a radix guest. However, the XICS-on-XIVE code depends on the XICS hypercalls being handled early even for virtual-mode exits, because the handlers need to be called before the XIVE vCPU state has been pulled off the hardware. Therefore that solution would have become quite invasive and complicated, and was rejected in favour of the simpler, though less elegant, solution presented here. Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Tested-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-05-10 14:39:41 +08:00
/* For radix, we might be in virtual mode, so punt */
if (kvm_is_radix(vcpu->kvm))
return H_TOO_HARD;
stt = kvmppc_find_table(vcpu->kvm, liobn);
if (!stt)
return H_TOO_HARD;
entry = ioba >> stt->page_shift;
/*
* The spec says that the maximum size of the list is 512 TCEs
* so the whole table addressed resides in 4K page
*/
if (npages > 512)
return H_PARAMETER;
if (tce_list & (SZ_4K - 1))
return H_PARAMETER;
ret = kvmppc_ioba_validate(stt, ioba, npages);
if (ret != H_SUCCESS)
return ret;
if (mm_iommu_preregistered(vcpu->kvm->mm)) {
/*
* We get here if guest memory was pre-registered which
* is normally VFIO case and gpa->hpa translation does not
* depend on hpt.
*/
struct mm_iommu_table_group_mem_t *mem;
if (kvmppc_gpa_to_ua(vcpu->kvm, tce_list, &ua, NULL))
return H_TOO_HARD;
mem = mm_iommu_lookup_rm(vcpu->kvm->mm, ua, IOMMU_PAGE_SIZE_4K);
if (mem)
prereg = mm_iommu_ua_to_hpa_rm(mem, ua, &tces) == 0;
}
if (!prereg) {
/*
* This is usually a case of a guest with emulated devices only
* when TCE list is not in preregistered memory.
* We do not require memory to be preregistered in this case
* so lock rmap and do __find_linux_pte_or_hugepte().
*/
if (kvmppc_gpa_to_ua(vcpu->kvm, tce_list, &ua, &rmap))
return H_TOO_HARD;
rmap = (void *) vmalloc_to_phys(rmap);
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
if (WARN_ON_ONCE_RM(!rmap))
return H_HARDWARE;
/*
* Synchronize with the MMU notifier callbacks in
* book3s_64_mmu_hv.c (kvm_unmap_hva_hv etc.).
* While we have the rmap lock, code running on other CPUs
* cannot finish unmapping the host real page that backs
* this guest real page, so we are OK to access the host
* real page.
*/
lock_rmap(rmap);
if (kvmppc_rm_ua_to_hpa(vcpu, ua, &tces)) {
ret = H_TOO_HARD;
goto unlock_exit;
}
}
for (i = 0; i < npages; ++i) {
unsigned long tce = be64_to_cpu(((u64 *)tces)[i]);
ret = kvmppc_tce_validate(stt, tce);
if (ret != H_SUCCESS)
goto unlock_exit;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
ua = 0;
if (kvmppc_gpa_to_ua(vcpu->kvm,
tce & ~(TCE_PCI_READ | TCE_PCI_WRITE),
&ua, NULL))
return H_PARAMETER;
list_for_each_entry_lockless(stit, &stt->iommu_tables, next) {
ret = kvmppc_rm_tce_iommu_map(vcpu->kvm,
stit->tbl, entry + i, ua,
iommu_tce_direction(tce));
if (ret == H_SUCCESS)
continue;
if (ret == H_TOO_HARD)
goto unlock_exit;
WARN_ON_ONCE_RM(1);
kvmppc_rm_clear_tce(stit->tbl, entry);
}
kvmppc_tce_put(stt, entry + i, tce);
}
unlock_exit:
if (rmap)
unlock_rmap(rmap);
return ret;
}
long kvmppc_rm_h_stuff_tce(struct kvm_vcpu *vcpu,
unsigned long liobn, unsigned long ioba,
unsigned long tce_value, unsigned long npages)
{
struct kvmppc_spapr_tce_table *stt;
long i, ret;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
struct kvmppc_spapr_tce_iommu_table *stit;
KVM: PPC: Book3S HV: Add radix checks in real-mode hypercall handlers POWER9 running a radix guest will take some hypervisor interrupts without going to real mode (turning off the MMU). This means that early hypercall handlers may now be called in virtual mode. Most of the handlers work just fine in both modes, but there are some that can crash the host if called in virtual mode, notably the TCE (IOMMU) hypercalls H_PUT_TCE, H_STUFF_TCE and H_PUT_TCE_INDIRECT. These already have both a real-mode and a virtual-mode version, so we arrange for the real-mode version to return H_TOO_HARD for radix guests, which will result in the virtual-mode version being called. The other hypercall which is sensitive to the MMU mode is H_RANDOM. It doesn't have a virtual-mode version, so this adds code to enable it to be called in either mode. An alternative solution was considered which would refuse to call any of the early hypercall handlers when doing a virtual-mode exit from a radix guest. However, the XICS-on-XIVE code depends on the XICS hypercalls being handled early even for virtual-mode exits, because the handlers need to be called before the XIVE vCPU state has been pulled off the hardware. Therefore that solution would have become quite invasive and complicated, and was rejected in favour of the simpler, though less elegant, solution presented here. Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Tested-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-05-10 14:39:41 +08:00
/* For radix, we might be in virtual mode, so punt */
if (kvm_is_radix(vcpu->kvm))
return H_TOO_HARD;
stt = kvmppc_find_table(vcpu->kvm, liobn);
if (!stt)
return H_TOO_HARD;
ret = kvmppc_ioba_validate(stt, ioba, npages);
if (ret != H_SUCCESS)
return ret;
/* Check permission bits only to allow userspace poison TCE for debug */
if (tce_value & (TCE_PCI_WRITE | TCE_PCI_READ))
return H_PARAMETER;
KVM: PPC: VFIO: Add in-kernel acceleration for VFIO This allows the host kernel to handle H_PUT_TCE, H_PUT_TCE_INDIRECT and H_STUFF_TCE requests targeted an IOMMU TCE table used for VFIO without passing them to user space which saves time on switching to user space and back. This adds H_PUT_TCE/H_PUT_TCE_INDIRECT/H_STUFF_TCE handlers to KVM. KVM tries to handle a TCE request in the real mode, if failed it passes the request to the virtual mode to complete the operation. If it a virtual mode handler fails, the request is passed to the user space; this is not expected to happen though. To avoid dealing with page use counters (which is tricky in real mode), this only accelerates SPAPR TCE IOMMU v2 clients which are required to pre-register the userspace memory. The very first TCE request will be handled in the VFIO SPAPR TCE driver anyway as the userspace view of the TCE table (iommu_table::it_userspace) is not allocated till the very first mapping happens and we cannot call vmalloc in real mode. If we fail to update a hardware IOMMU table unexpected reason, we just clear it and move on as there is nothing really we can do about it - for example, if we hot plug a VFIO device to a guest, existing TCE tables will be mirrored automatically to the hardware and there is no interface to report to the guest about possible failures. This adds new attribute - KVM_DEV_VFIO_GROUP_SET_SPAPR_TCE - to the VFIO KVM device. It takes a VFIO group fd and SPAPR TCE table fd and associates a physical IOMMU table with the SPAPR TCE table (which is a guest view of the hardware IOMMU table). The iommu_table object is cached and referenced so we do not have to look up for it in real mode. This does not implement the UNSET counterpart as there is no use for it - once the acceleration is enabled, the existing userspace won't disable it unless a VFIO container is destroyed; this adds necessary cleanup to the KVM_DEV_VFIO_GROUP_DEL handler. This advertises the new KVM_CAP_SPAPR_TCE_VFIO capability to the user space. This adds real mode version of WARN_ON_ONCE() as the generic version causes problems with rcu_sched. Since we testing what vmalloc_to_phys() returns in the code, this also adds a check for already existing vmalloc_to_phys() call in kvmppc_rm_h_put_tce_indirect(). This finally makes use of vfio_external_user_iommu_id() which was introduced quite some time ago and was considered for removal. Tests show that this patch increases transmission speed from 220MB/s to 750..1020MB/s on 10Gb network (Chelsea CXGB3 10Gb ethernet card). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-03-22 12:21:56 +08:00
list_for_each_entry_lockless(stit, &stt->iommu_tables, next) {
unsigned long entry = ioba >> stit->tbl->it_page_shift;
for (i = 0; i < npages; ++i) {
ret = kvmppc_rm_tce_iommu_unmap(vcpu->kvm,
stit->tbl, entry + i);
if (ret == H_SUCCESS)
continue;
if (ret == H_TOO_HARD)
return ret;
WARN_ON_ONCE_RM(1);
kvmppc_rm_clear_tce(stit->tbl, entry);
}
}
for (i = 0; i < npages; ++i, ioba += (1ULL << stt->page_shift))
kvmppc_tce_put(stt, ioba >> stt->page_shift, tce_value);
return H_SUCCESS;
}
KVM: PPC: Book3S HV: Add radix checks in real-mode hypercall handlers POWER9 running a radix guest will take some hypervisor interrupts without going to real mode (turning off the MMU). This means that early hypercall handlers may now be called in virtual mode. Most of the handlers work just fine in both modes, but there are some that can crash the host if called in virtual mode, notably the TCE (IOMMU) hypercalls H_PUT_TCE, H_STUFF_TCE and H_PUT_TCE_INDIRECT. These already have both a real-mode and a virtual-mode version, so we arrange for the real-mode version to return H_TOO_HARD for radix guests, which will result in the virtual-mode version being called. The other hypercall which is sensitive to the MMU mode is H_RANDOM. It doesn't have a virtual-mode version, so this adds code to enable it to be called in either mode. An alternative solution was considered which would refuse to call any of the early hypercall handlers when doing a virtual-mode exit from a radix guest. However, the XICS-on-XIVE code depends on the XICS hypercalls being handled early even for virtual-mode exits, because the handlers need to be called before the XIVE vCPU state has been pulled off the hardware. Therefore that solution would have become quite invasive and complicated, and was rejected in favour of the simpler, though less elegant, solution presented here. Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Tested-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-05-10 14:39:41 +08:00
/* This can be called in either virtual mode or real mode */
long kvmppc_h_get_tce(struct kvm_vcpu *vcpu, unsigned long liobn,
unsigned long ioba)
{
struct kvmppc_spapr_tce_table *stt;
long ret;
unsigned long idx;
struct page *page;
u64 *tbl;
stt = kvmppc_find_table(vcpu->kvm, liobn);
if (!stt)
return H_TOO_HARD;
ret = kvmppc_ioba_validate(stt, ioba, 1);
if (ret != H_SUCCESS)
return ret;
idx = (ioba >> stt->page_shift) - stt->offset;
page = stt->pages[idx / TCES_PER_PAGE];
tbl = (u64 *)page_address(page);
vcpu->arch.gpr[4] = tbl[idx % TCES_PER_PAGE];
return H_SUCCESS;
}
EXPORT_SYMBOL_GPL(kvmppc_h_get_tce);
#endif /* KVM_BOOK3S_HV_POSSIBLE */