diff options
author | Le Tan <tamlokveer@gmail.com> | 2014-08-16 13:55:38 +0800 |
---|---|---|
committer | Michael S. Tsirkin <mst@redhat.com> | 2014-08-28 23:10:22 +0200 |
commit | 1da12ec4c81330eb1ed1f29a78854bea87a23595 (patch) | |
tree | 25ff4496e3366f3892b6bb96d4b1d6a4abce5fe0 /hw/i386 | |
parent | 8d7b8cb9c22b7aa809050800a62e89f869054b35 (diff) |
intel-iommu: introduce Intel IOMMU (VT-d) emulation
Add support for emulating Intel IOMMU according to the VT-d specification for
the q35 chipset machine. Implement the logics for DMAR (DMA remapping) without
PASID support. The emulation supports register-based invalidation and primary
fault logging.
Signed-off-by: Le Tan <tamlokveer@gmail.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Diffstat (limited to 'hw/i386')
-rw-r--r-- | hw/i386/Makefile.objs | 1 | ||||
-rw-r--r-- | hw/i386/intel_iommu.c | 1257 | ||||
-rw-r--r-- | hw/i386/intel_iommu_internal.h | 333 |
3 files changed, 1591 insertions, 0 deletions
diff --git a/hw/i386/Makefile.objs b/hw/i386/Makefile.objs index 3688cf8243..9d419addb4 100644 --- a/hw/i386/Makefile.objs +++ b/hw/i386/Makefile.objs @@ -2,6 +2,7 @@ obj-$(CONFIG_KVM) += kvm/ obj-y += multiboot.o smbios.o obj-y += pc.o pc_piix.o pc_q35.o obj-y += pc_sysfw.o +obj-y += intel_iommu.o obj-$(CONFIG_XEN) += ../xenpv/ xen/ obj-y += kvmvapic.o diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c new file mode 100644 index 0000000000..8e67e04d19 --- /dev/null +++ b/hw/i386/intel_iommu.c @@ -0,0 +1,1257 @@ +/* + * QEMU emulation of an Intel IOMMU (VT-d) + * (DMA Remapping device) + * + * Copyright (C) 2013 Knut Omang, Oracle <knut.omang@oracle.com> + * Copyright (C) 2014 Le Tan, <tamlokveer@gmail.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + + * 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, see <http://www.gnu.org/licenses/>. + */ + +#include "hw/sysbus.h" +#include "exec/address-spaces.h" +#include "intel_iommu_internal.h" + +/*#define DEBUG_INTEL_IOMMU*/ +#ifdef DEBUG_INTEL_IOMMU +enum { + DEBUG_GENERAL, DEBUG_CSR, DEBUG_INV, DEBUG_MMU, DEBUG_FLOG, +}; +#define VTD_DBGBIT(x) (1 << DEBUG_##x) +static int vtd_dbgflags = VTD_DBGBIT(GENERAL) | VTD_DBGBIT(CSR); + +#define VTD_DPRINTF(what, fmt, ...) do { \ + if (vtd_dbgflags & VTD_DBGBIT(what)) { \ + fprintf(stderr, "(vtd)%s: " fmt "\n", __func__, \ + ## __VA_ARGS__); } \ + } while (0) +#else +#define VTD_DPRINTF(what, fmt, ...) do {} while (0) +#endif + +static void vtd_define_quad(IntelIOMMUState *s, hwaddr addr, uint64_t val, + uint64_t wmask, uint64_t w1cmask) +{ + stq_le_p(&s->csr[addr], val); + stq_le_p(&s->wmask[addr], wmask); + stq_le_p(&s->w1cmask[addr], w1cmask); +} + +static void vtd_define_quad_wo(IntelIOMMUState *s, hwaddr addr, uint64_t mask) +{ + stq_le_p(&s->womask[addr], mask); +} + +static void vtd_define_long(IntelIOMMUState *s, hwaddr addr, uint32_t val, + uint32_t wmask, uint32_t w1cmask) +{ + stl_le_p(&s->csr[addr], val); + stl_le_p(&s->wmask[addr], wmask); + stl_le_p(&s->w1cmask[addr], w1cmask); +} + +static void vtd_define_long_wo(IntelIOMMUState *s, hwaddr addr, uint32_t mask) +{ + stl_le_p(&s->womask[addr], mask); +} + +/* "External" get/set operations */ +static void vtd_set_quad(IntelIOMMUState *s, hwaddr addr, uint64_t val) +{ + uint64_t oldval = ldq_le_p(&s->csr[addr]); + uint64_t wmask = ldq_le_p(&s->wmask[addr]); + uint64_t w1cmask = ldq_le_p(&s->w1cmask[addr]); + stq_le_p(&s->csr[addr], + ((oldval & ~wmask) | (val & wmask)) & ~(w1cmask & val)); +} + +static void vtd_set_long(IntelIOMMUState *s, hwaddr addr, uint32_t val) +{ + uint32_t oldval = ldl_le_p(&s->csr[addr]); + uint32_t wmask = ldl_le_p(&s->wmask[addr]); + uint32_t w1cmask = ldl_le_p(&s->w1cmask[addr]); + stl_le_p(&s->csr[addr], + ((oldval & ~wmask) | (val & wmask)) & ~(w1cmask & val)); +} + +static uint64_t vtd_get_quad(IntelIOMMUState *s, hwaddr addr) +{ + uint64_t val = ldq_le_p(&s->csr[addr]); + uint64_t womask = ldq_le_p(&s->womask[addr]); + return val & ~womask; +} + +static uint32_t vtd_get_long(IntelIOMMUState *s, hwaddr addr) +{ + uint32_t val = ldl_le_p(&s->csr[addr]); + uint32_t womask = ldl_le_p(&s->womask[addr]); + return val & ~womask; +} + +/* "Internal" get/set operations */ +static uint64_t vtd_get_quad_raw(IntelIOMMUState *s, hwaddr addr) +{ + return ldq_le_p(&s->csr[addr]); +} + +static uint32_t vtd_get_long_raw(IntelIOMMUState *s, hwaddr addr) +{ + return ldl_le_p(&s->csr[addr]); +} + +static void vtd_set_quad_raw(IntelIOMMUState *s, hwaddr addr, uint64_t val) +{ + stq_le_p(&s->csr[addr], val); +} + +static uint32_t vtd_set_clear_mask_long(IntelIOMMUState *s, hwaddr addr, + uint32_t clear, uint32_t mask) +{ + uint32_t new_val = (ldl_le_p(&s->csr[addr]) & ~clear) | mask; + stl_le_p(&s->csr[addr], new_val); + return new_val; +} + +static uint64_t vtd_set_clear_mask_quad(IntelIOMMUState *s, hwaddr addr, + uint64_t clear, uint64_t mask) +{ + uint64_t new_val = (ldq_le_p(&s->csr[addr]) & ~clear) | mask; + stq_le_p(&s->csr[addr], new_val); + return new_val; +} + +/* Given the reg addr of both the message data and address, generate an + * interrupt via MSI. + */ +static void vtd_generate_interrupt(IntelIOMMUState *s, hwaddr mesg_addr_reg, + hwaddr mesg_data_reg) +{ + hwaddr addr; + uint32_t data; + + assert(mesg_data_reg < DMAR_REG_SIZE); + assert(mesg_addr_reg < DMAR_REG_SIZE); + + addr = vtd_get_long_raw(s, mesg_addr_reg); + data = vtd_get_long_raw(s, mesg_data_reg); + + VTD_DPRINTF(FLOG, "msi: addr 0x%"PRIx64 " data 0x%"PRIx32, addr, data); + stl_le_phys(&address_space_memory, addr, data); +} + +/* Generate a fault event to software via MSI if conditions are met. + * Notice that the value of FSTS_REG being passed to it should be the one + * before any update. + */ +static void vtd_generate_fault_event(IntelIOMMUState *s, uint32_t pre_fsts) +{ + if (pre_fsts & VTD_FSTS_PPF || pre_fsts & VTD_FSTS_PFO || + pre_fsts & VTD_FSTS_IQE) { + VTD_DPRINTF(FLOG, "there are previous interrupt conditions " + "to be serviced by software, fault event is not generated " + "(FSTS_REG 0x%"PRIx32 ")", pre_fsts); + return; + } + vtd_set_clear_mask_long(s, DMAR_FECTL_REG, 0, VTD_FECTL_IP); + if (vtd_get_long_raw(s, DMAR_FECTL_REG) & VTD_FECTL_IM) { + VTD_DPRINTF(FLOG, "Interrupt Mask set, fault event is not generated"); + } else { + vtd_generate_interrupt(s, DMAR_FEADDR_REG, DMAR_FEDATA_REG); + vtd_set_clear_mask_long(s, DMAR_FECTL_REG, VTD_FECTL_IP, 0); + } +} + +/* Check if the Fault (F) field of the Fault Recording Register referenced by + * @index is Set. + */ +static bool vtd_is_frcd_set(IntelIOMMUState *s, uint16_t index) +{ + /* Each reg is 128-bit */ + hwaddr addr = DMAR_FRCD_REG_OFFSET + (((uint64_t)index) << 4); + addr += 8; /* Access the high 64-bit half */ + + assert(index < DMAR_FRCD_REG_NR); + + return vtd_get_quad_raw(s, addr) & VTD_FRCD_F; +} + +/* Update the PPF field of Fault Status Register. + * Should be called whenever change the F field of any fault recording + * registers. + */ +static void vtd_update_fsts_ppf(IntelIOMMUState *s) +{ + uint32_t i; + uint32_t ppf_mask = 0; + + for (i = 0; i < DMAR_FRCD_REG_NR; i++) { + if (vtd_is_frcd_set(s, i)) { + ppf_mask = VTD_FSTS_PPF; + break; + } + } + vtd_set_clear_mask_long(s, DMAR_FSTS_REG, VTD_FSTS_PPF, ppf_mask); + VTD_DPRINTF(FLOG, "set PPF of FSTS_REG to %d", ppf_mask ? 1 : 0); +} + +static void vtd_set_frcd_and_update_ppf(IntelIOMMUState *s, uint16_t index) +{ + /* Each reg is 128-bit */ + hwaddr addr = DMAR_FRCD_REG_OFFSET + (((uint64_t)index) << 4); + addr += 8; /* Access the high 64-bit half */ + + assert(index < DMAR_FRCD_REG_NR); + + vtd_set_clear_mask_quad(s, addr, 0, VTD_FRCD_F); + vtd_update_fsts_ppf(s); +} + +/* Must not update F field now, should be done later */ +static void vtd_record_frcd(IntelIOMMUState *s, uint16_t index, + uint16_t source_id, hwaddr addr, + VTDFaultReason fault, bool is_write) +{ + uint64_t hi = 0, lo; + hwaddr frcd_reg_addr = DMAR_FRCD_REG_OFFSET + (((uint64_t)index) << 4); + + assert(index < DMAR_FRCD_REG_NR); + + lo = VTD_FRCD_FI(addr); + hi = VTD_FRCD_SID(source_id) | VTD_FRCD_FR(fault); + if (!is_write) { + hi |= VTD_FRCD_T; + } + vtd_set_quad_raw(s, frcd_reg_addr, lo); + vtd_set_quad_raw(s, frcd_reg_addr + 8, hi); + VTD_DPRINTF(FLOG, "record to FRCD_REG #%"PRIu16 ": hi 0x%"PRIx64 + ", lo 0x%"PRIx64, index, hi, lo); +} + +/* Try to collapse multiple pending faults from the same requester */ +static bool vtd_try_collapse_fault(IntelIOMMUState *s, uint16_t source_id) +{ + uint32_t i; + uint64_t frcd_reg; + hwaddr addr = DMAR_FRCD_REG_OFFSET + 8; /* The high 64-bit half */ + + for (i = 0; i < DMAR_FRCD_REG_NR; i++) { + frcd_reg = vtd_get_quad_raw(s, addr); + VTD_DPRINTF(FLOG, "frcd_reg #%d 0x%"PRIx64, i, frcd_reg); + if ((frcd_reg & VTD_FRCD_F) && + ((frcd_reg & VTD_FRCD_SID_MASK) == source_id)) { + return true; + } + addr += 16; /* 128-bit for each */ + } + return false; +} + +/* Log and report an DMAR (address translation) fault to software */ +static void vtd_report_dmar_fault(IntelIOMMUState *s, uint16_t source_id, + hwaddr addr, VTDFaultReason fault, + bool is_write) +{ + uint32_t fsts_reg = vtd_get_long_raw(s, DMAR_FSTS_REG); + + assert(fault < VTD_FR_MAX); + + if (fault == VTD_FR_RESERVED_ERR) { + /* This is not a normal fault reason case. Drop it. */ + return; + } + VTD_DPRINTF(FLOG, "sid 0x%"PRIx16 ", fault %d, addr 0x%"PRIx64 + ", is_write %d", source_id, fault, addr, is_write); + if (fsts_reg & VTD_FSTS_PFO) { + VTD_DPRINTF(FLOG, "new fault is not recorded due to " + "Primary Fault Overflow"); + return; + } + if (vtd_try_collapse_fault(s, source_id)) { + VTD_DPRINTF(FLOG, "new fault is not recorded due to " + "compression of faults"); + return; + } + if (vtd_is_frcd_set(s, s->next_frcd_reg)) { + VTD_DPRINTF(FLOG, "Primary Fault Overflow and " + "new fault is not recorded, set PFO field"); + vtd_set_clear_mask_long(s, DMAR_FSTS_REG, 0, VTD_FSTS_PFO); + return; + } + + vtd_record_frcd(s, s->next_frcd_reg, source_id, addr, fault, is_write); + + if (fsts_reg & VTD_FSTS_PPF) { + VTD_DPRINTF(FLOG, "there are pending faults already, " + "fault event is not generated"); + vtd_set_frcd_and_update_ppf(s, s->next_frcd_reg); + s->next_frcd_reg++; + if (s->next_frcd_reg == DMAR_FRCD_REG_NR) { + s->next_frcd_reg = 0; + } + } else { + vtd_set_clear_mask_long(s, DMAR_FSTS_REG, VTD_FSTS_FRI_MASK, + VTD_FSTS_FRI(s->next_frcd_reg)); + vtd_set_frcd_and_update_ppf(s, s->next_frcd_reg); /* Will set PPF */ + s->next_frcd_reg++; + if (s->next_frcd_reg == DMAR_FRCD_REG_NR) { + s->next_frcd_reg = 0; + } + /* This case actually cause the PPF to be Set. + * So generate fault event (interrupt). + */ + vtd_generate_fault_event(s, fsts_reg); + } +} + +static inline bool vtd_root_entry_present(VTDRootEntry *root) +{ + return root->val & VTD_ROOT_ENTRY_P; +} + +static int vtd_get_root_entry(IntelIOMMUState *s, uint8_t index, + VTDRootEntry *re) +{ + dma_addr_t addr; + + addr = s->root + index * sizeof(*re); + if (dma_memory_read(&address_space_memory, addr, re, sizeof(*re))) { + VTD_DPRINTF(GENERAL, "error: fail to access root-entry at 0x%"PRIx64 + " + %"PRIu8, s->root, index); + re->val = 0; + return -VTD_FR_ROOT_TABLE_INV; + } + re->val = le64_to_cpu(re->val); + return 0; +} + +static inline bool vtd_context_entry_present(VTDContextEntry *context) +{ + return context->lo & VTD_CONTEXT_ENTRY_P; +} + +static int vtd_get_context_entry_from_root(VTDRootEntry *root, uint8_t index, + VTDContextEntry *ce) +{ + dma_addr_t addr; + + if (!vtd_root_entry_present(root)) { + VTD_DPRINTF(GENERAL, "error: root-entry is not present"); + return -VTD_FR_ROOT_ENTRY_P; + } + addr = (root->val & VTD_ROOT_ENTRY_CTP) + index * sizeof(*ce); + if (dma_memory_read(&address_space_memory, addr, ce, sizeof(*ce))) { + VTD_DPRINTF(GENERAL, "error: fail to access context-entry at 0x%"PRIx64 + " + %"PRIu8, + (uint64_t)(root->val & VTD_ROOT_ENTRY_CTP), index); + return -VTD_FR_CONTEXT_TABLE_INV; + } + ce->lo = le64_to_cpu(ce->lo); + ce->hi = le64_to_cpu(ce->hi); + return 0; +} + +static inline dma_addr_t vtd_get_slpt_base_from_context(VTDContextEntry *ce) +{ + return ce->lo & VTD_CONTEXT_ENTRY_SLPTPTR; +} + +/* The shift of an addr for a certain level of paging structure */ +static inline uint32_t vtd_slpt_level_shift(uint32_t level) +{ + return VTD_PAGE_SHIFT_4K + (level - 1) * VTD_SL_LEVEL_BITS; +} + +static inline uint64_t vtd_get_slpte_addr(uint64_t slpte) +{ + return slpte & VTD_SL_PT_BASE_ADDR_MASK; +} + +/* Whether the pte indicates the address of the page frame */ +static inline bool vtd_is_last_slpte(uint64_t slpte, uint32_t level) +{ + return level == VTD_SL_PT_LEVEL || (slpte & VTD_SL_PT_PAGE_SIZE_MASK); +} + +/* Get the content of a spte located in @base_addr[@index] */ +static uint64_t vtd_get_slpte(dma_addr_t base_addr, uint32_t index) +{ + uint64_t slpte; + + assert(index < VTD_SL_PT_ENTRY_NR); + + if (dma_memory_read(&address_space_memory, + base_addr + index * sizeof(slpte), &slpte, + sizeof(slpte))) { + slpte = (uint64_t)-1; + return slpte; + } + slpte = le64_to_cpu(slpte); + return slpte; +} + +/* Given a gpa and the level of paging structure, return the offset of current + * level. + */ +static inline uint32_t vtd_gpa_level_offset(uint64_t gpa, uint32_t level) +{ + return (gpa >> vtd_slpt_level_shift(level)) & + ((1ULL << VTD_SL_LEVEL_BITS) - 1); +} + +/* Check Capability Register to see if the @level of page-table is supported */ +static inline bool vtd_is_level_supported(IntelIOMMUState *s, uint32_t level) +{ + return VTD_CAP_SAGAW_MASK & s->cap & + (1ULL << (level - 2 + VTD_CAP_SAGAW_SHIFT)); +} + +/* Get the page-table level that hardware should use for the second-level + * page-table walk from the Address Width field of context-entry. + */ +static inline uint32_t vtd_get_level_from_context_entry(VTDContextEntry *ce) +{ + return 2 + (ce->hi & VTD_CONTEXT_ENTRY_AW); +} + +static inline uint32_t vtd_get_agaw_from_context_entry(VTDContextEntry *ce) +{ + return 30 + (ce->hi & VTD_CONTEXT_ENTRY_AW) * 9; +} + +static const uint64_t vtd_paging_entry_rsvd_field[] = { + [0] = ~0ULL, + /* For not large page */ + [1] = 0x800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [2] = 0x800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [3] = 0x800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [4] = 0x880ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + /* For large page */ + [5] = 0x800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [6] = 0x1ff800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [7] = 0x3ffff800ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), + [8] = 0x880ULL | ~(VTD_HAW_MASK | VTD_SL_IGN_COM), +}; + +static bool vtd_slpte_nonzero_rsvd(uint64_t slpte, uint32_t level) +{ + if (slpte & VTD_SL_PT_PAGE_SIZE_MASK) { + /* Maybe large page */ + return slpte & vtd_paging_entry_rsvd_field[level + 4]; + } else { + return slpte & vtd_paging_entry_rsvd_field[level]; + } +} + +/* Given the @gpa, get relevant @slptep. @slpte_level will be the last level + * of the translation, can be used for deciding the size of large page. + */ +static int vtd_gpa_to_slpte(VTDContextEntry *ce, uint64_t gpa, bool is_write, + uint64_t *slptep, uint32_t *slpte_level, + bool *reads, bool *writes) +{ + dma_addr_t addr = vtd_get_slpt_base_from_context(ce); + uint32_t level = vtd_get_level_from_context_entry(ce); + uint32_t offset; + uint64_t slpte; + uint32_t ce_agaw = vtd_get_agaw_from_context_entry(ce); + uint64_t access_right_check; + + /* Check if @gpa is above 2^X-1, where X is the minimum of MGAW in CAP_REG + * and AW in context-entry. + */ + if (gpa & ~((1ULL << MIN(ce_agaw, VTD_MGAW)) - 1)) { + VTD_DPRINTF(GENERAL, "error: gpa 0x%"PRIx64 " exceeds limits", gpa); + return -VTD_FR_ADDR_BEYOND_MGAW; + } + + /* FIXME: what is the Atomics request here? */ + access_right_check = is_write ? VTD_SL_W : VTD_SL_R; + + while (true) { + offset = vtd_gpa_level_offset(gpa, level); + slpte = vtd_get_slpte(addr, offset); + + if (slpte == (uint64_t)-1) { + VTD_DPRINTF(GENERAL, "error: fail to access second-level paging " + "entry at level %"PRIu32 " for gpa 0x%"PRIx64, + level, gpa); + if (level == vtd_get_level_from_context_entry(ce)) { + /* Invalid programming of context-entry */ + return -VTD_FR_CONTEXT_ENTRY_INV; + } else { + return -VTD_FR_PAGING_ENTRY_INV; + } + } + *reads = (*reads) && (slpte & VTD_SL_R); + *writes = (*writes) && (slpte & VTD_SL_W); + if (!(slpte & access_right_check)) { + VTD_DPRINTF(GENERAL, "error: lack of %s permission for " + "gpa 0x%"PRIx64 " slpte 0x%"PRIx64, + (is_write ? "write" : "read"), gpa, slpte); + return is_write ? -VTD_FR_WRITE : -VTD_FR_READ; + } + if (vtd_slpte_nonzero_rsvd(slpte, level)) { + VTD_DPRINTF(GENERAL, "error: non-zero reserved field in second " + "level paging entry level %"PRIu32 " slpte 0x%"PRIx64, + level, slpte); + return -VTD_FR_PAGING_ENTRY_RSVD; + } + + if (vtd_is_last_slpte(slpte, level)) { + *slptep = slpte; + *slpte_level = level; + return 0; + } + addr = vtd_get_slpte_addr(slpte); + level--; + } +} + +/* Map a device to its corresponding domain (context-entry) */ +static int vtd_dev_to_context_entry(IntelIOMMUState *s, uint8_t bus_num, + uint8_t devfn, VTDContextEntry *ce) +{ + VTDRootEntry re; + int ret_fr; + + ret_fr = vtd_get_root_entry(s, bus_num, &re); + if (ret_fr) { + return ret_fr; + } + + if (!vtd_root_entry_present(&re)) { + VTD_DPRINTF(GENERAL, "error: root-entry #%"PRIu8 " is not present", + bus_num); + return -VTD_FR_ROOT_ENTRY_P; + } else if (re.rsvd || (re.val & VTD_ROOT_ENTRY_RSVD)) { + VTD_DPRINTF(GENERAL, "error: non-zero reserved field in root-entry " + "hi 0x%"PRIx64 " lo 0x%"PRIx64, re.rsvd, re.val); + return -VTD_FR_ROOT_ENTRY_RSVD; + } + + ret_fr = vtd_get_context_entry_from_root(&re, devfn, ce); + if (ret_fr) { + return ret_fr; + } + + if (!vtd_context_entry_present(ce)) { + VTD_DPRINTF(GENERAL, + "error: context-entry #%"PRIu8 "(bus #%"PRIu8 ") " + "is not present", devfn, bus_num); + return -VTD_FR_CONTEXT_ENTRY_P; + } else if ((ce->hi & VTD_CONTEXT_ENTRY_RSVD_HI) || + (ce->lo & VTD_CONTEXT_ENTRY_RSVD_LO)) { + VTD_DPRINTF(GENERAL, + "error: non-zero reserved field in context-entry " + "hi 0x%"PRIx64 " lo 0x%"PRIx64, ce->hi, ce->lo); + return -VTD_FR_CONTEXT_ENTRY_RSVD; + } + /* Check if the programming of context-entry is valid */ + if (!vtd_is_level_supported(s, vtd_get_level_from_context_entry(ce))) { + VTD_DPRINTF(GENERAL, "error: unsupported Address Width value in " + "context-entry hi 0x%"PRIx64 " lo 0x%"PRIx64, + ce->hi, ce->lo); + return -VTD_FR_CONTEXT_ENTRY_INV; + } else if (ce->lo & VTD_CONTEXT_ENTRY_TT) { + VTD_DPRINTF(GENERAL, "error: unsupported Translation Type in " + "context-entry hi 0x%"PRIx64 " lo 0x%"PRIx64, + ce->hi, ce->lo); + return -VTD_FR_CONTEXT_ENTRY_INV; + } + return 0; +} + +static inline uint16_t vtd_make_source_id(uint8_t bus_num, uint8_t devfn) +{ + return ((bus_num & 0xffUL) << 8) | (devfn & 0xffUL); +} + +static const bool vtd_qualified_faults[] = { + [VTD_FR_RESERVED] = false, + [VTD_FR_ROOT_ENTRY_P] = false, + [VTD_FR_CONTEXT_ENTRY_P] = true, + [VTD_FR_CONTEXT_ENTRY_INV] = true, + [VTD_FR_ADDR_BEYOND_MGAW] = true, + [VTD_FR_WRITE] = true, + [VTD_FR_READ] = true, + [VTD_FR_PAGING_ENTRY_INV] = true, + [VTD_FR_ROOT_TABLE_INV] = false, + [VTD_FR_CONTEXT_TABLE_INV] = false, + [VTD_FR_ROOT_ENTRY_RSVD] = false, + [VTD_FR_PAGING_ENTRY_RSVD] = true, + [VTD_FR_CONTEXT_ENTRY_TT] = true, + [VTD_FR_RESERVED_ERR] = false, + [VTD_FR_MAX] = false, +}; + +/* To see if a fault condition is "qualified", which is reported to software + * only if the FPD field in the context-entry used to process the faulting + * request is 0. + */ +static inline bool vtd_is_qualified_fault(VTDFaultReason fault) +{ + return vtd_qualified_faults[fault]; +} + +static inline bool vtd_is_interrupt_addr(hwaddr addr) +{ + return VTD_INTERRUPT_ADDR_FIRST <= addr && addr <= VTD_INTERRUPT_ADDR_LAST; +} + +/* Map dev to context-entry then do a paging-structures walk to do a iommu + * translation. + * @bus_num: The bus number + * @devfn: The devfn, which is the combined of device and function number + * @is_write: The access is a write operation + * @entry: IOMMUTLBEntry that contain the addr to be translated and result + */ +static void vtd_do_iommu_translate(IntelIOMMUState *s, uint8_t bus_num, + uint8_t devfn, hwaddr addr, bool is_write, + IOMMUTLBEntry *entry) +{ + VTDContextEntry ce; + uint64_t slpte; + uint32_t level; + uint16_t source_id = vtd_make_source_id(bus_num, devfn); + int ret_fr; + bool is_fpd_set = false; + bool reads = true; + bool writes = true; + + /* Check if the request is in interrupt address range */ + if (vtd_is_interrupt_addr(addr)) { + if (is_write) { + /* FIXME: since we don't know the length of the access here, we + * treat Non-DWORD length write requests without PASID as + * interrupt requests, too. Withoud interrupt remapping support, + * we just use 1:1 mapping. + */ + VTD_DPRINTF(MMU, "write request to interrupt address " + "gpa 0x%"PRIx64, addr); + entry->iova = addr & VTD_PAGE_MASK_4K; + entry->translated_addr = addr & VTD_PAGE_MASK_4K; + entry->addr_mask = ~VTD_PAGE_MASK_4K; + entry->perm = IOMMU_WO; + return; + } else { + VTD_DPRINTF(GENERAL, "error: read request from interrupt address " + "gpa 0x%"PRIx64, addr); + vtd_report_dmar_fault(s, source_id, addr, VTD_FR_READ, is_write); + return; + } + } + + ret_fr = vtd_dev_to_context_entry(s, bus_num, devfn, &ce); + is_fpd_set = ce.lo & VTD_CONTEXT_ENTRY_FPD; + if (ret_fr) { + ret_fr = -ret_fr; + if (is_fpd_set && vtd_is_qualified_fault(ret_fr)) { + VTD_DPRINTF(FLOG, "fault processing is disabled for DMA requests " + "through this context-entry (with FPD Set)"); + } else { + vtd_report_dmar_fault(s, source_id, addr, ret_fr, is_write); + } + return; + } + + ret_fr = vtd_gpa_to_slpte(&ce, addr, is_write, &slpte, &level, + &reads, &writes); + if (ret_fr) { + ret_fr = -ret_fr; + if (is_fpd_set && vtd_is_qualified_fault(ret_fr)) { + VTD_DPRINTF(FLOG, "fault processing is disabled for DMA requests " + "through this context-entry (with FPD Set)"); + } else { + vtd_report_dmar_fault(s, source_id, addr, ret_fr, is_write); + } + return; + } + + entry->iova = addr & VTD_PAGE_MASK_4K; + entry->translated_addr = vtd_get_slpte_addr(slpte) & VTD_PAGE_MASK_4K; + entry->addr_mask = ~VTD_PAGE_MASK_4K; + entry->perm = (writes ? 2 : 0) + (reads ? 1 : 0); +} + +static void vtd_root_table_setup(IntelIOMMUState *s) +{ + s->root = vtd_get_quad_raw(s, DMAR_RTADDR_REG); + s->root_extended = s->root & VTD_RTADDR_RTT; + s->root &= VTD_RTADDR_ADDR_MASK; + + VTD_DPRINTF(CSR, "root_table addr 0x%"PRIx64 " %s", s->root, + (s->root_extended ? "(extended)" : "")); +} + +/* Context-cache invalidation + * Returns the Context Actual Invalidation Granularity. + * @val: the content of the CCMD_REG + */ +static uint64_t vtd_context_cache_invalidate(IntelIOMMUState *s, uint64_t val) +{ + uint64_t caig; + uint64_t type = val & VTD_CCMD_CIRG_MASK; + + switch (type) { + case VTD_CCMD_GLOBAL_INVL: + VTD_DPRINTF(INV, "Global invalidation request"); + caig = VTD_CCMD_GLOBAL_INVL_A; + break; + + case VTD_CCMD_DOMAIN_INVL: + VTD_DPRINTF(INV, "Domain-selective invalidation request"); + caig = VTD_CCMD_DOMAIN_INVL_A; + break; + + case VTD_CCMD_DEVICE_INVL: + VTD_DPRINTF(INV, "Domain-selective invalidation request"); + caig = VTD_CCMD_DEVICE_INVL_A; + break; + + default: + VTD_DPRINTF(GENERAL, + "error: wrong context-cache invalidation granularity"); + caig = 0; + } + return caig; +} + +/* Flush IOTLB + * Returns the IOTLB Actual Invalidation Granularity. + * @val: the content of the IOTLB_REG + */ +static uint64_t vtd_iotlb_flush(IntelIOMMUState *s, uint64_t val) +{ + uint64_t iaig; + uint64_t type = val & VTD_TLB_FLUSH_GRANU_MASK; + + switch (type) { + case VTD_TLB_GLOBAL_FLUSH: + VTD_DPRINTF(INV, "Global IOTLB flush"); + iaig = VTD_TLB_GLOBAL_FLUSH_A; + break; + + case VTD_TLB_DSI_FLUSH: + VTD_DPRINTF(INV, "Domain-selective IOTLB flush"); + iaig = VTD_TLB_DSI_FLUSH_A; + break; + + case VTD_TLB_PSI_FLUSH: + VTD_DPRINTF(INV, "Page-selective-within-domain IOTLB flush"); + iaig = VTD_TLB_PSI_FLUSH_A; + break; + + default: + VTD_DPRINTF(GENERAL, "error: wrong iotlb flush granularity"); + iaig = 0; + } + return iaig; +} + +/* Set Root Table Pointer */ +static void vtd_handle_gcmd_srtp(IntelIOMMUState *s) +{ + VTD_DPRINTF(CSR, "set Root Table Pointer"); + + vtd_root_table_setup(s); + /* Ok - report back to driver */ + vtd_set_clear_mask_long(s, DMAR_GSTS_REG, 0, VTD_GSTS_RTPS); +} + +/* Handle Translation Enable/Disable */ +static void vtd_handle_gcmd_te(IntelIOMMUState *s, bool en) +{ + VTD_DPRINTF(CSR, "Translation Enable %s", (en ? "on" : "off")); + + if (en) { + s->dmar_enabled = true; + /* Ok - report back to driver */ + vtd_set_clear_mask_long(s, DMAR_GSTS_REG, 0, VTD_GSTS_TES); + } else { + s->dmar_enabled = false; + + /* Clear the index of Fault Recording Register */ + s->next_frcd_reg = 0; + /* Ok - report back to driver */ + vtd_set_clear_mask_long(s, DMAR_GSTS_REG, VTD_GSTS_TES, 0); + } +} + +/* Handle write to Global Command Register */ +static void vtd_handle_gcmd_write(IntelIOMMUState *s) +{ + uint32_t status = vtd_get_long_raw(s, DMAR_GSTS_REG); + uint32_t val = vtd_get_long_raw(s, DMAR_GCMD_REG); + uint32_t changed = status ^ val; + + VTD_DPRINTF(CSR, "value 0x%"PRIx32 " status 0x%"PRIx32, val, status); + if (changed & VTD_GCMD_TE) { + /* Translation enable/disable */ + vtd_handle_gcmd_te(s, val & VTD_GCMD_TE); + } + if (val & VTD_GCMD_SRTP) { + /* Set/update the root-table pointer */ + vtd_handle_gcmd_srtp(s); + } +} + +/* Handle write to Context Command Register */ +static void vtd_handle_ccmd_write(IntelIOMMUState *s) +{ + uint64_t ret; + uint64_t val = vtd_get_quad_raw(s, DMAR_CCMD_REG); + + /* Context-cache invalidation request */ + if (val & VTD_CCMD_ICC) { + ret = vtd_context_cache_invalidate(s, val); + /* Invalidation completed. Change something to show */ + vtd_set_clear_mask_quad(s, DMAR_CCMD_REG, VTD_CCMD_ICC, 0ULL); + ret = vtd_set_clear_mask_quad(s, DMAR_CCMD_REG, VTD_CCMD_CAIG_MASK, + ret); + VTD_DPRINTF(INV, "CCMD_REG write-back val: 0x%"PRIx64, ret); + } +} + +/* Handle write to IOTLB Invalidation Register */ +static void vtd_handle_iotlb_write(IntelIOMMUState *s) +{ + uint64_t ret; + uint64_t val = vtd_get_quad_raw(s, DMAR_IOTLB_REG); + + /* IOTLB invalidation request */ + if (val & VTD_TLB_IVT) { + ret = vtd_iotlb_flush(s, val); + /* Invalidation completed. Change something to show */ + vtd_set_clear_mask_quad(s, DMAR_IOTLB_REG, VTD_TLB_IVT, 0ULL); + ret = vtd_set_clear_mask_quad(s, DMAR_IOTLB_REG, + VTD_TLB_FLUSH_GRANU_MASK_A, ret); + VTD_DPRINTF(INV, "IOTLB_REG write-back val: 0x%"PRIx64, ret); + } +} + +static void vtd_handle_fsts_write(IntelIOMMUState *s) +{ + uint32_t fsts_reg = vtd_get_long_raw(s, DMAR_FSTS_REG); + uint32_t fectl_reg = vtd_get_long_raw(s, DMAR_FECTL_REG); + uint32_t status_fields = VTD_FSTS_PFO | VTD_FSTS_PPF | VTD_FSTS_IQE; + + if ((fectl_reg & VTD_FECTL_IP) && !(fsts_reg & status_fields)) { + vtd_set_clear_mask_long(s, DMAR_FECTL_REG, VTD_FECTL_IP, 0); + VTD_DPRINTF(FLOG, "all pending interrupt conditions serviced, clear " + "IP field of FECTL_REG"); + } +} + +static void vtd_handle_fectl_write(IntelIOMMUState *s) +{ + uint32_t fectl_reg; + /* FIXME: when software clears the IM field, check the IP field. But do we + * need to compare the old value and the new value to conclude that + * software clears the IM field? Or just check if the IM field is zero? + */ + fectl_reg = vtd_get_long_raw(s, DMAR_FECTL_REG); + if ((fectl_reg & VTD_FECTL_IP) && !(fectl_reg & VTD_FECTL_IM)) { + vtd_generate_interrupt(s, DMAR_FEADDR_REG, DMAR_FEDATA_REG); + vtd_set_clear_mask_long(s, DMAR_FECTL_REG, VTD_FECTL_IP, 0); + VTD_DPRINTF(FLOG, "IM field is cleared, generate " + "fault event interrupt"); + } +} + +static uint64_t vtd_mem_read(void *opaque, hwaddr addr, unsigned size) +{ + IntelIOMMUState *s = opaque; + uint64_t val; + + if (addr + size > DMAR_REG_SIZE) { + VTD_DPRINTF(GENERAL, "error: addr outside region: max 0x%"PRIx64 + ", got 0x%"PRIx64 " %d", + (uint64_t)DMAR_REG_SIZE, addr, size); + return (uint64_t)-1; + } + + switch (addr) { + /* Root Table Address Register, 64-bit */ + case DMAR_RTADDR_REG: + if (size == 4) { + val = s->root & ((1ULL << 32) - 1); + } else { + val = s->root; + } + break; + + case DMAR_RTADDR_REG_HI: + assert(size == 4); + val = s->root >> 32; + break; + + default: + if (size == 4) { + val = vtd_get_long(s, addr); + } else { + val = vtd_get_quad(s, addr); + } + } + VTD_DPRINTF(CSR, "addr 0x%"PRIx64 " size %d val 0x%"PRIx64, + addr, size, val); + return val; +} + +static void vtd_mem_write(void *opaque, hwaddr addr, + uint64_t val, unsigned size) +{ + IntelIOMMUState *s = opaque; + + if (addr + size > DMAR_REG_SIZE) { + VTD_DPRINTF(GENERAL, "error: addr outside region: max 0x%"PRIx64 + ", got 0x%"PRIx64 " %d", + (uint64_t)DMAR_REG_SIZE, addr, size); + return; + } + + switch (addr) { + /* Global Command Register, 32-bit */ + case DMAR_GCMD_REG: + VTD_DPRINTF(CSR, "DMAR_GCMD_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + vtd_set_long(s, addr, val); + vtd_handle_gcmd_write(s); + break; + + /* Context Command Register, 64-bit */ + case DMAR_CCMD_REG: + VTD_DPRINTF(CSR, "DMAR_CCMD_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + vtd_handle_ccmd_write(s); + } + break; + + case DMAR_CCMD_REG_HI: + VTD_DPRINTF(CSR, "DMAR_CCMD_REG_HI write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + vtd_handle_ccmd_write(s); + break; + + /* IOTLB Invalidation Register, 64-bit */ + case DMAR_IOTLB_REG: + VTD_DPRINTF(INV, "DMAR_IOTLB_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + vtd_handle_iotlb_write(s); + } + break; + + case DMAR_IOTLB_REG_HI: + VTD_DPRINTF(INV, "DMAR_IOTLB_REG_HI write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + vtd_handle_iotlb_write(s); + break; + + /* Fault Status Register, 32-bit */ + case DMAR_FSTS_REG: + VTD_DPRINTF(FLOG, "DMAR_FSTS_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + vtd_handle_fsts_write(s); + break; + + /* Fault Event Control Register, 32-bit */ + case DMAR_FECTL_REG: + VTD_DPRINTF(FLOG, "DMAR_FECTL_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + vtd_handle_fectl_write(s); + break; + + /* Fault Event Data Register, 32-bit */ + case DMAR_FEDATA_REG: + VTD_DPRINTF(FLOG, "DMAR_FEDATA_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + /* Fault Event Address Register, 32-bit */ + case DMAR_FEADDR_REG: + VTD_DPRINTF(FLOG, "DMAR_FEADDR_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + /* Fault Event Upper Address Register, 32-bit */ + case DMAR_FEUADDR_REG: + VTD_DPRINTF(FLOG, "DMAR_FEUADDR_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + /* Protected Memory Enable Register, 32-bit */ + case DMAR_PMEN_REG: + VTD_DPRINTF(CSR, "DMAR_PMEN_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + /* Root Table Address Register, 64-bit */ + case DMAR_RTADDR_REG: + VTD_DPRINTF(CSR, "DMAR_RTADDR_REG write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + } + break; + + case DMAR_RTADDR_REG_HI: + VTD_DPRINTF(CSR, "DMAR_RTADDR_REG_HI write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + /* Fault Recording Registers, 128-bit */ + case DMAR_FRCD_REG_0_0: + VTD_DPRINTF(FLOG, "DMAR_FRCD_REG_0_0 write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + } + break; + + case DMAR_FRCD_REG_0_1: + VTD_DPRINTF(FLOG, "DMAR_FRCD_REG_0_1 write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + break; + + case DMAR_FRCD_REG_0_2: + VTD_DPRINTF(FLOG, "DMAR_FRCD_REG_0_2 write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + /* May clear bit 127 (Fault), update PPF */ + vtd_update_fsts_ppf(s); + } + break; + + case DMAR_FRCD_REG_0_3: + VTD_DPRINTF(FLOG, "DMAR_FRCD_REG_0_3 write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + assert(size == 4); + vtd_set_long(s, addr, val); + /* May clear bit 127 (Fault), update PPF */ + vtd_update_fsts_ppf(s); + break; + + default: + VTD_DPRINTF(GENERAL, "error: unhandled reg write addr 0x%"PRIx64 + ", size %d, val 0x%"PRIx64, addr, size, val); + if (size == 4) { + vtd_set_long(s, addr, val); + } else { + vtd_set_quad(s, addr, val); + } + } +} + +static IOMMUTLBEntry vtd_iommu_translate(MemoryRegion *iommu, hwaddr addr, + bool is_write) +{ + VTDAddressSpace *vtd_as = container_of(iommu, VTDAddressSpace, iommu); + IntelIOMMUState *s = vtd_as->iommu_state; + uint8_t bus_num = vtd_as->bus_num; + uint8_t devfn = vtd_as->devfn; + IOMMUTLBEntry ret = { + .target_as = &address_space_memory, + .iova = addr, + .translated_addr = 0, + .addr_mask = ~(hwaddr)0, + .perm = IOMMU_NONE, + }; + + if (!s->dmar_enabled) { + /* DMAR disabled, passthrough, use 4k-page*/ + ret.iova = addr & VTD_PAGE_MASK_4K; + ret.translated_addr = addr & VTD_PAGE_MASK_4K; + ret.addr_mask = ~VTD_PAGE_MASK_4K; + ret.perm = IOMMU_RW; + return ret; + } + + vtd_do_iommu_translate(s, bus_num, devfn, addr, is_write, &ret); + + VTD_DPRINTF(MMU, + "bus %"PRIu8 " slot %"PRIu8 " func %"PRIu8 " devfn %"PRIu8 + " gpa 0x%"PRIx64 " hpa 0x%"PRIx64, bus_num, + VTD_PCI_SLOT(devfn), VTD_PCI_FUNC(devfn), devfn, addr, + ret.translated_addr); + return ret; +} + +static const VMStateDescription vtd_vmstate = { + .name = "iommu-intel", + .unmigratable = 1, +}; + +static const MemoryRegionOps vtd_mem_ops = { + .read = vtd_mem_read, + .write = vtd_mem_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .impl = { + .min_access_size = 4, + .max_access_size = 8, + }, + .valid = { + .min_access_size = 4, + .max_access_size = 8, + }, +}; + +static Property vtd_properties[] = { + DEFINE_PROP_UINT32("version", IntelIOMMUState, version, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +/* Do the initialization. It will also be called when reset, so pay + * attention when adding new initialization stuff. + */ +static void vtd_init(IntelIOMMUState *s) +{ + memset(s->csr, 0, DMAR_REG_SIZE); + memset(s->wmask, 0, DMAR_REG_SIZE); + memset(s->w1cmask, 0, DMAR_REG_SIZE); + memset(s->womask, 0, DMAR_REG_SIZE); + + s->iommu_ops.translate = vtd_iommu_translate; + s->root = 0; + s->root_extended = false; + s->dmar_enabled = false; + s->iq_head = 0; + s->iq_tail = 0; + s->iq = 0; + s->iq_size = 0; + s->qi_enabled = false; + s->iq_last_desc_type = VTD_INV_DESC_NONE; + s->next_frcd_reg = 0; + s->cap = VTD_CAP_FRO | VTD_CAP_NFR | VTD_CAP_ND | VTD_CAP_MGAW | + VTD_CAP_SAGAW; + s->ecap = VTD_ECAP_IRO; + + /* Define registers with default values and bit semantics */ + vtd_define_long(s, DMAR_VER_REG, 0x10UL, 0, 0); + vtd_define_quad(s, DMAR_CAP_REG, s->cap, 0, 0); + vtd_define_quad(s, DMAR_ECAP_REG, s->ecap, 0, 0); + vtd_define_long(s, DMAR_GCMD_REG, 0, 0xff800000UL, 0); + vtd_define_long_wo(s, DMAR_GCMD_REG, 0xff800000UL); + vtd_define_long(s, DMAR_GSTS_REG, 0, 0, 0); + vtd_define_quad(s, DMAR_RTADDR_REG, 0, 0xfffffffffffff000ULL, 0); + vtd_define_quad(s, DMAR_CCMD_REG, 0, 0xe0000003ffffffffULL, 0); + vtd_define_quad_wo(s, DMAR_CCMD_REG, 0x3ffff0000ULL); + + /* Advanced Fault Logging not supported */ + vtd_define_long(s, DMAR_FSTS_REG, 0, 0, 0x11UL); + vtd_define_long(s, DMAR_FECTL_REG, 0x80000000UL, 0x80000000UL, 0); + vtd_define_long(s, DMAR_FEDATA_REG, 0, 0x0000ffffUL, 0); + vtd_define_long(s, DMAR_FEADDR_REG, 0, 0xfffffffcUL, 0); + + /* Treated as RsvdZ when EIM in ECAP_REG is not supported + * vtd_define_long(s, DMAR_FEUADDR_REG, 0, 0xffffffffUL, 0); + */ + vtd_define_long(s, DMAR_FEUADDR_REG, 0, 0, 0); + + /* Treated as RO for implementations that PLMR and PHMR fields reported + * as Clear in the CAP_REG. + * vtd_define_long(s, DMAR_PMEN_REG, 0, 0x80000000UL, 0); + */ + vtd_define_long(s, DMAR_PMEN_REG, 0, 0, 0); + + /* IOTLB registers */ + vtd_define_quad(s, DMAR_IOTLB_REG, 0, 0Xb003ffff00000000ULL, 0); + vtd_define_quad(s, DMAR_IVA_REG, 0, 0xfffffffffffff07fULL, 0); + vtd_define_quad_wo(s, DMAR_IVA_REG, 0xfffffffffffff07fULL); + + /* Fault Recording Registers, 128-bit */ + vtd_define_quad(s, DMAR_FRCD_REG_0_0, 0, 0, 0); + vtd_define_quad(s, DMAR_FRCD_REG_0_2, 0, 0, 0x8000000000000000ULL); +} + +/* Should not reset address_spaces when reset because devices will still use + * the address space they got at first (won't ask the bus again). + */ +static void vtd_reset(DeviceState *dev) +{ + IntelIOMMUState *s = INTEL_IOMMU_DEVICE(dev); + + VTD_DPRINTF(GENERAL, ""); + vtd_init(s); +} + +static void vtd_realize(DeviceState *dev, Error **errp) +{ + IntelIOMMUState *s = INTEL_IOMMU_DEVICE(dev); + + VTD_DPRINTF(GENERAL, ""); + memset(s->address_spaces, 0, sizeof(s->address_spaces)); + memory_region_init_io(&s->csrmem, OBJECT(s), &vtd_mem_ops, s, + "intel_iommu", DMAR_REG_SIZE); + sysbus_init_mmio(SYS_BUS_DEVICE(s), &s->csrmem); + vtd_init(s); +} + +static void vtd_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->reset = vtd_reset; + dc->realize = vtd_realize; + dc->vmsd = &vtd_vmstate; + dc->props = vtd_properties; +} + +static const TypeInfo vtd_info = { + .name = TYPE_INTEL_IOMMU_DEVICE, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(IntelIOMMUState), + .class_init = vtd_class_init, +}; + +static void vtd_register_types(void) +{ + VTD_DPRINTF(GENERAL, ""); + type_register_static(&vtd_info); +} + +type_init(vtd_register_types) diff --git a/hw/i386/intel_iommu_internal.h b/hw/i386/intel_iommu_internal.h new file mode 100644 index 0000000000..7ca034d8e6 --- /dev/null +++ b/hw/i386/intel_iommu_internal.h @@ -0,0 +1,333 @@ +/* + * QEMU emulation of an Intel IOMMU (VT-d) + * (DMA Remapping device) + * + * Copyright (C) 2013 Knut Omang, Oracle <knut.omang@oracle.com> + * Copyright (C) 2014 Le Tan, <tamlokveer@gmail.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + + * 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, see <http://www.gnu.org/licenses/>. + * + * Lots of defines copied from kernel/include/linux/intel-iommu.h: + * Copyright (C) 2006-2008 Intel Corporation + * Author: Ashok Raj <ashok.raj@intel.com> + * Author: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> + * + */ + +#ifndef HW_I386_INTEL_IOMMU_INTERNAL_H +#define HW_I386_INTEL_IOMMU_INTERNAL_H +#include "hw/i386/intel_iommu.h" + +/* + * Intel IOMMU register specification + */ +#define DMAR_VER_REG 0x0 /* Arch version supported by this IOMMU */ +#define DMAR_CAP_REG 0x8 /* Hardware supported capabilities */ +#define DMAR_CAP_REG_HI 0xc /* High 32-bit of DMAR_CAP_REG */ +#define DMAR_ECAP_REG 0x10 /* Extended capabilities supported */ +#define DMAR_ECAP_REG_HI 0X14 +#define DMAR_GCMD_REG 0x18 /* Global command */ +#define DMAR_GSTS_REG 0x1c /* Global status */ +#define DMAR_RTADDR_REG 0x20 /* Root entry table */ +#define DMAR_RTADDR_REG_HI 0X24 +#define DMAR_CCMD_REG 0x28 /* Context command */ +#define DMAR_CCMD_REG_HI 0x2c +#define DMAR_FSTS_REG 0x34 /* Fault status */ +#define DMAR_FECTL_REG 0x38 /* Fault control */ +#define DMAR_FEDATA_REG 0x3c /* Fault event interrupt data */ +#define DMAR_FEADDR_REG 0x40 /* Fault event interrupt addr */ +#define DMAR_FEUADDR_REG 0x44 /* Upper address */ +#define DMAR_AFLOG_REG 0x58 /* Advanced fault control */ +#define DMAR_AFLOG_REG_HI 0X5c +#define DMAR_PMEN_REG 0x64 /* Enable protected memory region */ +#define DMAR_PLMBASE_REG 0x68 /* PMRR low addr */ +#define DMAR_PLMLIMIT_REG 0x6c /* PMRR low limit */ +#define DMAR_PHMBASE_REG 0x70 /* PMRR high base addr */ +#define DMAR_PHMBASE_REG_HI 0X74 +#define DMAR_PHMLIMIT_REG 0x78 /* PMRR high limit */ +#define DMAR_PHMLIMIT_REG_HI 0x7c +#define DMAR_IQH_REG 0x80 /* Invalidation queue head */ +#define DMAR_IQH_REG_HI 0X84 +#define DMAR_IQT_REG 0x88 /* Invalidation queue tail */ +#define DMAR_IQT_REG_HI 0X8c +#define DMAR_IQA_REG 0x90 /* Invalidation queue addr */ +#define DMAR_IQA_REG_HI 0x94 +#define DMAR_ICS_REG 0x9c /* Invalidation complete status */ +#define DMAR_IRTA_REG 0xb8 /* Interrupt remapping table addr */ +#define DMAR_IRTA_REG_HI 0xbc +#define DMAR_IECTL_REG 0xa0 /* Invalidation event control */ +#define DMAR_IEDATA_REG 0xa4 /* Invalidation event data */ +#define DMAR_IEADDR_REG 0xa8 /* Invalidation event address */ +#define DMAR_IEUADDR_REG 0xac /* Invalidation event address */ +#define DMAR_PQH_REG 0xc0 /* Page request queue head */ +#define DMAR_PQH_REG_HI 0xc4 +#define DMAR_PQT_REG 0xc8 /* Page request queue tail*/ +#define DMAR_PQT_REG_HI 0xcc +#define DMAR_PQA_REG 0xd0 /* Page request queue address */ +#define DMAR_PQA_REG_HI 0xd4 +#define DMAR_PRS_REG 0xdc /* Page request status */ +#define DMAR_PECTL_REG 0xe0 /* Page request event control */ +#define DMAR_PEDATA_REG 0xe4 /* Page request event data */ +#define DMAR_PEADDR_REG 0xe8 /* Page request event address */ +#define DMAR_PEUADDR_REG 0xec /* Page event upper address */ +#define DMAR_MTRRCAP_REG 0x100 /* MTRR capability */ +#define DMAR_MTRRCAP_REG_HI 0x104 +#define DMAR_MTRRDEF_REG 0x108 /* MTRR default type */ +#define DMAR_MTRRDEF_REG_HI 0x10c + +/* IOTLB registers */ +#define DMAR_IOTLB_REG_OFFSET 0xf0 /* Offset to the IOTLB registers */ +#define DMAR_IVA_REG DMAR_IOTLB_REG_OFFSET /* Invalidate address */ +#define DMAR_IVA_REG_HI (DMAR_IVA_REG + 4) +/* IOTLB invalidate register */ +#define DMAR_IOTLB_REG (DMAR_IOTLB_REG_OFFSET + 0x8) +#define DMAR_IOTLB_REG_HI (DMAR_IOTLB_REG + 4) + +/* FRCD */ +#define DMAR_FRCD_REG_OFFSET 0x220 /* Offset to the fault recording regs */ +/* NOTICE: If you change the DMAR_FRCD_REG_NR, please remember to change the + * DMAR_REG_SIZE in include/hw/i386/intel_iommu.h. + * #define DMAR_REG_SIZE (DMAR_FRCD_REG_OFFSET + 16 * DMAR_FRCD_REG_NR) + */ +#define DMAR_FRCD_REG_NR 1ULL /* Num of fault recording regs */ + +#define DMAR_FRCD_REG_0_0 0x220 /* The 0th fault recording regs */ +#define DMAR_FRCD_REG_0_1 0x224 +#define DMAR_FRCD_REG_0_2 0x228 +#define DMAR_FRCD_REG_0_3 0x22c + +/* Interrupt Address Range */ +#define VTD_INTERRUPT_ADDR_FIRST 0xfee00000ULL +#define VTD_INTERRUPT_ADDR_LAST 0xfeefffffULL + +/* IOTLB_REG */ +#define VTD_TLB_GLOBAL_FLUSH (1ULL << 60) /* Global invalidation */ +#define VTD_TLB_DSI_FLUSH (2ULL << 60) /* Domain-selective */ +#define VTD_TLB_PSI_FLUSH (3ULL << 60) /* Page-selective */ +#define VTD_TLB_FLUSH_GRANU_MASK (3ULL << 60) +#define VTD_TLB_GLOBAL_FLUSH_A (1ULL << 57) +#define VTD_TLB_DSI_FLUSH_A (2ULL << 57) +#define VTD_TLB_PSI_FLUSH_A (3ULL << 57) +#define VTD_TLB_FLUSH_GRANU_MASK_A (3ULL << 57) +#define VTD_TLB_IVT (1ULL << 63) + +/* GCMD_REG */ +#define VTD_GCMD_TE (1UL << 31) +#define VTD_GCMD_SRTP (1UL << 30) +#define VTD_GCMD_SFL (1UL << 29) +#define VTD_GCMD_EAFL (1UL << 28) +#define VTD_GCMD_WBF (1UL << 27) +#define VTD_GCMD_QIE (1UL << 26) +#define VTD_GCMD_IRE (1UL << 25) +#define VTD_GCMD_SIRTP (1UL << 24) +#define VTD_GCMD_CFI (1UL << 23) + +/* GSTS_REG */ +#define VTD_GSTS_TES (1UL << 31) +#define VTD_GSTS_RTPS (1UL << 30) +#define VTD_GSTS_FLS (1UL << 29) +#define VTD_GSTS_AFLS (1UL << 28) +#define VTD_GSTS_WBFS (1UL << 27) +#define VTD_GSTS_QIES (1UL << 26) +#define VTD_GSTS_IRES (1UL << 25) +#define VTD_GSTS_IRTPS (1UL << 24) +#define VTD_GSTS_CFIS (1UL << 23) + +/* CCMD_REG */ +#define VTD_CCMD_ICC (1ULL << 63) +#define VTD_CCMD_GLOBAL_INVL (1ULL << 61) +#define VTD_CCMD_DOMAIN_INVL (2ULL << 61) +#define VTD_CCMD_DEVICE_INVL (3ULL << 61) +#define VTD_CCMD_CIRG_MASK (3ULL << 61) +#define VTD_CCMD_GLOBAL_INVL_A (1ULL << 59) +#define VTD_CCMD_DOMAIN_INVL_A (2ULL << 59) +#define VTD_CCMD_DEVICE_INVL_A (3ULL << 59) +#define VTD_CCMD_CAIG_MASK (3ULL << 59) + +/* RTADDR_REG */ +#define VTD_RTADDR_RTT (1ULL << 11) +#define VTD_RTADDR_ADDR_MASK (VTD_HAW_MASK ^ 0xfffULL) + +/* ECAP_REG */ +/* (offset >> 4) << 8 */ +#define VTD_ECAP_IRO (DMAR_IOTLB_REG_OFFSET << 4) +#define VTD_ECAP_QI (1ULL << 1) + +/* CAP_REG */ +/* (offset >> 4) << 24 */ +#define VTD_CAP_FRO (DMAR_FRCD_REG_OFFSET << 20) +#define VTD_CAP_NFR ((DMAR_FRCD_REG_NR - 1) << 40) +#define VTD_DOMAIN_ID_SHIFT 16 /* 16-bit domain id for 64K domains */ +#define VTD_CAP_ND (((VTD_DOMAIN_ID_SHIFT - 4) / 2) & 7ULL) +#define VTD_MGAW 39 /* Maximum Guest Address Width */ +#define VTD_CAP_MGAW (((VTD_MGAW - 1) & 0x3fULL) << 16) + +/* Supported Adjusted Guest Address Widths */ +#define VTD_CAP_SAGAW_SHIFT 8 +#define VTD_CAP_SAGAW_MASK (0x1fULL << VTD_CAP_SAGAW_SHIFT) + /* 39-bit AGAW, 3-level page-table */ +#define VTD_CAP_SAGAW_39bit (0x2ULL << VTD_CAP_SAGAW_SHIFT) + /* 48-bit AGAW, 4-level page-table */ +#define VTD_CAP_SAGAW_48bit (0x4ULL << VTD_CAP_SAGAW_SHIFT) +#define VTD_CAP_SAGAW VTD_CAP_SAGAW_39bit + +/* IQT_REG */ +#define VTD_IQT_QT(val) (((val) >> 4) & 0x7fffULL) + +/* IQA_REG */ +#define VTD_IQA_IQA_MASK (VTD_HAW_MASK ^ 0xfffULL) +#define VTD_IQA_QS 0x7ULL + +/* IQH_REG */ +#define VTD_IQH_QH_SHIFT 4 +#define VTD_IQH_QH_MASK 0x7fff0ULL + +/* ICS_REG */ +#define VTD_ICS_IWC 1UL + +/* IECTL_REG */ +#define VTD_IECTL_IM (1UL << 31) +#define VTD_IECTL_IP (1UL << 30) + +/* FSTS_REG */ +#define VTD_FSTS_FRI_MASK 0xff00UL +#define VTD_FSTS_FRI(val) ((((uint32_t)(val)) << 8) & VTD_FSTS_FRI_MASK) +#define VTD_FSTS_IQE (1UL << 4) +#define VTD_FSTS_PPF (1UL << 1) +#define VTD_FSTS_PFO 1UL + +/* FECTL_REG */ +#define VTD_FECTL_IM (1UL << 31) +#define VTD_FECTL_IP (1UL << 30) + +/* Fault Recording Register */ +/* For the high 64-bit of 128-bit */ +#define VTD_FRCD_F (1ULL << 63) +#define VTD_FRCD_T (1ULL << 62) +#define VTD_FRCD_FR(val) (((val) & 0xffULL) << 32) +#define VTD_FRCD_SID_MASK 0xffffULL +#define VTD_FRCD_SID(val) ((val) & VTD_FRCD_SID_MASK) +/* For the low 64-bit of 128-bit */ +#define VTD_FRCD_FI(val) ((val) & (((1ULL << VTD_MGAW) - 1) ^ 0xfffULL)) + +/* DMA Remapping Fault Conditions */ +typedef enum VTDFaultReason { + VTD_FR_RESERVED = 0, /* Reserved for Advanced Fault logging */ + VTD_FR_ROOT_ENTRY_P = 1, /* The Present(P) field of root-entry is 0 */ + VTD_FR_CONTEXT_ENTRY_P, /* The Present(P) field of context-entry is 0 */ + VTD_FR_CONTEXT_ENTRY_INV, /* Invalid programming of a context-entry */ + VTD_FR_ADDR_BEYOND_MGAW, /* Input-address above (2^x-1) */ + VTD_FR_WRITE, /* No write permission */ + VTD_FR_READ, /* No read permission */ + /* Fail to access a second-level paging entry (not SL_PML4E) */ + VTD_FR_PAGING_ENTRY_INV, + VTD_FR_ROOT_TABLE_INV, /* Fail to access a root-entry */ + VTD_FR_CONTEXT_TABLE_INV, /* Fail to access a context-entry */ + /* Non-zero reserved field in a present root-entry */ + VTD_FR_ROOT_ENTRY_RSVD, + /* Non-zero reserved field in a present context-entry */ + VTD_FR_CONTEXT_ENTRY_RSVD, + /* Non-zero reserved field in a second-level paging entry with at lease one + * Read(R) and Write(W) or Execute(E) field is Set. + */ + VTD_FR_PAGING_ENTRY_RSVD, + /* Translation request or translated request explicitly blocked dut to the + * programming of the Translation Type (T) field in the present + * context-entry. + */ + VTD_FR_CONTEXT_ENTRY_TT, + /* This is not a normal fault reason. We use this to indicate some faults + * that are not referenced by the VT-d specification. + * Fault event with such reason should not be recorded. + */ + VTD_FR_RESERVED_ERR, + VTD_FR_MAX, /* Guard */ +} VTDFaultReason; + +/* Masks for Queued Invalidation Descriptor */ +#define VTD_INV_DESC_TYPE 0xf +#define VTD_INV_DESC_CC 0x1 /* Context-cache Invalidate Desc */ +#define VTD_INV_DESC_IOTLB 0x2 +#define VTD_INV_DESC_WAIT 0x5 /* Invalidation Wait Descriptor */ +#define VTD_INV_DESC_NONE 0 /* Not an Invalidate Descriptor */ + +/* Pagesize of VTD paging structures, including root and context tables */ +#define VTD_PAGE_SHIFT 12 +#define VTD_PAGE_SIZE (1ULL << VTD_PAGE_SHIFT) + +#define VTD_PAGE_SHIFT_4K 12 +#define VTD_PAGE_MASK_4K (~((1ULL << VTD_PAGE_SHIFT_4K) - 1)) +#define VTD_PAGE_SHIFT_2M 21 +#define VTD_PAGE_MASK_2M (~((1ULL << VTD_PAGE_SHIFT_2M) - 1)) +#define VTD_PAGE_SHIFT_1G 30 +#define VTD_PAGE_MASK_1G (~((1ULL << VTD_PAGE_SHIFT_1G) - 1)) + +struct VTDRootEntry { + uint64_t val; + uint64_t rsvd; +}; +typedef struct VTDRootEntry VTDRootEntry; + +/* Masks for struct VTDRootEntry */ +#define VTD_ROOT_ENTRY_P 1ULL +#define VTD_ROOT_ENTRY_CTP (~0xfffULL) + +#define VTD_ROOT_ENTRY_NR (VTD_PAGE_SIZE / sizeof(VTDRootEntry)) +#define VTD_ROOT_ENTRY_RSVD (0xffeULL | ~VTD_HAW_MASK) + +/* Context-Entry */ +struct VTDContextEntry { + uint64_t lo; + uint64_t hi; +}; +typedef struct VTDContextEntry VTDContextEntry; + +/* Masks for struct VTDContextEntry */ +/* lo */ +#define VTD_CONTEXT_ENTRY_P (1ULL << 0) +#define VTD_CONTEXT_ENTRY_FPD (1ULL << 1) /* Fault Processing Disable */ +#define VTD_CONTEXT_ENTRY_TT (3ULL << 2) /* Translation Type */ +#define VTD_CONTEXT_TT_MULTI_LEVEL 0 +#define VTD_CONTEXT_TT_DEV_IOTLB 1 +#define VTD_CONTEXT_TT_PASS_THROUGH 2 +/* Second Level Page Translation Pointer*/ +#define VTD_CONTEXT_ENTRY_SLPTPTR (~0xfffULL) +#define VTD_CONTEXT_ENTRY_RSVD_LO (0xff0ULL | ~VTD_HAW_MASK) +/* hi */ +#define VTD_CONTEXT_ENTRY_AW 7ULL /* Adjusted guest-address-width */ +#define VTD_CONTEXT_ENTRY_DID (0xffffULL << 8) /* Domain Identifier */ +#define VTD_CONTEXT_ENTRY_RSVD_HI 0xffffffffff000080ULL + +#define VTD_CONTEXT_ENTRY_NR (VTD_PAGE_SIZE / sizeof(VTDContextEntry)) + +/* Paging Structure common */ +#define VTD_SL_PT_PAGE_SIZE_MASK (1ULL << 7) +/* Bits to decide the offset for each level */ +#define VTD_SL_LEVEL_BITS 9 + +/* Second Level Paging Structure */ +#define VTD_SL_PML4_LEVEL 4 +#define VTD_SL_PDP_LEVEL 3 +#define VTD_SL_PD_LEVEL 2 +#define VTD_SL_PT_LEVEL 1 +#define VTD_SL_PT_ENTRY_NR 512 + +/* Masks for Second Level Paging Entry */ +#define VTD_SL_RW_MASK 3ULL +#define VTD_SL_R 1ULL +#define VTD_SL_W (1ULL << 1) +#define VTD_SL_PT_BASE_ADDR_MASK (~(VTD_PAGE_SIZE - 1) & VTD_HAW_MASK) +#define VTD_SL_IGN_COM 0xbff0000000000000ULL + +#endif |