diff options
Diffstat (limited to 'arch/x86/coco/tdx/tdx.c')
| -rw-r--r-- | arch/x86/coco/tdx/tdx.c | 750 |
1 files changed, 579 insertions, 171 deletions
diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c index b8998cf0508a..7b2833705d47 100644 --- a/arch/x86/coco/tdx/tdx.c +++ b/arch/x86/coco/tdx/tdx.c @@ -5,20 +5,19 @@ #define pr_fmt(fmt) "tdx: " fmt #include <linux/cpufeature.h> +#include <linux/export.h> +#include <linux/io.h> +#include <linux/kexec.h> #include <asm/coco.h> #include <asm/tdx.h> #include <asm/vmx.h> +#include <asm/ia32.h> #include <asm/insn.h> #include <asm/insn-eval.h> +#include <asm/paravirt_types.h> #include <asm/pgtable.h> - -/* TDX module Call Leaf IDs */ -#define TDX_GET_INFO 1 -#define TDX_GET_VEINFO 3 -#define TDX_ACCEPT_PAGE 6 - -/* TDX hypercall Leaf IDs */ -#define TDVMCALL_MAP_GPA 0x10001 +#include <asm/set_memory.h> +#include <asm/traps.h> /* MMIO direction */ #define EPT_READ 0 @@ -34,48 +33,27 @@ #define VE_GET_PORT_NUM(e) ((e) >> 16) #define VE_IS_IO_STRING(e) ((e) & BIT(4)) -#define ATTR_SEPT_VE_DISABLE BIT(28) +/* TDX Module call error codes */ +#define TDCALL_RETURN_CODE(a) ((a) >> 32) +#define TDCALL_INVALID_OPERAND 0xc0000100 +#define TDCALL_OPERAND_BUSY 0x80000200 -/* - * Wrapper for standard use of __tdx_hypercall with no output aside from - * return code. - */ -static inline u64 _tdx_hypercall(u64 fn, u64 r12, u64 r13, u64 r14, u64 r15) -{ - struct tdx_hypercall_args args = { - .r10 = TDX_HYPERCALL_STANDARD, - .r11 = fn, - .r12 = r12, - .r13 = r13, - .r14 = r14, - .r15 = r15, - }; +#define TDREPORT_SUBTYPE_0 0 - return __tdx_hypercall(&args, 0); -} +static atomic_long_t nr_shared; /* Called from __tdx_hypercall() for unrecoverable failure */ -void __tdx_hypercall_failed(void) +noinstr void __noreturn __tdx_hypercall_failed(void) { + instrumentation_begin(); panic("TDVMCALL failed. TDX module bug?"); } -/* - * The TDG.VP.VMCALL-Instruction-execution sub-functions are defined - * independently from but are currently matched 1:1 with VMX EXIT_REASONs. - * Reusing the KVM EXIT_REASON macros makes it easier to connect the host and - * guest sides of these calls. - */ -static u64 hcall_func(u64 exit_reason) -{ - return exit_reason; -} - #ifdef CONFIG_KVM_GUEST long tdx_kvm_hypercall(unsigned int nr, unsigned long p1, unsigned long p2, unsigned long p3, unsigned long p4) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = nr, .r11 = p1, .r12 = p2, @@ -83,7 +61,7 @@ long tdx_kvm_hypercall(unsigned int nr, unsigned long p1, unsigned long p2, .r14 = p4, }; - return __tdx_hypercall(&args, 0); + return __tdx_hypercall(&args); } EXPORT_SYMBOL_GPL(tdx_kvm_hypercall); #endif @@ -93,16 +71,264 @@ EXPORT_SYMBOL_GPL(tdx_kvm_hypercall); * should only be used for calls that have no legitimate reason to fail * or where the kernel can not survive the call failing. */ -static inline void tdx_module_call(u64 fn, u64 rcx, u64 rdx, u64 r8, u64 r9, - struct tdx_module_output *out) +static inline void tdcall(u64 fn, struct tdx_module_args *args) { - if (__tdx_module_call(fn, rcx, rdx, r8, r9, out)) + if (__tdcall_ret(fn, args)) panic("TDCALL %lld failed (Buggy TDX module!)\n", fn); } -static void tdx_parse_tdinfo(u64 *cc_mask) +/* Read TD-scoped metadata */ +static inline u64 tdg_vm_rd(u64 field, u64 *value) { - struct tdx_module_output out; + struct tdx_module_args args = { + .rdx = field, + }; + u64 ret; + + ret = __tdcall_ret(TDG_VM_RD, &args); + *value = args.r8; + + return ret; +} + +/* Write TD-scoped metadata */ +static inline u64 tdg_vm_wr(u64 field, u64 value, u64 mask) +{ + struct tdx_module_args args = { + .rdx = field, + .r8 = value, + .r9 = mask, + }; + + return __tdcall(TDG_VM_WR, &args); +} + +/** + * tdx_mcall_get_report0() - Wrapper to get TDREPORT0 (a.k.a. TDREPORT + * subtype 0) using TDG.MR.REPORT TDCALL. + * @reportdata: Address of the input buffer which contains user-defined + * REPORTDATA to be included into TDREPORT. + * @tdreport: Address of the output buffer to store TDREPORT. + * + * Refer to section titled "TDG.MR.REPORT leaf" in the TDX Module v1.0 + * specification for more information on TDG.MR.REPORT TDCALL. + * + * It is used in the TDX guest driver module to get the TDREPORT0. + * + * Return 0 on success, -ENXIO for invalid operands, -EBUSY for busy operation, + * or -EIO on other TDCALL failures. + */ +int tdx_mcall_get_report0(u8 *reportdata, u8 *tdreport) +{ + struct tdx_module_args args = { + .rcx = virt_to_phys(tdreport), + .rdx = virt_to_phys(reportdata), + .r8 = TDREPORT_SUBTYPE_0, + }; + u64 ret; + + ret = __tdcall(TDG_MR_REPORT, &args); + if (ret) { + if (TDCALL_RETURN_CODE(ret) == TDCALL_INVALID_OPERAND) + return -ENXIO; + else if (TDCALL_RETURN_CODE(ret) == TDCALL_OPERAND_BUSY) + return -EBUSY; + return -EIO; + } + + return 0; +} +EXPORT_SYMBOL_GPL(tdx_mcall_get_report0); + +/** + * tdx_mcall_extend_rtmr() - Wrapper to extend RTMR registers using + * TDG.MR.RTMR.EXTEND TDCALL. + * @index: Index of RTMR register to be extended. + * @data: Address of the input buffer with RTMR register extend data. + * + * Refer to section titled "TDG.MR.RTMR.EXTEND leaf" in the TDX Module v1.0 + * specification for more information on TDG.MR.RTMR.EXTEND TDCALL. + * + * It is used in the TDX guest driver module to allow user to extend the RTMR + * registers. + * + * Return 0 on success, -ENXIO for invalid operands, -EBUSY for busy operation, + * or -EIO on other TDCALL failures. + */ +int tdx_mcall_extend_rtmr(u8 index, u8 *data) +{ + struct tdx_module_args args = { + .rcx = virt_to_phys(data), + .rdx = index, + }; + u64 ret; + + ret = __tdcall(TDG_MR_RTMR_EXTEND, &args); + if (ret) { + if (TDCALL_RETURN_CODE(ret) == TDCALL_INVALID_OPERAND) + return -ENXIO; + if (TDCALL_RETURN_CODE(ret) == TDCALL_OPERAND_BUSY) + return -EBUSY; + return -EIO; + } + + return 0; +} +EXPORT_SYMBOL_GPL(tdx_mcall_extend_rtmr); + +/** + * tdx_hcall_get_quote() - Wrapper to request TD Quote using GetQuote + * hypercall. + * @buf: Address of the directly mapped shared kernel buffer which + * contains TDREPORT. The same buffer will be used by VMM to + * store the generated TD Quote output. + * @size: size of the tdquote buffer (4KB-aligned). + * + * Refer to section titled "TDG.VP.VMCALL<GetQuote>" in the TDX GHCI + * v1.0 specification for more information on GetQuote hypercall. + * It is used in the TDX guest driver module to get the TD Quote. + * + * Return 0 on success or error code on failure. + */ +u64 tdx_hcall_get_quote(u8 *buf, size_t size) +{ + /* Since buf is a shared memory, set the shared (decrypted) bits */ + return _tdx_hypercall(TDVMCALL_GET_QUOTE, cc_mkdec(virt_to_phys(buf)), size, 0, 0); +} +EXPORT_SYMBOL_GPL(tdx_hcall_get_quote); + +static void __noreturn tdx_panic(const char *msg) +{ + struct tdx_module_args args = { + .r10 = TDX_HYPERCALL_STANDARD, + .r11 = TDVMCALL_REPORT_FATAL_ERROR, + .r12 = 0, /* Error code: 0 is Panic */ + }; + union { + /* Define register order according to the GHCI */ + struct { u64 r14, r15, rbx, rdi, rsi, r8, r9, rdx; }; + + char bytes[64] __nonstring; + } message; + + /* VMM assumes '\0' in byte 65, if the message took all 64 bytes */ + strtomem_pad(message.bytes, msg, '\0'); + + args.r8 = message.r8; + args.r9 = message.r9; + args.r14 = message.r14; + args.r15 = message.r15; + args.rdi = message.rdi; + args.rsi = message.rsi; + args.rbx = message.rbx; + args.rdx = message.rdx; + + /* + * This hypercall should never return and it is not safe + * to keep the guest running. Call it forever if it + * happens to return. + */ + while (1) + __tdx_hypercall(&args); +} + +/* + * The kernel cannot handle #VEs when accessing normal kernel memory. Ensure + * that no #VE will be delivered for accesses to TD-private memory. + * + * TDX 1.0 does not allow the guest to disable SEPT #VE on its own. The VMM + * controls if the guest will receive such #VE with TD attribute + * TDX_ATTR_SEPT_VE_DISABLE. + * + * Newer TDX modules allow the guest to control if it wants to receive SEPT + * violation #VEs. + * + * Check if the feature is available and disable SEPT #VE if possible. + * + * If the TD is allowed to disable/enable SEPT #VEs, the TDX_ATTR_SEPT_VE_DISABLE + * attribute is no longer reliable. It reflects the initial state of the + * control for the TD, but it will not be updated if someone (e.g. bootloader) + * changes it before the kernel starts. Kernel must check TDCS_TD_CTLS bit to + * determine if SEPT #VEs are enabled or disabled. + */ +static void disable_sept_ve(u64 td_attr) +{ + const char *msg = "TD misconfiguration: SEPT #VE has to be disabled"; + bool debug = td_attr & TDX_ATTR_DEBUG; + u64 config, controls; + + /* Is this TD allowed to disable SEPT #VE */ + tdg_vm_rd(TDCS_CONFIG_FLAGS, &config); + if (!(config & TDCS_CONFIG_FLEXIBLE_PENDING_VE)) { + /* No SEPT #VE controls for the guest: check the attribute */ + if (td_attr & TDX_ATTR_SEPT_VE_DISABLE) + return; + + /* Relax SEPT_VE_DISABLE check for debug TD for backtraces */ + if (debug) + pr_warn("%s\n", msg); + else + tdx_panic(msg); + return; + } + + /* Check if SEPT #VE has been disabled before us */ + tdg_vm_rd(TDCS_TD_CTLS, &controls); + if (controls & TD_CTLS_PENDING_VE_DISABLE) + return; + + /* Keep #VEs enabled for splats in debugging environments */ + if (debug) + return; + + /* Disable SEPT #VEs */ + tdg_vm_wr(TDCS_TD_CTLS, TD_CTLS_PENDING_VE_DISABLE, + TD_CTLS_PENDING_VE_DISABLE); +} + +/* + * TDX 1.0 generates a #VE when accessing topology-related CPUID leafs (0xB and + * 0x1F) and the X2APIC_APICID MSR. The kernel returns all zeros on CPUID #VEs. + * In practice, this means that the kernel can only boot with a plain topology. + * Any complications will cause problems. + * + * The ENUM_TOPOLOGY feature allows the VMM to provide topology information. + * Enabling the feature eliminates topology-related #VEs: the TDX module + * virtualizes accesses to the CPUID leafs and the MSR. + * + * Enable ENUM_TOPOLOGY if it is available. + */ +static void enable_cpu_topology_enumeration(void) +{ + u64 configured; + + /* Has the VMM provided a valid topology configuration? */ + tdg_vm_rd(TDCS_TOPOLOGY_ENUM_CONFIGURED, &configured); + if (!configured) { + pr_err("VMM did not configure X2APIC_IDs properly\n"); + return; + } + + tdg_vm_wr(TDCS_TD_CTLS, TD_CTLS_ENUM_TOPOLOGY, TD_CTLS_ENUM_TOPOLOGY); +} + +static void reduce_unnecessary_ve(void) +{ + u64 err = tdg_vm_wr(TDCS_TD_CTLS, TD_CTLS_REDUCE_VE, TD_CTLS_REDUCE_VE); + + if (err == TDX_SUCCESS) + return; + + /* + * Enabling REDUCE_VE includes ENUM_TOPOLOGY. Only try to + * enable ENUM_TOPOLOGY if REDUCE_VE was not successful. + */ + enable_cpu_topology_enumeration(); +} + +static void tdx_setup(u64 *cc_mask) +{ + struct tdx_module_args args = {}; unsigned int gpa_width; u64 td_attr; @@ -113,7 +339,7 @@ static void tdx_parse_tdinfo(u64 *cc_mask) * Guest-Host-Communication Interface (GHCI), section 2.4.2 TDCALL * [TDG.VP.INFO]. */ - tdx_module_call(TDX_GET_INFO, 0, 0, 0, 0, &out); + tdcall(TDG_VP_INFO, &args); /* * The highest bit of a guest physical address is the "sharing" bit. @@ -122,17 +348,17 @@ static void tdx_parse_tdinfo(u64 *cc_mask) * The GPA width that comes out of this call is critical. TDX guests * can not meaningfully run without it. */ - gpa_width = out.rcx & GENMASK(5, 0); + gpa_width = args.rcx & GENMASK(5, 0); *cc_mask = BIT_ULL(gpa_width - 1); - /* - * The kernel can not handle #VE's when accessing normal kernel - * memory. Ensure that no #VE will be delivered for accesses to - * TD-private memory. Only VMM-shared memory (MMIO) will #VE. - */ - td_attr = out.rdx; - if (!(td_attr & ATTR_SEPT_VE_DISABLE)) - panic("TD misconfiguration: SEPT_VE_DISABLE attibute must be set.\n"); + td_attr = args.rdx; + + /* Kernel does not use NOTIFY_ENABLES and does not need random #VEs */ + tdg_vm_wr(TDCS_NOTIFY_ENABLES, 0, -1ULL); + + disable_sept_ve(td_attr); + + reduce_unnecessary_ve(); } /* @@ -180,9 +406,9 @@ static int ve_instr_len(struct ve_info *ve) } } -static u64 __cpuidle __halt(const bool irq_disabled, const bool do_sti) +static u64 __cpuidle __halt(const bool irq_disabled) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_HLT), .r12 = irq_disabled, @@ -200,45 +426,51 @@ static u64 __cpuidle __halt(const bool irq_disabled, const bool do_sti) * can keep the vCPU in virtual HLT, even if an IRQ is * pending, without hanging/breaking the guest. */ - return __tdx_hypercall(&args, do_sti ? TDX_HCALL_ISSUE_STI : 0); + return __tdx_hypercall(&args); } static int handle_halt(struct ve_info *ve) { + const bool irq_disabled = irqs_disabled(); + /* - * Since non safe halt is mainly used in CPU offlining - * and the guest will always stay in the halt state, don't - * call the STI instruction (set do_sti as false). + * HLT with IRQs enabled is unsafe, as an IRQ that is intended to be a + * wake event may be consumed before requesting HLT emulation, leaving + * the vCPU blocking indefinitely. */ - const bool irq_disabled = irqs_disabled(); - const bool do_sti = false; + if (WARN_ONCE(!irq_disabled, "HLT emulation with IRQs enabled")) + return -EIO; - if (__halt(irq_disabled, do_sti)) + if (__halt(irq_disabled)) return -EIO; return ve_instr_len(ve); } -void __cpuidle tdx_safe_halt(void) +void __cpuidle tdx_halt(void) { - /* - * For do_sti=true case, __tdx_hypercall() function enables - * interrupts using the STI instruction before the TDCALL. So - * set irq_disabled as false. - */ const bool irq_disabled = false; - const bool do_sti = true; /* * Use WARN_ONCE() to report the failure. */ - if (__halt(irq_disabled, do_sti)) + if (__halt(irq_disabled)) WARN_ONCE(1, "HLT instruction emulation failed\n"); } +static void __cpuidle tdx_safe_halt(void) +{ + tdx_halt(); + /* + * "__cpuidle" section doesn't support instrumentation, so stick + * with raw_* variant that avoids tracing hooks. + */ + raw_local_irq_enable(); +} + static int read_msr(struct pt_regs *regs, struct ve_info *ve) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_MSR_READ), .r12 = regs->cx, @@ -249,7 +481,7 @@ static int read_msr(struct pt_regs *regs, struct ve_info *ve) * can be found in TDX Guest-Host-Communication Interface * (GHCI), section titled "TDG.VP.VMCALL<Instruction.RDMSR>". */ - if (__tdx_hypercall(&args, TDX_HCALL_HAS_OUTPUT)) + if (__tdx_hypercall(&args)) return -EIO; regs->ax = lower_32_bits(args.r11); @@ -259,7 +491,7 @@ static int read_msr(struct pt_regs *regs, struct ve_info *ve) static int write_msr(struct pt_regs *regs, struct ve_info *ve) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_MSR_WRITE), .r12 = regs->cx, @@ -271,7 +503,7 @@ static int write_msr(struct pt_regs *regs, struct ve_info *ve) * can be found in TDX Guest-Host-Communication Interface * (GHCI) section titled "TDG.VP.VMCALL<Instruction.WRMSR>". */ - if (__tdx_hypercall(&args, 0)) + if (__tdx_hypercall(&args)) return -EIO; return ve_instr_len(ve); @@ -279,7 +511,7 @@ static int write_msr(struct pt_regs *regs, struct ve_info *ve) static int handle_cpuid(struct pt_regs *regs, struct ve_info *ve) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_CPUID), .r12 = regs->ax, @@ -303,7 +535,7 @@ static int handle_cpuid(struct pt_regs *regs, struct ve_info *ve) * ABI can be found in TDX Guest-Host-Communication Interface * (GHCI), section titled "VP.VMCALL<Instruction.CPUID>". */ - if (__tdx_hypercall(&args, TDX_HCALL_HAS_OUTPUT)) + if (__tdx_hypercall(&args)) return -EIO; /* @@ -321,17 +553,17 @@ static int handle_cpuid(struct pt_regs *regs, struct ve_info *ve) static bool mmio_read(int size, unsigned long addr, unsigned long *val) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_EPT_VIOLATION), .r12 = size, .r13 = EPT_READ, .r14 = addr, - .r15 = *val, }; - if (__tdx_hypercall(&args, TDX_HCALL_HAS_OUTPUT)) + if (__tdx_hypercall(&args)) return false; + *val = args.r11; return true; } @@ -346,8 +578,8 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve) { unsigned long *reg, val, vaddr; char buffer[MAX_INSN_SIZE]; + enum insn_mmio_type mmio; struct insn insn = {}; - enum mmio_type mmio; int size, extend_size; u8 extend_val = 0; @@ -362,15 +594,20 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve) return -EINVAL; mmio = insn_decode_mmio(&insn, &size); - if (WARN_ON_ONCE(mmio == MMIO_DECODE_FAILED)) + if (WARN_ON_ONCE(mmio == INSN_MMIO_DECODE_FAILED)) return -EINVAL; - if (mmio != MMIO_WRITE_IMM && mmio != MMIO_MOVS) { + if (mmio != INSN_MMIO_WRITE_IMM && mmio != INSN_MMIO_MOVS) { reg = insn_get_modrm_reg_ptr(&insn, regs); if (!reg) return -EINVAL; } + if (!fault_in_kernel_space(ve->gla)) { + WARN_ONCE(1, "Access to userspace address is not supported"); + return -EINVAL; + } + /* * Reject EPT violation #VEs that split pages. * @@ -386,23 +623,23 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve) /* Handle writes first */ switch (mmio) { - case MMIO_WRITE: + case INSN_MMIO_WRITE: memcpy(&val, reg, size); if (!mmio_write(size, ve->gpa, val)) return -EIO; return insn.length; - case MMIO_WRITE_IMM: + case INSN_MMIO_WRITE_IMM: val = insn.immediate.value; if (!mmio_write(size, ve->gpa, val)) return -EIO; return insn.length; - case MMIO_READ: - case MMIO_READ_ZERO_EXTEND: - case MMIO_READ_SIGN_EXTEND: + case INSN_MMIO_READ: + case INSN_MMIO_READ_ZERO_EXTEND: + case INSN_MMIO_READ_SIGN_EXTEND: /* Reads are handled below */ break; - case MMIO_MOVS: - case MMIO_DECODE_FAILED: + case INSN_MMIO_MOVS: + case INSN_MMIO_DECODE_FAILED: /* * MMIO was accessed with an instruction that could not be * decoded or handled properly. It was likely not using io.h @@ -419,15 +656,15 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve) return -EIO; switch (mmio) { - case MMIO_READ: + case INSN_MMIO_READ: /* Zero-extend for 32-bit operation */ extend_size = size == 4 ? sizeof(*reg) : 0; break; - case MMIO_READ_ZERO_EXTEND: + case INSN_MMIO_READ_ZERO_EXTEND: /* Zero extend based on operand size */ extend_size = insn.opnd_bytes; break; - case MMIO_READ_SIGN_EXTEND: + case INSN_MMIO_READ_SIGN_EXTEND: /* Sign extend based on operand size */ extend_size = insn.opnd_bytes; if (size == 1 && val & BIT(7)) @@ -449,7 +686,7 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve) static bool handle_in(struct pt_regs *regs, int size, int port) { - struct tdx_hypercall_args args = { + struct tdx_module_args args = { .r10 = TDX_HYPERCALL_STANDARD, .r11 = hcall_func(EXIT_REASON_IO_INSTRUCTION), .r12 = size, @@ -464,7 +701,7 @@ static bool handle_in(struct pt_regs *regs, int size, int port) * in TDX Guest-Host-Communication Interface (GHCI) section titled * "TDG.VP.VMCALL<Instruction.IO>". */ - success = !__tdx_hypercall(&args, TDX_HCALL_HAS_OUTPUT); + success = !__tdx_hypercall(&args); /* Update part of the register affected by the emulated instruction */ regs->ax &= ~mask; @@ -543,7 +780,7 @@ __init bool tdx_early_handle_ve(struct pt_regs *regs) void tdx_get_ve_info(struct ve_info *ve) { - struct tdx_module_output out; + struct tdx_module_args args = {}; /* * Called during #VE handling to retrieve the #VE info from the @@ -560,15 +797,15 @@ void tdx_get_ve_info(struct ve_info *ve) * Note, the TDX module treats virtual NMIs as inhibited if the #VE * valid flag is set. It means that NMI=>#VE will not result in a #DF. */ - tdx_module_call(TDX_GET_VEINFO, 0, 0, 0, 0, &out); + tdcall(TDG_VP_VEINFO_GET, &args); /* Transfer the output parameters */ - ve->exit_reason = out.rcx; - ve->exit_qual = out.rdx; - ve->gla = out.r8; - ve->gpa = out.r9; - ve->instr_len = lower_32_bits(out.r10); - ve->instr_info = upper_32_bits(out.r10); + ve->exit_reason = args.rcx; + ve->exit_qual = args.rdx; + ve->gla = args.r8; + ve->gpa = args.r9; + ve->instr_len = lower_32_bits(args.r10); + ve->instr_info = upper_32_bits(args.r10); } /* @@ -588,6 +825,11 @@ static int virt_exception_user(struct pt_regs *regs, struct ve_info *ve) } } +static inline bool is_private_gpa(u64 gpa) +{ + return gpa == cc_mkenc(gpa); +} + /* * Handle the kernel #VE. * @@ -606,6 +848,8 @@ static int virt_exception_kernel(struct pt_regs *regs, struct ve_info *ve) case EXIT_REASON_CPUID: return handle_cpuid(regs, ve); case EXIT_REASON_EPT_VIOLATION: + if (is_private_gpa(ve->gpa)) + panic("Unexpected EPT-violation on private memory."); return handle_mmio(regs, ve); case EXIT_REASON_IO_INSTRUCTION: return handle_io(regs, ve); @@ -661,45 +905,55 @@ static bool tdx_cache_flush_required(void) return true; } -static bool try_accept_one(phys_addr_t *start, unsigned long len, - enum pg_level pg_level) +/* + * Notify the VMM about page mapping conversion. More info about ABI + * can be found in TDX Guest-Host-Communication Interface (GHCI), + * section "TDG.VP.VMCALL<MapGPA>". + */ +static bool tdx_map_gpa(phys_addr_t start, phys_addr_t end, bool enc) { - unsigned long accept_size = page_level_size(pg_level); - u64 tdcall_rcx; - u8 page_size; + /* Retrying the hypercall a second time should succeed; use 3 just in case */ + const int max_retries_per_page = 3; + int retry_count = 0; - if (!IS_ALIGNED(*start, accept_size)) - return false; + if (!enc) { + /* Set the shared (decrypted) bits: */ + start |= cc_mkdec(0); + end |= cc_mkdec(0); + } - if (len < accept_size) - return false; + while (retry_count < max_retries_per_page) { + struct tdx_module_args args = { + .r10 = TDX_HYPERCALL_STANDARD, + .r11 = TDVMCALL_MAP_GPA, + .r12 = start, + .r13 = end - start }; - /* - * Pass the page physical address to the TDX module to accept the - * pending, private page. - * - * Bits 2:0 of RCX encode page size: 0 - 4K, 1 - 2M, 2 - 1G. - */ - switch (pg_level) { - case PG_LEVEL_4K: - page_size = 0; - break; - case PG_LEVEL_2M: - page_size = 1; - break; - case PG_LEVEL_1G: - page_size = 2; - break; - default: - return false; - } + u64 map_fail_paddr; + u64 ret = __tdx_hypercall(&args); - tdcall_rcx = *start | page_size; - if (__tdx_module_call(TDX_ACCEPT_PAGE, tdcall_rcx, 0, 0, 0, NULL)) - return false; + if (ret != TDVMCALL_STATUS_RETRY) + return !ret; + /* + * The guest must retry the operation for the pages in the + * region starting at the GPA specified in R11. R11 comes + * from the untrusted VMM. Sanity check it. + */ + map_fail_paddr = args.r11; + if (map_fail_paddr < start || map_fail_paddr >= end) + return false; - *start += accept_size; - return true; + /* "Consume" a retry without forward progress */ + if (map_fail_paddr == start) { + retry_count++; + continue; + } + + start = map_fail_paddr; + retry_count = 0; + } + + return false; } /* @@ -712,48 +966,148 @@ static bool tdx_enc_status_changed(unsigned long vaddr, int numpages, bool enc) phys_addr_t start = __pa(vaddr); phys_addr_t end = __pa(vaddr + numpages * PAGE_SIZE); - if (!enc) { - /* Set the shared (decrypted) bits: */ - start |= cc_mkdec(0); - end |= cc_mkdec(0); - } + if (!tdx_map_gpa(start, end, enc)) + return false; + /* shared->private conversion requires memory to be accepted before use */ + if (enc) + return tdx_accept_memory(start, end); + + return true; +} + +static int tdx_enc_status_change_prepare(unsigned long vaddr, int numpages, + bool enc) +{ /* - * Notify the VMM about page mapping conversion. More info about ABI - * can be found in TDX Guest-Host-Communication Interface (GHCI), - * section "TDG.VP.VMCALL<MapGPA>" + * Only handle shared->private conversion here. + * See the comment in tdx_early_init(). */ - if (_tdx_hypercall(TDVMCALL_MAP_GPA, start, end - start, 0, 0)) - return false; + if (enc && !tdx_enc_status_changed(vaddr, numpages, enc)) + return -EIO; - /* private->shared conversion requires only MapGPA call */ - if (!enc) - return true; + return 0; +} +static int tdx_enc_status_change_finish(unsigned long vaddr, int numpages, + bool enc) +{ /* - * For shared->private conversion, accept the page using - * TDX_ACCEPT_PAGE TDX module call. + * Only handle private->shared conversion here. + * See the comment in tdx_early_init(). */ - while (start < end) { - unsigned long len = end - start; + if (!enc && !tdx_enc_status_changed(vaddr, numpages, enc)) + return -EIO; - /* - * Try larger accepts first. It gives chance to VMM to keep - * 1G/2M SEPT entries where possible and speeds up process by - * cutting number of hypercalls (if successful). - */ + if (enc) + atomic_long_sub(numpages, &nr_shared); + else + atomic_long_add(numpages, &nr_shared); - if (try_accept_one(&start, len, PG_LEVEL_1G)) - continue; + return 0; +} - if (try_accept_one(&start, len, PG_LEVEL_2M)) - continue; +/* Stop new private<->shared conversions */ +static void tdx_kexec_begin(void) +{ + if (!IS_ENABLED(CONFIG_KEXEC_CORE)) + return; - if (!try_accept_one(&start, len, PG_LEVEL_4K)) - return false; + /* + * Crash kernel reaches here with interrupts disabled: can't wait for + * conversions to finish. + * + * If race happened, just report and proceed. + */ + if (!set_memory_enc_stop_conversion()) + pr_warn("Failed to stop shared<->private conversions\n"); +} + +/* Walk direct mapping and convert all shared memory back to private */ +static void tdx_kexec_finish(void) +{ + unsigned long addr, end; + long found = 0, shared; + + if (!IS_ENABLED(CONFIG_KEXEC_CORE)) + return; + + lockdep_assert_irqs_disabled(); + + addr = PAGE_OFFSET; + end = PAGE_OFFSET + get_max_mapped(); + + while (addr < end) { + unsigned long size; + unsigned int level; + pte_t *pte; + + pte = lookup_address(addr, &level); + size = page_level_size(level); + + if (pte && pte_decrypted(*pte)) { + int pages = size / PAGE_SIZE; + + /* + * Touching memory with shared bit set triggers implicit + * conversion to shared. + * + * Make sure nobody touches the shared range from + * now on. + */ + set_pte(pte, __pte(0)); + + /* + * Memory encryption state persists across kexec. + * If tdx_enc_status_changed() fails in the first + * kernel, it leaves memory in an unknown state. + * + * If that memory remains shared, accessing it in the + * *next* kernel through a private mapping will result + * in an unrecoverable guest shutdown. + * + * The kdump kernel boot is not impacted as it uses + * a pre-reserved memory range that is always private. + * However, gathering crash information could lead to + * a crash if it accesses unconverted memory through + * a private mapping which is possible when accessing + * that memory through /proc/vmcore, for example. + * + * In all cases, print error info in order to leave + * enough bread crumbs for debugging. + */ + if (!tdx_enc_status_changed(addr, pages, true)) { + pr_err("Failed to unshare range %#lx-%#lx\n", + addr, addr + size); + } + + found += pages; + } + + addr += size; } - return true; + __flush_tlb_all(); + + shared = atomic_long_read(&nr_shared); + if (shared != found) { + pr_err("shared page accounting is off\n"); + pr_err("nr_shared = %ld, nr_found = %ld\n", shared, found); + } +} + +static __init void tdx_announce(void) +{ + struct tdx_module_args args = {}; + u64 controls; + + pr_info("Guest detected\n"); + + tdcall(TDG_VP_INFO, &args); + tdx_dump_attributes(args.rdx); + + tdg_vm_rd(TDCS_TD_CTLS, &controls); + tdx_dump_td_ctls(controls); } void __init tdx_early_init(void) @@ -768,8 +1122,14 @@ void __init tdx_early_init(void) setup_force_cpu_cap(X86_FEATURE_TDX_GUEST); - cc_set_vendor(CC_VENDOR_INTEL); - tdx_parse_tdinfo(&cc_mask); + /* TSC is the only reliable clock in TDX guest */ + setup_force_cpu_cap(X86_FEATURE_TSC_RELIABLE); + + cc_vendor = CC_VENDOR_INTEL; + + /* Configure the TD */ + tdx_setup(&cc_mask); + cc_set_mask(cc_mask); /* @@ -780,9 +1140,57 @@ void __init tdx_early_init(void) */ physical_mask &= cc_mask - 1; - x86_platform.guest.enc_cache_flush_required = tdx_cache_flush_required; - x86_platform.guest.enc_tlb_flush_required = tdx_tlb_flush_required; - x86_platform.guest.enc_status_change_finish = tdx_enc_status_changed; + /* + * The kernel mapping should match the TDX metadata for the page. + * load_unaligned_zeropad() can touch memory *adjacent* to that which is + * owned by the caller and can catch even _momentary_ mismatches. Bad + * things happen on mismatch: + * + * - Private mapping => Shared Page == Guest shutdown + * - Shared mapping => Private Page == Recoverable #VE + * + * guest.enc_status_change_prepare() converts the page from + * shared=>private before the mapping becomes private. + * + * guest.enc_status_change_finish() converts the page from + * private=>shared after the mapping becomes private. + * + * In both cases there is a temporary shared mapping to a private page, + * which can result in a #VE. But, there is never a private mapping to + * a shared page. + */ + x86_platform.guest.enc_status_change_prepare = tdx_enc_status_change_prepare; + x86_platform.guest.enc_status_change_finish = tdx_enc_status_change_finish; + + x86_platform.guest.enc_cache_flush_required = tdx_cache_flush_required; + x86_platform.guest.enc_tlb_flush_required = tdx_tlb_flush_required; - pr_info("Guest detected\n"); + x86_platform.guest.enc_kexec_begin = tdx_kexec_begin; + x86_platform.guest.enc_kexec_finish = tdx_kexec_finish; + + /* + * Avoid "sti;hlt" execution in TDX guests as HLT induces a #VE that + * will enable interrupts before HLT TDCALL invocation if executed + * in STI-shadow, possibly resulting in missed wakeup events. + * + * Modify all possible HLT execution paths to use TDX specific routines + * that directly execute TDCALL and toggle the interrupt state as + * needed after TDCALL completion. This also reduces HLT related #VEs + * in addition to having a reliable halt logic execution. + */ + pv_ops.irq.safe_halt = tdx_safe_halt; + pv_ops.irq.halt = tdx_halt; + + /* + * TDX intercepts the RDMSR to read the X2APIC ID in the parallel + * bringup low level code. That raises #VE which cannot be handled + * there. + * + * Intel-TDX has a secure RDMSR hypercall, but that needs to be + * implemented separately in the low level startup ASM code. + * Until that is in place, disable parallel bringup for TDX. + */ + x86_cpuinit.parallel_bringup = false; + + tdx_announce(); } |