Two small bugfixes, plus most of RTH's refactoring of cpregs handling. -- PMM The following changes since commit 1fba9dc71a170b3a05b9d3272dd8ecfe7f26e215: Merge tag 'pull-request-2022-05-04' of https://gitlab.com/thuth/qemu into staging (2022-05-04 08:07:02 -0700) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20220505 for you to fetch changes up to 99a50d1a67c602126fc2b3a4812d3000eba9bf34: target/arm: read access to performance counters from EL0 (2022-05-05 09:36:22 +0100) ---------------------------------------------------------------- target-arm queue: * Enable read access to performance counters from EL0 * Enable SCTLR_EL1.BT0 for aarch64-linux-user * Refactoring of cpreg handling ---------------------------------------------------------------- Alex Zuepke (1): target/arm: read access to performance counters from EL0 Richard Henderson (22): target/arm: Enable SCTLR_EL1.BT0 for aarch64-linux-user target/arm: Split out cpregs.h target/arm: Reorg CPAccessResult and access_check_cp_reg target/arm: Replace sentinels with ARRAY_SIZE in cpregs.h target/arm: Make some more cpreg data static const target/arm: Reorg ARMCPRegInfo type field bits target/arm: Avoid bare abort() or assert(0) target/arm: Change cpreg access permissions to enum target/arm: Name CPState type target/arm: Name CPSecureState type target/arm: Drop always-true test in define_arm_vh_e2h_redirects_aliases target/arm: Store cpregs key in the hash table directly target/arm: Merge allocation of the cpreg and its name target/arm: Hoist computation of key in add_cpreg_to_hashtable target/arm: Consolidate cpreg updates in add_cpreg_to_hashtable target/arm: Use bool for is64 and ns in add_cpreg_to_hashtable target/arm: Hoist isbanked computation in add_cpreg_to_hashtable target/arm: Perform override check early in add_cpreg_to_hashtable target/arm: Reformat comments in add_cpreg_to_hashtable target/arm: Remove HOST_BIG_ENDIAN ifdef in add_cpreg_to_hashtable target/arm: Add isar predicates for FEAT_Debugv8p2 target/arm: Add isar_feature_{aa64,any}_ras target/arm/cpregs.h | 453 ++++++++++++++++++++++++++++++++++++++ target/arm/cpu.h | 393 +++------------------------------ hw/arm/pxa2xx.c | 2 +- hw/arm/pxa2xx_pic.c | 2 +- hw/intc/arm_gicv3_cpuif.c | 6 +- hw/intc/arm_gicv3_kvm.c | 3 +- target/arm/cpu.c | 25 +-- target/arm/cpu64.c | 2 +- target/arm/cpu_tcg.c | 5 +- target/arm/gdbstub.c | 5 +- target/arm/helper.c | 358 +++++++++++++----------------- target/arm/hvf/hvf.c | 2 +- target/arm/kvm-stub.c | 4 +- target/arm/kvm.c | 4 +- target/arm/machine.c | 4 +- target/arm/op_helper.c | 57 ++--- target/arm/translate-a64.c | 14 +- target/arm/translate-neon.c | 2 +- target/arm/translate.c | 13 +- tests/tcg/aarch64/bti-3.c | 42 ++++ tests/tcg/aarch64/Makefile.target | 6 +- 21 files changed, 738 insertions(+), 664 deletions(-) create mode 100644 target/arm/cpregs.h create mode 100644 tests/tcg/aarch64/bti-3.c From: Richard Henderson <richard.henderson@linaro.org> This controls whether the PACI{A,B}SP instructions trap with BTYPE=3 (indirect branch from register other than x16/x17). The linux kernel sets this in bti_enable(). Resolves: https://gitlab.com/qemu-project/qemu/-/issues/998 Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220427042312.294300-1-richard.henderson@linaro.org [PMM: remove stray change to makefile comment] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 2 ++ tests/tcg/aarch64/bti-3.c | 42 +++++++++++++++++++++++++++++++ tests/tcg/aarch64/Makefile.target | 6 ++--- 3 files changed, 47 insertions(+), 3 deletions(-) create mode 100644 tests/tcg/aarch64/bti-3.c diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(DeviceState *dev) /* Enable all PAC keys. */ env->cp15.sctlr_el[1] |= (SCTLR_EnIA | SCTLR_EnIB | SCTLR_EnDA | SCTLR_EnDB); + /* Trap on btype=3 for PACIxSP. */ + env->cp15.sctlr_el[1] |= SCTLR_BT0; /* and to the FP/Neon instructions */ env->cp15.cpacr_el1 = deposit64(env->cp15.cpacr_el1, 20, 2, 3); /* and to the SVE instructions */ diff --git a/tests/tcg/aarch64/bti-3.c b/tests/tcg/aarch64/bti-3.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/tcg/aarch64/bti-3.c @@ -XXX,XX +XXX,XX @@ +/* + * BTI vs PACIASP + */ + +#include "bti-crt.inc.c" + +static void skip2_sigill(int sig, siginfo_t *info, ucontext_t *uc) +{ + uc->uc_mcontext.pc += 8; + uc->uc_mcontext.pstate = 1; +} + +#define BTYPE_1() \ + asm("mov %0,#1; adr x16, 1f; br x16; 1: hint #25; mov %0,#0" \ + : "=r"(skipped) : : "x16", "x30") + +#define BTYPE_2() \ + asm("mov %0,#1; adr x16, 1f; blr x16; 1: hint #25; mov %0,#0" \ + : "=r"(skipped) : : "x16", "x30") + +#define BTYPE_3() \ + asm("mov %0,#1; adr x15, 1f; br x15; 1: hint #25; mov %0,#0" \ + : "=r"(skipped) : : "x15", "x30") + +#define TEST(WHICH, EXPECT) \ + do { WHICH(); fail += skipped ^ EXPECT; } while (0) + +int main() +{ + int fail = 0; + int skipped; + + /* Signal-like with SA_SIGINFO. */ + signal_info(SIGILL, skip2_sigill); + + /* With SCTLR_EL1.BT0 set, PACIASP is not compatible with type=3. */ + TEST(BTYPE_1, 0); + TEST(BTYPE_2, 0); + TEST(BTYPE_3, 1); + + return fail; +} diff --git a/tests/tcg/aarch64/Makefile.target b/tests/tcg/aarch64/Makefile.target index XXXXXXX..XXXXXXX 100644 --- a/tests/tcg/aarch64/Makefile.target +++ b/tests/tcg/aarch64/Makefile.target @@ -XXX,XX +XXX,XX @@ endif # BTI Tests # bti-1 tests the elf notes, so we require special compiler support. ifneq ($(CROSS_CC_HAS_ARMV8_BTI),) -AARCH64_TESTS += bti-1 -bti-1: CFLAGS += -mbranch-protection=standard -bti-1: LDFLAGS += -nostdlib +AARCH64_TESTS += bti-1 bti-3 +bti-1 bti-3: CFLAGS += -mbranch-protection=standard +bti-1 bti-3: LDFLAGS += -nostdlib endif # bti-2 tests PROT_BTI, so no special compiler support required. AARCH64_TESTS += bti-2 -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Move ARMCPRegInfo and all related declarations to a new internal header, out of the public cpu.h. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 413 +++++++++++++++++++++++++++++++++++++ target/arm/cpu.h | 368 --------------------------------- hw/arm/pxa2xx.c | 1 + hw/arm/pxa2xx_pic.c | 1 + hw/intc/arm_gicv3_cpuif.c | 1 + hw/intc/arm_gicv3_kvm.c | 2 + target/arm/cpu.c | 1 + target/arm/cpu64.c | 1 + target/arm/cpu_tcg.c | 1 + target/arm/gdbstub.c | 3 +- target/arm/helper.c | 1 + target/arm/op_helper.c | 1 + target/arm/translate-a64.c | 4 +- target/arm/translate.c | 3 +- 14 files changed, 427 insertions(+), 374 deletions(-) create mode 100644 target/arm/cpregs.h diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU ARM CP Register access and descriptions + * + * Copyright (c) 2022 Linaro Ltd + * + * 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/gpl-2.0.html> + */ + +#ifndef TARGET_ARM_CPREGS_H +#define TARGET_ARM_CPREGS_H + +/* + * ARMCPRegInfo type field bits. If the SPECIAL bit is set this is a + * special-behaviour cp reg and bits [11..8] indicate what behaviour + * it has. Otherwise it is a simple cp reg, where CONST indicates that + * TCG can assume the value to be constant (ie load at translate time) + * and 64BIT indicates a 64 bit wide coprocessor register. SUPPRESS_TB_END + * indicates that the TB should not be ended after a write to this register + * (the default is that the TB ends after cp writes). OVERRIDE permits + * a register definition to override a previous definition for the + * same (cp, is64, crn, crm, opc1, opc2) tuple: either the new or the + * old must have the OVERRIDE bit set. + * ALIAS indicates that this register is an alias view of some underlying + * state which is also visible via another register, and that the other + * register is handling migration and reset; registers marked ALIAS will not be + * migrated but may have their state set by syncing of register state from KVM. + * NO_RAW indicates that this register has no underlying state and does not + * support raw access for state saving/loading; it will not be used for either + * migration or KVM state synchronization. (Typically this is for "registers" + * which are actually used as instructions for cache maintenance and so on.) + * IO indicates that this register does I/O and therefore its accesses + * need to be marked with gen_io_start() and also end the TB. In particular, + * registers which implement clocks or timers require this. + * RAISES_EXC is for when the read or write hook might raise an exception; + * the generated code will synchronize the CPU state before calling the hook + * so that it is safe for the hook to call raise_exception(). + * NEWEL is for writes to registers that might change the exception + * level - typically on older ARM chips. For those cases we need to + * re-read the new el when recomputing the translation flags. + */ +#define ARM_CP_SPECIAL 0x0001 +#define ARM_CP_CONST 0x0002 +#define ARM_CP_64BIT 0x0004 +#define ARM_CP_SUPPRESS_TB_END 0x0008 +#define ARM_CP_OVERRIDE 0x0010 +#define ARM_CP_ALIAS 0x0020 +#define ARM_CP_IO 0x0040 +#define ARM_CP_NO_RAW 0x0080 +#define ARM_CP_NOP (ARM_CP_SPECIAL | 0x0100) +#define ARM_CP_WFI (ARM_CP_SPECIAL | 0x0200) +#define ARM_CP_NZCV (ARM_CP_SPECIAL | 0x0300) +#define ARM_CP_CURRENTEL (ARM_CP_SPECIAL | 0x0400) +#define ARM_CP_DC_ZVA (ARM_CP_SPECIAL | 0x0500) +#define ARM_CP_DC_GVA (ARM_CP_SPECIAL | 0x0600) +#define ARM_CP_DC_GZVA (ARM_CP_SPECIAL | 0x0700) +#define ARM_LAST_SPECIAL ARM_CP_DC_GZVA +#define ARM_CP_FPU 0x1000 +#define ARM_CP_SVE 0x2000 +#define ARM_CP_NO_GDB 0x4000 +#define ARM_CP_RAISES_EXC 0x8000 +#define ARM_CP_NEWEL 0x10000 +/* Used only as a terminator for ARMCPRegInfo lists */ +#define ARM_CP_SENTINEL 0xfffff +/* Mask of only the flag bits in a type field */ +#define ARM_CP_FLAG_MASK 0x1f0ff + +/* + * Valid values for ARMCPRegInfo state field, indicating which of + * the AArch32 and AArch64 execution states this register is visible in. + * If the reginfo doesn't explicitly specify then it is AArch32 only. + * If the reginfo is declared to be visible in both states then a second + * reginfo is synthesised for the AArch32 view of the AArch64 register, + * such that the AArch32 view is the lower 32 bits of the AArch64 one. + * Note that we rely on the values of these enums as we iterate through + * the various states in some places. + */ +enum { + ARM_CP_STATE_AA32 = 0, + ARM_CP_STATE_AA64 = 1, + ARM_CP_STATE_BOTH = 2, +}; + +/* + * ARM CP register secure state flags. These flags identify security state + * attributes for a given CP register entry. + * The existence of both or neither secure and non-secure flags indicates that + * the register has both a secure and non-secure hash entry. A single one of + * these flags causes the register to only be hashed for the specified + * security state. + * Although definitions may have any combination of the S/NS bits, each + * registered entry will only have one to identify whether the entry is secure + * or non-secure. + */ +enum { + ARM_CP_SECSTATE_S = (1 << 0), /* bit[0]: Secure state register */ + ARM_CP_SECSTATE_NS = (1 << 1), /* bit[1]: Non-secure state register */ +}; + +/* + * Return true if cptype is a valid type field. This is used to try to + * catch errors where the sentinel has been accidentally left off the end + * of a list of registers. + */ +static inline bool cptype_valid(int cptype) +{ + return ((cptype & ~ARM_CP_FLAG_MASK) == 0) + || ((cptype & ARM_CP_SPECIAL) && + ((cptype & ~ARM_CP_FLAG_MASK) <= ARM_LAST_SPECIAL)); +} + +/* + * Access rights: + * We define bits for Read and Write access for what rev C of the v7-AR ARM ARM + * defines as PL0 (user), PL1 (fiq/irq/svc/abt/und/sys, ie privileged), and + * PL2 (hyp). The other level which has Read and Write bits is Secure PL1 + * (ie any of the privileged modes in Secure state, or Monitor mode). + * If a register is accessible in one privilege level it's always accessible + * in higher privilege levels too. Since "Secure PL1" also follows this rule + * (ie anything visible in PL2 is visible in S-PL1, some things are only + * visible in S-PL1) but "Secure PL1" is a bit of a mouthful, we bend the + * terminology a little and call this PL3. + * In AArch64 things are somewhat simpler as the PLx bits line up exactly + * with the ELx exception levels. + * + * If access permissions for a register are more complex than can be + * described with these bits, then use a laxer set of restrictions, and + * do the more restrictive/complex check inside a helper function. + */ +#define PL3_R 0x80 +#define PL3_W 0x40 +#define PL2_R (0x20 | PL3_R) +#define PL2_W (0x10 | PL3_W) +#define PL1_R (0x08 | PL2_R) +#define PL1_W (0x04 | PL2_W) +#define PL0_R (0x02 | PL1_R) +#define PL0_W (0x01 | PL1_W) + +/* + * For user-mode some registers are accessible to EL0 via a kernel + * trap-and-emulate ABI. In this case we define the read permissions + * as actually being PL0_R. However some bits of any given register + * may still be masked. + */ +#ifdef CONFIG_USER_ONLY +#define PL0U_R PL0_R +#else +#define PL0U_R PL1_R +#endif + +#define PL3_RW (PL3_R | PL3_W) +#define PL2_RW (PL2_R | PL2_W) +#define PL1_RW (PL1_R | PL1_W) +#define PL0_RW (PL0_R | PL0_W) + +typedef enum CPAccessResult { + /* Access is permitted */ + CP_ACCESS_OK = 0, + /* + * Access fails due to a configurable trap or enable which would + * result in a categorized exception syndrome giving information about + * the failing instruction (ie syndrome category 0x3, 0x4, 0x5, 0x6, + * 0xc or 0x18). The exception is taken to the usual target EL (EL1 or + * PL1 if in EL0, otherwise to the current EL). + */ + CP_ACCESS_TRAP = 1, + /* + * Access fails and results in an exception syndrome 0x0 ("uncategorized"). + * Note that this is not a catch-all case -- the set of cases which may + * result in this failure is specifically defined by the architecture. + */ + CP_ACCESS_TRAP_UNCATEGORIZED = 2, + /* As CP_ACCESS_TRAP, but for traps directly to EL2 or EL3 */ + CP_ACCESS_TRAP_EL2 = 3, + CP_ACCESS_TRAP_EL3 = 4, + /* As CP_ACCESS_UNCATEGORIZED, but for traps directly to EL2 or EL3 */ + CP_ACCESS_TRAP_UNCATEGORIZED_EL2 = 5, + CP_ACCESS_TRAP_UNCATEGORIZED_EL3 = 6, +} CPAccessResult; + +typedef struct ARMCPRegInfo ARMCPRegInfo; + +/* + * Access functions for coprocessor registers. These cannot fail and + * may not raise exceptions. + */ +typedef uint64_t CPReadFn(CPUARMState *env, const ARMCPRegInfo *opaque); +typedef void CPWriteFn(CPUARMState *env, const ARMCPRegInfo *opaque, + uint64_t value); +/* Access permission check functions for coprocessor registers. */ +typedef CPAccessResult CPAccessFn(CPUARMState *env, + const ARMCPRegInfo *opaque, + bool isread); +/* Hook function for register reset */ +typedef void CPResetFn(CPUARMState *env, const ARMCPRegInfo *opaque); + +#define CP_ANY 0xff + +/* Definition of an ARM coprocessor register */ +struct ARMCPRegInfo { + /* Name of register (useful mainly for debugging, need not be unique) */ + const char *name; + /* + * Location of register: coprocessor number and (crn,crm,opc1,opc2) + * tuple. Any of crm, opc1 and opc2 may be CP_ANY to indicate a + * 'wildcard' field -- any value of that field in the MRC/MCR insn + * will be decoded to this register. The register read and write + * callbacks will be passed an ARMCPRegInfo with the crn/crm/opc1/opc2 + * used by the program, so it is possible to register a wildcard and + * then behave differently on read/write if necessary. + * For 64 bit registers, only crm and opc1 are relevant; crn and opc2 + * must both be zero. + * For AArch64-visible registers, opc0 is also used. + * Since there are no "coprocessors" in AArch64, cp is purely used as a + * way to distinguish (for KVM's benefit) guest-visible system registers + * from demuxed ones provided to preserve the "no side effects on + * KVM register read/write from QEMU" semantics. cp==0x13 is guest + * visible (to match KVM's encoding); cp==0 will be converted to + * cp==0x13 when the ARMCPRegInfo is registered, for convenience. + */ + uint8_t cp; + uint8_t crn; + uint8_t crm; + uint8_t opc0; + uint8_t opc1; + uint8_t opc2; + /* Execution state in which this register is visible: ARM_CP_STATE_* */ + int state; + /* Register type: ARM_CP_* bits/values */ + int type; + /* Access rights: PL*_[RW] */ + int access; + /* Security state: ARM_CP_SECSTATE_* bits/values */ + int secure; + /* + * The opaque pointer passed to define_arm_cp_regs_with_opaque() when + * this register was defined: can be used to hand data through to the + * register read/write functions, since they are passed the ARMCPRegInfo*. + */ + void *opaque; + /* + * Value of this register, if it is ARM_CP_CONST. Otherwise, if + * fieldoffset is non-zero, the reset value of the register. + */ + uint64_t resetvalue; + /* + * Offset of the field in CPUARMState for this register. + * This is not needed if either: + * 1. type is ARM_CP_CONST or one of the ARM_CP_SPECIALs + * 2. both readfn and writefn are specified + */ + ptrdiff_t fieldoffset; /* offsetof(CPUARMState, field) */ + + /* + * Offsets of the secure and non-secure fields in CPUARMState for the + * register if it is banked. These fields are only used during the static + * registration of a register. During hashing the bank associated + * with a given security state is copied to fieldoffset which is used from + * there on out. + * + * It is expected that register definitions use either fieldoffset or + * bank_fieldoffsets in the definition but not both. It is also expected + * that both bank offsets are set when defining a banked register. This + * use indicates that a register is banked. + */ + ptrdiff_t bank_fieldoffsets[2]; + + /* + * Function for making any access checks for this register in addition to + * those specified by the 'access' permissions bits. If NULL, no extra + * checks required. The access check is performed at runtime, not at + * translate time. + */ + CPAccessFn *accessfn; + /* + * Function for handling reads of this register. If NULL, then reads + * will be done by loading from the offset into CPUARMState specified + * by fieldoffset. + */ + CPReadFn *readfn; + /* + * Function for handling writes of this register. If NULL, then writes + * will be done by writing to the offset into CPUARMState specified + * by fieldoffset. + */ + CPWriteFn *writefn; + /* + * Function for doing a "raw" read; used when we need to copy + * coprocessor state to the kernel for KVM or out for + * migration. This only needs to be provided if there is also a + * readfn and it has side effects (for instance clear-on-read bits). + */ + CPReadFn *raw_readfn; + /* + * Function for doing a "raw" write; used when we need to copy KVM + * kernel coprocessor state into userspace, or for inbound + * migration. This only needs to be provided if there is also a + * writefn and it masks out "unwritable" bits or has write-one-to-clear + * or similar behaviour. + */ + CPWriteFn *raw_writefn; + /* + * Function for resetting the register. If NULL, then reset will be done + * by writing resetvalue to the field specified in fieldoffset. If + * fieldoffset is 0 then no reset will be done. + */ + CPResetFn *resetfn; + + /* + * "Original" writefn and readfn. + * For ARMv8.1-VHE register aliases, we overwrite the read/write + * accessor functions of various EL1/EL0 to perform the runtime + * check for which sysreg should actually be modified, and then + * forwards the operation. Before overwriting the accessors, + * the original function is copied here, so that accesses that + * really do go to the EL1/EL0 version proceed normally. + * (The corresponding EL2 register is linked via opaque.) + */ + CPReadFn *orig_readfn; + CPWriteFn *orig_writefn; +}; + +/* + * Macros which are lvalues for the field in CPUARMState for the + * ARMCPRegInfo *ri. + */ +#define CPREG_FIELD32(env, ri) \ + (*(uint32_t *)((char *)(env) + (ri)->fieldoffset)) +#define CPREG_FIELD64(env, ri) \ + (*(uint64_t *)((char *)(env) + (ri)->fieldoffset)) + +#define REGINFO_SENTINEL { .type = ARM_CP_SENTINEL } + +void define_arm_cp_regs_with_opaque(ARMCPU *cpu, + const ARMCPRegInfo *regs, void *opaque); +void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, + const ARMCPRegInfo *regs, void *opaque); +static inline void define_arm_cp_regs(ARMCPU *cpu, const ARMCPRegInfo *regs) +{ + define_arm_cp_regs_with_opaque(cpu, regs, 0); +} +static inline void define_one_arm_cp_reg(ARMCPU *cpu, const ARMCPRegInfo *regs) +{ + define_one_arm_cp_reg_with_opaque(cpu, regs, 0); +} +const ARMCPRegInfo *get_arm_cp_reginfo(GHashTable *cpregs, uint32_t encoded_cp); + +/* + * Definition of an ARM co-processor register as viewed from + * userspace. This is used for presenting sanitised versions of + * registers to userspace when emulating the Linux AArch64 CPU + * ID/feature ABI (advertised as HWCAP_CPUID). + */ +typedef struct ARMCPRegUserSpaceInfo { + /* Name of register */ + const char *name; + + /* Is the name actually a glob pattern */ + bool is_glob; + + /* Only some bits are exported to user space */ + uint64_t exported_bits; + + /* Fixed bits are applied after the mask */ + uint64_t fixed_bits; +} ARMCPRegUserSpaceInfo; + +#define REGUSERINFO_SENTINEL { .name = NULL } + +void modify_arm_cp_regs(ARMCPRegInfo *regs, const ARMCPRegUserSpaceInfo *mods); + +/* CPWriteFn that can be used to implement writes-ignored behaviour */ +void arm_cp_write_ignore(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value); +/* CPReadFn that can be used for read-as-zero behaviour */ +uint64_t arm_cp_read_zero(CPUARMState *env, const ARMCPRegInfo *ri); + +/* + * CPResetFn that does nothing, for use if no reset is required even + * if fieldoffset is non zero. + */ +void arm_cp_reset_ignore(CPUARMState *env, const ARMCPRegInfo *opaque); + +/* + * Return true if this reginfo struct's field in the cpu state struct + * is 64 bits wide. + */ +static inline bool cpreg_field_is_64bit(const ARMCPRegInfo *ri) +{ + return (ri->state == ARM_CP_STATE_AA64) || (ri->type & ARM_CP_64BIT); +} + +static inline bool cp_access_ok(int current_el, + const ARMCPRegInfo *ri, int isread) +{ + return (ri->access >> ((current_el * 2) + isread)) & 1; +} + +/* Raw read of a coprocessor register (as needed for migration, etc) */ +uint64_t read_raw_cp_reg(CPUARMState *env, const ARMCPRegInfo *ri); + +#endif /* TARGET_ARM_CPREGS_H */ diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline uint64_t cpreg_to_kvm_id(uint32_t cpregid) return kvmid; } -/* ARMCPRegInfo type field bits. If the SPECIAL bit is set this is a - * special-behaviour cp reg and bits [11..8] indicate what behaviour - * it has. Otherwise it is a simple cp reg, where CONST indicates that - * TCG can assume the value to be constant (ie load at translate time) - * and 64BIT indicates a 64 bit wide coprocessor register. SUPPRESS_TB_END - * indicates that the TB should not be ended after a write to this register - * (the default is that the TB ends after cp writes). OVERRIDE permits - * a register definition to override a previous definition for the - * same (cp, is64, crn, crm, opc1, opc2) tuple: either the new or the - * old must have the OVERRIDE bit set. - * ALIAS indicates that this register is an alias view of some underlying - * state which is also visible via another register, and that the other - * register is handling migration and reset; registers marked ALIAS will not be - * migrated but may have their state set by syncing of register state from KVM. - * NO_RAW indicates that this register has no underlying state and does not - * support raw access for state saving/loading; it will not be used for either - * migration or KVM state synchronization. (Typically this is for "registers" - * which are actually used as instructions for cache maintenance and so on.) - * IO indicates that this register does I/O and therefore its accesses - * need to be marked with gen_io_start() and also end the TB. In particular, - * registers which implement clocks or timers require this. - * RAISES_EXC is for when the read or write hook might raise an exception; - * the generated code will synchronize the CPU state before calling the hook - * so that it is safe for the hook to call raise_exception(). - * NEWEL is for writes to registers that might change the exception - * level - typically on older ARM chips. For those cases we need to - * re-read the new el when recomputing the translation flags. - */ -#define ARM_CP_SPECIAL 0x0001 -#define ARM_CP_CONST 0x0002 -#define ARM_CP_64BIT 0x0004 -#define ARM_CP_SUPPRESS_TB_END 0x0008 -#define ARM_CP_OVERRIDE 0x0010 -#define ARM_CP_ALIAS 0x0020 -#define ARM_CP_IO 0x0040 -#define ARM_CP_NO_RAW 0x0080 -#define ARM_CP_NOP (ARM_CP_SPECIAL | 0x0100) -#define ARM_CP_WFI (ARM_CP_SPECIAL | 0x0200) -#define ARM_CP_NZCV (ARM_CP_SPECIAL | 0x0300) -#define ARM_CP_CURRENTEL (ARM_CP_SPECIAL | 0x0400) -#define ARM_CP_DC_ZVA (ARM_CP_SPECIAL | 0x0500) -#define ARM_CP_DC_GVA (ARM_CP_SPECIAL | 0x0600) -#define ARM_CP_DC_GZVA (ARM_CP_SPECIAL | 0x0700) -#define ARM_LAST_SPECIAL ARM_CP_DC_GZVA -#define ARM_CP_FPU 0x1000 -#define ARM_CP_SVE 0x2000 -#define ARM_CP_NO_GDB 0x4000 -#define ARM_CP_RAISES_EXC 0x8000 -#define ARM_CP_NEWEL 0x10000 -/* Used only as a terminator for ARMCPRegInfo lists */ -#define ARM_CP_SENTINEL 0xfffff -/* Mask of only the flag bits in a type field */ -#define ARM_CP_FLAG_MASK 0x1f0ff - -/* Valid values for ARMCPRegInfo state field, indicating which of - * the AArch32 and AArch64 execution states this register is visible in. - * If the reginfo doesn't explicitly specify then it is AArch32 only. - * If the reginfo is declared to be visible in both states then a second - * reginfo is synthesised for the AArch32 view of the AArch64 register, - * such that the AArch32 view is the lower 32 bits of the AArch64 one. - * Note that we rely on the values of these enums as we iterate through - * the various states in some places. - */ -enum { - ARM_CP_STATE_AA32 = 0, - ARM_CP_STATE_AA64 = 1, - ARM_CP_STATE_BOTH = 2, -}; - -/* ARM CP register secure state flags. These flags identify security state - * attributes for a given CP register entry. - * The existence of both or neither secure and non-secure flags indicates that - * the register has both a secure and non-secure hash entry. A single one of - * these flags causes the register to only be hashed for the specified - * security state. - * Although definitions may have any combination of the S/NS bits, each - * registered entry will only have one to identify whether the entry is secure - * or non-secure. - */ -enum { - ARM_CP_SECSTATE_S = (1 << 0), /* bit[0]: Secure state register */ - ARM_CP_SECSTATE_NS = (1 << 1), /* bit[1]: Non-secure state register */ -}; - -/* Return true if cptype is a valid type field. This is used to try to - * catch errors where the sentinel has been accidentally left off the end - * of a list of registers. - */ -static inline bool cptype_valid(int cptype) -{ - return ((cptype & ~ARM_CP_FLAG_MASK) == 0) - || ((cptype & ARM_CP_SPECIAL) && - ((cptype & ~ARM_CP_FLAG_MASK) <= ARM_LAST_SPECIAL)); -} - -/* Access rights: - * We define bits for Read and Write access for what rev C of the v7-AR ARM ARM - * defines as PL0 (user), PL1 (fiq/irq/svc/abt/und/sys, ie privileged), and - * PL2 (hyp). The other level which has Read and Write bits is Secure PL1 - * (ie any of the privileged modes in Secure state, or Monitor mode). - * If a register is accessible in one privilege level it's always accessible - * in higher privilege levels too. Since "Secure PL1" also follows this rule - * (ie anything visible in PL2 is visible in S-PL1, some things are only - * visible in S-PL1) but "Secure PL1" is a bit of a mouthful, we bend the - * terminology a little and call this PL3. - * In AArch64 things are somewhat simpler as the PLx bits line up exactly - * with the ELx exception levels. - * - * If access permissions for a register are more complex than can be - * described with these bits, then use a laxer set of restrictions, and - * do the more restrictive/complex check inside a helper function. - */ -#define PL3_R 0x80 -#define PL3_W 0x40 -#define PL2_R (0x20 | PL3_R) -#define PL2_W (0x10 | PL3_W) -#define PL1_R (0x08 | PL2_R) -#define PL1_W (0x04 | PL2_W) -#define PL0_R (0x02 | PL1_R) -#define PL0_W (0x01 | PL1_W) - -/* - * For user-mode some registers are accessible to EL0 via a kernel - * trap-and-emulate ABI. In this case we define the read permissions - * as actually being PL0_R. However some bits of any given register - * may still be masked. - */ -#ifdef CONFIG_USER_ONLY -#define PL0U_R PL0_R -#else -#define PL0U_R PL1_R -#endif - -#define PL3_RW (PL3_R | PL3_W) -#define PL2_RW (PL2_R | PL2_W) -#define PL1_RW (PL1_R | PL1_W) -#define PL0_RW (PL0_R | PL0_W) - /* Return the highest implemented Exception Level */ static inline int arm_highest_el(CPUARMState *env) { @@ -XXX,XX +XXX,XX @@ static inline int arm_current_el(CPUARMState *env) } } -typedef struct ARMCPRegInfo ARMCPRegInfo; - -typedef enum CPAccessResult { - /* Access is permitted */ - CP_ACCESS_OK = 0, - /* Access fails due to a configurable trap or enable which would - * result in a categorized exception syndrome giving information about - * the failing instruction (ie syndrome category 0x3, 0x4, 0x5, 0x6, - * 0xc or 0x18). The exception is taken to the usual target EL (EL1 or - * PL1 if in EL0, otherwise to the current EL). - */ - CP_ACCESS_TRAP = 1, - /* Access fails and results in an exception syndrome 0x0 ("uncategorized"). - * Note that this is not a catch-all case -- the set of cases which may - * result in this failure is specifically defined by the architecture. - */ - CP_ACCESS_TRAP_UNCATEGORIZED = 2, - /* As CP_ACCESS_TRAP, but for traps directly to EL2 or EL3 */ - CP_ACCESS_TRAP_EL2 = 3, - CP_ACCESS_TRAP_EL3 = 4, - /* As CP_ACCESS_UNCATEGORIZED, but for traps directly to EL2 or EL3 */ - CP_ACCESS_TRAP_UNCATEGORIZED_EL2 = 5, - CP_ACCESS_TRAP_UNCATEGORIZED_EL3 = 6, -} CPAccessResult; - -/* Access functions for coprocessor registers. These cannot fail and - * may not raise exceptions. - */ -typedef uint64_t CPReadFn(CPUARMState *env, const ARMCPRegInfo *opaque); -typedef void CPWriteFn(CPUARMState *env, const ARMCPRegInfo *opaque, - uint64_t value); -/* Access permission check functions for coprocessor registers. */ -typedef CPAccessResult CPAccessFn(CPUARMState *env, - const ARMCPRegInfo *opaque, - bool isread); -/* Hook function for register reset */ -typedef void CPResetFn(CPUARMState *env, const ARMCPRegInfo *opaque); - -#define CP_ANY 0xff - -/* Definition of an ARM coprocessor register */ -struct ARMCPRegInfo { - /* Name of register (useful mainly for debugging, need not be unique) */ - const char *name; - /* Location of register: coprocessor number and (crn,crm,opc1,opc2) - * tuple. Any of crm, opc1 and opc2 may be CP_ANY to indicate a - * 'wildcard' field -- any value of that field in the MRC/MCR insn - * will be decoded to this register. The register read and write - * callbacks will be passed an ARMCPRegInfo with the crn/crm/opc1/opc2 - * used by the program, so it is possible to register a wildcard and - * then behave differently on read/write if necessary. - * For 64 bit registers, only crm and opc1 are relevant; crn and opc2 - * must both be zero. - * For AArch64-visible registers, opc0 is also used. - * Since there are no "coprocessors" in AArch64, cp is purely used as a - * way to distinguish (for KVM's benefit) guest-visible system registers - * from demuxed ones provided to preserve the "no side effects on - * KVM register read/write from QEMU" semantics. cp==0x13 is guest - * visible (to match KVM's encoding); cp==0 will be converted to - * cp==0x13 when the ARMCPRegInfo is registered, for convenience. - */ - uint8_t cp; - uint8_t crn; - uint8_t crm; - uint8_t opc0; - uint8_t opc1; - uint8_t opc2; - /* Execution state in which this register is visible: ARM_CP_STATE_* */ - int state; - /* Register type: ARM_CP_* bits/values */ - int type; - /* Access rights: PL*_[RW] */ - int access; - /* Security state: ARM_CP_SECSTATE_* bits/values */ - int secure; - /* The opaque pointer passed to define_arm_cp_regs_with_opaque() when - * this register was defined: can be used to hand data through to the - * register read/write functions, since they are passed the ARMCPRegInfo*. - */ - void *opaque; - /* Value of this register, if it is ARM_CP_CONST. Otherwise, if - * fieldoffset is non-zero, the reset value of the register. - */ - uint64_t resetvalue; - /* Offset of the field in CPUARMState for this register. - * - * This is not needed if either: - * 1. type is ARM_CP_CONST or one of the ARM_CP_SPECIALs - * 2. both readfn and writefn are specified - */ - ptrdiff_t fieldoffset; /* offsetof(CPUARMState, field) */ - - /* Offsets of the secure and non-secure fields in CPUARMState for the - * register if it is banked. These fields are only used during the static - * registration of a register. During hashing the bank associated - * with a given security state is copied to fieldoffset which is used from - * there on out. - * - * It is expected that register definitions use either fieldoffset or - * bank_fieldoffsets in the definition but not both. It is also expected - * that both bank offsets are set when defining a banked register. This - * use indicates that a register is banked. - */ - ptrdiff_t bank_fieldoffsets[2]; - - /* Function for making any access checks for this register in addition to - * those specified by the 'access' permissions bits. If NULL, no extra - * checks required. The access check is performed at runtime, not at - * translate time. - */ - CPAccessFn *accessfn; - /* Function for handling reads of this register. If NULL, then reads - * will be done by loading from the offset into CPUARMState specified - * by fieldoffset. - */ - CPReadFn *readfn; - /* Function for handling writes of this register. If NULL, then writes - * will be done by writing to the offset into CPUARMState specified - * by fieldoffset. - */ - CPWriteFn *writefn; - /* Function for doing a "raw" read; used when we need to copy - * coprocessor state to the kernel for KVM or out for - * migration. This only needs to be provided if there is also a - * readfn and it has side effects (for instance clear-on-read bits). - */ - CPReadFn *raw_readfn; - /* Function for doing a "raw" write; used when we need to copy KVM - * kernel coprocessor state into userspace, or for inbound - * migration. This only needs to be provided if there is also a - * writefn and it masks out "unwritable" bits or has write-one-to-clear - * or similar behaviour. - */ - CPWriteFn *raw_writefn; - /* Function for resetting the register. If NULL, then reset will be done - * by writing resetvalue to the field specified in fieldoffset. If - * fieldoffset is 0 then no reset will be done. - */ - CPResetFn *resetfn; - - /* - * "Original" writefn and readfn. - * For ARMv8.1-VHE register aliases, we overwrite the read/write - * accessor functions of various EL1/EL0 to perform the runtime - * check for which sysreg should actually be modified, and then - * forwards the operation. Before overwriting the accessors, - * the original function is copied here, so that accesses that - * really do go to the EL1/EL0 version proceed normally. - * (The corresponding EL2 register is linked via opaque.) - */ - CPReadFn *orig_readfn; - CPWriteFn *orig_writefn; -}; - -/* Macros which are lvalues for the field in CPUARMState for the - * ARMCPRegInfo *ri. - */ -#define CPREG_FIELD32(env, ri) \ - (*(uint32_t *)((char *)(env) + (ri)->fieldoffset)) -#define CPREG_FIELD64(env, ri) \ - (*(uint64_t *)((char *)(env) + (ri)->fieldoffset)) - -#define REGINFO_SENTINEL { .type = ARM_CP_SENTINEL } - -void define_arm_cp_regs_with_opaque(ARMCPU *cpu, - const ARMCPRegInfo *regs, void *opaque); -void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, - const ARMCPRegInfo *regs, void *opaque); -static inline void define_arm_cp_regs(ARMCPU *cpu, const ARMCPRegInfo *regs) -{ - define_arm_cp_regs_with_opaque(cpu, regs, 0); -} -static inline void define_one_arm_cp_reg(ARMCPU *cpu, const ARMCPRegInfo *regs) -{ - define_one_arm_cp_reg_with_opaque(cpu, regs, 0); -} -const ARMCPRegInfo *get_arm_cp_reginfo(GHashTable *cpregs, uint32_t encoded_cp); - -/* - * Definition of an ARM co-processor register as viewed from - * userspace. This is used for presenting sanitised versions of - * registers to userspace when emulating the Linux AArch64 CPU - * ID/feature ABI (advertised as HWCAP_CPUID). - */ -typedef struct ARMCPRegUserSpaceInfo { - /* Name of register */ - const char *name; - - /* Is the name actually a glob pattern */ - bool is_glob; - - /* Only some bits are exported to user space */ - uint64_t exported_bits; - - /* Fixed bits are applied after the mask */ - uint64_t fixed_bits; -} ARMCPRegUserSpaceInfo; - -#define REGUSERINFO_SENTINEL { .name = NULL } - -void modify_arm_cp_regs(ARMCPRegInfo *regs, const ARMCPRegUserSpaceInfo *mods); - -/* CPWriteFn that can be used to implement writes-ignored behaviour */ -void arm_cp_write_ignore(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value); -/* CPReadFn that can be used for read-as-zero behaviour */ -uint64_t arm_cp_read_zero(CPUARMState *env, const ARMCPRegInfo *ri); - -/* CPResetFn that does nothing, for use if no reset is required even - * if fieldoffset is non zero. - */ -void arm_cp_reset_ignore(CPUARMState *env, const ARMCPRegInfo *opaque); - -/* Return true if this reginfo struct's field in the cpu state struct - * is 64 bits wide. - */ -static inline bool cpreg_field_is_64bit(const ARMCPRegInfo *ri) -{ - return (ri->state == ARM_CP_STATE_AA64) || (ri->type & ARM_CP_64BIT); -} - -static inline bool cp_access_ok(int current_el, - const ARMCPRegInfo *ri, int isread) -{ - return (ri->access >> ((current_el * 2) + isread)) & 1; -} - -/* Raw read of a coprocessor register (as needed for migration, etc) */ -uint64_t read_raw_cp_reg(CPUARMState *env, const ARMCPRegInfo *ri); - /** * write_list_to_cpustate * @cpu: ARMCPU diff --git a/hw/arm/pxa2xx.c b/hw/arm/pxa2xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/pxa2xx.c +++ b/hw/arm/pxa2xx.c @@ -XXX,XX +XXX,XX @@ #include "qemu/cutils.h" #include "qemu/log.h" #include "qom/object.h" +#include "target/arm/cpregs.h" static struct { hwaddr io_base; diff --git a/hw/arm/pxa2xx_pic.c b/hw/arm/pxa2xx_pic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/pxa2xx_pic.c +++ b/hw/arm/pxa2xx_pic.c @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "migration/vmstate.h" #include "qom/object.h" +#include "target/arm/cpregs.h" #define ICIP 0x00 /* Interrupt Controller IRQ Pending register */ #define ICMR 0x04 /* Interrupt Controller Mask register */ diff --git a/hw/intc/arm_gicv3_cpuif.c b/hw/intc/arm_gicv3_cpuif.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/arm_gicv3_cpuif.c +++ b/hw/intc/arm_gicv3_cpuif.c @@ -XXX,XX +XXX,XX @@ #include "gicv3_internal.h" #include "hw/irq.h" #include "cpu.h" +#include "target/arm/cpregs.h" /* * Special case return value from hppvi_index(); must be larger than diff --git a/hw/intc/arm_gicv3_kvm.c b/hw/intc/arm_gicv3_kvm.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/arm_gicv3_kvm.c +++ b/hw/intc/arm_gicv3_kvm.c @@ -XXX,XX +XXX,XX @@ #include "vgic_common.h" #include "migration/blocker.h" #include "qom/object.h" +#include "target/arm/cpregs.h" + #ifdef DEBUG_GICV3_KVM #define DPRINTF(fmt, ...) \ diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ #include "kvm_arm.h" #include "disas/capstone.h" #include "fpu/softfloat.h" +#include "cpregs.h" static void arm_cpu_set_pc(CPUState *cs, vaddr value) { diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ #include "hvf_arm.h" #include "qapi/visitor.h" #include "hw/qdev-properties.h" +#include "cpregs.h" #ifndef CONFIG_USER_ONLY diff --git a/target/arm/cpu_tcg.c b/target/arm/cpu_tcg.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu_tcg.c +++ b/target/arm/cpu_tcg.c @@ -XXX,XX +XXX,XX @@ #if !defined(CONFIG_USER_ONLY) #include "hw/boards.h" #endif +#include "cpregs.h" /* CPU models. These are not needed for the AArch64 linux-user build. */ #if !defined(CONFIG_USER_ONLY) || !defined(TARGET_AARCH64) diff --git a/target/arm/gdbstub.c b/target/arm/gdbstub.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/gdbstub.c +++ b/target/arm/gdbstub.c @@ -XXX,XX +XXX,XX @@ */ #include "qemu/osdep.h" #include "cpu.h" -#include "internals.h" #include "exec/gdbstub.h" +#include "internals.h" +#include "cpregs.h" typedef struct RegisterSysregXmlParam { CPUState *cs; diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "semihosting/common-semi.h" #endif +#include "cpregs.h" #define ARM_CPU_FREQ 1000000000 /* FIXME: 1 GHz, should be configurable */ #define PMCR_NUM_COUNTERS 4 /* QEMU IMPDEF choice */ diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "exec/exec-all.h" #include "exec/cpu_ldst.h" +#include "cpregs.h" #define SIGNBIT (uint32_t)0x80000000 #define SIGNBIT64 ((uint64_t)1 << 63) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ #include "translate.h" #include "internals.h" #include "qemu/host-utils.h" - #include "semihosting/semihost.h" #include "exec/gen-icount.h" - #include "exec/helper-proto.h" #include "exec/helper-gen.h" #include "exec/log.h" - +#include "cpregs.h" #include "translate-a64.h" #include "qemu/atomic128.h" diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/bitops.h" #include "arm_ldst.h" #include "semihosting/semihost.h" - #include "exec/helper-proto.h" #include "exec/helper-gen.h" - #include "exec/log.h" +#include "cpregs.h" #define ENABLE_ARCH_4T arm_dc_feature(s, ARM_FEATURE_V4T) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Rearrange the values of the enumerators of CPAccessResult so that we may directly extract the target el. For the two special cases in access_check_cp_reg, use CPAccessResult. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 26 ++++++++++++-------- target/arm/op_helper.c | 56 +++++++++++++++++++++--------------------- 2 files changed, 44 insertions(+), 38 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ static inline bool cptype_valid(int cptype) typedef enum CPAccessResult { /* Access is permitted */ CP_ACCESS_OK = 0, + + /* + * Combined with one of the following, the low 2 bits indicate the + * target exception level. If 0, the exception is taken to the usual + * target EL (EL1 or PL1 if in EL0, otherwise to the current EL). + */ + CP_ACCESS_EL_MASK = 3, + /* * Access fails due to a configurable trap or enable which would * result in a categorized exception syndrome giving information about * the failing instruction (ie syndrome category 0x3, 0x4, 0x5, 0x6, - * 0xc or 0x18). The exception is taken to the usual target EL (EL1 or - * PL1 if in EL0, otherwise to the current EL). + * 0xc or 0x18). */ - CP_ACCESS_TRAP = 1, + CP_ACCESS_TRAP = (1 << 2), + CP_ACCESS_TRAP_EL2 = CP_ACCESS_TRAP | 2, + CP_ACCESS_TRAP_EL3 = CP_ACCESS_TRAP | 3, + /* * Access fails and results in an exception syndrome 0x0 ("uncategorized"). * Note that this is not a catch-all case -- the set of cases which may * result in this failure is specifically defined by the architecture. */ - CP_ACCESS_TRAP_UNCATEGORIZED = 2, - /* As CP_ACCESS_TRAP, but for traps directly to EL2 or EL3 */ - CP_ACCESS_TRAP_EL2 = 3, - CP_ACCESS_TRAP_EL3 = 4, - /* As CP_ACCESS_UNCATEGORIZED, but for traps directly to EL2 or EL3 */ - CP_ACCESS_TRAP_UNCATEGORIZED_EL2 = 5, - CP_ACCESS_TRAP_UNCATEGORIZED_EL3 = 6, + CP_ACCESS_TRAP_UNCATEGORIZED = (2 << 2), + CP_ACCESS_TRAP_UNCATEGORIZED_EL2 = CP_ACCESS_TRAP_UNCATEGORIZED | 2, + CP_ACCESS_TRAP_UNCATEGORIZED_EL3 = CP_ACCESS_TRAP_UNCATEGORIZED | 3, } CPAccessResult; typedef struct ARMCPRegInfo ARMCPRegInfo; diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(access_check_cp_reg)(CPUARMState *env, void *rip, uint32_t syndrome, uint32_t isread) { const ARMCPRegInfo *ri = rip; + CPAccessResult res = CP_ACCESS_OK; int target_el; if (arm_feature(env, ARM_FEATURE_XSCALE) && ri->cp < 14 && extract32(env->cp15.c15_cpar, ri->cp, 1) == 0) { - raise_exception(env, EXCP_UDEF, syndrome, exception_target_el(env)); + res = CP_ACCESS_TRAP; + goto fail; } /* @@ -XXX,XX +XXX,XX @@ void HELPER(access_check_cp_reg)(CPUARMState *env, void *rip, uint32_t syndrome, mask &= ~((1 << 4) | (1 << 14)); if (env->cp15.hstr_el2 & mask) { - target_el = 2; - goto exept; + res = CP_ACCESS_TRAP_EL2; + goto fail; } } - if (!ri->accessfn) { + if (ri->accessfn) { + res = ri->accessfn(env, ri, isread); + } + if (likely(res == CP_ACCESS_OK)) { return; } - switch (ri->accessfn(env, ri, isread)) { - case CP_ACCESS_OK: - return; + fail: + switch (res & ~CP_ACCESS_EL_MASK) { case CP_ACCESS_TRAP: - target_el = exception_target_el(env); - break; - case CP_ACCESS_TRAP_EL2: - /* Requesting a trap to EL2 when we're in EL3 is - * a bug in the access function. - */ - assert(arm_current_el(env) != 3); - target_el = 2; - break; - case CP_ACCESS_TRAP_EL3: - target_el = 3; break; case CP_ACCESS_TRAP_UNCATEGORIZED: - target_el = exception_target_el(env); - syndrome = syn_uncategorized(); - break; - case CP_ACCESS_TRAP_UNCATEGORIZED_EL2: - target_el = 2; - syndrome = syn_uncategorized(); - break; - case CP_ACCESS_TRAP_UNCATEGORIZED_EL3: - target_el = 3; syndrome = syn_uncategorized(); break; default: g_assert_not_reached(); } -exept: + target_el = res & CP_ACCESS_EL_MASK; + switch (target_el) { + case 0: + target_el = exception_target_el(env); + break; + case 2: + assert(arm_current_el(env) != 3); + assert(arm_is_el2_enabled(env)); + break; + case 3: + assert(arm_feature(env, ARM_FEATURE_EL3)); + break; + default: + /* No "direct" traps to EL1 */ + g_assert_not_reached(); + } + raise_exception(env, EXCP_UDEF, syndrome, target_el); } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Remove a possible source of error by removing REGINFO_SENTINEL and using ARRAY_SIZE (convinently hidden inside a macro) to find the end of the set of regs being registered or modified. The space saved by not having the extra array element reduces the executable's .data.rel.ro section by about 9k. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 53 +++++++++--------- hw/arm/pxa2xx.c | 1 - hw/arm/pxa2xx_pic.c | 1 - hw/intc/arm_gicv3_cpuif.c | 5 -- hw/intc/arm_gicv3_kvm.c | 1 - target/arm/cpu64.c | 1 - target/arm/cpu_tcg.c | 4 -- target/arm/helper.c | 111 ++++++++------------------------------ 8 files changed, 48 insertions(+), 129 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ #define ARM_CP_NO_GDB 0x4000 #define ARM_CP_RAISES_EXC 0x8000 #define ARM_CP_NEWEL 0x10000 -/* Used only as a terminator for ARMCPRegInfo lists */ -#define ARM_CP_SENTINEL 0xfffff /* Mask of only the flag bits in a type field */ #define ARM_CP_FLAG_MASK 0x1f0ff @@ -XXX,XX +XXX,XX @@ enum { ARM_CP_SECSTATE_NS = (1 << 1), /* bit[1]: Non-secure state register */ }; -/* - * Return true if cptype is a valid type field. This is used to try to - * catch errors where the sentinel has been accidentally left off the end - * of a list of registers. - */ -static inline bool cptype_valid(int cptype) -{ - return ((cptype & ~ARM_CP_FLAG_MASK) == 0) - || ((cptype & ARM_CP_SPECIAL) && - ((cptype & ~ARM_CP_FLAG_MASK) <= ARM_LAST_SPECIAL)); -} - /* * Access rights: * We define bits for Read and Write access for what rev C of the v7-AR ARM ARM @@ -XXX,XX +XXX,XX @@ struct ARMCPRegInfo { #define CPREG_FIELD64(env, ri) \ (*(uint64_t *)((char *)(env) + (ri)->fieldoffset)) -#define REGINFO_SENTINEL { .type = ARM_CP_SENTINEL } +void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, const ARMCPRegInfo *reg, + void *opaque); -void define_arm_cp_regs_with_opaque(ARMCPU *cpu, - const ARMCPRegInfo *regs, void *opaque); -void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, - const ARMCPRegInfo *regs, void *opaque); -static inline void define_arm_cp_regs(ARMCPU *cpu, const ARMCPRegInfo *regs) -{ - define_arm_cp_regs_with_opaque(cpu, regs, 0); -} static inline void define_one_arm_cp_reg(ARMCPU *cpu, const ARMCPRegInfo *regs) { - define_one_arm_cp_reg_with_opaque(cpu, regs, 0); + define_one_arm_cp_reg_with_opaque(cpu, regs, NULL); } + +void define_arm_cp_regs_with_opaque_len(ARMCPU *cpu, const ARMCPRegInfo *regs, + void *opaque, size_t len); + +#define define_arm_cp_regs_with_opaque(CPU, REGS, OPAQUE) \ + do { \ + QEMU_BUILD_BUG_ON(ARRAY_SIZE(REGS) == 0); \ + define_arm_cp_regs_with_opaque_len(CPU, REGS, OPAQUE, \ + ARRAY_SIZE(REGS)); \ + } while (0) + +#define define_arm_cp_regs(CPU, REGS) \ + define_arm_cp_regs_with_opaque(CPU, REGS, NULL) + const ARMCPRegInfo *get_arm_cp_reginfo(GHashTable *cpregs, uint32_t encoded_cp); /* @@ -XXX,XX +XXX,XX @@ typedef struct ARMCPRegUserSpaceInfo { uint64_t fixed_bits; } ARMCPRegUserSpaceInfo; -#define REGUSERINFO_SENTINEL { .name = NULL } +void modify_arm_cp_regs_with_len(ARMCPRegInfo *regs, size_t regs_len, + const ARMCPRegUserSpaceInfo *mods, + size_t mods_len); -void modify_arm_cp_regs(ARMCPRegInfo *regs, const ARMCPRegUserSpaceInfo *mods); +#define modify_arm_cp_regs(REGS, MODS) \ + do { \ + QEMU_BUILD_BUG_ON(ARRAY_SIZE(REGS) == 0); \ + QEMU_BUILD_BUG_ON(ARRAY_SIZE(MODS) == 0); \ + modify_arm_cp_regs_with_len(REGS, ARRAY_SIZE(REGS), \ + MODS, ARRAY_SIZE(MODS)); \ + } while (0) /* CPWriteFn that can be used to implement writes-ignored behaviour */ void arm_cp_write_ignore(CPUARMState *env, const ARMCPRegInfo *ri, diff --git a/hw/arm/pxa2xx.c b/hw/arm/pxa2xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/pxa2xx.c +++ b/hw/arm/pxa2xx.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pxa_cp_reginfo[] = { { .name = "PWRMODE", .cp = 14, .crn = 7, .crm = 0, .opc1 = 0, .opc2 = 0, .access = PL1_RW, .type = ARM_CP_IO, .readfn = arm_cp_read_zero, .writefn = pxa2xx_pwrmode_write }, - REGINFO_SENTINEL }; static void pxa2xx_setup_cp14(PXA2xxState *s) diff --git a/hw/arm/pxa2xx_pic.c b/hw/arm/pxa2xx_pic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/pxa2xx_pic.c +++ b/hw/arm/pxa2xx_pic.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pxa_pic_cp_reginfo[] = { REGINFO_FOR_PIC_CP("ICLR2", 8), REGINFO_FOR_PIC_CP("ICFP2", 9), REGINFO_FOR_PIC_CP("ICPR2", 0xa), - REGINFO_SENTINEL }; static const MemoryRegionOps pxa2xx_pic_ops = { diff --git a/hw/intc/arm_gicv3_cpuif.c b/hw/intc/arm_gicv3_cpuif.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/arm_gicv3_cpuif.c +++ b/hw/intc/arm_gicv3_cpuif.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gicv3_cpuif_reginfo[] = { .readfn = icc_igrpen1_el3_read, .writefn = icc_igrpen1_el3_write, }, - REGINFO_SENTINEL }; static uint64_t ich_ap_read(CPUARMState *env, const ARMCPRegInfo *ri) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gicv3_cpuif_hcr_reginfo[] = { .readfn = ich_vmcr_read, .writefn = ich_vmcr_write, }, - REGINFO_SENTINEL }; static const ARMCPRegInfo gicv3_cpuif_ich_apxr1_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gicv3_cpuif_ich_apxr1_reginfo[] = { .readfn = ich_ap_read, .writefn = ich_ap_write, }, - REGINFO_SENTINEL }; static const ARMCPRegInfo gicv3_cpuif_ich_apxr23_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gicv3_cpuif_ich_apxr23_reginfo[] = { .readfn = ich_ap_read, .writefn = ich_ap_write, }, - REGINFO_SENTINEL }; static void gicv3_cpuif_el_change_hook(ARMCPU *cpu, void *opaque) @@ -XXX,XX +XXX,XX @@ void gicv3_init_cpuif(GICv3State *s) .readfn = ich_lr_read, .writefn = ich_lr_write, }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, lr_regset); } diff --git a/hw/intc/arm_gicv3_kvm.c b/hw/intc/arm_gicv3_kvm.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/arm_gicv3_kvm.c +++ b/hw/intc/arm_gicv3_kvm.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo gicv3_cpuif_reginfo[] = { */ .resetfn = arm_gicv3_icc_reset, }, - REGINFO_SENTINEL }; /** diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cortex_a72_a57_a53_cp_reginfo[] = { { .name = "L2MERRSR", .cp = 15, .opc1 = 3, .crm = 15, .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_64BIT, .resetvalue = 0 }, - REGINFO_SENTINEL }; static void aarch64_a57_initfn(Object *obj) diff --git a/target/arm/cpu_tcg.c b/target/arm/cpu_tcg.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu_tcg.c +++ b/target/arm/cpu_tcg.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cortexa8_cp_reginfo[] = { .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, { .name = "L2AUXCR", .cp = 15, .crn = 9, .crm = 0, .opc1 = 1, .opc2 = 2, .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; static void cortex_a8_initfn(Object *obj) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cortexa9_cp_reginfo[] = { .access = PL1_RW, .resetvalue = 0, .type = ARM_CP_CONST }, { .name = "TLB_ATTR", .cp = 15, .crn = 15, .crm = 7, .opc1 = 5, .opc2 = 2, .access = PL1_RW, .resetvalue = 0, .type = ARM_CP_CONST }, - REGINFO_SENTINEL }; static void cortex_a9_initfn(Object *obj) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cortexa15_cp_reginfo[] = { #endif { .name = "L2ECTLR", .cp = 15, .crn = 9, .crm = 0, .opc1 = 1, .opc2 = 3, .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; static void cortex_a7_initfn(Object *obj) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cortexr5_cp_reginfo[] = { .access = PL1_RW, .type = ARM_CP_CONST }, { .name = "DCACHE_INVAL", .cp = 15, .opc1 = 0, .crn = 15, .crm = 5, .opc2 = 0, .access = PL1_W, .type = ARM_CP_NOP }, - REGINFO_SENTINEL }; static void cortex_r5_initfn(Object *obj) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cp_reginfo[] = { .secure = ARM_CP_SECSTATE_S, .fieldoffset = offsetof(CPUARMState, cp15.contextidr_s), .resetvalue = 0, .writefn = contextidr_write, .raw_writefn = raw_write, }, - REGINFO_SENTINEL }; static const ARMCPRegInfo not_v8_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo not_v8_cp_reginfo[] = { { .name = "CACHEMAINT", .cp = 15, .crn = 7, .crm = CP_ANY, .opc1 = 0, .opc2 = CP_ANY, .access = PL1_W, .type = ARM_CP_NOP | ARM_CP_OVERRIDE }, - REGINFO_SENTINEL }; static const ARMCPRegInfo not_v6_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo not_v6_cp_reginfo[] = { */ { .name = "WFI_v5", .cp = 15, .crn = 7, .crm = 8, .opc1 = 0, .opc2 = 2, .access = PL1_W, .type = ARM_CP_WFI }, - REGINFO_SENTINEL }; static const ARMCPRegInfo not_v7_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo not_v7_cp_reginfo[] = { .opc1 = 0, .opc2 = 0, .access = PL1_RW, .type = ARM_CP_NOP }, { .name = "NMRR", .cp = 15, .crn = 10, .crm = 2, .opc1 = 0, .opc2 = 1, .access = PL1_RW, .type = ARM_CP_NOP }, - REGINFO_SENTINEL }; static void cpacr_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v6_cp_reginfo[] = { .crn = 1, .crm = 0, .opc1 = 0, .opc2 = 2, .accessfn = cpacr_access, .access = PL1_RW, .fieldoffset = offsetof(CPUARMState, cp15.cpacr_el1), .resetfn = cpacr_reset, .writefn = cpacr_write, .readfn = cpacr_read }, - REGINFO_SENTINEL }; typedef struct pm_event { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v7_cp_reginfo[] = { { .name = "TLBIMVAA", .cp = 15, .opc1 = 0, .crn = 8, .crm = 7, .opc2 = 3, .type = ARM_CP_NO_RAW, .access = PL1_W, .accessfn = access_ttlb, .writefn = tlbimvaa_write }, - REGINFO_SENTINEL }; static const ARMCPRegInfo v7mp_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v7mp_cp_reginfo[] = { { .name = "TLBIMVAAIS", .cp = 15, .opc1 = 0, .crn = 8, .crm = 3, .opc2 = 3, .type = ARM_CP_NO_RAW, .access = PL1_W, .accessfn = access_ttlb, .writefn = tlbimvaa_is_write }, - REGINFO_SENTINEL }; static const ARMCPRegInfo pmovsset_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pmovsset_cp_reginfo[] = { .fieldoffset = offsetof(CPUARMState, cp15.c9_pmovsr), .writefn = pmovsset_write, .raw_writefn = raw_write }, - REGINFO_SENTINEL }; static void teecr_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo t2ee_cp_reginfo[] = { { .name = "TEEHBR", .cp = 14, .crn = 1, .crm = 0, .opc1 = 6, .opc2 = 0, .access = PL0_RW, .fieldoffset = offsetof(CPUARMState, teehbr), .accessfn = teehbr_access, .resetvalue = 0 }, - REGINFO_SENTINEL }; static const ARMCPRegInfo v6k_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v6k_cp_reginfo[] = { .bank_fieldoffsets = { offsetoflow32(CPUARMState, cp15.tpidrprw_s), offsetoflow32(CPUARMState, cp15.tpidrprw_ns) }, .resetvalue = 0 }, - REGINFO_SENTINEL }; #ifndef CONFIG_USER_ONLY @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo generic_timer_cp_reginfo[] = { .fieldoffset = offsetof(CPUARMState, cp15.c14_timer[GTIMER_SEC].cval), .writefn = gt_sec_cval_write, .raw_writefn = raw_write, }, - REGINFO_SENTINEL }; static CPAccessResult e2h_access(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo generic_timer_cp_reginfo[] = { .access = PL0_R, .type = ARM_CP_NO_RAW | ARM_CP_IO, .readfn = gt_virt_cnt_read, }, - REGINFO_SENTINEL }; #endif @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vapa_cp_reginfo[] = { .access = PL1_W, .accessfn = ats_access, .writefn = ats_write, .type = ARM_CP_NO_RAW | ARM_CP_RAISES_EXC }, #endif - REGINFO_SENTINEL }; /* Return basic MPU access permission bits. */ @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pmsav7_cp_reginfo[] = { .fieldoffset = offsetof(CPUARMState, pmsav7.rnr[M_REG_NS]), .writefn = pmsav7_rgnr_write, .resetfn = arm_cp_reset_ignore }, - REGINFO_SENTINEL }; static const ARMCPRegInfo pmsav5_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pmsav5_cp_reginfo[] = { { .name = "946_PRBS7", .cp = 15, .crn = 6, .crm = 7, .opc1 = 0, .opc2 = CP_ANY, .access = PL1_RW, .resetvalue = 0, .fieldoffset = offsetof(CPUARMState, cp15.c6_region[7]) }, - REGINFO_SENTINEL }; static void vmsa_ttbcr_raw_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vmsa_pmsa_cp_reginfo[] = { .access = PL1_RW, .accessfn = access_tvm_trvm, .fieldoffset = offsetof(CPUARMState, cp15.far_el[1]), .resetvalue = 0, }, - REGINFO_SENTINEL }; static const ARMCPRegInfo vmsa_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vmsa_cp_reginfo[] = { /* No offsetoflow32 -- pass the entire TCR to writefn/raw_writefn. */ .bank_fieldoffsets = { offsetof(CPUARMState, cp15.tcr_el[3]), offsetof(CPUARMState, cp15.tcr_el[1])} }, - REGINFO_SENTINEL }; /* Note that unlike TTBCR, writing to TTBCR2 does not require flushing @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo omap_cp_reginfo[] = { { .name = "C9", .cp = 15, .crn = 9, .crm = CP_ANY, .opc1 = CP_ANY, .opc2 = CP_ANY, .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_OVERRIDE, .resetvalue = 0 }, - REGINFO_SENTINEL }; static void xscale_cpar_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo xscale_cp_reginfo[] = { { .name = "XSCALE_UNLOCK_DCACHE", .cp = 15, .opc1 = 0, .crn = 9, .crm = 2, .opc2 = 1, .access = PL1_W, .type = ARM_CP_NOP }, - REGINFO_SENTINEL }; static const ARMCPRegInfo dummy_c15_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo dummy_c15_cp_reginfo[] = { .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_NO_RAW | ARM_CP_OVERRIDE, .resetvalue = 0 }, - REGINFO_SENTINEL }; static const ARMCPRegInfo cache_dirty_status_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cache_dirty_status_cp_reginfo[] = { { .name = "CDSR", .cp = 15, .crn = 7, .crm = 10, .opc1 = 0, .opc2 = 6, .access = PL1_R, .type = ARM_CP_CONST | ARM_CP_NO_RAW, .resetvalue = 0 }, - REGINFO_SENTINEL }; static const ARMCPRegInfo cache_block_ops_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cache_block_ops_cp_reginfo[] = { .access = PL0_W, .type = ARM_CP_NOP|ARM_CP_64BIT }, { .name = "CIDCR", .cp = 15, .crm = 14, .opc1 = 0, .access = PL1_W, .type = ARM_CP_NOP|ARM_CP_64BIT }, - REGINFO_SENTINEL }; static const ARMCPRegInfo cache_test_clean_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo cache_test_clean_cp_reginfo[] = { { .name = "TCI_DCACHE", .cp = 15, .crn = 7, .crm = 14, .opc1 = 0, .opc2 = 3, .access = PL0_R, .type = ARM_CP_CONST | ARM_CP_NO_RAW, .resetvalue = (1 << 30) }, - REGINFO_SENTINEL }; static const ARMCPRegInfo strongarm_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo strongarm_cp_reginfo[] = { .crm = CP_ANY, .opc1 = CP_ANY, .opc2 = CP_ANY, .access = PL1_RW, .resetvalue = 0, .type = ARM_CP_CONST | ARM_CP_OVERRIDE | ARM_CP_NO_RAW }, - REGINFO_SENTINEL }; static uint64_t midr_read(CPUARMState *env, const ARMCPRegInfo *ri) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo lpae_cp_reginfo[] = { .bank_fieldoffsets = { offsetof(CPUARMState, cp15.ttbr1_s), offsetof(CPUARMState, cp15.ttbr1_ns) }, .writefn = vmsa_ttbr_write, }, - REGINFO_SENTINEL }; static uint64_t aa64_fpcr_read(CPUARMState *env, const ARMCPRegInfo *ri) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .access = PL1_RW, .accessfn = access_trap_aa32s_el1, .writefn = sdcr_write, .fieldoffset = offsetoflow32(CPUARMState, cp15.mdcr_el3) }, - REGINFO_SENTINEL }; /* Used to describe the behaviour of EL2 regs when EL2 does not exist. */ @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_no_el2_cp_reginfo[] = { .type = ARM_CP_CONST, .cp = 15, .opc1 = 4, .crn = 6, .crm = 0, .opc2 = 2, .access = PL2_RW, .resetvalue = 0 }, - REGINFO_SENTINEL }; /* Ditto, but for registers which exist in ARMv8 but not v7 */ @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_no_el2_v8_cp_reginfo[] = { .cp = 15, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 4, .access = PL2_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; static void do_hcr_write(CPUARMState *env, uint64_t value, uint64_t valid_mask) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el2_cp_reginfo[] = { .cp = 15, .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 3, .access = PL2_RW, .fieldoffset = offsetof(CPUARMState, cp15.hstr_el2) }, - REGINFO_SENTINEL }; static const ARMCPRegInfo el2_v8_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el2_v8_cp_reginfo[] = { .access = PL2_RW, .fieldoffset = offsetofhigh32(CPUARMState, cp15.hcr_el2), .writefn = hcr_writehigh }, - REGINFO_SENTINEL }; static CPAccessResult sel2_access(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el2_sec_cp_reginfo[] = { .opc0 = 3, .opc1 = 4, .crn = 2, .crm = 6, .opc2 = 2, .access = PL2_RW, .accessfn = sel2_access, .fieldoffset = offsetof(CPUARMState, cp15.vstcr_el2) }, - REGINFO_SENTINEL }; static CPAccessResult nsacr_access(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_cp_reginfo[] = { .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 7, .opc2 = 5, .access = PL3_W, .type = ARM_CP_NO_RAW, .writefn = tlbi_aa64_vae3_write }, - REGINFO_SENTINEL }; #ifndef CONFIG_USER_ONLY @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo debug_cp_reginfo[] = { .cp = 14, .opc0 = 2, .opc1 = 0, .crn = 0, .crm = 2, .opc2 = 0, .access = PL1_RW, .accessfn = access_tda, .type = ARM_CP_NOP }, - REGINFO_SENTINEL }; static const ARMCPRegInfo debug_lpae_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo debug_lpae_cp_reginfo[] = { .access = PL0_R, .type = ARM_CP_CONST|ARM_CP_64BIT, .resetvalue = 0 }, { .name = "DBGDSAR", .cp = 14, .crm = 2, .opc1 = 0, .access = PL0_R, .type = ARM_CP_CONST|ARM_CP_64BIT, .resetvalue = 0 }, - REGINFO_SENTINEL }; /* Return the exception level to which exceptions should be taken @@ -XXX,XX +XXX,XX @@ static void define_debug_regs(ARMCPU *cpu) .fieldoffset = offsetof(CPUARMState, cp15.dbgbcr[i]), .writefn = dbgbcr_write, .raw_writefn = raw_write }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, dbgregs); } @@ -XXX,XX +XXX,XX @@ static void define_debug_regs(ARMCPU *cpu) .fieldoffset = offsetof(CPUARMState, cp15.dbgwcr[i]), .writefn = dbgwcr_write, .raw_writefn = raw_write }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, dbgregs); } @@ -XXX,XX +XXX,XX @@ static void define_pmu_regs(ARMCPU *cpu) .type = ARM_CP_IO, .readfn = pmevtyper_readfn, .writefn = pmevtyper_writefn, .raw_writefn = pmevtyper_rawwrite }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, pmev_regs); g_free(pmevcntr_name); @@ -XXX,XX +XXX,XX @@ static void define_pmu_regs(ARMCPU *cpu) .cp = 15, .opc1 = 0, .crn = 9, .crm = 14, .opc2 = 5, .access = PL0_R, .accessfn = pmreg_access, .type = ARM_CP_CONST, .resetvalue = extract64(cpu->pmceid1, 32, 32) }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, v81_pmu_regs); } @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo lor_reginfo[] = { .opc0 = 3, .opc1 = 0, .crn = 10, .crm = 4, .opc2 = 7, .access = PL1_R, .accessfn = access_lor_ns, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; #ifdef TARGET_AARCH64 @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo pauth_reginfo[] = { .opc0 = 3, .opc1 = 0, .crn = 2, .crm = 1, .opc2 = 3, .access = PL1_RW, .accessfn = access_pauth, .fieldoffset = offsetof(CPUARMState, keys.apib.hi) }, - REGINFO_SENTINEL }; static const ARMCPRegInfo tlbirange_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo tlbirange_reginfo[] = { .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 6, .opc2 = 5, .access = PL3_W, .type = ARM_CP_NO_RAW, .writefn = tlbi_aa64_rvae3_write }, - REGINFO_SENTINEL }; static const ARMCPRegInfo tlbios_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo tlbios_reginfo[] = { .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 1, .opc2 = 5, .access = PL3_W, .type = ARM_CP_NO_RAW, .writefn = tlbi_aa64_vae3is_write }, - REGINFO_SENTINEL }; static uint64_t rndr_readfn(CPUARMState *env, const ARMCPRegInfo *ri) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo rndr_reginfo[] = { .type = ARM_CP_NO_RAW | ARM_CP_SUPPRESS_TB_END | ARM_CP_IO, .opc0 = 3, .opc1 = 3, .crn = 2, .crm = 4, .opc2 = 1, .access = PL0_R, .readfn = rndr_readfn }, - REGINFO_SENTINEL }; #ifndef CONFIG_USER_ONLY @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo dcpop_reg[] = { .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 12, .opc2 = 1, .access = PL0_W, .type = ARM_CP_NO_RAW | ARM_CP_SUPPRESS_TB_END, .accessfn = aa64_cacheop_poc_access, .writefn = dccvap_writefn }, - REGINFO_SENTINEL }; static const ARMCPRegInfo dcpodp_reg[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo dcpodp_reg[] = { .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 13, .opc2 = 1, .access = PL0_W, .type = ARM_CP_NO_RAW | ARM_CP_SUPPRESS_TB_END, .accessfn = aa64_cacheop_poc_access, .writefn = dccvap_writefn }, - REGINFO_SENTINEL }; #endif /*CONFIG_USER_ONLY*/ @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_reginfo[] = { { .name = "DC_CIGDSW", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 14, .opc2 = 6, .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, - REGINFO_SENTINEL }; static const ARMCPRegInfo mte_tco_ro_reginfo[] = { { .name = "TCO", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, .type = ARM_CP_CONST, .access = PL0_RW, }, - REGINFO_SENTINEL }; static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { .accessfn = aa64_zva_access, #endif }, - REGINFO_SENTINEL }; #endif @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo predinv_reginfo[] = { { .name = "CPPRCTX", .state = ARM_CP_STATE_AA32, .cp = 15, .opc1 = 0, .crn = 7, .crm = 3, .opc2 = 7, .type = ARM_CP_NOP, .access = PL0_W, .accessfn = access_predinv }, - REGINFO_SENTINEL }; static uint64_t ccsidr2_read(CPUARMState *env, const ARMCPRegInfo *ri) @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo ccsidr2_reginfo[] = { .access = PL1_R, .accessfn = access_aa64_tid2, .readfn = ccsidr2_read, .type = ARM_CP_NO_RAW }, - REGINFO_SENTINEL }; static CPAccessResult access_aa64_tid3(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo jazelle_regs[] = { .cp = 14, .crn = 2, .crm = 0, .opc1 = 7, .opc2 = 0, .accessfn = access_joscr_jmcr, .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; static const ARMCPRegInfo vhe_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo vhe_reginfo[] = { .access = PL2_RW, .accessfn = e2h_access, .writefn = gt_virt_cval_write, .raw_writefn = raw_write }, #endif - REGINFO_SENTINEL }; #ifndef CONFIG_USER_ONLY @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo ats1e1_reginfo[] = { .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 9, .opc2 = 1, .access = PL1_W, .type = ARM_CP_NO_RAW | ARM_CP_RAISES_EXC, .writefn = ats_write64 }, - REGINFO_SENTINEL }; static const ARMCPRegInfo ats1cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo ats1cp_reginfo[] = { .cp = 15, .opc1 = 0, .crn = 7, .crm = 9, .opc2 = 1, .access = PL1_W, .type = ARM_CP_NO_RAW | ARM_CP_RAISES_EXC, .writefn = ats_write }, - REGINFO_SENTINEL }; #endif @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo actlr2_hactlr2_reginfo[] = { .cp = 15, .opc1 = 4, .crn = 1, .crm = 0, .opc2 = 3, .access = PL2_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; void register_cp_regs_for_features(ARMCPU *cpu) @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL1_R, .type = ARM_CP_CONST, .accessfn = access_aa32_tid3, .resetvalue = cpu->isar.id_isar6 }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, v6_idregs); define_arm_cp_regs(cpu, v6_cp_reginfo); @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .opc0 = 3, .opc1 = 3, .crn = 9, .crm = 12, .opc2 = 7, .access = PL0_R, .accessfn = pmreg_access, .type = ARM_CP_CONST, .resetvalue = cpu->pmceid1 }, - REGINFO_SENTINEL }; #ifdef CONFIG_USER_ONLY ARMCPRegUserSpaceInfo v8_user_idregs[] = { @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .exported_bits = 0x000000f0ffffffff }, { .name = "ID_AA64ISAR*_EL1_RESERVED", .is_glob = true }, - REGUSERINFO_SENTINEL }; modify_arm_cp_regs(v8_idregs, v8_user_idregs); #endif @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL2_RW, .resetvalue = vmpidr_def, .fieldoffset = offsetof(CPUARMState, cp15.vmpidr_el2) }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, vpidr_regs); define_arm_cp_regs(cpu, el2_cp_reginfo); @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL2_RW, .accessfn = access_el3_aa32ns, .type = ARM_CP_NO_RAW, .writefn = arm_cp_write_ignore, .readfn = mpidr_read }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, vpidr_regs); define_arm_cp_regs(cpu, el3_no_el2_cp_reginfo); @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .raw_writefn = raw_write, .writefn = sctlr_write, .fieldoffset = offsetof(CPUARMState, cp15.sctlr_el[3]), .resetvalue = cpu->reset_sctlr }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, el3_regs); @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) { .name = "DUMMY", .cp = 15, .crn = 0, .crm = 7, .opc1 = 0, .opc2 = CP_ANY, .access = PL1_R, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; ARMCPRegInfo id_v8_midr_cp_reginfo[] = { { .name = "MIDR_EL1", .state = ARM_CP_STATE_BOTH, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL1_R, .accessfn = access_aa64_tid1, .type = ARM_CP_CONST, .resetvalue = cpu->revidr }, - REGINFO_SENTINEL }; ARMCPRegInfo id_cp_reginfo[] = { /* These are common to v8 and pre-v8 */ @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL1_R, .accessfn = access_aa32_tid1, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; /* TLBTR is specific to VMSA */ ARMCPRegInfo id_tlbtr_reginfo = { @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) { .name = "MIDR_EL1", .exported_bits = 0x00000000ffffffff }, { .name = "REVIDR_EL1" }, - REGUSERINFO_SENTINEL }; modify_arm_cp_regs(id_v8_midr_cp_reginfo, id_v8_user_midr_cp_reginfo); #endif if (arm_feature(env, ARM_FEATURE_OMAPCP) || arm_feature(env, ARM_FEATURE_STRONGARM)) { - ARMCPRegInfo *r; + size_t i; /* Register the blanket "writes ignored" value first to cover the * whole space. Then update the specific ID registers to allow write * access, so that they ignore writes rather than causing them to * UNDEF. */ define_one_arm_cp_reg(cpu, &crn0_wi_reginfo); - for (r = id_pre_v8_midr_cp_reginfo; - r->type != ARM_CP_SENTINEL; r++) { - r->access = PL1_RW; + for (i = 0; i < ARRAY_SIZE(id_pre_v8_midr_cp_reginfo); ++i) { + id_pre_v8_midr_cp_reginfo[i].access = PL1_RW; } - for (r = id_cp_reginfo; r->type != ARM_CP_SENTINEL; r++) { - r->access = PL1_RW; + for (i = 0; i < ARRAY_SIZE(id_cp_reginfo); ++i) { + id_cp_reginfo[i].access = PL1_RW; } id_mpuir_reginfo.access = PL1_RW; id_tlbtr_reginfo.access = PL1_RW; @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) { .name = "MPIDR_EL1", .state = ARM_CP_STATE_BOTH, .opc0 = 3, .crn = 0, .crm = 0, .opc1 = 0, .opc2 = 5, .access = PL1_R, .readfn = mpidr_read, .type = ARM_CP_NO_RAW }, - REGINFO_SENTINEL }; #ifdef CONFIG_USER_ONLY ARMCPRegUserSpaceInfo mpidr_user_cp_reginfo[] = { { .name = "MPIDR_EL1", .fixed_bits = 0x0000000080000000 }, - REGUSERINFO_SENTINEL }; modify_arm_cp_regs(mpidr_cp_reginfo, mpidr_user_cp_reginfo); #endif @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .opc0 = 3, .opc1 = 6, .crn = 1, .crm = 0, .opc2 = 1, .access = PL3_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, auxcr_reginfo); if (cpu_isar_feature(aa32_ac2, cpu)) { @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .type = ARM_CP_CONST, .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 3, .opc2 = 0, .access = PL1_R, .resetvalue = cpu->reset_cbar }, - REGINFO_SENTINEL }; /* We don't implement a r/w 64 bit CBAR currently */ assert(arm_feature(env, ARM_FEATURE_CBAR_RO)); @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .bank_fieldoffsets = { offsetof(CPUARMState, cp15.vbar_s), offsetof(CPUARMState, cp15.vbar_ns) }, .resetvalue = 0 }, - REGINFO_SENTINEL }; define_arm_cp_regs(cpu, vbar_cp_reginfo); } @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, r->writefn); } } - /* Bad type field probably means missing sentinel at end of reg list */ - assert(cptype_valid(r->type)); + for (crm = crmmin; crm <= crmmax; crm++) { for (opc1 = opc1min; opc1 <= opc1max; opc1++) { for (opc2 = opc2min; opc2 <= opc2max; opc2++) { @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, } } -void define_arm_cp_regs_with_opaque(ARMCPU *cpu, - const ARMCPRegInfo *regs, void *opaque) +/* Define a whole list of registers */ +void define_arm_cp_regs_with_opaque_len(ARMCPU *cpu, const ARMCPRegInfo *regs, + void *opaque, size_t len) { - /* Define a whole list of registers */ - const ARMCPRegInfo *r; - for (r = regs; r->type != ARM_CP_SENTINEL; r++) { - define_one_arm_cp_reg_with_opaque(cpu, r, opaque); + size_t i; + for (i = 0; i < len; ++i) { + define_one_arm_cp_reg_with_opaque(cpu, regs + i, opaque); } } @@ -XXX,XX +XXX,XX @@ void define_arm_cp_regs_with_opaque(ARMCPU *cpu, * user-space cannot alter any values and dynamic values pertaining to * execution state are hidden from user space view anyway. */ -void modify_arm_cp_regs(ARMCPRegInfo *regs, const ARMCPRegUserSpaceInfo *mods) +void modify_arm_cp_regs_with_len(ARMCPRegInfo *regs, size_t regs_len, + const ARMCPRegUserSpaceInfo *mods, + size_t mods_len) { - const ARMCPRegUserSpaceInfo *m; - ARMCPRegInfo *r; - - for (m = mods; m->name; m++) { + for (size_t mi = 0; mi < mods_len; ++mi) { + const ARMCPRegUserSpaceInfo *m = mods + mi; GPatternSpec *pat = NULL; + if (m->is_glob) { pat = g_pattern_spec_new(m->name); } - for (r = regs; r->type != ARM_CP_SENTINEL; r++) { + for (size_t ri = 0; ri < regs_len; ++ri) { + ARMCPRegInfo *r = regs + ri; + if (pat && g_pattern_match_string(pat, r->name)) { r->type = ARM_CP_CONST; r->access = PL0U_R; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These particular data structures are not modified at runtime. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .resetvalue = cpu->pmceid1 }, }; #ifdef CONFIG_USER_ONLY - ARMCPRegUserSpaceInfo v8_user_idregs[] = { + static const ARMCPRegUserSpaceInfo v8_user_idregs[] = { { .name = "ID_AA64PFR0_EL1", .exported_bits = 0x000f000f00ff0000, .fixed_bits = 0x0000000000000011 }, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) */ if (arm_feature(env, ARM_FEATURE_EL3)) { if (arm_feature(env, ARM_FEATURE_AARCH64)) { - ARMCPRegInfo nsacr = { + static const ARMCPRegInfo nsacr = { .name = "NSACR", .type = ARM_CP_CONST, .cp = 15, .opc1 = 0, .crn = 1, .crm = 1, .opc2 = 2, .access = PL1_RW, .accessfn = nsacr_access, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) }; define_one_arm_cp_reg(cpu, &nsacr); } else { - ARMCPRegInfo nsacr = { + static const ARMCPRegInfo nsacr = { .name = "NSACR", .cp = 15, .opc1 = 0, .crn = 1, .crm = 1, .opc2 = 2, .access = PL3_RW | PL1_R, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) } } else { if (arm_feature(env, ARM_FEATURE_V8)) { - ARMCPRegInfo nsacr = { + static const ARMCPRegInfo nsacr = { .name = "NSACR", .type = ARM_CP_CONST, .cp = 15, .opc1 = 0, .crn = 1, .crm = 1, .opc2 = 2, .access = PL1_R, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL1_R, .type = ARM_CP_CONST, .resetvalue = cpu->pmsav7_dregion << 8 }; - ARMCPRegInfo crn0_wi_reginfo = { + static const ARMCPRegInfo crn0_wi_reginfo = { .name = "CRN0_WI", .cp = 15, .crn = 0, .crm = CP_ANY, .opc1 = CP_ANY, .opc2 = CP_ANY, .access = PL1_W, .type = ARM_CP_NOP | ARM_CP_OVERRIDE }; #ifdef CONFIG_USER_ONLY - ARMCPRegUserSpaceInfo id_v8_user_midr_cp_reginfo[] = { + static const ARMCPRegUserSpaceInfo id_v8_user_midr_cp_reginfo[] = { { .name = "MIDR_EL1", .exported_bits = 0x00000000ffffffff }, { .name = "REVIDR_EL1" }, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .access = PL1_R, .readfn = mpidr_read, .type = ARM_CP_NO_RAW }, }; #ifdef CONFIG_USER_ONLY - ARMCPRegUserSpaceInfo mpidr_user_cp_reginfo[] = { + static const ARMCPRegUserSpaceInfo mpidr_user_cp_reginfo[] = { { .name = "MPIDR_EL1", .fixed_bits = 0x0000000080000000 }, }; @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) } if (arm_feature(env, ARM_FEATURE_VBAR)) { - ARMCPRegInfo vbar_cp_reginfo[] = { + static const ARMCPRegInfo vbar_cp_reginfo[] = { { .name = "VBAR", .state = ARM_CP_STATE_BOTH, .opc0 = 3, .crn = 12, .crm = 0, .opc1 = 0, .opc2 = 0, .access = PL1_RW, .writefn = vbar_write, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Instead of defining ARM_CP_FLAG_MASK to remove flags, define ARM_CP_SPECIAL_MASK to isolate special cases. Sort the specials to the low bits. Use an enum. Split the large comment block so as to document each value separately. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 130 +++++++++++++++++++++++-------------- target/arm/cpu.c | 4 +- target/arm/helper.c | 4 +- target/arm/translate-a64.c | 6 +- target/arm/translate.c | 6 +- 5 files changed, 92 insertions(+), 58 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ #define TARGET_ARM_CPREGS_H /* - * ARMCPRegInfo type field bits. If the SPECIAL bit is set this is a - * special-behaviour cp reg and bits [11..8] indicate what behaviour - * it has. Otherwise it is a simple cp reg, where CONST indicates that - * TCG can assume the value to be constant (ie load at translate time) - * and 64BIT indicates a 64 bit wide coprocessor register. SUPPRESS_TB_END - * indicates that the TB should not be ended after a write to this register - * (the default is that the TB ends after cp writes). OVERRIDE permits - * a register definition to override a previous definition for the - * same (cp, is64, crn, crm, opc1, opc2) tuple: either the new or the - * old must have the OVERRIDE bit set. - * ALIAS indicates that this register is an alias view of some underlying - * state which is also visible via another register, and that the other - * register is handling migration and reset; registers marked ALIAS will not be - * migrated but may have their state set by syncing of register state from KVM. - * NO_RAW indicates that this register has no underlying state and does not - * support raw access for state saving/loading; it will not be used for either - * migration or KVM state synchronization. (Typically this is for "registers" - * which are actually used as instructions for cache maintenance and so on.) - * IO indicates that this register does I/O and therefore its accesses - * need to be marked with gen_io_start() and also end the TB. In particular, - * registers which implement clocks or timers require this. - * RAISES_EXC is for when the read or write hook might raise an exception; - * the generated code will synchronize the CPU state before calling the hook - * so that it is safe for the hook to call raise_exception(). - * NEWEL is for writes to registers that might change the exception - * level - typically on older ARM chips. For those cases we need to - * re-read the new el when recomputing the translation flags. + * ARMCPRegInfo type field bits: */ -#define ARM_CP_SPECIAL 0x0001 -#define ARM_CP_CONST 0x0002 -#define ARM_CP_64BIT 0x0004 -#define ARM_CP_SUPPRESS_TB_END 0x0008 -#define ARM_CP_OVERRIDE 0x0010 -#define ARM_CP_ALIAS 0x0020 -#define ARM_CP_IO 0x0040 -#define ARM_CP_NO_RAW 0x0080 -#define ARM_CP_NOP (ARM_CP_SPECIAL | 0x0100) -#define ARM_CP_WFI (ARM_CP_SPECIAL | 0x0200) -#define ARM_CP_NZCV (ARM_CP_SPECIAL | 0x0300) -#define ARM_CP_CURRENTEL (ARM_CP_SPECIAL | 0x0400) -#define ARM_CP_DC_ZVA (ARM_CP_SPECIAL | 0x0500) -#define ARM_CP_DC_GVA (ARM_CP_SPECIAL | 0x0600) -#define ARM_CP_DC_GZVA (ARM_CP_SPECIAL | 0x0700) -#define ARM_LAST_SPECIAL ARM_CP_DC_GZVA -#define ARM_CP_FPU 0x1000 -#define ARM_CP_SVE 0x2000 -#define ARM_CP_NO_GDB 0x4000 -#define ARM_CP_RAISES_EXC 0x8000 -#define ARM_CP_NEWEL 0x10000 -/* Mask of only the flag bits in a type field */ -#define ARM_CP_FLAG_MASK 0x1f0ff +enum { + /* + * Register must be handled specially during translation. + * The method is one of the values below: + */ + ARM_CP_SPECIAL_MASK = 0x000f, + /* Special: no change to PE state: writes ignored, reads ignored. */ + ARM_CP_NOP = 0x0001, + /* Special: sysreg is WFI, for v5 and v6. */ + ARM_CP_WFI = 0x0002, + /* Special: sysreg is NZCV. */ + ARM_CP_NZCV = 0x0003, + /* Special: sysreg is CURRENTEL. */ + ARM_CP_CURRENTEL = 0x0004, + /* Special: sysreg is DC ZVA or similar. */ + ARM_CP_DC_ZVA = 0x0005, + ARM_CP_DC_GVA = 0x0006, + ARM_CP_DC_GZVA = 0x0007, + + /* Flag: reads produce resetvalue; writes ignored. */ + ARM_CP_CONST = 1 << 4, + /* Flag: For ARM_CP_STATE_AA32, sysreg is 64-bit. */ + ARM_CP_64BIT = 1 << 5, + /* + * Flag: TB should not be ended after a write to this register + * (the default is that the TB ends after cp writes). + */ + ARM_CP_SUPPRESS_TB_END = 1 << 6, + /* + * Flag: Permit a register definition to override a previous definition + * for the same (cp, is64, crn, crm, opc1, opc2) tuple: either the new + * or the old must have the ARM_CP_OVERRIDE bit set. + */ + ARM_CP_OVERRIDE = 1 << 7, + /* + * Flag: Register is an alias view of some underlying state which is also + * visible via another register, and that the other register is handling + * migration and reset; registers marked ARM_CP_ALIAS will not be migrated + * but may have their state set by syncing of register state from KVM. + */ + ARM_CP_ALIAS = 1 << 8, + /* + * Flag: Register does I/O and therefore its accesses need to be marked + * with gen_io_start() and also end the TB. In particular, registers which + * implement clocks or timers require this. + */ + ARM_CP_IO = 1 << 9, + /* + * Flag: Register has no underlying state and does not support raw access + * for state saving/loading; it will not be used for either migration or + * KVM state synchronization. Typically this is for "registers" which are + * actually used as instructions for cache maintenance and so on. + */ + ARM_CP_NO_RAW = 1 << 10, + /* + * Flag: The read or write hook might raise an exception; the generated + * code will synchronize the CPU state before calling the hook so that it + * is safe for the hook to call raise_exception(). + */ + ARM_CP_RAISES_EXC = 1 << 11, + /* + * Flag: Writes to the sysreg might change the exception level - typically + * on older ARM chips. For those cases we need to re-read the new el when + * recomputing the translation flags. + */ + ARM_CP_NEWEL = 1 << 12, + /* + * Flag: Access check for this sysreg is identical to accessing FPU state + * from an instruction: use translation fp_access_check(). + */ + ARM_CP_FPU = 1 << 13, + /* + * Flag: Access check for this sysreg is identical to accessing SVE state + * from an instruction: use translation sve_access_check(). + */ + ARM_CP_SVE = 1 << 14, + /* Flag: Do not expose in gdb sysreg xml. */ + ARM_CP_NO_GDB = 1 << 15, +}; /* * Valid values for ARMCPRegInfo state field, indicating which of diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void cp_reg_reset(gpointer key, gpointer value, gpointer opaque) ARMCPRegInfo *ri = value; ARMCPU *cpu = opaque; - if (ri->type & (ARM_CP_SPECIAL | ARM_CP_ALIAS)) { + if (ri->type & (ARM_CP_SPECIAL_MASK | ARM_CP_ALIAS)) { return; } @@ -XXX,XX +XXX,XX @@ static void cp_reg_check_reset(gpointer key, gpointer value, gpointer opaque) ARMCPU *cpu = opaque; uint64_t oldvalue, newvalue; - if (ri->type & (ARM_CP_SPECIAL | ARM_CP_ALIAS | ARM_CP_NO_RAW)) { + if (ri->type & (ARM_CP_SPECIAL_MASK | ARM_CP_ALIAS | ARM_CP_NO_RAW)) { return; } diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, * multiple times. Special registers (ie NOP/WFI) are * never migratable and not even raw-accessible. */ - if ((r->type & ARM_CP_SPECIAL)) { + if (r->type & ARM_CP_SPECIAL_MASK) { r2->type |= ARM_CP_NO_RAW; } if (((r->crm == CP_ANY) && crm != 0) || @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, /* Check that the register definition has enough info to handle * reads and writes if they are permitted. */ - if (!(r->type & (ARM_CP_SPECIAL|ARM_CP_CONST))) { + if (!(r->type & (ARM_CP_SPECIAL_MASK | ARM_CP_CONST))) { if (r->access & PL3_R) { assert((r->fieldoffset || (r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1])) || diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, } /* Handle special cases first */ - switch (ri->type & ~(ARM_CP_FLAG_MASK & ~ARM_CP_SPECIAL)) { + switch (ri->type & ARM_CP_SPECIAL_MASK) { + case 0: + break; case ARM_CP_NOP: return; case ARM_CP_NZCV: @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, } return; default: - break; + g_assert_not_reached(); } if ((ri->type & ARM_CP_FPU) && !fp_access_check(s)) { return; diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void do_coproc_insn(DisasContext *s, int cpnum, int is64, } /* Handle special cases first */ - switch (ri->type & ~(ARM_CP_FLAG_MASK & ~ARM_CP_SPECIAL)) { + switch (ri->type & ARM_CP_SPECIAL_MASK) { + case 0: + break; case ARM_CP_NOP: return; case ARM_CP_WFI: @@ -XXX,XX +XXX,XX @@ static void do_coproc_insn(DisasContext *s, int cpnum, int is64, s->base.is_jmp = DISAS_WFI; return; default: - break; + g_assert_not_reached(); } if ((tb_cflags(s->base.tb) & CF_USE_ICOUNT) && (ri->type & ARM_CP_IO)) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Standardize on g_assert_not_reached() for "should not happen". Retain abort() when preceeded by fprintf or error_report. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 7 +++---- target/arm/hvf/hvf.c | 2 +- target/arm/kvm-stub.c | 4 ++-- target/arm/kvm.c | 4 ++-- target/arm/machine.c | 4 ++-- target/arm/translate-a64.c | 4 ++-- target/arm/translate-neon.c | 2 +- target/arm/translate.c | 4 ++-- 8 files changed, 15 insertions(+), 16 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, break; default: /* broken reginfo with out-of-range opc1 */ - assert(false); - break; + g_assert_not_reached(); } /* assert our permissions are not too lax (stricter is fine) */ assert((r->access & ~mask) == 0); @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_v5(CPUARMState *env, uint32_t address, break; default: /* Never happens, but compiler isn't smart enough to tell. */ - abort(); + g_assert_not_reached(); } } *prot = ap_to_rw_prot(env, mmu_idx, ap, domain_prot); @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_v6(CPUARMState *env, uint32_t address, break; default: /* Never happens, but compiler isn't smart enough to tell. */ - abort(); + g_assert_not_reached(); } } if (domain_prot == 3) { diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/hvf/hvf.c +++ b/target/arm/hvf/hvf.c @@ -XXX,XX +XXX,XX @@ int hvf_vcpu_exec(CPUState *cpu) /* we got kicked, no exit to process */ return 0; default: - assert(0); + g_assert_not_reached(); } hvf_sync_vtimer(cpu); diff --git a/target/arm/kvm-stub.c b/target/arm/kvm-stub.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm-stub.c +++ b/target/arm/kvm-stub.c @@ -XXX,XX +XXX,XX @@ bool write_kvmstate_to_list(ARMCPU *cpu) { - abort(); + g_assert_not_reached(); } bool write_list_to_kvmstate(ARMCPU *cpu, int level) { - abort(); + g_assert_not_reached(); } diff --git a/target/arm/kvm.c b/target/arm/kvm.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm.c +++ b/target/arm/kvm.c @@ -XXX,XX +XXX,XX @@ bool write_kvmstate_to_list(ARMCPU *cpu) ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); break; default: - abort(); + g_assert_not_reached(); } if (ret) { ok = false; @@ -XXX,XX +XXX,XX @@ bool write_list_to_kvmstate(ARMCPU *cpu, int level) r.addr = (uintptr_t)(cpu->cpreg_values + i); break; default: - abort(); + g_assert_not_reached(); } ret = kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, &r); if (ret) { diff --git a/target/arm/machine.c b/target/arm/machine.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/machine.c +++ b/target/arm/machine.c @@ -XXX,XX +XXX,XX @@ static int cpu_pre_save(void *opaque) if (kvm_enabled()) { if (!write_kvmstate_to_list(cpu)) { /* This should never fail */ - abort(); + g_assert_not_reached(); } /* @@ -XXX,XX +XXX,XX @@ static int cpu_pre_save(void *opaque) } else { if (!write_cpustate_to_list(cpu, false)) { /* This should never fail. */ - abort(); + g_assert_not_reached(); } } diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_half(DisasContext *s, int opcode, int rd, int rn) gen_helper_advsimd_rinth(tcg_res, tcg_op, fpst); break; default: - abort(); + g_assert_not_reached(); } write_fp_sreg(s, rd, tcg_res); @@ -XXX,XX +XXX,XX @@ static void handle_fp_fcvt(DisasContext *s, int opcode, break; } default: - abort(); + g_assert_not_reached(); } } diff --git a/target/arm/translate-neon.c b/target/arm/translate-neon.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.c +++ b/target/arm/translate-neon.c @@ -XXX,XX +XXX,XX @@ static bool trans_VLDST_single(DisasContext *s, arg_VLDST_single *a) } break; default: - abort(); + g_assert_not_reached(); } if ((vd + a->stride * (nregs - 1)) > 31) { /* diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_srs(DisasContext *s, offset = 4; break; default: - abort(); + g_assert_not_reached(); } tcg_gen_addi_i32(addr, addr, offset); tmp = load_reg(s, 14); @@ -XXX,XX +XXX,XX @@ static void gen_srs(DisasContext *s, offset = 0; break; default: - abort(); + g_assert_not_reached(); } tcg_gen_addi_i32(addr, addr, offset); gen_helper_set_r13_banked(cpu_env, tcg_constant_i32(mode), addr); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Create a typedef as well, and use it in ARMCPRegInfo. This won't be perfect for debugging, but it'll nicely display the most common cases. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 44 +++++++++++++++++++++++--------------------- target/arm/helper.c | 2 +- 2 files changed, 24 insertions(+), 22 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ enum { * described with these bits, then use a laxer set of restrictions, and * do the more restrictive/complex check inside a helper function. */ -#define PL3_R 0x80 -#define PL3_W 0x40 -#define PL2_R (0x20 | PL3_R) -#define PL2_W (0x10 | PL3_W) -#define PL1_R (0x08 | PL2_R) -#define PL1_W (0x04 | PL2_W) -#define PL0_R (0x02 | PL1_R) -#define PL0_W (0x01 | PL1_W) +typedef enum { + PL3_R = 0x80, + PL3_W = 0x40, + PL2_R = 0x20 | PL3_R, + PL2_W = 0x10 | PL3_W, + PL1_R = 0x08 | PL2_R, + PL1_W = 0x04 | PL2_W, + PL0_R = 0x02 | PL1_R, + PL0_W = 0x01 | PL1_W, -/* - * For user-mode some registers are accessible to EL0 via a kernel - * trap-and-emulate ABI. In this case we define the read permissions - * as actually being PL0_R. However some bits of any given register - * may still be masked. - */ + /* + * For user-mode some registers are accessible to EL0 via a kernel + * trap-and-emulate ABI. In this case we define the read permissions + * as actually being PL0_R. However some bits of any given register + * may still be masked. + */ #ifdef CONFIG_USER_ONLY -#define PL0U_R PL0_R + PL0U_R = PL0_R, #else -#define PL0U_R PL1_R + PL0U_R = PL1_R, #endif -#define PL3_RW (PL3_R | PL3_W) -#define PL2_RW (PL2_R | PL2_W) -#define PL1_RW (PL1_R | PL1_W) -#define PL0_RW (PL0_R | PL0_W) + PL3_RW = PL3_R | PL3_W, + PL2_RW = PL2_R | PL2_W, + PL1_RW = PL1_R | PL1_W, + PL0_RW = PL0_R | PL0_W, +} CPAccessRights; typedef enum CPAccessResult { /* Access is permitted */ @@ -XXX,XX +XXX,XX @@ struct ARMCPRegInfo { /* Register type: ARM_CP_* bits/values */ int type; /* Access rights: PL*_[RW] */ - int access; + CPAccessRights access; /* Security state: ARM_CP_SECSTATE_* bits/values */ int secure; /* diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, * to encompass the generic architectural permission check. */ if (r->state != ARM_CP_STATE_AA32) { - int mask = 0; + CPAccessRights mask; switch (r->opc1) { case 0: /* min_EL EL1, but some accessible to EL0 via kernel ABI */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Give this enum a name and use in ARMCPRegInfo, add_cpreg_to_hashtable and define_one_arm_cp_reg_with_opaque. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 6 +++--- target/arm/helper.c | 6 ++++-- 2 files changed, 7 insertions(+), 5 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ enum { * Note that we rely on the values of these enums as we iterate through * the various states in some places. */ -enum { +typedef enum { ARM_CP_STATE_AA32 = 0, ARM_CP_STATE_AA64 = 1, ARM_CP_STATE_BOTH = 2, -}; +} CPState; /* * ARM CP register secure state flags. These flags identify security state @@ -XXX,XX +XXX,XX @@ struct ARMCPRegInfo { uint8_t opc1; uint8_t opc2; /* Execution state in which this register is visible: ARM_CP_STATE_* */ - int state; + CPState state; /* Register type: ARM_CP_* bits/values */ int type; /* Access rights: PL*_[RW] */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp) } static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, - void *opaque, int state, int secstate, + void *opaque, CPState state, int secstate, int crm, int opc1, int opc2, const char *name) { @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, * bits; the ARM_CP_64BIT* flag applies only to the AArch32 view of * the register, if any. */ - int crm, opc1, opc2, state; + int crm, opc1, opc2; int crmmin = (r->crm == CP_ANY) ? 0 : r->crm; int crmmax = (r->crm == CP_ANY) ? 15 : r->crm; int opc1min = (r->opc1 == CP_ANY) ? 0 : r->opc1; int opc1max = (r->opc1 == CP_ANY) ? 7 : r->opc1; int opc2min = (r->opc2 == CP_ANY) ? 0 : r->opc2; int opc2max = (r->opc2 == CP_ANY) ? 7 : r->opc2; + CPState state; + /* 64 bit registers have only CRm and Opc1 fields */ assert(!((r->type & ARM_CP_64BIT) && (r->opc2 || r->crn))); /* op0 only exists in the AArch64 encodings */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Give this enum a name and use in ARMCPRegInfo and add_cpreg_to_hashtable. Add the enumerator ARM_CP_SECSTATE_BOTH to clarify how 0 is handled in define_one_arm_cp_reg_with_opaque. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 7 ++++--- target/arm/helper.c | 7 +++++-- 2 files changed, 9 insertions(+), 5 deletions(-) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ typedef enum { * registered entry will only have one to identify whether the entry is secure * or non-secure. */ -enum { +typedef enum { + ARM_CP_SECSTATE_BOTH = 0, /* define one cpreg for each secstate */ ARM_CP_SECSTATE_S = (1 << 0), /* bit[0]: Secure state register */ ARM_CP_SECSTATE_NS = (1 << 1), /* bit[1]: Non-secure state register */ -}; +} CPSecureState; /* * Access rights: @@ -XXX,XX +XXX,XX @@ struct ARMCPRegInfo { /* Access rights: PL*_[RW] */ CPAccessRights access; /* Security state: ARM_CP_SECSTATE_* bits/values */ - int secure; + CPSecureState secure; /* * The opaque pointer passed to define_arm_cp_regs_with_opaque() when * this register was defined: can be used to hand data through to the diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp) } static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, - void *opaque, CPState state, int secstate, + void *opaque, CPState state, + CPSecureState secstate, int crm, int opc1, int opc2, const char *name) { @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, r->secure, crm, opc1, opc2, r->name); break; - default: + case ARM_CP_SECSTATE_BOTH: name = g_strdup_printf("%s_S", r->name); add_cpreg_to_hashtable(cpu, r, opaque, state, ARM_CP_SECSTATE_S, @@ -XXX,XX +XXX,XX @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu, ARM_CP_SECSTATE_NS, crm, opc1, opc2, r->name); break; + default: + g_assert_not_reached(); } } else { /* AArch64 registers get mapped to non-secure instance -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> The new_key field is always non-zero -- drop the if. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-11-richard.henderson@linaro.org [PMM: reinstated dropped PL3_RW mask] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) for (i = 0; i < ARRAY_SIZE(aliases); i++) { const struct E2HAlias *a = &aliases[i]; - ARMCPRegInfo *src_reg, *dst_reg; + ARMCPRegInfo *src_reg, *dst_reg, *new_reg; + uint32_t *new_key; + bool ok; if (a->feature && !a->feature(&cpu->isar)) { continue; @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) g_assert(src_reg->opaque == NULL); /* Create alias before redirection so we dup the right data. */ - if (a->new_key) { - ARMCPRegInfo *new_reg = g_memdup(src_reg, sizeof(ARMCPRegInfo)); - uint32_t *new_key = g_memdup(&a->new_key, sizeof(uint32_t)); - bool ok; + new_reg = g_memdup(src_reg, sizeof(ARMCPRegInfo)); + new_key = g_memdup(&a->new_key, sizeof(uint32_t)); - new_reg->name = a->new_name; - new_reg->type |= ARM_CP_ALIAS; - /* Remove PL1/PL0 access, leaving PL2/PL3 R/W in place. */ - new_reg->access &= PL2_RW | PL3_RW; + new_reg->name = a->new_name; + new_reg->type |= ARM_CP_ALIAS; + /* Remove PL1/PL0 access, leaving PL2/PL3 R/W in place. */ + new_reg->access &= PL2_RW | PL3_RW; - ok = g_hash_table_insert(cpu->cp_regs, new_key, new_reg); - g_assert(ok); - } + ok = g_hash_table_insert(cpu->cp_regs, new_key, new_reg); + g_assert(ok); src_reg->opaque = dst_reg; src_reg->orig_readfn = src_reg->readfn ?: raw_read; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Cast the uint32_t key into a gpointer directly, which allows us to avoid allocating storage for each key. Use g_hash_table_lookup when we already have a gpointer (e.g. for callbacks like count_cpreg), or when using get_arm_cp_reginfo would require casting away const. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 4 ++-- target/arm/gdbstub.c | 2 +- target/arm/helper.c | 41 ++++++++++++++++++----------------------- 3 files changed, 21 insertions(+), 26 deletions(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_initfn(Object *obj) ARMCPU *cpu = ARM_CPU(obj); cpu_set_cpustate_pointers(cpu); - cpu->cp_regs = g_hash_table_new_full(g_int_hash, g_int_equal, - g_free, cpreg_hashtable_data_destroy); + cpu->cp_regs = g_hash_table_new_full(g_direct_hash, g_direct_equal, + NULL, cpreg_hashtable_data_destroy); QLIST_INIT(&cpu->pre_el_change_hooks); QLIST_INIT(&cpu->el_change_hooks); diff --git a/target/arm/gdbstub.c b/target/arm/gdbstub.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/gdbstub.c +++ b/target/arm/gdbstub.c @@ -XXX,XX +XXX,XX @@ static void arm_gen_one_xml_sysreg_tag(GString *s, DynamicGDBXMLInfo *dyn_xml, static void arm_register_sysreg_for_xml(gpointer key, gpointer value, gpointer p) { - uint32_t ri_key = *(uint32_t *)key; + uint32_t ri_key = (uintptr_t)key; ARMCPRegInfo *ri = value; RegisterSysregXmlParam *param = (RegisterSysregXmlParam *)p; GString *s = param->s; diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ bool write_list_to_cpustate(ARMCPU *cpu) static void add_cpreg_to_list(gpointer key, gpointer opaque) { ARMCPU *cpu = opaque; - uint64_t regidx; - const ARMCPRegInfo *ri; - - regidx = *(uint32_t *)key; - ri = get_arm_cp_reginfo(cpu->cp_regs, regidx); + uint32_t regidx = (uintptr_t)key; + const ARMCPRegInfo *ri = get_arm_cp_reginfo(cpu->cp_regs, regidx); if (!(ri->type & (ARM_CP_NO_RAW|ARM_CP_ALIAS))) { cpu->cpreg_indexes[cpu->cpreg_array_len] = cpreg_to_kvm_id(regidx); @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_list(gpointer key, gpointer opaque) static void count_cpreg(gpointer key, gpointer opaque) { ARMCPU *cpu = opaque; - uint64_t regidx; const ARMCPRegInfo *ri; - regidx = *(uint32_t *)key; - ri = get_arm_cp_reginfo(cpu->cp_regs, regidx); + ri = g_hash_table_lookup(cpu->cp_regs, key); if (!(ri->type & (ARM_CP_NO_RAW|ARM_CP_ALIAS))) { cpu->cpreg_array_len++; @@ -XXX,XX +XXX,XX @@ static void count_cpreg(gpointer key, gpointer opaque) static gint cpreg_key_compare(gconstpointer a, gconstpointer b) { - uint64_t aidx = cpreg_to_kvm_id(*(uint32_t *)a); - uint64_t bidx = cpreg_to_kvm_id(*(uint32_t *)b); + uint64_t aidx = cpreg_to_kvm_id((uintptr_t)a); + uint64_t bidx = cpreg_to_kvm_id((uintptr_t)b); if (aidx > bidx) { return 1; @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) for (i = 0; i < ARRAY_SIZE(aliases); i++) { const struct E2HAlias *a = &aliases[i]; ARMCPRegInfo *src_reg, *dst_reg, *new_reg; - uint32_t *new_key; bool ok; if (a->feature && !a->feature(&cpu->isar)) { continue; } - src_reg = g_hash_table_lookup(cpu->cp_regs, &a->src_key); - dst_reg = g_hash_table_lookup(cpu->cp_regs, &a->dst_key); + src_reg = g_hash_table_lookup(cpu->cp_regs, + (gpointer)(uintptr_t)a->src_key); + dst_reg = g_hash_table_lookup(cpu->cp_regs, + (gpointer)(uintptr_t)a->dst_key); g_assert(src_reg != NULL); g_assert(dst_reg != NULL); @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) /* Create alias before redirection so we dup the right data. */ new_reg = g_memdup(src_reg, sizeof(ARMCPRegInfo)); - new_key = g_memdup(&a->new_key, sizeof(uint32_t)); new_reg->name = a->new_name; new_reg->type |= ARM_CP_ALIAS; /* Remove PL1/PL0 access, leaving PL2/PL3 R/W in place. */ new_reg->access &= PL2_RW | PL3_RW; - ok = g_hash_table_insert(cpu->cp_regs, new_key, new_reg); + ok = g_hash_table_insert(cpu->cp_regs, + (gpointer)(uintptr_t)a->new_key, new_reg); g_assert(ok); src_reg->opaque = dst_reg; @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, /* Private utility function for define_one_arm_cp_reg_with_opaque(): * add a single reginfo struct to the hash table. */ - uint32_t *key = g_new(uint32_t, 1); + uint32_t key; ARMCPRegInfo *r2 = g_memdup(r, sizeof(ARMCPRegInfo)); int is64 = (r->type & ARM_CP_64BIT) ? 1 : 0; int ns = (secstate & ARM_CP_SECSTATE_NS) ? 1 : 0; @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, if (r->cp == 0 || r->state == ARM_CP_STATE_BOTH) { r2->cp = CP_REG_ARM64_SYSREG_CP; } - *key = ENCODE_AA64_CP_REG(r2->cp, r2->crn, crm, - r2->opc0, opc1, opc2); + key = ENCODE_AA64_CP_REG(r2->cp, r2->crn, crm, + r2->opc0, opc1, opc2); } else { - *key = ENCODE_CP_REG(r2->cp, is64, ns, r2->crn, crm, opc1, opc2); + key = ENCODE_CP_REG(r2->cp, is64, ns, r2->crn, crm, opc1, opc2); } if (opaque) { r2->opaque = opaque; @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, * requested. */ if (!(r->type & ARM_CP_OVERRIDE)) { - ARMCPRegInfo *oldreg; - oldreg = g_hash_table_lookup(cpu->cp_regs, key); + const ARMCPRegInfo *oldreg = get_arm_cp_reginfo(cpu->cp_regs, key); if (oldreg && !(oldreg->type & ARM_CP_OVERRIDE)) { fprintf(stderr, "Register redefined: cp=%d %d bit " "crn=%d crm=%d opc1=%d opc2=%d, " @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, g_assert_not_reached(); } } - g_hash_table_insert(cpu->cp_regs, key, r2); + g_hash_table_insert(cpu->cp_regs, (gpointer)(uintptr_t)key, r2); } @@ -XXX,XX +XXX,XX @@ void modify_arm_cp_regs_with_len(ARMCPRegInfo *regs, size_t regs_len, const ARMCPRegInfo *get_arm_cp_reginfo(GHashTable *cpregs, uint32_t encoded_cp) { - return g_hash_table_lookup(cpregs, &encoded_cp); + return g_hash_table_lookup(cpregs, (gpointer)(uintptr_t)encoded_cp); } void arm_cp_write_ignore(CPUARMState *env, const ARMCPRegInfo *ri, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Simplify freeing cp_regs hash table entries by using a single allocation for the entire value. This fixes a theoretical bug if we were to ever free the entire hash table, because we've been installing string literal constants into the cpreg structure in define_arm_vh_e2h_redirects_aliases. However, at present we only free entries created for AArch32 wildcard cpregs which get overwritten by more specific cpregs, so this bug is never exposed. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 16 +--------------- target/arm/helper.c | 10 ++++++++-- 2 files changed, 9 insertions(+), 17 deletions(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ uint64_t arm_cpu_mp_affinity(int idx, uint8_t clustersz) return (Aff1 << ARM_AFF1_SHIFT) | Aff0; } -static void cpreg_hashtable_data_destroy(gpointer data) -{ - /* - * Destroy function for cpu->cp_regs hashtable data entries. - * We must free the name string because it was g_strdup()ed in - * add_cpreg_to_hashtable(). It's OK to cast away the 'const' - * from r->name because we know we definitely allocated it. - */ - ARMCPRegInfo *r = data; - - g_free((void *)r->name); - g_free(r); -} - static void arm_cpu_initfn(Object *obj) { ARMCPU *cpu = ARM_CPU(obj); cpu_set_cpustate_pointers(cpu); cpu->cp_regs = g_hash_table_new_full(g_direct_hash, g_direct_equal, - NULL, cpreg_hashtable_data_destroy); + NULL, g_free); QLIST_INIT(&cpu->pre_el_change_hooks); QLIST_INIT(&cpu->el_change_hooks); diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, * add a single reginfo struct to the hash table. */ uint32_t key; - ARMCPRegInfo *r2 = g_memdup(r, sizeof(ARMCPRegInfo)); + ARMCPRegInfo *r2; int is64 = (r->type & ARM_CP_64BIT) ? 1 : 0; int ns = (secstate & ARM_CP_SECSTATE_NS) ? 1 : 0; + size_t name_len; + + /* Combine cpreg and name into one allocation. */ + name_len = strlen(name) + 1; + r2 = g_malloc(sizeof(*r2) + name_len); + *r2 = *r; + r2->name = memcpy(r2 + 1, name, name_len); - r2->name = g_strdup(name); /* Reset the secure state to the specific incoming state. This is * necessary as the register may have been defined with both states. */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Move the computation of key to the top of the function. Hoist the resolution of cp as well, as an input to the computation of key. This will be required by a subsequent patch. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 49 +++++++++++++++++++++++++-------------------- 1 file changed, 27 insertions(+), 22 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, ARMCPRegInfo *r2; int is64 = (r->type & ARM_CP_64BIT) ? 1 : 0; int ns = (secstate & ARM_CP_SECSTATE_NS) ? 1 : 0; + int cp = r->cp; size_t name_len; + switch (state) { + case ARM_CP_STATE_AA32: + /* We assume it is a cp15 register if the .cp field is left unset. */ + if (cp == 0 && r->state == ARM_CP_STATE_BOTH) { + cp = 15; + } + key = ENCODE_CP_REG(cp, is64, ns, r->crn, crm, opc1, opc2); + break; + case ARM_CP_STATE_AA64: + /* + * To allow abbreviation of ARMCPRegInfo definitions, we treat + * cp == 0 as equivalent to the value for "standard guest-visible + * sysreg". STATE_BOTH definitions are also always "standard sysreg" + * in their AArch64 view (the .cp value may be non-zero for the + * benefit of the AArch32 view). + */ + if (cp == 0 || r->state == ARM_CP_STATE_BOTH) { + cp = CP_REG_ARM64_SYSREG_CP; + } + key = ENCODE_AA64_CP_REG(cp, r->crn, crm, r->opc0, opc1, opc2); + break; + default: + g_assert_not_reached(); + } + /* Combine cpreg and name into one allocation. */ name_len = strlen(name) + 1; r2 = g_malloc(sizeof(*r2) + name_len); @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, } if (r->state == ARM_CP_STATE_BOTH) { - /* We assume it is a cp15 register if the .cp field is left unset. - */ - if (r2->cp == 0) { - r2->cp = 15; - } - #if HOST_BIG_ENDIAN if (r2->fieldoffset) { r2->fieldoffset += sizeof(uint32_t); @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, #endif } } - if (state == ARM_CP_STATE_AA64) { - /* To allow abbreviation of ARMCPRegInfo - * definitions, we treat cp == 0 as equivalent to - * the value for "standard guest-visible sysreg". - * STATE_BOTH definitions are also always "standard - * sysreg" in their AArch64 view (the .cp value may - * be non-zero for the benefit of the AArch32 view). - */ - if (r->cp == 0 || r->state == ARM_CP_STATE_BOTH) { - r2->cp = CP_REG_ARM64_SYSREG_CP; - } - key = ENCODE_AA64_CP_REG(r2->cp, r2->crn, crm, - r2->opc0, opc1, opc2); - } else { - key = ENCODE_CP_REG(r2->cp, is64, ns, r2->crn, crm, opc1, opc2); - } if (opaque) { r2->opaque = opaque; } @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, /* Make sure reginfo passed to helpers for wildcarded regs * has the correct crm/opc1/opc2 for this reg, not CP_ANY: */ + r2->cp = cp; r2->crm = crm; r2->opc1 = opc1; r2->opc2 = opc2; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Put most of the value writeback to the same place, and improve the comment that goes with them. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 28 ++++++++++++---------------- 1 file changed, 12 insertions(+), 16 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, *r2 = *r; r2->name = memcpy(r2 + 1, name, name_len); - /* Reset the secure state to the specific incoming state. This is - * necessary as the register may have been defined with both states. + /* + * Update fields to match the instantiation, overwiting wildcards + * such as CP_ANY, ARM_CP_STATE_BOTH, or ARM_CP_SECSTATE_BOTH. */ + r2->cp = cp; + r2->crm = crm; + r2->opc1 = opc1; + r2->opc2 = opc2; + r2->state = state; r2->secure = secstate; + if (opaque) { + r2->opaque = opaque; + } if (r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1]) { /* Register is banked (using both entries in array). @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, #endif } } - if (opaque) { - r2->opaque = opaque; - } - /* reginfo passed to helpers is correct for the actual access, - * and is never ARM_CP_STATE_BOTH: - */ - r2->state = state; - /* Make sure reginfo passed to helpers for wildcarded regs - * has the correct crm/opc1/opc2 for this reg, not CP_ANY: - */ - r2->cp = cp; - r2->crm = crm; - r2->opc1 = opc1; - r2->opc2 = opc2; + /* By convention, for wildcarded registers only the first * entry is used for migration; the others are marked as * ALIAS so we don't try to transfer the register -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Bool is a more appropriate type for these variables. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, */ uint32_t key; ARMCPRegInfo *r2; - int is64 = (r->type & ARM_CP_64BIT) ? 1 : 0; - int ns = (secstate & ARM_CP_SECSTATE_NS) ? 1 : 0; + bool is64 = r->type & ARM_CP_64BIT; + bool ns = secstate & ARM_CP_SECSTATE_NS; int cp = r->cp; size_t name_len; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Computing isbanked only once makes the code a bit easier to read. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, bool is64 = r->type & ARM_CP_64BIT; bool ns = secstate & ARM_CP_SECSTATE_NS; int cp = r->cp; + bool isbanked; size_t name_len; switch (state) { @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, r2->opaque = opaque; } - if (r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1]) { + isbanked = r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1]; + if (isbanked) { /* Register is banked (using both entries in array). * Overwriting fieldoffset as the array is only used to define * banked registers but later only fieldoffset is used. @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, } if (state == ARM_CP_STATE_AA32) { - if (r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1]) { + if (isbanked) { /* If the register is banked then we don't need to migrate or * reset the 32-bit instance in certain cases: * -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Perform the override check early, so that it is still done even when we decide to discard an unreachable cpreg. Use assert not printf+abort. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, g_assert_not_reached(); } + /* Overriding of an existing definition must be explicitly requested. */ + if (!(r->type & ARM_CP_OVERRIDE)) { + const ARMCPRegInfo *oldreg = get_arm_cp_reginfo(cpu->cp_regs, key); + if (oldreg) { + assert(oldreg->type & ARM_CP_OVERRIDE); + } + } + /* Combine cpreg and name into one allocation. */ name_len = strlen(name) + 1; r2 = g_malloc(sizeof(*r2) + name_len); @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, assert(!raw_accessors_invalid(r2)); } - /* Overriding of an existing definition must be explicitly - * requested. - */ - if (!(r->type & ARM_CP_OVERRIDE)) { - const ARMCPRegInfo *oldreg = get_arm_cp_reginfo(cpu->cp_regs, key); - if (oldreg && !(oldreg->type & ARM_CP_OVERRIDE)) { - fprintf(stderr, "Register redefined: cp=%d %d bit " - "crn=%d crm=%d opc1=%d opc2=%d, " - "was %s, now %s\n", r2->cp, 32 + 32 * is64, - r2->crn, r2->crm, r2->opc1, r2->opc2, - oldreg->name, r2->name); - g_assert_not_reached(); - } - } g_hash_table_insert(cpu->cp_regs, (gpointer)(uintptr_t)key, r2); } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Put the block comments into the current coding style. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp) return cpu_list; } +/* + * Private utility function for define_one_arm_cp_reg_with_opaque(): + * add a single reginfo struct to the hash table. + */ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, void *opaque, CPState state, CPSecureState secstate, int crm, int opc1, int opc2, const char *name) { - /* Private utility function for define_one_arm_cp_reg_with_opaque(): - * add a single reginfo struct to the hash table. - */ uint32_t key; ARMCPRegInfo *r2; bool is64 = r->type & ARM_CP_64BIT; @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, isbanked = r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1]; if (isbanked) { - /* Register is banked (using both entries in array). + /* + * Register is banked (using both entries in array). * Overwriting fieldoffset as the array is only used to define * banked registers but later only fieldoffset is used. */ @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, if (state == ARM_CP_STATE_AA32) { if (isbanked) { - /* If the register is banked then we don't need to migrate or + /* + * If the register is banked then we don't need to migrate or * reset the 32-bit instance in certain cases: * * 1) If the register has both 32-bit and 64-bit instances then we @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, r2->type |= ARM_CP_ALIAS; } } else if ((secstate != r->secure) && !ns) { - /* The register is not banked so we only want to allow migration of - * the non-secure instance. + /* + * The register is not banked so we only want to allow migration + * of the non-secure instance. */ r2->type |= ARM_CP_ALIAS; } @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, } } - /* By convention, for wildcarded registers only the first + /* + * By convention, for wildcarded registers only the first * entry is used for migration; the others are marked as * ALIAS so we don't try to transfer the register * multiple times. Special registers (ie NOP/WFI) are @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, r2->type |= ARM_CP_ALIAS | ARM_CP_NO_GDB; } - /* Check that raw accesses are either forbidden or handled. Note that + /* + * Check that raw accesses are either forbidden or handled. Note that * we can't assert this earlier because the setup of fieldoffset for * banked registers has to be done first. */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Since e03b56863d2bc, our host endian indicator is unconditionally set, which means that we can use a normal C condition. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220501055028.646596-20-richard.henderson@linaro.org [PMM: quote correct git hash in commit message] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r, r2->type |= ARM_CP_ALIAS; } - if (r->state == ARM_CP_STATE_BOTH) { -#if HOST_BIG_ENDIAN - if (r2->fieldoffset) { - r2->fieldoffset += sizeof(uint32_t); - } -#endif + if (HOST_BIG_ENDIAN && + r->state == ARM_CP_STATE_BOTH && r2->fieldoffset) { + r2->fieldoffset += sizeof(uint32_t); } } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-24-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa32_ssbs(const ARMISARegisters *id) return FIELD_EX32(id->id_pfr2, ID_PFR2, SSBS) != 0; } +static inline bool isar_feature_aa32_debugv8p2(const ARMISARegisters *id) +{ + return FIELD_EX32(id->id_dfr0, ID_DFR0, COPDBG) >= 8; +} + /* * 64-bit feature tests via id registers. */ @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_ssbs(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, SSBS) != 0; } +static inline bool isar_feature_aa64_debugv8p2(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64dfr0, ID_AA64DFR0, DEBUGVER) >= 8; +} + static inline bool isar_feature_aa64_sve2(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64zfr0, ID_AA64ZFR0, SVEVER) != 0; @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_any_tts2uxn(const ARMISARegisters *id) return isar_feature_aa64_tts2uxn(id) || isar_feature_aa32_tts2uxn(id); } +static inline bool isar_feature_any_debugv8p2(const ARMISARegisters *id) +{ + return isar_feature_aa64_debugv8p2(id) || isar_feature_aa32_debugv8p2(id); +} + /* * Forward to the above feature tests given an ARMCPU pointer. */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Add the aa64 predicate for detecting RAS support from id registers. We already have the aa32 version from the M-profile work. Add the 'any' predicate for testing both aa64 and aa32. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220501055028.646596-34-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_aa32_el1(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, EL1) >= 2; } +static inline bool isar_feature_aa64_ras(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, RAS) != 0; +} + static inline bool isar_feature_aa64_sve(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64pfr0, ID_AA64PFR0, SVE) != 0; @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_any_debugv8p2(const ARMISARegisters *id) return isar_feature_aa64_debugv8p2(id) || isar_feature_aa32_debugv8p2(id); } +static inline bool isar_feature_any_ras(const ARMISARegisters *id) +{ + return isar_feature_aa64_ras(id) || isar_feature_aa32_ras(id); +} + /* * Forward to the above feature tests given an ARMCPU pointer. */ -- 2.25.1 From: Alex Zuepke <alex.zuepke@tum.de> The ARMv8 manual defines that PMUSERENR_EL0.ER enables read-access to both PMXEVCNTR_EL0 and PMEVCNTR<n>_EL0 registers, however, we only use it for PMXEVCNTR_EL0. Extend to PMEVCNTR<n>_EL0 as well. Signed-off-by: Alex Zuepke <alex.zuepke@tum.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220428132717.84190-1-alex.zuepke@tum.de Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void define_pmu_regs(ARMCPU *cpu) .crm = 8 | (3 & (i >> 3)), .opc1 = 0, .opc2 = i & 7, .access = PL0_RW, .type = ARM_CP_IO | ARM_CP_ALIAS, .readfn = pmevcntr_readfn, .writefn = pmevcntr_writefn, - .accessfn = pmreg_access }, + .accessfn = pmreg_access_xevcntr }, { .name = pmevcntr_el0_name, .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 3, .crn = 14, .crm = 8 | (3 & (i >> 3)), - .opc2 = i & 7, .access = PL0_RW, .accessfn = pmreg_access, + .opc2 = i & 7, .access = PL0_RW, .accessfn = pmreg_access_xevcntr, .type = ARM_CP_IO, .readfn = pmevcntr_readfn, .writefn = pmevcntr_writefn, .raw_readfn = pmevcntr_rawread, -- 2.25.1 Hi; here's another Arm pullreq: the big thing in here is Bernhard's imx8mp-evk board model; there's also various cleanup type patches from me, as well as some bugfixes. thanks -- PMM The following changes since commit b69801dd6b1eb4d107f7c2f643adf0a4e3ec9124: Merge tag 'for_upstream' of https://git.kernel.org/pub/scm/virt/kvm/mst/qemu into staging (2025-02-22 05:06:39 +0800) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20250225 for you to fetch changes up to 1aaf3478684ff1cd02d1b36c32a00bfac9a5dbd5: hw/arm/fsl-imx8mp: Add on-chip RAM (2025-02-25 17:24:00 +0000) ---------------------------------------------------------------- target-arm queue: * hw/arm/smmuv3: Fill u.f_cd_fetch.addr for SMMU_EVT_F_CD_FETCH * hw/arm/virt: Support larger highmem MMIO regions * machine: Centralize -machine dumpdtb option handling and report attempt to dump nonexistent DTB as an error * fpu: remove target ifdefs and build it only once * target/arm: Refactor to move TCG-only vfp_helper code into tcg/ * target/arm/hvf: Disable SME feature * target/arm/hvf: sign extend the data for a load operation when SSE=1 * hw/misc/npcm_clk: fix buffer-overflow * hw/arm: Add i.MX 8M Plus EVK board ("imx8mp-evk") ---------------------------------------------------------------- Bernhard Beschow (16): hw/usb/hcd-dwc3: Align global registers size with Linux hw/pci-host/designware: Prevent device attachment on internal PCIe root bus hw/gpio/pca955*: Move Kconfig switches next to implementations hw/arm: Add i.MX 8M Plus EVK board hw/arm/fsl-imx8mp: Implement clock tree hw/arm/fsl-imx8mp: Add SNVS hw/arm/fsl-imx8mp: Add USDHC storage controllers hw/arm/fsl-imx8mp: Add PCIe support hw/arm/fsl-imx8mp: Add GPIO controllers hw/arm/fsl-imx8mp: Add I2C controllers hw/arm/fsl-imx8mp: Add SPI controllers hw/arm/fsl-imx8mp: Add watchdog support hw/arm/fsl-imx8mp: Implement general purpose timers hw/arm/fsl-imx8mp: Add Ethernet controller hw/arm/fsl-imx8mp: Add USB support hw/arm/fsl-imx8mp: Add on-chip RAM Joelle van Dyne (2): target/arm/hvf: Disable SME feature target/arm/hvf: sign extend the data for a load operation when SSE=1 Matthew R. Ochs (1): hw/arm/virt: Support larger highmem MMIO regions Nicolin Chen (1): hw/arm/smmuv3: Fill u.f_cd_fetch.addr for SMMU_EVT_F_CD_FETCH Peter Maydell (22): monitor/hmp-cmds.c: Clean up hmp_dumpdtb printf hw/openrisc: Support monitor dumpdtb command hw/mips/boston: Check for error return from boston_fdt_filter() hw/mips/boston: Support dumpdtb monitor commands hw: Centralize handling of -machine dumpdtb option hw/core/machine.c: Make -machine dumpdtb=file.dtb with no DTB an error fpu: Make targets specify floatx80 default Inf at runtime target/m68k: Avoid using floatx80_infinity global const target/i386: Avoid using floatx80_infinity global const fpu: Pass float_status to floatx80_is_infinity() fpu: Make targets specify whether floatx80 Inf can have Int bit clear fpu: Pass float_status to floatx80_invalid_encoding() fpu: Make floatx80 invalid encoding settable at runtime fpu: Move m68k_denormal fmt flag into floatx80_behaviour fpu: Always decide no_signaling_nans() at runtime fpu: Always decide snan_bit_is_one() at runtime fpu: Don't compile-time disable hardfloat for PPC targets fpu: Build only once target/arm: Move TCG-only VFP code into tcg/ subdir target/arm: Move FPSCR get/set helpers to tcg/vfp_helper.c target/arm: Move softfloat specific FPCR/FPSR handling to tcg/ target/arm: Rename vfp_helper.c to vfp_fpscr.c Pierrick Bouvier (1): hw/misc/npcm_clk: fix buffer-overflow MAINTAINERS | 13 + docs/system/arm/imx8mp-evk.rst | 70 ++++ docs/system/arm/virt.rst | 4 + docs/system/target-arm.rst | 1 + include/fpu/softfloat-helpers.h | 12 + include/fpu/softfloat-types.h | 51 +++ include/fpu/softfloat.h | 91 ++--- include/hw/arm/fsl-imx8mp.h | 284 ++++++++++++++ include/hw/loader-fit.h | 21 +- include/hw/misc/imx8mp_analog.h | 81 ++++ include/hw/misc/imx8mp_ccm.h | 30 ++ include/hw/openrisc/boot.h | 3 +- include/hw/pci-host/designware.h | 7 + include/hw/pci-host/fsl_imx8m_phy.h | 28 ++ include/hw/timer/imx_gpt.h | 1 + include/hw/usb/hcd-dwc3.h | 2 +- include/system/device_tree.h | 2 - target/arm/internals.h | 9 + fpu/softfloat.c | 23 +- hw/arm/boot.c | 2 - hw/arm/fsl-imx8mp.c | 714 ++++++++++++++++++++++++++++++++++++ hw/arm/imx8mp-evk.c | 74 ++++ hw/arm/smmuv3.c | 2 +- hw/arm/virt.c | 52 ++- hw/core/loader-fit.c | 38 +- hw/core/machine.c | 23 ++ hw/loongarch/virt-fdt-build.c | 1 - hw/mips/boston.c | 16 +- hw/misc/imx8mp_analog.c | 160 ++++++++ hw/misc/imx8mp_ccm.c | 175 +++++++++ hw/misc/npcm_clk.c | 5 +- hw/openrisc/boot.c | 8 +- hw/openrisc/openrisc_sim.c | 2 +- hw/openrisc/virt.c | 2 +- hw/pci-host/designware.c | 18 +- hw/pci-host/fsl_imx8m_phy.c | 98 +++++ hw/ppc/e500.c | 1 - hw/ppc/pegasos2.c | 1 - hw/ppc/pnv.c | 1 - hw/ppc/spapr.c | 1 - hw/riscv/boot.c | 2 - hw/timer/imx_gpt.c | 25 ++ hw/usb/hcd-dwc3.c | 5 + monitor/hmp-cmds.c | 2 +- system/device_tree-stub.c | 5 +- system/device_tree.c | 22 +- target/arm/hvf/hvf.c | 16 + target/arm/tcg-stubs.c | 22 ++ target/arm/{ => tcg}/vfp_helper.c | 189 +--------- target/arm/vfp_fpscr.c | 155 ++++++++ target/hppa/fpu_helper.c | 1 + target/i386/tcg/fpu_helper.c | 51 +-- target/m68k/cpu.c | 35 ++ target/m68k/fpu_helper.c | 2 +- target/m68k/softfloat.c | 47 +-- target/sh4/cpu.c | 1 + fpu/softfloat-parts.c.inc | 27 +- fpu/softfloat-specialize.c.inc | 29 +- fpu/meson.build | 2 +- hw/arm/Kconfig | 24 ++ hw/arm/meson.build | 2 + hw/gpio/Kconfig | 8 + hw/misc/Kconfig | 14 +- hw/misc/meson.build | 2 + hw/pci-host/Kconfig | 3 + hw/pci-host/meson.build | 1 + target/arm/meson.build | 2 +- target/arm/tcg/meson.build | 1 + 68 files changed, 2439 insertions(+), 383 deletions(-) create mode 100644 docs/system/arm/imx8mp-evk.rst create mode 100644 include/hw/arm/fsl-imx8mp.h create mode 100644 include/hw/misc/imx8mp_analog.h create mode 100644 include/hw/misc/imx8mp_ccm.h create mode 100644 include/hw/pci-host/fsl_imx8m_phy.h create mode 100644 hw/arm/fsl-imx8mp.c create mode 100644 hw/arm/imx8mp-evk.c create mode 100644 hw/misc/imx8mp_analog.c create mode 100644 hw/misc/imx8mp_ccm.c create mode 100644 hw/pci-host/fsl_imx8m_phy.c rename target/arm/{ => tcg}/vfp_helper.c (90%) create mode 100644 target/arm/vfp_fpscr.c From: Nicolin Chen <nicolinc@nvidia.com> When we fill in the SMMUEventInfo for SMMU_EVT_F_CD_FETCH we write the address into the f_ste_fetch member of the union, but then when we come to read it back in smmuv3_record_event() we will (correctly) be using the f_cd_fetch member. This is more like a cosmetics fix since the f_cd_fetch and f_ste_fetch are basically the same field since they are in the exact same union with exact same type, but it's conceptually wrong. Use the correct union member. Signed-off-by: Nicolin Chen <nicolinc@nvidia.com> Message-id: 20250220213832.80289-1-nicolinc@nvidia.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/smmuv3.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/smmuv3.c +++ b/hw/arm/smmuv3.c @@ -XXX,XX +XXX,XX @@ static int smmu_get_cd(SMMUv3State *s, STE *ste, SMMUTransCfg *cfg, qemu_log_mask(LOG_GUEST_ERROR, "Cannot fetch pte at address=0x%"PRIx64"\n", addr); event->type = SMMU_EVT_F_CD_FETCH; - event->u.f_ste_fetch.addr = addr; + event->u.f_cd_fetch.addr = addr; return -EINVAL; } for (i = 0; i < ARRAY_SIZE(buf->word); i++) { -- 2.43.0 From: "Matthew R. Ochs" <mochs@nvidia.com> The MMIO region size required to support virtualized environments with large PCI BAR regions can exceed the hardcoded limit configured in QEMU. For example, a VM with multiple NVIDIA Grace-Hopper GPUs passed through requires more MMIO memory than the amount provided by VIRT_HIGH_PCIE_MMIO (currently 512GB). Instead of updating VIRT_HIGH_PCIE_MMIO, introduce a new parameter, highmem-mmio-size, that specifies the MMIO size required to support the VM configuration. Example usage with 1TB MMIO region size: -machine virt,gic-version=3,highmem-mmio-size=1T Signed-off-by: Matthew R. Ochs <mochs@nvidia.com> Reviewed-by: Gavin Shan <gshan@redhat.com> Reviewed-by: Shameer Kolothum <shameerali.kolothum.thodi@huawei.com> Reviewed-by: Eric Auger <eric.auger@redhat.com> Reviewed-by: Nicolin Chen <nicolinc@nvidia.com> Message-id: 20250221145419.1281890-1-mochs@nvidia.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/virt.rst | 4 ++++ hw/arm/virt.c | 52 +++++++++++++++++++++++++++++++++++++++- 2 files changed, 55 insertions(+), 1 deletion(-) diff --git a/docs/system/arm/virt.rst b/docs/system/arm/virt.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/virt.rst +++ b/docs/system/arm/virt.rst @@ -XXX,XX +XXX,XX @@ highmem-mmio Set ``on``/``off`` to enable/disable the high memory region for PCI MMIO. The default is ``on``. +highmem-mmio-size + Set the high memory region size for PCI MMIO. Must be a power of 2 and + greater than or equal to the default size (512G). + gic-version Specify the version of the Generic Interrupt Controller (GIC) to provide. Valid values are: diff --git a/hw/arm/virt.c b/hw/arm/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -XXX,XX +XXX,XX @@ #include "hw/loader.h" #include "qapi/error.h" #include "qemu/bitops.h" +#include "qemu/cutils.h" #include "qemu/error-report.h" #include "qemu/module.h" #include "hw/pci-host/gpex.h" @@ -XXX,XX +XXX,XX @@ static const MemMapEntry base_memmap[] = { [VIRT_MEM] = { GiB, LEGACY_RAMLIMIT_BYTES }, }; +/* Update the docs for highmem-mmio-size when changing this default */ +#define DEFAULT_HIGH_PCIE_MMIO_SIZE_GB 512 +#define DEFAULT_HIGH_PCIE_MMIO_SIZE (DEFAULT_HIGH_PCIE_MMIO_SIZE_GB * GiB) + /* * Highmem IO Regions: This memory map is floating, located after the RAM. * Each MemMapEntry base (GPA) will be dynamically computed, depending on the @@ -XXX,XX +XXX,XX @@ static const MemMapEntry base_memmap[] = { * PA space for one specific region is always reserved, even if the region * has been disabled or doesn't fit into the PA space. However, the PA space * for the region won't be reserved in these circumstances with compact layout. + * + * Note that the highmem-mmio-size property will update the high PCIE MMIO size + * field in this array. */ static MemMapEntry extended_memmap[] = { /* Additional 64 MB redist region (can contain up to 512 redistributors) */ [VIRT_HIGH_GIC_REDIST2] = { 0x0, 64 * MiB }, [VIRT_HIGH_PCIE_ECAM] = { 0x0, 256 * MiB }, /* Second PCIe window */ - [VIRT_HIGH_PCIE_MMIO] = { 0x0, 512 * GiB }, + [VIRT_HIGH_PCIE_MMIO] = { 0x0, DEFAULT_HIGH_PCIE_MMIO_SIZE }, }; static const int a15irqmap[] = { @@ -XXX,XX +XXX,XX @@ static void virt_set_highmem_mmio(Object *obj, bool value, Error **errp) vms->highmem_mmio = value; } +static void virt_get_highmem_mmio_size(Object *obj, Visitor *v, + const char *name, void *opaque, + Error **errp) +{ + uint64_t size = extended_memmap[VIRT_HIGH_PCIE_MMIO].size; + + visit_type_size(v, name, &size, errp); +} + +static void virt_set_highmem_mmio_size(Object *obj, Visitor *v, + const char *name, void *opaque, + Error **errp) +{ + uint64_t size; + + if (!visit_type_size(v, name, &size, errp)) { + return; + } + + if (!is_power_of_2(size)) { + error_setg(errp, "highmem-mmio-size is not a power of 2"); + return; + } + + if (size < DEFAULT_HIGH_PCIE_MMIO_SIZE) { + char *sz = size_to_str(DEFAULT_HIGH_PCIE_MMIO_SIZE); + error_setg(errp, "highmem-mmio-size cannot be set to a lower value " + "than the default (%s)", sz); + g_free(sz); + return; + } + + extended_memmap[VIRT_HIGH_PCIE_MMIO].size = size; +} static bool virt_get_its(Object *obj, Error **errp) { @@ -XXX,XX +XXX,XX @@ static void virt_machine_class_init(ObjectClass *oc, void *data) "Set on/off to enable/disable high " "memory region for PCI MMIO"); + object_class_property_add(oc, "highmem-mmio-size", "size", + virt_get_highmem_mmio_size, + virt_set_highmem_mmio_size, + NULL, NULL); + object_class_property_set_description(oc, "highmem-mmio-size", + "Set the high memory region size " + "for PCI MMIO"); + object_class_property_add_str(oc, "gic-version", virt_get_gic_version, virt_set_gic_version); object_class_property_set_description(oc, "gic-version", -- 2.43.0 In hmp_dumpdtb(), we print a message when the command succeeds. This message is missing the trailing \n, so the HMP command prompt is printed immediately after it. We also weren't capitalizing 'DTB', or quoting the filename in the message. Fix these nits. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250206151214.2947842-2-peter.maydell@linaro.org --- monitor/hmp-cmds.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/monitor/hmp-cmds.c +++ b/monitor/hmp-cmds.c @@ -XXX,XX +XXX,XX @@ void hmp_dumpdtb(Monitor *mon, const QDict *qdict) return; } - monitor_printf(mon, "dtb dumped to %s", filename); + monitor_printf(mon, "DTB dumped to '%s'\n", filename); } #endif -- 2.43.0 The openrisc machines don't set MachineState::fdt to point to their DTB blob. This means that although the command line '-machine dumpdtb=file.dtb' option works, the equivalent QMP and HMP monitor commands do not, but instead produce the error "This machine doesn't have a FDT". Set MachineState::fdt in openrisc_load_fdt(), when we write it to guest memory. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250206151214.2947842-3-peter.maydell@linaro.org --- include/hw/openrisc/boot.h | 3 ++- hw/openrisc/boot.c | 7 +++++-- hw/openrisc/openrisc_sim.c | 2 +- hw/openrisc/virt.c | 2 +- 4 files changed, 9 insertions(+), 5 deletions(-) diff --git a/include/hw/openrisc/boot.h b/include/hw/openrisc/boot.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/openrisc/boot.h +++ b/include/hw/openrisc/boot.h @@ -XXX,XX +XXX,XX @@ #define OPENRISC_BOOT_H #include "exec/cpu-defs.h" +#include "hw/boards.h" hwaddr openrisc_load_kernel(ram_addr_t ram_size, const char *kernel_filename, @@ -XXX,XX +XXX,XX @@ hwaddr openrisc_load_kernel(ram_addr_t ram_size, hwaddr openrisc_load_initrd(void *fdt, const char *filename, hwaddr load_start, uint64_t mem_size); -uint32_t openrisc_load_fdt(void *fdt, hwaddr load_start, +uint32_t openrisc_load_fdt(MachineState *ms, void *fdt, hwaddr load_start, uint64_t mem_size); #endif /* OPENRISC_BOOT_H */ diff --git a/hw/openrisc/boot.c b/hw/openrisc/boot.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/boot.c +++ b/hw/openrisc/boot.c @@ -XXX,XX +XXX,XX @@ hwaddr openrisc_load_initrd(void *fdt, const char *filename, return start + size; } -uint32_t openrisc_load_fdt(void *fdt, hwaddr load_start, - uint64_t mem_size) +uint32_t openrisc_load_fdt(MachineState *ms, void *fdt, + hwaddr load_start, uint64_t mem_size) { uint32_t fdt_addr; int ret; @@ -XXX,XX +XXX,XX @@ uint32_t openrisc_load_fdt(void *fdt, hwaddr load_start, /* copy in the device tree */ qemu_fdt_dumpdtb(fdt, fdtsize); + /* Save FDT for dumpdtb monitor command */ + ms->fdt = fdt; + rom_add_blob_fixed_as("fdt", fdt, fdtsize, fdt_addr, &address_space_memory); qemu_register_reset_nosnapshotload(qemu_fdt_randomize_seeds, diff --git a/hw/openrisc/openrisc_sim.c b/hw/openrisc/openrisc_sim.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/openrisc_sim.c +++ b/hw/openrisc/openrisc_sim.c @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_init(MachineState *machine) machine->initrd_filename, load_addr, machine->ram_size); } - boot_info.fdt_addr = openrisc_load_fdt(state->fdt, load_addr, + boot_info.fdt_addr = openrisc_load_fdt(machine, state->fdt, load_addr, machine->ram_size); } } diff --git a/hw/openrisc/virt.c b/hw/openrisc/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/virt.c +++ b/hw/openrisc/virt.c @@ -XXX,XX +XXX,XX @@ static void openrisc_virt_init(MachineState *machine) machine->initrd_filename, load_addr, machine->ram_size); } - boot_info.fdt_addr = openrisc_load_fdt(state->fdt, load_addr, + boot_info.fdt_addr = openrisc_load_fdt(machine, state->fdt, load_addr, machine->ram_size); } } -- 2.43.0 The function boston_fdt_filter() can return NULL on errors (in which case it will print an error message). When we call this from the non-FIT-image codepath, we aren't checking the return value, so we will plough on with a NULL pointer, and segfault in fdt_totalsize(). Check for errors here. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250206151214.2947842-4-peter.maydell@linaro.org --- hw/mips/boston.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/hw/mips/boston.c b/hw/mips/boston.c index XXXXXXX..XXXXXXX 100644 --- a/hw/mips/boston.c +++ b/hw/mips/boston.c @@ -XXX,XX +XXX,XX @@ static void boston_mach_init(MachineState *machine) dtb_load_data = boston_fdt_filter(s, dtb_file_data, NULL, &dtb_vaddr); + if (!dtb_load_data) { + /* boston_fdt_filter() already printed the error for us */ + exit(1); + } /* Calculate real fdt size after filter */ dt_size = fdt_totalsize(dtb_load_data); -- 2.43.0 The boston machine doesn't set MachineState::fdt to the DTB blob that it has loaded or created, which means that the QMP/HMP dumpdtb monitor commands don't work. Setting MachineState::fdt is easy in the non-FIT codepath: we can simply do so immediately before loading the DTB into guest memory. The FIT codepath is a bit more awkward as currently the FIT loader throws away the memory that the FDT was in after it loads it into guest memory. So we add a void *pfdt argument to load_fit() for it to store the FDT pointer into. There is some readjustment required of the pointer handling in loader-fit.c, so that it applies 'const' only where it should (e.g. the data pointer we get back from fdt_getprop() is const, because it's into the middle of the input FDT data, but the pointer that fit_load_image_alloc() should not be const, because it's freshly allocated memory that the caller can change if it likes). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250206151214.2947842-5-peter.maydell@linaro.org --- include/hw/loader-fit.h | 21 ++++++++++++++++++--- hw/core/loader-fit.c | 38 +++++++++++++++++++++----------------- hw/mips/boston.c | 11 +++++++---- 3 files changed, 46 insertions(+), 24 deletions(-) diff --git a/include/hw/loader-fit.h b/include/hw/loader-fit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/loader-fit.h +++ b/include/hw/loader-fit.h @@ -XXX,XX +XXX,XX @@ struct fit_loader_match { struct fit_loader { const struct fit_loader_match *matches; hwaddr (*addr_to_phys)(void *opaque, uint64_t addr); - const void *(*fdt_filter)(void *opaque, const void *fdt, - const void *match_data, hwaddr *load_addr); + void *(*fdt_filter)(void *opaque, const void *fdt, + const void *match_data, hwaddr *load_addr); const void *(*kernel_filter)(void *opaque, const void *kernel, hwaddr *load_addr, hwaddr *entry_addr); }; -int load_fit(const struct fit_loader *ldr, const char *filename, void *opaque); +/** + * load_fit: load a FIT format image + * @ldr: structure defining board specific properties and hooks + * @filename: image to load + * @pfdt: pointer to update with address of FDT blob + * @opaque: opaque value passed back to the hook functions in @ldr + * Returns: 0 on success, or a negative errno on failure + * + * @pfdt is used to tell the caller about the FDT blob. On return, it + * has been set to point to the FDT blob, and it is now the caller's + * responsibility to free that memory with g_free(). Usually the caller + * will want to pass in &machine->fdt here, to record the FDT blob for + * the dumpdtb option and QMP/HMP commands. + */ +int load_fit(const struct fit_loader *ldr, const char *filename, void **pfdt, + void *opaque); #endif /* HW_LOADER_FIT_H */ diff --git a/hw/core/loader-fit.c b/hw/core/loader-fit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/loader-fit.c +++ b/hw/core/loader-fit.c @@ -XXX,XX +XXX,XX @@ #define FIT_LOADER_MAX_PATH (128) -static const void *fit_load_image_alloc(const void *itb, const char *name, - int *poff, size_t *psz, Error **errp) +static void *fit_load_image_alloc(const void *itb, const char *name, + int *poff, size_t *psz, Error **errp) { const void *data; const char *comp; @@ -XXX,XX +XXX,XX @@ static const void *fit_load_image_alloc(const void *itb, const char *name, return NULL; } - data = g_realloc(uncomp_data, uncomp_len); + uncomp_data = g_realloc(uncomp_data, uncomp_len); if (psz) { *psz = uncomp_len; } - return data; + return uncomp_data; } error_setg(errp, "unknown compression '%s'", comp); @@ -XXX,XX +XXX,XX @@ out: static int fit_load_fdt(const struct fit_loader *ldr, const void *itb, int cfg, void *opaque, const void *match_data, - hwaddr kernel_end, Error **errp) + hwaddr kernel_end, void **pfdt, Error **errp) { ERRP_GUARD(); Error *err = NULL; const char *name; - const void *data; - const void *load_data; + void *data; hwaddr load_addr; int img_off; size_t sz; @@ -XXX,XX +XXX,XX @@ static int fit_load_fdt(const struct fit_loader *ldr, const void *itb, return 0; } - load_data = data = fit_load_image_alloc(itb, name, &img_off, &sz, errp); + data = fit_load_image_alloc(itb, name, &img_off, &sz, errp); if (!data) { error_prepend(errp, "unable to load FDT image from FIT: "); return -EINVAL; @@ -XXX,XX +XXX,XX @@ static int fit_load_fdt(const struct fit_loader *ldr, const void *itb, } if (ldr->fdt_filter) { - load_data = ldr->fdt_filter(opaque, data, match_data, &load_addr); + void *filtered_data; + + filtered_data = ldr->fdt_filter(opaque, data, match_data, &load_addr); + if (filtered_data != data) { + g_free(data); + data = filtered_data; + } } load_addr = ldr->addr_to_phys(opaque, load_addr); - sz = fdt_totalsize(load_data); - rom_add_blob_fixed(name, load_data, sz, load_addr); + sz = fdt_totalsize(data); + rom_add_blob_fixed(name, data, sz, load_addr); - ret = 0; + *pfdt = data; + return 0; out: g_free((void *) data); - if (data != load_data) { - g_free((void *) load_data); - } return ret; } @@ -XXX,XX +XXX,XX @@ out: return ret; } -int load_fit(const struct fit_loader *ldr, const char *filename, void *opaque) +int load_fit(const struct fit_loader *ldr, const char *filename, + void **pfdt, void *opaque) { Error *err = NULL; const struct fit_loader_match *match; @@ -XXX,XX +XXX,XX @@ int load_fit(const struct fit_loader *ldr, const char *filename, void *opaque) goto out; } - ret = fit_load_fdt(ldr, itb, cfg_off, opaque, match_data, kernel_end, + ret = fit_load_fdt(ldr, itb, cfg_off, opaque, match_data, kernel_end, pfdt, &err); if (ret) { error_report_err(err); diff --git a/hw/mips/boston.c b/hw/mips/boston.c index XXXXXXX..XXXXXXX 100644 --- a/hw/mips/boston.c +++ b/hw/mips/boston.c @@ -XXX,XX +XXX,XX @@ static void gen_firmware(void *p, hwaddr kernel_entry, hwaddr fdt_addr) kernel_entry); } -static const void *boston_fdt_filter(void *opaque, const void *fdt_orig, - const void *match_data, hwaddr *load_addr) +static void *boston_fdt_filter(void *opaque, const void *fdt_orig, + const void *match_data, hwaddr *load_addr) { BostonState *s = BOSTON(opaque); MachineState *machine = s->mach; @@ -XXX,XX +XXX,XX @@ static void boston_mach_init(MachineState *machine) if (kernel_size > 0) { int dt_size; g_autofree const void *dtb_file_data = NULL; - g_autofree const void *dtb_load_data = NULL; + void *dtb_load_data = NULL; hwaddr dtb_paddr = QEMU_ALIGN_UP(kernel_high, 64 * KiB); hwaddr dtb_vaddr = cpu_mips_phys_to_kseg0(NULL, dtb_paddr); @@ -XXX,XX +XXX,XX @@ static void boston_mach_init(MachineState *machine) exit(1); } + machine->fdt = dtb_load_data; + /* Calculate real fdt size after filter */ dt_size = fdt_totalsize(dtb_load_data); rom_add_blob_fixed("dtb", dtb_load_data, dt_size, dtb_paddr); @@ -XXX,XX +XXX,XX @@ static void boston_mach_init(MachineState *machine) rom_ptr(dtb_paddr, dt_size)); } else { /* Try to load file as FIT */ - fit_err = load_fit(&boston_fit_loader, machine->kernel_filename, s); + fit_err = load_fit(&boston_fit_loader, machine->kernel_filename, + &machine->fdt, s); if (fit_err) { error_report("unable to load kernel image"); exit(1); -- 2.43.0 Currently we handle the 'dumpdtb' machine sub-option ad-hoc in every board model that has an FDT. It's up to the board code to make sure it calls qemu_fdt_dumpdtb() in the right place. This means we're inconsistent and often just ignore the user's command line argument: * if the board doesn't have an FDT at all * if the board supports FDT, but there happens not to be one present (usually because of a missing -fdt option) This isn't very helpful because it gives the user no clue why their option was ignored. However, in order to support the QMP/HMP dumpdtb commands we require now that every FDT machine stores a pointer to the FDT in MachineState::fdt. This means we can handle -machine dumpdtb centrally by calling the qmp_dumpdtb() function, unifying its handling with the QMP/HMP commands. All the board code calls to qemu_fdt_dumpdtb() can then be removed. For this commit we retain the existing behaviour that if there is no FDT we silently ignore the -machine dumpdtb option. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- include/system/device_tree.h | 2 -- hw/arm/boot.c | 2 -- hw/core/machine.c | 25 +++++++++++++++++++++++++ hw/loongarch/virt-fdt-build.c | 1 - hw/mips/boston.c | 1 - hw/openrisc/boot.c | 1 - hw/ppc/e500.c | 1 - hw/ppc/pegasos2.c | 1 - hw/ppc/pnv.c | 1 - hw/ppc/spapr.c | 1 - hw/riscv/boot.c | 2 -- system/device_tree.c | 15 --------------- 12 files changed, 25 insertions(+), 28 deletions(-) diff --git a/include/system/device_tree.h b/include/system/device_tree.h index XXXXXXX..XXXXXXX 100644 --- a/include/system/device_tree.h +++ b/include/system/device_tree.h @@ -XXX,XX +XXX,XX @@ int qemu_fdt_add_path(void *fdt, const char *path); sizeof(qdt_tmp)); \ } while (0) -void qemu_fdt_dumpdtb(void *fdt, int size); - /** * qemu_fdt_setprop_sized_cells_from_array: * @fdt: device tree blob diff --git a/hw/arm/boot.c b/hw/arm/boot.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/boot.c +++ b/hw/arm/boot.c @@ -XXX,XX +XXX,XX @@ int arm_load_dtb(hwaddr addr, const struct arm_boot_info *binfo, binfo->modify_dtb(binfo, fdt); } - qemu_fdt_dumpdtb(fdt, size); - /* Put the DTB into the memory map as a ROM image: this will ensure * the DTB is copied again upon reset, even if addr points into RAM. */ diff --git a/hw/core/machine.c b/hw/core/machine.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/machine.c +++ b/hw/core/machine.c @@ -XXX,XX +XXX,XX @@ #include "qemu/error-report.h" #include "qapi/error.h" #include "qapi/qapi-visit-machine.h" +#include "qapi/qapi-commands-machine.h" #include "qemu/madvise.h" #include "qom/object_interfaces.h" #include "system/cpus.h" @@ -XXX,XX +XXX,XX @@ void qemu_remove_machine_init_done_notifier(Notifier *notify) notifier_remove(notify); } +static void handle_machine_dumpdtb(MachineState *ms) +{ + if (!ms->dumpdtb) { + return; + } + if (!ms->fdt) { + /* Silently ignore dumpdtb option if there is nothing to dump */ + return; + } +#ifdef CONFIG_FDT + qmp_dumpdtb(ms->dumpdtb, &error_fatal); + exit(0); +#else + error_report("This machine doesn't have an FDT"); + exit(1); +#endif +} + void qdev_machine_creation_done(void) { cpu_synchronize_all_post_init(); @@ -XXX,XX +XXX,XX @@ void qdev_machine_creation_done(void) phase_advance(PHASE_MACHINE_READY); qdev_assert_realized_properly(); + /* + * If the user used -machine dumpdtb=file.dtb to request that we + * dump the DTB to a file, do it now, and exit. + */ + handle_machine_dumpdtb(current_machine); + /* TODO: once all bus devices are qdevified, this should be done * when bus is created by qdev.c */ /* diff --git a/hw/loongarch/virt-fdt-build.c b/hw/loongarch/virt-fdt-build.c index XXXXXXX..XXXXXXX 100644 --- a/hw/loongarch/virt-fdt-build.c +++ b/hw/loongarch/virt-fdt-build.c @@ -XXX,XX +XXX,XX @@ void virt_fdt_setup(LoongArchVirtMachineState *lvms) * Put the FDT into the memory map as a ROM image: this will ensure * the FDT is copied again upon reset, even if addr points into RAM. */ - qemu_fdt_dumpdtb(machine->fdt, lvms->fdt_size); rom_add_blob_fixed_as("fdt", machine->fdt, lvms->fdt_size, FDT_BASE, &address_space_memory); qemu_register_reset_nosnapshotload(qemu_fdt_randomize_seeds, diff --git a/hw/mips/boston.c b/hw/mips/boston.c index XXXXXXX..XXXXXXX 100644 --- a/hw/mips/boston.c +++ b/hw/mips/boston.c @@ -XXX,XX +XXX,XX @@ static void *boston_fdt_filter(void *opaque, const void *fdt_orig, 1, ram_high_sz); fdt = g_realloc(fdt, fdt_totalsize(fdt)); - qemu_fdt_dumpdtb(fdt, fdt_sz); s->fdt_base = *load_addr; diff --git a/hw/openrisc/boot.c b/hw/openrisc/boot.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/boot.c +++ b/hw/openrisc/boot.c @@ -XXX,XX +XXX,XX @@ uint32_t openrisc_load_fdt(MachineState *ms, void *fdt, /* Should only fail if we've built a corrupted tree */ g_assert(ret == 0); /* copy in the device tree */ - qemu_fdt_dumpdtb(fdt, fdtsize); /* Save FDT for dumpdtb monitor command */ ms->fdt = fdt; diff --git a/hw/ppc/e500.c b/hw/ppc/e500.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ppc/e500.c +++ b/hw/ppc/e500.c @@ -XXX,XX +XXX,XX @@ static int ppce500_load_device_tree(PPCE500MachineState *pms, done: if (!dry_run) { - qemu_fdt_dumpdtb(fdt, fdt_size); cpu_physical_memory_write(addr, fdt, fdt_size); /* Set machine->fdt for 'dumpdtb' QMP/HMP command */ diff --git a/hw/ppc/pegasos2.c b/hw/ppc/pegasos2.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ppc/pegasos2.c +++ b/hw/ppc/pegasos2.c @@ -XXX,XX +XXX,XX @@ static void pegasos2_machine_reset(MachineState *machine, ResetType type) d[1] = cpu_to_be64(pm->kernel_size - (pm->kernel_entry - pm->kernel_addr)); qemu_fdt_setprop(fdt, "/chosen", "qemu,boot-kernel", d, sizeof(d)); - qemu_fdt_dumpdtb(fdt, fdt_totalsize(fdt)); g_free(pm->fdt_blob); pm->fdt_blob = fdt; diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ppc/pnv.c +++ b/hw/ppc/pnv.c @@ -XXX,XX +XXX,XX @@ static void pnv_reset(MachineState *machine, ResetType type) _FDT((fdt_pack(fdt))); } - qemu_fdt_dumpdtb(fdt, fdt_totalsize(fdt)); cpu_physical_memory_write(PNV_FDT_ADDR, fdt, fdt_totalsize(fdt)); /* Update machine->fdt with latest fdt */ diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ppc/spapr.c +++ b/hw/ppc/spapr.c @@ -XXX,XX +XXX,XX @@ static void spapr_machine_reset(MachineState *machine, ResetType type) 0, fdt_addr, 0); cpu_physical_memory_write(fdt_addr, fdt, fdt_totalsize(fdt)); } - qemu_fdt_dumpdtb(fdt, fdt_totalsize(fdt)); g_free(spapr->fdt_blob); spapr->fdt_size = fdt_totalsize(fdt); diff --git a/hw/riscv/boot.c b/hw/riscv/boot.c index XXXXXXX..XXXXXXX 100644 --- a/hw/riscv/boot.c +++ b/hw/riscv/boot.c @@ -XXX,XX +XXX,XX @@ void riscv_load_fdt(hwaddr fdt_addr, void *fdt) uint32_t fdtsize = fdt_totalsize(fdt); /* copy in the device tree */ - qemu_fdt_dumpdtb(fdt, fdtsize); - rom_add_blob_fixed_as("fdt", fdt, fdtsize, fdt_addr, &address_space_memory); qemu_register_reset_nosnapshotload(qemu_fdt_randomize_seeds, diff --git a/system/device_tree.c b/system/device_tree.c index XXXXXXX..XXXXXXX 100644 --- a/system/device_tree.c +++ b/system/device_tree.c @@ -XXX,XX +XXX,XX @@ int qemu_fdt_add_path(void *fdt, const char *path) return retval; } -void qemu_fdt_dumpdtb(void *fdt, int size) -{ - const char *dumpdtb = current_machine->dumpdtb; - - if (dumpdtb) { - /* Dump the dtb to a file and quit */ - if (g_file_set_contents(dumpdtb, fdt, size, NULL)) { - info_report("dtb dumped to %s. Exiting.", dumpdtb); - exit(0); - } - error_report("%s: Failed dumping dtb to %s", __func__, dumpdtb); - exit(1); - } -} - int qemu_fdt_setprop_sized_cells_from_array(void *fdt, const char *node_path, const char *property, -- 2.43.0 Currently if the user requests via -machine dumpdtb=file.dtb that we dump the DTB, but the machine doesn't have a DTB, we silently ignore the option. This is confusing to users, and is a legacy of the old board-specific implementation of the option, where if the execution codepath didn't go via a call to qemu_fdt_dumpdtb() we would never handle the option. Now we handle the option in one place in machine.c, we can provide the user with a useful message if they asked us to dump a DTB when none exists. qmp_dumpdtb() already produces this error; remove the logic in handle_machine_dumpdtb() that was there specifically to avoid hitting it. While we're here, beef up the error message a bit with a hint, and make it consistent about "an FDT" rather than "a FDT". (In the qmp_dumpdtb() case this needs an ERRP_GUARD to make error_append_hint() work when the caller passes error_fatal.) Note that the three places where we might report "doesn't have an FDT" are hit in different situations: (1) in handle_machine_dumpdtb(), if CONFIG_FDT is not set: this is because the QEMU binary was built without libfdt at all. The build system will not let you build with a machine type that needs an FDT but no libfdt, so here we know both that the machine doesn't use FDT and that QEMU doesn't have the support: (2) in the device_tree-stub.c qmp_dumpdtb(): this is used when we had libfdt at build time but the target architecture didn't enable any machines which did "select DEVICE_TREE", so here we know that the machine doesn't use FDT. (3) in qmp_dumpdtb(), if current_machine->fdt is NULL all we know is that this machine never set it. That might be because it doesn't use FDT, or it might be because the user didn't pass an FDT on the command line and the machine doesn't autogenerate an FDT. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2733 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250206151214.2947842-7-peter.maydell@linaro.org --- hw/core/machine.c | 6 ++---- system/device_tree-stub.c | 5 ++++- system/device_tree.c | 7 ++++++- 3 files changed, 12 insertions(+), 6 deletions(-) diff --git a/hw/core/machine.c b/hw/core/machine.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/machine.c +++ b/hw/core/machine.c @@ -XXX,XX +XXX,XX @@ static void handle_machine_dumpdtb(MachineState *ms) if (!ms->dumpdtb) { return; } - if (!ms->fdt) { - /* Silently ignore dumpdtb option if there is nothing to dump */ - return; - } #ifdef CONFIG_FDT qmp_dumpdtb(ms->dumpdtb, &error_fatal); exit(0); #else error_report("This machine doesn't have an FDT"); + error_printf("(this machine type definitely doesn't use FDT, and " + "this QEMU doesn't have FDT support compiled in)\n"); exit(1); #endif } diff --git a/system/device_tree-stub.c b/system/device_tree-stub.c index XXXXXXX..XXXXXXX 100644 --- a/system/device_tree-stub.c +++ b/system/device_tree-stub.c @@ -XXX,XX +XXX,XX @@ #ifdef CONFIG_FDT void qmp_dumpdtb(const char *filename, Error **errp) { - error_setg(errp, "This machine doesn't have a FDT"); + ERRP_GUARD(); + + error_setg(errp, "This machine doesn't have an FDT"); + error_append_hint(errp, "(this machine type definitely doesn't use FDT)\n"); } #endif diff --git a/system/device_tree.c b/system/device_tree.c index XXXXXXX..XXXXXXX 100644 --- a/system/device_tree.c +++ b/system/device_tree.c @@ -XXX,XX +XXX,XX @@ out: void qmp_dumpdtb(const char *filename, Error **errp) { + ERRP_GUARD(); + g_autoptr(GError) err = NULL; uint32_t size; if (!current_machine->fdt) { - error_setg(errp, "This machine doesn't have a FDT"); + error_setg(errp, "This machine doesn't have an FDT"); + error_append_hint(errp, + "(Perhaps it doesn't support FDT at all, or perhaps " + "you need to provide an FDT with the -fdt option?)\n"); return; } -- 2.43.0 Currently we hardcode at compile time whether the floatx80 default Infinity value has the explicit integer bit set or not (x86 sets it; m68k does not). To be able to compile softfloat once for all targets we'd like to move this setting to runtime. Define a new FloatX80Behaviour enum which is a set of flags that define the target's floatx80 handling. Initially we define just one flag, for whether the default Infinity has the Integer bit set or not, but we will expand this in future commits to cover the other floatx80 target specifics that we currently make compile-time settings. Define a new function floatx80_default_inf() which returns the appropriate default Infinity value of the given sign, and use it in the code that was previously directly using the compile-time constant floatx80_infinity_{low,high} values when packing an infinity into a floatx80. Since floatx80 is highly unlikely to be supported in any new architecture, and the existing code is generally written as "default to like x87, with an ifdef for m68k", we make the default value for the floatx80 behaviour flags be "what x87 does". This means we only need to change the m68k target to specify the behaviour flags. (Other users of floatx80 are the Arm NWFPE emulation, which is obsolete and probably not actually doing the right thing anyway, and the PPC xsrqpxp insn. Making the default be "like x87" avoids our needing to review and test for behaviour changes there.) We will clean up the remaining uses of the floatx80_infinity global constant in subsequent commits. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-2-peter.maydell@linaro.org Message-id: 20250217125055.160887-2-peter.maydell@linaro.org --- include/fpu/softfloat-helpers.h | 12 ++++++++++++ include/fpu/softfloat-types.h | 13 +++++++++++++ include/fpu/softfloat.h | 1 + fpu/softfloat.c | 7 +++---- target/m68k/cpu.c | 6 ++++++ fpu/softfloat-specialize.c.inc | 10 ++++++++++ 6 files changed, 45 insertions(+), 4 deletions(-) diff --git a/include/fpu/softfloat-helpers.h b/include/fpu/softfloat-helpers.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-helpers.h +++ b/include/fpu/softfloat-helpers.h @@ -XXX,XX +XXX,XX @@ static inline void set_floatx80_rounding_precision(FloatX80RoundPrec val, status->floatx80_rounding_precision = val; } +static inline void set_floatx80_behaviour(FloatX80Behaviour b, + float_status *status) +{ + status->floatx80_behaviour = b; +} + static inline void set_float_2nan_prop_rule(Float2NaNPropRule rule, float_status *status) { @@ -XXX,XX +XXX,XX @@ get_floatx80_rounding_precision(const float_status *status) return status->floatx80_rounding_precision; } +static inline FloatX80Behaviour +get_floatx80_behaviour(const float_status *status) +{ + return status->floatx80_behaviour; +} + static inline Float2NaNPropRule get_float_2nan_prop_rule(const float_status *status) { diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { float_ftz_before_rounding = 1, } FloatFTZDetection; +/* + * floatx80 is primarily used by x86 and m68k, and there are + * differences in the handling, largely related to the explicit + * Integer bit which floatx80 has and the other float formats do not. + * These flag values allow specification of the target's requirements + * and can be ORed together to set floatx80_behaviour. + */ +typedef enum __attribute__((__packed__)) { + /* In the default Infinity value, is the Integer bit 0 ? */ + floatx80_default_inf_int_bit_is_zero = 1, +} FloatX80Behaviour; + /* * Floating Point Status. Individual architectures may maintain * several versions of float_status for different functions. The @@ -XXX,XX +XXX,XX @@ typedef struct float_status { uint16_t float_exception_flags; FloatRoundMode float_rounding_mode; FloatX80RoundPrec floatx80_rounding_precision; + FloatX80Behaviour floatx80_behaviour; Float2NaNPropRule float_2nan_prop_rule; Float3NaNPropRule float_3nan_prop_rule; FloatInfZeroNaNRule float_infzeronan_rule; diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ float128 floatx80_to_float128(floatx80, float_status *status); | The pattern for an extended double-precision inf. *----------------------------------------------------------------------------*/ extern const floatx80 floatx80_infinity; +floatx80 floatx80_default_inf(bool zSign, float_status *status); /*---------------------------------------------------------------------------- | Software IEC/IEEE extended double-precision operations. diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ static floatx80 floatx80_round_pack_canonical(FloatParts128 *p, case float_class_inf: /* x86 and m68k differ in the setting of the integer bit. */ - frac = floatx80_infinity_low; + frac = s->floatx80_behaviour & floatx80_default_inf_int_bit_is_zero ? + 0 : (1ULL << 63); exp = fmt->exp_max; break; @@ -XXX,XX +XXX,XX @@ floatx80 roundAndPackFloatx80(FloatX80RoundPrec roundingPrecision, bool zSign, ) { return packFloatx80( zSign, 0x7FFE, ~ roundMask ); } - return packFloatx80(zSign, - floatx80_infinity_high, - floatx80_infinity_low); + return floatx80_default_inf(zSign, status); } if ( zExp <= 0 ) { isTiny = status->tininess_before_rounding diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static void m68k_cpu_reset_hold(Object *obj, ResetType type) set_float_2nan_prop_rule(float_2nan_prop_ab, &env->fp_status); /* Default NaN: sign bit clear, all frac bits set */ set_float_default_nan_pattern(0b01111111, &env->fp_status); + /* + * m68k-specific floatx80 behaviour: + * * default Infinity values have a zero Integer bit + */ + set_floatx80_behaviour(floatx80_default_inf_int_bit_is_zero, + &env->fp_status); nan = floatx80_default_nan(&env->fp_status); for (i = 0; i < 8; i++) { diff --git a/fpu/softfloat-specialize.c.inc b/fpu/softfloat-specialize.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-specialize.c.inc +++ b/fpu/softfloat-specialize.c.inc @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_default_nan(float_status *status) | The pattern for a default generated extended double-precision inf. *----------------------------------------------------------------------------*/ +floatx80 floatx80_default_inf(bool zSign, float_status *status) +{ + /* + * Whether the Integer bit is set in the default Infinity is + * target dependent. + */ + bool z = status->floatx80_behaviour & floatx80_default_inf_int_bit_is_zero; + return packFloatx80(zSign, 0x7fff, z ? 0 : (1ULL << 63)); +} + #define floatx80_infinity_high 0x7FFF #if defined(TARGET_M68K) #define floatx80_infinity_low UINT64_C(0x0000000000000000) -- 2.43.0 The global const floatx80_infinity is (unlike all the other float*_infinity values) target-specific, because whether the explicit Integer bit is set or not varies between m68k and i386. We want to be able to compile softfloat once for multiple targets, so we can't continue to use a single global whose value needs to be different between targets. Replace the direct uses of floatx80_infinity in target/m68k with calls to the new floatx80_default_inf() function. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-3-peter.maydell@linaro.org Message-id: 20250217125055.160887-3-peter.maydell@linaro.org --- target/m68k/softfloat.c | 47 ++++++++++++++--------------------------- 1 file changed, 16 insertions(+), 31 deletions(-) diff --git a/target/m68k/softfloat.c b/target/m68k/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/softfloat.c +++ b/target/m68k/softfloat.c @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_scale(floatx80 a, floatx80 b, float_status *status) if ((uint64_t) (aSig << 1)) { return propagateFloatx80NaN(a, b, status); } - return packFloatx80(aSign, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(aSign, status); } if (aExp == 0) { if (aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_lognp1(floatx80 a, float_status *status) float_raise(float_flag_invalid, status); return floatx80_default_nan(status); } - return packFloatx80(0, floatx80_infinity.high, floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_lognp1(floatx80 a, float_status *status) if (aSign && aExp >= one_exp) { if (aExp == one_exp && aSig == one_sig) { float_raise(float_flag_divbyzero, status); - return packFloatx80(aSign, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(aSign, status); } float_raise(float_flag_invalid, status); return floatx80_default_nan(status); @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_logn(floatx80 a, float_status *status) propagateFloatx80NaNOneArg(a, status); } if (aSign == 0) { - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } } @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_logn(floatx80 a, float_status *status) if (aExp == 0) { if (aSig == 0) { /* zero */ float_raise(float_flag_divbyzero, status); - return packFloatx80(1, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(1, status); } if ((aSig & one_sig) == 0) { /* denormal */ normalizeFloatx80Subnormal(aSig, &aExp, &aSig); @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_log10(floatx80 a, float_status *status) propagateFloatx80NaNOneArg(a, status); } if (aSign == 0) { - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } } if (aExp == 0 && aSig == 0) { float_raise(float_flag_divbyzero, status); - return packFloatx80(1, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(1, status); } if (aSign) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_log2(floatx80 a, float_status *status) propagateFloatx80NaNOneArg(a, status); } if (aSign == 0) { - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } } if (aExp == 0) { if (aSig == 0) { float_raise(float_flag_divbyzero, status); - return packFloatx80(1, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(1, status); } normalizeFloatx80Subnormal(aSig, &aExp, &aSig); } @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_etox(floatx80 a, float_status *status) if (aSign) { return packFloatx80(0, 0, 0); } - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_twotox(floatx80 a, float_status *status) if (aSign) { return packFloatx80(0, 0, 0); } - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_tentox(floatx80 a, float_status *status) if (aSign) { return packFloatx80(0, 0, 0); } - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_atanh(floatx80 a, float_status *status) if (compact >= 0x3FFF8000) { /* |X| >= 1 */ if (aExp == one_exp && aSig == one_sig) { /* |X| == 1 */ float_raise(float_flag_divbyzero, status); - return packFloatx80(aSign, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(aSign, status); } else { /* |X| > 1 */ float_raise(float_flag_invalid, status); return floatx80_default_nan(status); @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_etoxm1(floatx80 a, float_status *status) if (aSign) { return packFloatx80(aSign, one_exp, one_sig); } - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_sinh(floatx80 a, float_status *status) if ((uint64_t) (aSig << 1)) { return propagateFloatx80NaNOneArg(a, status); } - return packFloatx80(aSign, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(aSign, status); } if (aExp == 0 && aSig == 0) { @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_cosh(floatx80 a, float_status *status) if ((uint64_t) (aSig << 1)) { return propagateFloatx80NaNOneArg(a, status); } - return packFloatx80(0, floatx80_infinity.high, - floatx80_infinity.low); + return floatx80_default_inf(0, status); } if (aExp == 0 && aSig == 0) { -- 2.43.0 The global const floatx80_infinity is (unlike all the other float*_infinity values) target-specific, because whether the explicit Integer bit is set or not varies between m68k and i386. We want to be able to compile softfloat once for multiple targets, so we can't continue to use a single global whose value needs to be different between targets. Replace the direct uses of floatx80_infinity in target/i386 with calls to the new floatx80_default_inf() function. Note that because we can ask the function for either a negative or positive infinity, we don't need to change the sign of a positive infinity via floatx80_chs() for the negative-Inf case. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-4-peter.maydell@linaro.org Message-id: 20250217125055.160887-4-peter.maydell@linaro.org --- target/i386/tcg/fpu_helper.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/target/i386/tcg/fpu_helper.c b/target/i386/tcg/fpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/fpu_helper.c +++ b/target/i386/tcg/fpu_helper.c @@ -XXX,XX +XXX,XX @@ void helper_fxtract(CPUX86State *env) } else if (floatx80_is_infinity(ST0)) { fpush(env); ST0 = ST1; - ST1 = floatx80_infinity; + ST1 = floatx80_default_inf(0, &env->fp_status); } else { int expdif; @@ -XXX,XX +XXX,XX @@ void helper_fscale(CPUX86State *env) float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_default_nan(&env->fp_status); } else { - ST0 = (floatx80_is_neg(ST0) ? - floatx80_chs(floatx80_infinity) : - floatx80_infinity); + ST0 = floatx80_default_inf(floatx80_is_neg(ST0), + &env->fp_status); } } } else { -- 2.43.0 Unlike the other float formats, whether a floatx80 value is considered to be an Infinity is target-dependent. (On x86 if the explicit integer bit is clear this is a "pseudo-infinity" and not a valid infinity; m68k does not care about the value of the integer bit.) Currently we select this target-specific logic at compile time with an ifdef. We're going to want to do this at runtime, so change the floatx80_is_infinity() function to take a float_status. This commit doesn't change any logic; we'll do that in the next commit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-5-peter.maydell@linaro.org --- include/fpu/softfloat.h | 2 +- target/i386/tcg/fpu_helper.c | 20 +++++++++++--------- target/m68k/fpu_helper.c | 2 +- 3 files changed, 13 insertions(+), 11 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ static inline floatx80 floatx80_chs(floatx80 a) return a; } -static inline bool floatx80_is_infinity(floatx80 a) +static inline bool floatx80_is_infinity(floatx80 a, float_status *status) { #if defined(TARGET_M68K) return (a.high & 0x7fff) == floatx80_infinity.high && !(a.low << 1); diff --git a/target/i386/tcg/fpu_helper.c b/target/i386/tcg/fpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/fpu_helper.c +++ b/target/i386/tcg/fpu_helper.c @@ -XXX,XX +XXX,XX @@ void helper_fpatan(CPUX86State *env) /* Pass this NaN through. */ } else if (floatx80_is_zero(ST1) && !arg0_sign) { /* Pass this zero through. */ - } else if (((floatx80_is_infinity(ST0) && !floatx80_is_infinity(ST1)) || + } else if (((floatx80_is_infinity(ST0, &env->fp_status) && + !floatx80_is_infinity(ST1, &env->fp_status)) || arg0_exp - arg1_exp >= 80) && !arg0_sign) { /* @@ -XXX,XX +XXX,XX @@ void helper_fpatan(CPUX86State *env) rexp = pi_exp; rsig0 = pi_sig_high; rsig1 = pi_sig_low; - } else if (floatx80_is_infinity(ST1)) { - if (floatx80_is_infinity(ST0)) { + } else if (floatx80_is_infinity(ST1, &env->fp_status)) { + if (floatx80_is_infinity(ST0, &env->fp_status)) { if (arg0_sign) { rexp = pi_34_exp; rsig0 = pi_34_sig_high; @@ -XXX,XX +XXX,XX @@ void helper_fpatan(CPUX86State *env) rexp = pi_2_exp; rsig0 = pi_2_sig_high; rsig1 = pi_2_sig_low; - } else if (floatx80_is_infinity(ST0) || arg0_exp - arg1_exp >= 80) { + } else if (floatx80_is_infinity(ST0, &env->fp_status) || + arg0_exp - arg1_exp >= 80) { /* ST0 is negative. */ rexp = pi_exp; rsig0 = pi_sig_high; @@ -XXX,XX +XXX,XX @@ void helper_fxtract(CPUX86State *env) } fpush(env); ST0 = ST1; - } else if (floatx80_is_infinity(ST0)) { + } else if (floatx80_is_infinity(ST0, &env->fp_status)) { fpush(env); ST0 = ST1; ST1 = floatx80_default_inf(0, &env->fp_status); @@ -XXX,XX +XXX,XX @@ void helper_fyl2x(CPUX86State *env) } else if (arg0_sign && !floatx80_is_zero(ST0)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_default_nan(&env->fp_status); - } else if (floatx80_is_infinity(ST1)) { + } else if (floatx80_is_infinity(ST1, &env->fp_status)) { FloatRelation cmp = floatx80_compare(ST0, floatx80_one, &env->fp_status); switch (cmp) { @@ -XXX,XX +XXX,XX @@ void helper_fyl2x(CPUX86State *env) ST1 = floatx80_default_nan(&env->fp_status); break; } - } else if (floatx80_is_infinity(ST0)) { + } else if (floatx80_is_infinity(ST0, &env->fp_status)) { if (floatx80_is_zero(ST1)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_default_nan(&env->fp_status); @@ -XXX,XX +XXX,XX @@ void helper_fscale(CPUX86State *env) float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_silence_nan(ST0, &env->fp_status); } - } else if (floatx80_is_infinity(ST1) && + } else if (floatx80_is_infinity(ST1, &env->fp_status) && !floatx80_invalid_encoding(ST0) && !floatx80_is_any_nan(ST0)) { if (floatx80_is_neg(ST1)) { - if (floatx80_is_infinity(ST0)) { + if (floatx80_is_infinity(ST0, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_default_nan(&env->fp_status); } else { diff --git a/target/m68k/fpu_helper.c b/target/m68k/fpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/fpu_helper.c +++ b/target/m68k/fpu_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(ftst)(CPUM68KState *env, FPReg *val) if (floatx80_is_any_nan(val->d)) { cc |= FPSR_CC_A; - } else if (floatx80_is_infinity(val->d)) { + } else if (floatx80_is_infinity(val->d, &env->fp_status)) { cc |= FPSR_CC_I; } else if (floatx80_is_zero(val->d)) { cc |= FPSR_CC_Z; -- 2.43.0 In Intel terminology, a floatx80 Infinity with the explicit integer bit clear is a "pseudo-infinity"; for x86 these are not valid infinity values. m68k is looser and does not care whether the Integer bit is set or clear in an infinity. Move this setting to runtime rather than using an ifdef in floatx80_is_infinity(). Since this was the last use of the floatx80_infinity global constant, we remove it and its definition here. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-6-peter.maydell@linaro.org Message-id: 20250217125055.160887-5-peter.maydell@linaro.org --- include/fpu/softfloat-types.h | 5 +++++ include/fpu/softfloat.h | 18 +++++++++++------- target/m68k/cpu.c | 4 +++- fpu/softfloat-specialize.c.inc | 10 ---------- 4 files changed, 19 insertions(+), 18 deletions(-) diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { typedef enum __attribute__((__packed__)) { /* In the default Infinity value, is the Integer bit 0 ? */ floatx80_default_inf_int_bit_is_zero = 1, + /* + * Are Pseudo-infinities (Inf with the Integer bit zero) valid? + * If so, floatx80_is_infinity() will return true for them. + */ + floatx80_pseudo_inf_valid = 2, } FloatX80Behaviour; /* diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ float128 floatx80_to_float128(floatx80, float_status *status); /*---------------------------------------------------------------------------- | The pattern for an extended double-precision inf. *----------------------------------------------------------------------------*/ -extern const floatx80 floatx80_infinity; floatx80 floatx80_default_inf(bool zSign, float_status *status); /*---------------------------------------------------------------------------- @@ -XXX,XX +XXX,XX @@ static inline floatx80 floatx80_chs(floatx80 a) static inline bool floatx80_is_infinity(floatx80 a, float_status *status) { -#if defined(TARGET_M68K) - return (a.high & 0x7fff) == floatx80_infinity.high && !(a.low << 1); -#else - return (a.high & 0x7fff) == floatx80_infinity.high && - a.low == floatx80_infinity.low; -#endif + /* + * It's target-specific whether the Integer bit is permitted + * to be 0 in a valid Infinity value. (x86 says no, m68k says yes). + */ + bool intbit = a.low >> 63; + + if (!intbit && + !(status->floatx80_behaviour & floatx80_pseudo_inf_valid)) { + return false; + } + return (a.high & 0x7fff) == 0x7fff && !(a.low << 1); } static inline bool floatx80_is_neg(floatx80 a) diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static void m68k_cpu_reset_hold(Object *obj, ResetType type) /* * m68k-specific floatx80 behaviour: * * default Infinity values have a zero Integer bit + * * input Infinities may have the Integer bit either 0 or 1 */ - set_floatx80_behaviour(floatx80_default_inf_int_bit_is_zero, + set_floatx80_behaviour(floatx80_default_inf_int_bit_is_zero | + floatx80_pseudo_inf_valid, &env->fp_status); nan = floatx80_default_nan(&env->fp_status); diff --git a/fpu/softfloat-specialize.c.inc b/fpu/softfloat-specialize.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-specialize.c.inc +++ b/fpu/softfloat-specialize.c.inc @@ -XXX,XX +XXX,XX @@ floatx80 floatx80_default_inf(bool zSign, float_status *status) return packFloatx80(zSign, 0x7fff, z ? 0 : (1ULL << 63)); } -#define floatx80_infinity_high 0x7FFF -#if defined(TARGET_M68K) -#define floatx80_infinity_low UINT64_C(0x0000000000000000) -#else -#define floatx80_infinity_low UINT64_C(0x8000000000000000) -#endif - -const floatx80 floatx80_infinity - = make_floatx80_init(floatx80_infinity_high, floatx80_infinity_low); - /*---------------------------------------------------------------------------- | Returns 1 if the half-precision floating-point value `a' is a quiet | NaN; otherwise returns 0. -- 2.43.0 The definition of which floatx80 encodings are invalid is target-specific. Currently we handle this with an ifdef, but we would like to defer this decision to runtime. In preparation, pass a float_status argument to floatx80_invalid_encoding(). We will change the implementation from ifdef to looking at the status argument in the following commit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-7-peter.maydell@linaro.org --- include/fpu/softfloat.h | 2 +- fpu/softfloat.c | 2 +- target/i386/tcg/fpu_helper.c | 24 +++++++++++++----------- 3 files changed, 15 insertions(+), 13 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ static inline bool floatx80_unordered_quiet(floatx80 a, floatx80 b, | pseudo-denormals, which must still be correctly handled as inputs even | if they are never generated as outputs. *----------------------------------------------------------------------------*/ -static inline bool floatx80_invalid_encoding(floatx80 a) +static inline bool floatx80_invalid_encoding(floatx80 a, float_status *s) { #if defined(TARGET_M68K) /*------------------------------------------------------------------------- diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ static bool floatx80_unpack_canonical(FloatParts128 *p, floatx80 f, g_assert_not_reached(); } - if (unlikely(floatx80_invalid_encoding(f))) { + if (unlikely(floatx80_invalid_encoding(f, s))) { float_raise(float_flag_invalid, s); return false; } diff --git a/target/i386/tcg/fpu_helper.c b/target/i386/tcg/fpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/fpu_helper.c +++ b/target/i386/tcg/fpu_helper.c @@ -XXX,XX +XXX,XX @@ void helper_f2xm1(CPUX86State *env) int32_t exp = extractFloatx80Exp(ST0); bool sign = extractFloatx80Sign(ST0); - if (floatx80_invalid_encoding(ST0)) { + if (floatx80_invalid_encoding(ST0, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_default_nan(&env->fp_status); } else if (floatx80_is_any_nan(ST0)) { @@ -XXX,XX +XXX,XX @@ void helper_fpatan(CPUX86State *env) } else if (floatx80_is_signaling_nan(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_silence_nan(ST1, &env->fp_status); - } else if (floatx80_invalid_encoding(ST0) || - floatx80_invalid_encoding(ST1)) { + } else if (floatx80_invalid_encoding(ST0, &env->fp_status) || + floatx80_invalid_encoding(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_default_nan(&env->fp_status); } else if (floatx80_is_any_nan(ST0)) { @@ -XXX,XX +XXX,XX @@ void helper_fxtract(CPUX86State *env) &env->fp_status); fpush(env); ST0 = temp.d; - } else if (floatx80_invalid_encoding(ST0)) { + } else if (floatx80_invalid_encoding(ST0, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_default_nan(&env->fp_status); fpush(env); @@ -XXX,XX +XXX,XX @@ static void helper_fprem_common(CPUX86State *env, bool mod) env->fpus &= ~0x4700; /* (C3,C2,C1,C0) <-- 0000 */ if (floatx80_is_zero(ST0) || floatx80_is_zero(ST1) || exp0 == 0x7fff || exp1 == 0x7fff || - floatx80_invalid_encoding(ST0) || floatx80_invalid_encoding(ST1)) { + floatx80_invalid_encoding(ST0, &env->fp_status) || + floatx80_invalid_encoding(ST1, &env->fp_status)) { ST0 = floatx80_modrem(ST0, ST1, mod, &quotient, &env->fp_status); } else { if (exp0 == 0) { @@ -XXX,XX +XXX,XX @@ void helper_fyl2xp1(CPUX86State *env) } else if (floatx80_is_signaling_nan(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_silence_nan(ST1, &env->fp_status); - } else if (floatx80_invalid_encoding(ST0) || - floatx80_invalid_encoding(ST1)) { + } else if (floatx80_invalid_encoding(ST0, &env->fp_status) || + floatx80_invalid_encoding(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_default_nan(&env->fp_status); } else if (floatx80_is_any_nan(ST0)) { @@ -XXX,XX +XXX,XX @@ void helper_fyl2x(CPUX86State *env) } else if (floatx80_is_signaling_nan(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_silence_nan(ST1, &env->fp_status); - } else if (floatx80_invalid_encoding(ST0) || - floatx80_invalid_encoding(ST1)) { + } else if (floatx80_invalid_encoding(ST0, &env->fp_status) || + floatx80_invalid_encoding(ST1, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST1 = floatx80_default_nan(&env->fp_status); } else if (floatx80_is_any_nan(ST0)) { @@ -XXX,XX +XXX,XX @@ void helper_frndint(CPUX86State *env) void helper_fscale(CPUX86State *env) { uint8_t old_flags = save_exception_flags(env); - if (floatx80_invalid_encoding(ST1) || floatx80_invalid_encoding(ST0)) { + if (floatx80_invalid_encoding(ST1, &env->fp_status) || + floatx80_invalid_encoding(ST0, &env->fp_status)) { float_raise(float_flag_invalid, &env->fp_status); ST0 = floatx80_default_nan(&env->fp_status); } else if (floatx80_is_any_nan(ST1)) { @@ -XXX,XX +XXX,XX @@ void helper_fscale(CPUX86State *env) ST0 = floatx80_silence_nan(ST0, &env->fp_status); } } else if (floatx80_is_infinity(ST1, &env->fp_status) && - !floatx80_invalid_encoding(ST0) && + !floatx80_invalid_encoding(ST0, &env->fp_status) && !floatx80_is_any_nan(ST0)) { if (floatx80_is_neg(ST1)) { if (floatx80_is_infinity(ST0, &env->fp_status)) { -- 2.43.0 Because floatx80 has an explicit integer bit, this permits some odd encodings where the integer bit is not set correctly for the floating point value type. In In Intel terminology the categories are: exp == 0, int = 0, mantissa == 0 : zeroes exp == 0, int = 0, mantissa != 0 : denormals exp == 0, int = 1 : pseudo-denormals 0 < exp < 0x7fff, int = 0 : unnormals 0 < exp < 0x7fff, int = 1 : normals exp == 0x7fff, int = 0, mantissa == 0 : pseudo-infinities exp == 0x7fff, int = 1, mantissa == 0 : infinities exp == 0x7fff, int = 0, mantissa != 0 : pseudo-NaNs exp == 0x7fff, int = 1, mantissa == 0 : NaNs The usual IEEE cases of zero, denormal, normal, inf and NaN are always valid. x87 permits as input also pseudo-denormals. m68k permits all those and also pseudo-infinities, pseudo-NaNs and unnormals. Currently we have an ifdef in floatx80_invalid_encoding() to select the x86 vs m68k behaviour. Add new floatx80_behaviour flags to select whether pseudo-NaN and unnormal are valid, and use these (plus the existing pseudo_inf_valid flag) to decide whether these encodings are invalid at runtime. We leave pseudo-denormals as always-valid, since both x86 and m68k accept them. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-8-peter.maydell@linaro.org Message-id: 20250217125055.160887-6-peter.maydell@linaro.org --- include/fpu/softfloat-types.h | 14 ++++++++ include/fpu/softfloat.h | 68 ++++++++++++++++++----------------- target/m68k/cpu.c | 28 ++++++++++++++- 3 files changed, 77 insertions(+), 33 deletions(-) diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { /* * Are Pseudo-infinities (Inf with the Integer bit zero) valid? * If so, floatx80_is_infinity() will return true for them. + * If not, floatx80_invalid_encoding will return false for them, + * and using them as inputs to a float op will raise Invalid. */ floatx80_pseudo_inf_valid = 2, + /* + * Are Pseudo-NaNs (NaNs where the Integer bit is zero) valid? + * If not, floatx80_invalid_encoding() will return false for them, + * and using them as inputs to a float op will raise Invalid. + */ + floatx80_pseudo_nan_valid = 4, + /* + * Are Unnormals (0 < exp < 0x7fff, Integer bit zero) valid? + * If not, floatx80_invalid_encoding() will return false for them, + * and using them as inputs to a float op will raise Invalid. + */ + floatx80_unnormal_valid = 8, } FloatX80Behaviour; /* diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ static inline bool floatx80_unordered_quiet(floatx80 a, floatx80 b, /*---------------------------------------------------------------------------- | Return whether the given value is an invalid floatx80 encoding. -| Invalid floatx80 encodings arise when the integer bit is not set, but -| the exponent is not zero. The only times the integer bit is permitted to -| be zero is in subnormal numbers and the value zero. -| This includes what the Intel software developer's manual calls pseudo-NaNs, -| pseudo-infinities and un-normal numbers. It does not include -| pseudo-denormals, which must still be correctly handled as inputs even -| if they are never generated as outputs. +| Invalid floatx80 encodings may arise when the integer bit is not set +| correctly; this is target-specific. In Intel terminology the +| categories are: +| exp == 0, int = 0, mantissa == 0 : zeroes +| exp == 0, int = 0, mantissa != 0 : denormals +| exp == 0, int = 1 : pseudo-denormals +| 0 < exp < 0x7fff, int = 0 : unnormals +| 0 < exp < 0x7fff, int = 1 : normals +| exp == 0x7fff, int = 0, mantissa == 0 : pseudo-infinities +| exp == 0x7fff, int = 1, mantissa == 0 : infinities +| exp == 0x7fff, int = 0, mantissa != 0 : pseudo-NaNs +| exp == 0x7fff, int = 1, mantissa == 0 : NaNs +| +| The usual IEEE cases of zero, denormal, normal, inf and NaN are always valid. +| x87 permits as input also pseudo-denormals. +| m68k permits all those and also pseudo-infinities, pseudo-NaNs and unnormals. +| +| Since we don't have a target that handles floatx80 but prohibits +| pseudo-denormals in input, we don't currently have a floatx80_behaviour +| flag for that case, but instead always accept it. Conveniently this +| means that all cases with either exponent 0 or the integer bit set are +| valid for all targets. *----------------------------------------------------------------------------*/ static inline bool floatx80_invalid_encoding(floatx80 a, float_status *s) { -#if defined(TARGET_M68K) - /*------------------------------------------------------------------------- - | With m68k, the explicit integer bit can be zero in the case of: - | - zeros (exp == 0, mantissa == 0) - | - denormalized numbers (exp == 0, mantissa != 0) - | - unnormalized numbers (exp != 0, exp < 0x7FFF) - | - infinities (exp == 0x7FFF, mantissa == 0) - | - not-a-numbers (exp == 0x7FFF, mantissa != 0) - | - | For infinities and NaNs, the explicit integer bit can be either one or - | zero. - | - | The IEEE 754 standard does not define a zero integer bit. Such a number - | is an unnormalized number. Hardware does not directly support - | denormalized and unnormalized numbers, but implicitly supports them by - | trapping them as unimplemented data types, allowing efficient conversion - | in software. - | - | See "M68000 FAMILY PROGRAMMER’S REFERENCE MANUAL", - | "1.6 FLOATING-POINT DATA TYPES" - *------------------------------------------------------------------------*/ - return false; -#else - return (a.low & (1ULL << 63)) == 0 && (a.high & 0x7FFF) != 0; -#endif + if ((a.low >> 63) || (a.high & 0x7fff) == 0) { + /* Anything with the Integer bit set or the exponent 0 is valid */ + return false; + } + + if ((a.high & 0x7fff) == 0x7fff) { + if (a.low) { + return !(s->floatx80_behaviour & floatx80_pseudo_nan_valid); + } else { + return !(s->floatx80_behaviour & floatx80_pseudo_inf_valid); + } + } else { + return !(s->floatx80_behaviour & floatx80_unnormal_valid); + } } #define floatx80_zero make_floatx80(0x0000, 0x0000000000000000LL) diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static void m68k_cpu_reset_hold(Object *obj, ResetType type) * m68k-specific floatx80 behaviour: * * default Infinity values have a zero Integer bit * * input Infinities may have the Integer bit either 0 or 1 + * * pseudo-denormals supported for input and output + * * don't raise Invalid for pseudo-NaN/pseudo-Inf/Unnormal + * + * With m68k, the explicit integer bit can be zero in the case of: + * - zeros (exp == 0, mantissa == 0) + * - denormalized numbers (exp == 0, mantissa != 0) + * - unnormalized numbers (exp != 0, exp < 0x7FFF) + * - infinities (exp == 0x7FFF, mantissa == 0) + * - not-a-numbers (exp == 0x7FFF, mantissa != 0) + * + * For infinities and NaNs, the explicit integer bit can be either one or + * zero. + * + * The IEEE 754 standard does not define a zero integer bit. Such a number + * is an unnormalized number. Hardware does not directly support + * denormalized and unnormalized numbers, but implicitly supports them by + * trapping them as unimplemented data types, allowing efficient conversion + * in software. + * + * See "M68000 FAMILY PROGRAMMER’S REFERENCE MANUAL", + * "1.6 FLOATING-POINT DATA TYPES" + * + * Note though that QEMU's fp emulation does directly handle both + * denormal and unnormal values, and does not trap to guest software. */ set_floatx80_behaviour(floatx80_default_inf_int_bit_is_zero | - floatx80_pseudo_inf_valid, + floatx80_pseudo_inf_valid | + floatx80_pseudo_nan_valid | + floatx80_unnormal_valid, &env->fp_status); nan = floatx80_default_nan(&env->fp_status); -- 2.43.0 Currently we compile-time set an 'm68k_denormal' flag in the FloatFmt for floatx80 for m68k. This controls our handling of what the Intel documentation calls a "pseudo-denormal": a value where the exponent field is zero and the explicit integer bit is set. For x86, the x87 FPU is supposed to accept a pseudo-denormal as input, but never generate one on output. For m68k, these values are permitted on input and may be produced on output. Replace the flag in the FloatFmt with a flag indicating whether the float format has an explicit bit (which will be true for floatx80 for all targets, and false for every other float type). Then we can gate the handling of these pseudo-denormals on the setting of a floatx80_behaviour flag. As far as I can see from the code we don't actually handle the x86-mandated "accept on input but don't generate" behaviour, because the handling in partsN(canonicalize) looked at fmt->m68k_denormal. So I have added TODO comments to that effect. This commit doesn't change any behaviour for any target. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-id: 20250224111524.1101196-9-peter.maydell@linaro.org Message-id: 20250217125055.160887-7-peter.maydell@linaro.org --- include/fpu/softfloat-types.h | 19 +++++++++++++++++++ fpu/softfloat.c | 9 ++++----- target/m68k/cpu.c | 3 ++- fpu/softfloat-parts.c.inc | 27 ++++++++++++++++++++++++--- 4 files changed, 49 insertions(+), 9 deletions(-) diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { * and using them as inputs to a float op will raise Invalid. */ floatx80_unnormal_valid = 8, + + /* + * If the exponent is 0 and the Integer bit is set, Intel call + * this a "pseudo-denormal"; x86 supports that only on input + * (treating them as denormals by ignoring the Integer bit). + * For m68k, the integer bit is considered validly part of the + * input value when the exponent is 0, and may be 0 or 1, + * giving extra range. They may also be generated as outputs. + * (The m68k manual actually calls these values part of the + * normalized number range, not the denormalized number range.) + * + * By default you get the Intel behaviour where the Integer + * bit is ignored; if this is set then the Integer bit value + * is honoured, m68k-style. + * + * Either way, floatx80_invalid_encoding() will always accept + * pseudo-denormals. + */ + floatx80_pseudo_denormal_valid = 16, } FloatX80Behaviour; /* diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ typedef struct { * round_mask: bits below lsb which must be rounded * The following optional modifiers are available: * arm_althp: handle ARM Alternative Half Precision - * m68k_denormal: explicit integer bit for extended precision may be 1 + * has_explicit_bit: has an explicit integer bit; this affects whether + * the float_status floatx80_behaviour handling applies */ typedef struct { int exp_size; @@ -XXX,XX +XXX,XX @@ typedef struct { int frac_size; int frac_shift; bool arm_althp; - bool m68k_denormal; + bool has_explicit_bit; uint64_t round_mask; } FloatFmt; @@ -XXX,XX +XXX,XX @@ static const FloatFmt floatx80_params[3] = { [floatx80_precision_d] = { FLOATX80_PARAMS(52) }, [floatx80_precision_x] = { FLOATX80_PARAMS(64), -#ifdef TARGET_M68K - .m68k_denormal = true, -#endif + .has_explicit_bit = true, }, }; diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static void m68k_cpu_reset_hold(Object *obj, ResetType type) set_floatx80_behaviour(floatx80_default_inf_int_bit_is_zero | floatx80_pseudo_inf_valid | floatx80_pseudo_nan_valid | - floatx80_unnormal_valid, + floatx80_unnormal_valid | + floatx80_pseudo_denormal_valid, &env->fp_status); nan = floatx80_default_nan(&env->fp_status); diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(pick_nan_muladd)(FloatPartsN *a, FloatPartsN *b, static void partsN(canonicalize)(FloatPartsN *p, float_status *status, const FloatFmt *fmt) { + /* + * It's target-dependent how to handle the case of exponent 0 + * and Integer bit set. Intel calls these "pseudodenormals", + * and treats them as if the integer bit was 0, and never + * produces them on output. This is the default behaviour for QEMU. + * For m68k, the integer bit is considered validly part of the + * input value when the exponent is 0, and may be 0 or 1, + * giving extra range. They may also be generated as outputs. + * (The m68k manual actually calls these values part of the + * normalized number range, not the denormalized number range, + * but that distinction is not important for us, because + * m68k doesn't care about the input_denormal_used status flag.) + * floatx80_pseudo_denormal_valid selects the m68k behaviour, + * which changes both how we canonicalize such a value and + * how we uncanonicalize results. + */ + bool has_pseudo_denormals = fmt->has_explicit_bit && + (status->floatx80_behaviour & floatx80_pseudo_denormal_valid); + if (unlikely(p->exp == 0)) { if (likely(frac_eqz(p))) { p->cls = float_class_zero; @@ -XXX,XX +XXX,XX @@ static void partsN(canonicalize)(FloatPartsN *p, float_status *status, int shift = frac_normalize(p); p->cls = float_class_denormal; p->exp = fmt->frac_shift - fmt->exp_bias - - shift + !fmt->m68k_denormal; + - shift + !has_pseudo_denormals; } } else if (likely(p->exp < fmt->exp_max) || fmt->arm_althp) { p->cls = float_class_normal; @@ -XXX,XX +XXX,XX @@ static void partsN(uncanon_normal)(FloatPartsN *p, float_status *s, frac_clear(p); } else { bool is_tiny = s->tininess_before_rounding || exp < 0; + bool has_pseudo_denormals = fmt->has_explicit_bit && + (s->floatx80_behaviour & floatx80_pseudo_denormal_valid); if (!is_tiny) { FloatPartsN discard; is_tiny = !frac_addi(&discard, p, inc); } - frac_shrjam(p, !fmt->m68k_denormal - exp); + frac_shrjam(p, !has_pseudo_denormals - exp); if (p->frac_lo & round_mask) { /* Need to recompute round-to-even/round-to-odd. */ @@ -XXX,XX +XXX,XX @@ static void partsN(uncanon_normal)(FloatPartsN *p, float_status *s, p->frac_lo &= ~round_mask; } - exp = (p->frac_hi & DECOMPOSED_IMPLICIT_BIT) && !fmt->m68k_denormal; + exp = (p->frac_hi & DECOMPOSED_IMPLICIT_BIT) && !has_pseudo_denormals; frac_shr(p, frac_shift); if (is_tiny) { -- 2.43.0 Currently we have a compile-time shortcut where we return false from no_signaling_nans() on everything except Xtensa, because we know that's the only target that might ever set status->no_signaling_nans. Remove the ifdef, so we always look at the status flag; this has no behavioural change, but will be necessary if we want to build softfloat once for all targets. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250224111524.1101196-10-peter.maydell@linaro.org Message-id: 20250217125055.160887-8-peter.maydell@linaro.org --- fpu/softfloat-specialize.c.inc | 4 ---- 1 file changed, 4 deletions(-) diff --git a/fpu/softfloat-specialize.c.inc b/fpu/softfloat-specialize.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-specialize.c.inc +++ b/fpu/softfloat-specialize.c.inc @@ -XXX,XX +XXX,XX @@ this code that are retained. */ static inline bool no_signaling_nans(float_status *status) { -#if defined(TARGET_XTENSA) return status->no_signaling_nans; -#else - return false; -#endif } /* Define how the architecture discriminates signaling NaNs. -- 2.43.0 Currently we have a compile-time shortcut where we return a hardcode value from snan_bit_is_one() on everything except MIPS, because we know that's the only target that needs to change status->no_signaling_nans at runtime. Remove the ifdef, so we always look at the status flag. This means we must update the two targets (HPPA and SH4) that were previously hardcoded to return true so that they set the status flag correctly. This has no behavioural change, but will be necessary if we want to build softfloat once for all targets. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250224111524.1101196-11-peter.maydell@linaro.org Message-id: 20250217125055.160887-9-peter.maydell@linaro.org --- target/hppa/fpu_helper.c | 1 + target/sh4/cpu.c | 1 + fpu/softfloat-specialize.c.inc | 7 ------- 3 files changed, 2 insertions(+), 7 deletions(-) diff --git a/target/hppa/fpu_helper.c b/target/hppa/fpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/fpu_helper.c +++ b/target/hppa/fpu_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(loaded_fr0)(CPUHPPAState *env) set_float_infzeronan_rule(float_infzeronan_dnan_never, &env->fp_status); /* Default NaN: sign bit clear, msb-1 frac bit set */ set_float_default_nan_pattern(0b00100000, &env->fp_status); + set_snan_bit_is_one(true, &env->fp_status); /* * "PA-RISC 2.0 Architecture" says it is IMPDEF whether the flushing * enabled by FPSR.D happens before or after rounding. We pick "before" diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/cpu.c +++ b/target/sh4/cpu.c @@ -XXX,XX +XXX,XX @@ static void superh_cpu_reset_hold(Object *obj, ResetType type) set_flush_to_zero(1, &env->fp_status); #endif set_default_nan_mode(1, &env->fp_status); + set_snan_bit_is_one(true, &env->fp_status); /* sign bit clear, set all frac bits other than msb */ set_float_default_nan_pattern(0b00111111, &env->fp_status); /* diff --git a/fpu/softfloat-specialize.c.inc b/fpu/softfloat-specialize.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-specialize.c.inc +++ b/fpu/softfloat-specialize.c.inc @@ -XXX,XX +XXX,XX @@ static inline bool no_signaling_nans(float_status *status) * In IEEE 754-1985 this was implementation defined, but in IEEE 754-2008 * the msb must be zero. MIPS is (so far) unique in supporting both the * 2008 revision and backward compatibility with their original choice. - * Thus for MIPS we must make the choice at runtime. */ static inline bool snan_bit_is_one(float_status *status) { -#if defined(TARGET_MIPS) return status->snan_bit_is_one; -#elif defined(TARGET_HPPA) || defined(TARGET_SH4) - return 1; -#else - return 0; -#endif } /*---------------------------------------------------------------------------- -- 2.43.0 We happen to know that for the PPC target the FP status flags (and in particular float_flag_inexact) will always be cleared before a floating point operation, and so can_use_fpu() will always return false. So we speed things up a little by forcing QEMU_NO_HARDFLOAT to true on that target. We would like to build softfloat once for all targets; that means removing target-specific ifdefs. Remove the check for TARGET_PPC; this won't change behaviour because can_use_fpu() will see that float_flag_inexact is clear and take the softfloat path anyway. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250224111524.1101196-12-peter.maydell@linaro.org Message-id: 20250217125055.160887-10-peter.maydell@linaro.org --- fpu/softfloat.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ GEN_INPUT_FLUSH3(float64_input_flush3, float64) * the use of hardfloat, since hardfloat relies on the inexact flag being * already set. */ -#if defined(TARGET_PPC) || defined(__FAST_MATH__) # if defined(__FAST_MATH__) # warning disabling hardfloat due to -ffast-math: hardfloat requires an exact \ IEEE implementation -# endif # define QEMU_NO_HARDFLOAT 1 # define QEMU_SOFTFLOAT_ATTR QEMU_FLATTEN #else -- 2.43.0 Now we have removed all the target-specifics from the softfloat code, we can switch to building it once for the whole system rather than once per target. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250224111524.1101196-13-peter.maydell@linaro.org Message-id: 20250217125055.160887-11-peter.maydell@linaro.org --- fpu/softfloat.c | 3 --- fpu/meson.build | 2 +- 2 files changed, 1 insertion(+), 4 deletions(-) diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ this code that are retained. * version 2 or later. See the COPYING file in the top-level directory. */ -/* softfloat (and in particular the code in softfloat-specialize.h) is - * target-dependent and needs the TARGET_* macros. - */ #include "qemu/osdep.h" #include <math.h> #include "qemu/bitops.h" diff --git a/fpu/meson.build b/fpu/meson.build index XXXXXXX..XXXXXXX 100644 --- a/fpu/meson.build +++ b/fpu/meson.build @@ -1 +1 @@ -specific_ss.add(when: 'CONFIG_TCG', if_true: files('softfloat.c')) +common_ss.add(when: 'CONFIG_TCG', if_true: files('softfloat.c')) -- 2.43.0 Most of the target/arm/vfp_helper.c file is purely TCG helper code, guarded by #ifdef CONFIG_TCG. Move this into a new file in target/arm/tcg/. This leaves only the code relating to getting and setting the FPCR/FPSR/FPSCR in the original file. (Some of this also is TCG-only, but that needs more careful disentangling.) Having two vfp_helper.c files might seem a bit confusing, but once we've finished moving all the helper code out of the old file we are going to rename it to vfp_fpscr.c. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250221190957.811948-2-peter.maydell@linaro.org --- target/arm/{ => tcg}/vfp_helper.c | 399 +---------- target/arm/vfp_helper.c | 1109 ----------------------------- target/arm/tcg/meson.build | 1 + 3 files changed, 4 insertions(+), 1505 deletions(-) copy target/arm/{ => tcg}/vfp_helper.c (71%) diff --git a/target/arm/vfp_helper.c b/target/arm/tcg/vfp_helper.c similarity index 71% copy from target/arm/vfp_helper.c copy to target/arm/tcg/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/tcg/vfp_helper.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "cpu-features.h" #include "fpu/softfloat.h" -#ifdef CONFIG_TCG #include "qemu/log.h" -#endif - -/* VFP support. We follow the convention used for VFP instructions: - Single precision routines have a "s" suffix, double precision a - "d" suffix. */ /* - * Set the float_status behaviour to match the Arm defaults: - * * tininess-before-rounding - * * 2-input NaN propagation prefers SNaN over QNaN, and then - * operand A over operand B (see FPProcessNaNs() pseudocode) - * * 3-input NaN propagation prefers SNaN over QNaN, and then - * operand C over A over B (see FPProcessNaNs3() pseudocode, - * but note that for QEMU muladd is a * b + c, whereas for - * the pseudocode function the arguments are in the order c, a, b. - * * 0 * Inf + NaN returns the default NaN if the input NaN is quiet, - * and the input NaN if it is signalling - * * Default NaN has sign bit clear, msb frac bit set + * VFP support. We follow the convention used for VFP instructions: + * Single precision routines have a "s" suffix, double precision a + * "d" suffix. */ -void arm_set_default_fp_behaviours(float_status *s) -{ - set_float_detect_tininess(float_tininess_before_rounding, s); - set_float_ftz_detection(float_ftz_before_rounding, s); - set_float_2nan_prop_rule(float_2nan_prop_s_ab, s); - set_float_3nan_prop_rule(float_3nan_prop_s_cab, s); - set_float_infzeronan_rule(float_infzeronan_dnan_if_qnan, s); - set_float_default_nan_pattern(0b01000000, s); -} - -/* - * Set the float_status behaviour to match the FEAT_AFP - * FPCR.AH=1 requirements: - * * tininess-after-rounding - * * 2-input NaN propagation prefers the first NaN - * * 3-input NaN propagation prefers a over b over c - * * 0 * Inf + NaN always returns the input NaN and doesn't - * set Invalid for a QNaN - * * default NaN has sign bit set, msb frac bit set - */ -void arm_set_ah_fp_behaviours(float_status *s) -{ - set_float_detect_tininess(float_tininess_after_rounding, s); - set_float_ftz_detection(float_ftz_after_rounding, s); - set_float_2nan_prop_rule(float_2nan_prop_ab, s); - set_float_3nan_prop_rule(float_3nan_prop_abc, s); - set_float_infzeronan_rule(float_infzeronan_dnan_never | - float_infzeronan_suppress_invalid, s); - set_float_default_nan_pattern(0b11000000, s); -} - -#ifdef CONFIG_TCG - -/* Convert host exception flags to vfp form. */ -static inline uint32_t vfp_exceptbits_from_host(int host_bits, bool ah) -{ - uint32_t target_bits = 0; - - if (host_bits & float_flag_invalid) { - target_bits |= FPSR_IOC; - } - if (host_bits & float_flag_divbyzero) { - target_bits |= FPSR_DZC; - } - if (host_bits & float_flag_overflow) { - target_bits |= FPSR_OFC; - } - if (host_bits & (float_flag_underflow | float_flag_output_denormal_flushed)) { - target_bits |= FPSR_UFC; - } - if (host_bits & float_flag_inexact) { - target_bits |= FPSR_IXC; - } - if (host_bits & float_flag_input_denormal_flushed) { - target_bits |= FPSR_IDC; - } - /* - * With FPCR.AH, IDC is set when an input denormal is used, - * and flushing an output denormal to zero sets both IXC and UFC. - */ - if (ah && (host_bits & float_flag_input_denormal_used)) { - target_bits |= FPSR_IDC; - } - if (ah && (host_bits & float_flag_output_denormal_flushed)) { - target_bits |= FPSR_IXC; - } - return target_bits; -} - -static uint32_t vfp_get_fpsr_from_host(CPUARMState *env) -{ - uint32_t a32_flags = 0, a64_flags = 0; - - a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A32]); - a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_STD]); - /* FZ16 does not generate an input denormal exception. */ - a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A32_F16]) - & ~float_flag_input_denormal_flushed); - a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_STD_F16]) - & ~float_flag_input_denormal_flushed); - - a64_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A64]); - a64_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A64_F16]) - & ~(float_flag_input_denormal_flushed | float_flag_input_denormal_used)); - /* - * We do not merge in flags from FPST_AH or FPST_AH_F16, because - * they are used for insns that must not set the cumulative exception bits. - */ - - /* - * Flushing an input denormal *only* because FPCR.FIZ == 1 does - * not set FPSR.IDC; if FPCR.FZ is also set then this takes - * precedence and IDC is set (see the FPUnpackBase pseudocode). - * So squash it unless (FPCR.AH == 0 && FPCR.FZ == 1). - * We only do this for the a64 flags because FIZ has no effect - * on AArch32 even if it is set. - */ - if ((env->vfp.fpcr & (FPCR_FZ | FPCR_AH)) != FPCR_FZ) { - a64_flags &= ~float_flag_input_denormal_flushed; - } - return vfp_exceptbits_from_host(a64_flags, env->vfp.fpcr & FPCR_AH) | - vfp_exceptbits_from_host(a32_flags, false); -} - -static void vfp_clear_float_status_exc_flags(CPUARMState *env) -{ - /* - * Clear out all the exception-flag information in the float_status - * values. The caller should have arranged for env->vfp.fpsr to - * be the architecturally up-to-date exception flag information first. - */ - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH_F16]); -} - -static void vfp_sync_and_clear_float_status_exc_flags(CPUARMState *env) -{ - /* - * Synchronize any pending exception-flag information in the - * float_status values into env->vfp.fpsr, and then clear out - * the float_status data. - */ - env->vfp.fpsr |= vfp_get_fpsr_from_host(env); - vfp_clear_float_status_exc_flags(env); -} - -static void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) -{ - uint64_t changed = env->vfp.fpcr; - - changed ^= val; - changed &= mask; - if (changed & (3 << 22)) { - int i = (val >> 22) & 3; - switch (i) { - case FPROUNDING_TIEEVEN: - i = float_round_nearest_even; - break; - case FPROUNDING_POSINF: - i = float_round_up; - break; - case FPROUNDING_NEGINF: - i = float_round_down; - break; - case FPROUNDING_ZERO: - i = float_round_to_zero; - break; - } - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32_F16]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64_F16]); - } - if (changed & FPCR_FZ16) { - bool ftz_enabled = val & FPCR_FZ16; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); - } - if (changed & FPCR_FZ) { - bool ftz_enabled = val & FPCR_FZ; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64]); - /* FIZ is A64 only so FZ always makes A32 code flush inputs to zero */ - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); - } - if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { - /* - * A64: Flush denormalized inputs to zero if FPCR.FIZ = 1, or - * both FPCR.AH = 0 and FPCR.FZ = 1. - */ - bool fitz_enabled = (val & FPCR_FIZ) || - (val & (FPCR_FZ | FPCR_AH)) == FPCR_FZ; - set_flush_inputs_to_zero(fitz_enabled, &env->vfp.fp_status[FPST_A64]); - } - if (changed & FPCR_DN) { - bool dnan_enabled = val & FPCR_DN; - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH_F16]); - } - if (changed & FPCR_AH) { - bool ah_enabled = val & FPCR_AH; - - if (ah_enabled) { - /* Change behaviours for A64 FP operations */ - arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64]); - arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); - } else { - arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64]); - arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); - } - } - /* - * If any bits changed that we look at in vfp_get_fpsr_from_host(), - * we must sync the float_status flags into vfp.fpsr now (under the - * old regime) before we update vfp.fpcr. - */ - if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { - vfp_sync_and_clear_float_status_exc_flags(env); - } -} - -#else - -static uint32_t vfp_get_fpsr_from_host(CPUARMState *env) -{ - return 0; -} - -static void vfp_clear_float_status_exc_flags(CPUARMState *env) -{ -} - -static void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) -{ -} - -#endif - -uint32_t vfp_get_fpcr(CPUARMState *env) -{ - uint32_t fpcr = env->vfp.fpcr - | (env->vfp.vec_len << 16) - | (env->vfp.vec_stride << 20); - - /* - * M-profile LTPSIZE is the same bits [18:16] as A-profile Len; whichever - * of the two is not applicable to this CPU will always be zero. - */ - fpcr |= env->v7m.ltpsize << 16; - - return fpcr; -} - -uint32_t vfp_get_fpsr(CPUARMState *env) -{ - uint32_t fpsr = env->vfp.fpsr; - uint32_t i; - - fpsr |= vfp_get_fpsr_from_host(env); - - i = env->vfp.qc[0] | env->vfp.qc[1] | env->vfp.qc[2] | env->vfp.qc[3]; - fpsr |= i ? FPSR_QC : 0; - return fpsr; -} - -uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) -{ - return (vfp_get_fpcr(env) & FPSCR_FPCR_MASK) | - (vfp_get_fpsr(env) & FPSCR_FPSR_MASK); -} - -uint32_t vfp_get_fpscr(CPUARMState *env) -{ - return HELPER(vfp_get_fpscr)(env); -} - -void vfp_set_fpsr(CPUARMState *env, uint32_t val) -{ - ARMCPU *cpu = env_archcpu(env); - - if (arm_feature(env, ARM_FEATURE_NEON) || - cpu_isar_feature(aa32_mve, cpu)) { - /* - * The bit we set within vfp.qc[] is arbitrary; the array as a - * whole being zero/non-zero is what counts. - */ - env->vfp.qc[0] = val & FPSR_QC; - env->vfp.qc[1] = 0; - env->vfp.qc[2] = 0; - env->vfp.qc[3] = 0; - } - - /* - * NZCV lives only in env->vfp.fpsr. The cumulative exception flags - * IOC|DZC|OFC|UFC|IXC|IDC also live in env->vfp.fpsr, with possible - * extra pending exception information that hasn't yet been folded in - * living in the float_status values (for TCG). - * Since this FPSR write gives us the up to date values of the exception - * flags, we want to store into vfp.fpsr the NZCV and CEXC bits, zeroing - * anything else. We also need to clear out the float_status exception - * information so that the next vfp_get_fpsr does not fold in stale data. - */ - val &= FPSR_NZCV_MASK | FPSR_CEXC_MASK; - env->vfp.fpsr = val; - vfp_clear_float_status_exc_flags(env); -} - -static void vfp_set_fpcr_masked(CPUARMState *env, uint32_t val, uint32_t mask) -{ - /* - * We only set FPCR bits defined by mask, and leave the others alone. - * We assume the mask is sensible (e.g. doesn't try to set only - * part of a field) - */ - ARMCPU *cpu = env_archcpu(env); - - /* When ARMv8.2-FP16 is not supported, FZ16 is RES0. */ - if (!cpu_isar_feature(any_fp16, cpu)) { - val &= ~FPCR_FZ16; - } - if (!cpu_isar_feature(aa64_afp, cpu)) { - val &= ~(FPCR_FIZ | FPCR_AH | FPCR_NEP); - } - - if (!cpu_isar_feature(aa64_ebf16, cpu)) { - val &= ~FPCR_EBF; - } - - vfp_set_fpcr_to_host(env, val, mask); - - if (mask & (FPCR_LEN_MASK | FPCR_STRIDE_MASK)) { - if (!arm_feature(env, ARM_FEATURE_M)) { - /* - * Short-vector length and stride; on M-profile these bits - * are used for different purposes. - * We can't make this conditional be "if MVFR0.FPShVec != 0", - * because in v7A no-short-vector-support cores still had to - * allow Stride/Len to be written with the only effect that - * some insns are required to UNDEF if the guest sets them. - */ - env->vfp.vec_len = extract32(val, 16, 3); - env->vfp.vec_stride = extract32(val, 20, 2); - } else if (cpu_isar_feature(aa32_mve, cpu)) { - env->v7m.ltpsize = extract32(val, FPCR_LTPSIZE_SHIFT, - FPCR_LTPSIZE_LENGTH); - } - } - - /* - * We don't implement trapped exception handling, so the - * trap enable bits, IDE|IXE|UFE|OFE|DZE|IOE are all RAZ/WI (not RES0!) - * - * The FPCR bits we keep in vfp.fpcr are AHP, DN, FZ, RMode, EBF, FZ16, - * FIZ, AH, and NEP. - * Len, Stride and LTPSIZE we just handled. Store those bits - * there, and zero any of the other FPCR bits and the RES0 and RAZ/WI - * bits. - */ - val &= FPCR_AHP | FPCR_DN | FPCR_FZ | FPCR_RMODE_MASK | FPCR_FZ16 | - FPCR_EBF | FPCR_FIZ | FPCR_AH | FPCR_NEP; - env->vfp.fpcr &= ~mask; - env->vfp.fpcr |= val; -} - -void vfp_set_fpcr(CPUARMState *env, uint32_t val) -{ - vfp_set_fpcr_masked(env, val, MAKE_64BIT_MASK(0, 32)); -} - -void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) -{ - vfp_set_fpcr_masked(env, val, FPSCR_FPCR_MASK); - vfp_set_fpsr(env, val & FPSCR_FPSR_MASK); -} - -void vfp_set_fpscr(CPUARMState *env, uint32_t val) -{ - HELPER(vfp_set_fpscr)(env, val); -} - -#ifdef CONFIG_TCG #define VFP_HELPER(name, p) HELPER(glue(glue(vfp_,name),p)) @@ -XXX,XX +XXX,XX @@ void HELPER(check_hcr_el2_trap)(CPUARMState *env, uint32_t rt, uint32_t reg) raise_exception(env, EXCP_HYP_TRAP, syndrome, 2); } - -#endif diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "cpu-features.h" #include "fpu/softfloat.h" -#ifdef CONFIG_TCG -#include "qemu/log.h" -#endif - -/* VFP support. We follow the convention used for VFP instructions: - Single precision routines have a "s" suffix, double precision a - "d" suffix. */ /* * Set the float_status behaviour to match the Arm defaults: @@ -XXX,XX +XXX,XX @@ void vfp_set_fpscr(CPUARMState *env, uint32_t val) { HELPER(vfp_set_fpscr)(env, val); } - -#ifdef CONFIG_TCG - -#define VFP_HELPER(name, p) HELPER(glue(glue(vfp_,name),p)) - -#define VFP_BINOP(name) \ -dh_ctype_f16 VFP_HELPER(name, h)(dh_ctype_f16 a, dh_ctype_f16 b, float_status *fpst) \ -{ \ - return float16_ ## name(a, b, fpst); \ -} \ -float32 VFP_HELPER(name, s)(float32 a, float32 b, float_status *fpst) \ -{ \ - return float32_ ## name(a, b, fpst); \ -} \ -float64 VFP_HELPER(name, d)(float64 a, float64 b, float_status *fpst) \ -{ \ - return float64_ ## name(a, b, fpst); \ -} -VFP_BINOP(add) -VFP_BINOP(sub) -VFP_BINOP(mul) -VFP_BINOP(div) -VFP_BINOP(min) -VFP_BINOP(max) -VFP_BINOP(minnum) -VFP_BINOP(maxnum) -#undef VFP_BINOP - -dh_ctype_f16 VFP_HELPER(sqrt, h)(dh_ctype_f16 a, float_status *fpst) -{ - return float16_sqrt(a, fpst); -} - -float32 VFP_HELPER(sqrt, s)(float32 a, float_status *fpst) -{ - return float32_sqrt(a, fpst); -} - -float64 VFP_HELPER(sqrt, d)(float64 a, float_status *fpst) -{ - return float64_sqrt(a, fpst); -} - -static void softfloat_to_vfp_compare(CPUARMState *env, FloatRelation cmp) -{ - uint32_t flags; - switch (cmp) { - case float_relation_equal: - flags = 0x6; - break; - case float_relation_less: - flags = 0x8; - break; - case float_relation_greater: - flags = 0x2; - break; - case float_relation_unordered: - flags = 0x3; - break; - default: - g_assert_not_reached(); - } - env->vfp.fpsr = deposit64(env->vfp.fpsr, 28, 4, flags); /* NZCV */ -} - -/* XXX: check quiet/signaling case */ -#define DO_VFP_cmp(P, FLOATTYPE, ARGTYPE, FPST) \ -void VFP_HELPER(cmp, P)(ARGTYPE a, ARGTYPE b, CPUARMState *env) \ -{ \ - softfloat_to_vfp_compare(env, \ - FLOATTYPE ## _compare_quiet(a, b, &env->vfp.fp_status[FPST])); \ -} \ -void VFP_HELPER(cmpe, P)(ARGTYPE a, ARGTYPE b, CPUARMState *env) \ -{ \ - softfloat_to_vfp_compare(env, \ - FLOATTYPE ## _compare(a, b, &env->vfp.fp_status[FPST])); \ -} -DO_VFP_cmp(h, float16, dh_ctype_f16, FPST_A32_F16) -DO_VFP_cmp(s, float32, float32, FPST_A32) -DO_VFP_cmp(d, float64, float64, FPST_A32) -#undef DO_VFP_cmp - -/* Integer to float and float to integer conversions */ - -#define CONV_ITOF(name, ftype, fsz, sign) \ -ftype HELPER(name)(uint32_t x, float_status *fpst) \ -{ \ - return sign##int32_to_##float##fsz((sign##int32_t)x, fpst); \ -} - -#define CONV_FTOI(name, ftype, fsz, sign, round) \ -sign##int32_t HELPER(name)(ftype x, float_status *fpst) \ -{ \ - if (float##fsz##_is_any_nan(x)) { \ - float_raise(float_flag_invalid, fpst); \ - return 0; \ - } \ - return float##fsz##_to_##sign##int32##round(x, fpst); \ -} - -#define FLOAT_CONVS(name, p, ftype, fsz, sign) \ - CONV_ITOF(vfp_##name##to##p, ftype, fsz, sign) \ - CONV_FTOI(vfp_to##name##p, ftype, fsz, sign, ) \ - CONV_FTOI(vfp_to##name##z##p, ftype, fsz, sign, _round_to_zero) - -FLOAT_CONVS(si, h, uint32_t, 16, ) -FLOAT_CONVS(si, s, float32, 32, ) -FLOAT_CONVS(si, d, float64, 64, ) -FLOAT_CONVS(ui, h, uint32_t, 16, u) -FLOAT_CONVS(ui, s, float32, 32, u) -FLOAT_CONVS(ui, d, float64, 64, u) - -#undef CONV_ITOF -#undef CONV_FTOI -#undef FLOAT_CONVS - -/* floating point conversion */ -float64 VFP_HELPER(fcvtd, s)(float32 x, float_status *status) -{ - return float32_to_float64(x, status); -} - -float32 VFP_HELPER(fcvts, d)(float64 x, float_status *status) -{ - return float64_to_float32(x, status); -} - -uint32_t HELPER(bfcvt)(float32 x, float_status *status) -{ - return float32_to_bfloat16(x, status); -} - -uint32_t HELPER(bfcvt_pair)(uint64_t pair, float_status *status) -{ - bfloat16 lo = float32_to_bfloat16(extract64(pair, 0, 32), status); - bfloat16 hi = float32_to_bfloat16(extract64(pair, 32, 32), status); - return deposit32(lo, 16, 16, hi); -} - -/* - * VFP3 fixed point conversion. The AArch32 versions of fix-to-float - * must always round-to-nearest; the AArch64 ones honour the FPSCR - * rounding mode. (For AArch32 Neon the standard-FPSCR is set to - * round-to-nearest so either helper will work.) AArch32 float-to-fix - * must round-to-zero. - */ -#define VFP_CONV_FIX_FLOAT(name, p, fsz, ftype, isz, itype) \ -ftype HELPER(vfp_##name##to##p)(uint##isz##_t x, uint32_t shift, \ - float_status *fpst) \ -{ return itype##_to_##float##fsz##_scalbn(x, -shift, fpst); } - -#define VFP_CONV_FIX_FLOAT_ROUND(name, p, fsz, ftype, isz, itype) \ - ftype HELPER(vfp_##name##to##p##_round_to_nearest)(uint##isz##_t x, \ - uint32_t shift, \ - float_status *fpst) \ - { \ - ftype ret; \ - FloatRoundMode oldmode = fpst->float_rounding_mode; \ - fpst->float_rounding_mode = float_round_nearest_even; \ - ret = itype##_to_##float##fsz##_scalbn(x, -shift, fpst); \ - fpst->float_rounding_mode = oldmode; \ - return ret; \ - } - -#define VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, ftype, isz, itype, ROUND, suff) \ -uint##isz##_t HELPER(vfp_to##name##p##suff)(ftype x, uint32_t shift, \ - float_status *fpst) \ -{ \ - if (unlikely(float##fsz##_is_any_nan(x))) { \ - float_raise(float_flag_invalid, fpst); \ - return 0; \ - } \ - return float##fsz##_to_##itype##_scalbn(x, ROUND, shift, fpst); \ -} - -#define VFP_CONV_FIX(name, p, fsz, ftype, isz, itype) \ -VFP_CONV_FIX_FLOAT(name, p, fsz, ftype, isz, itype) \ -VFP_CONV_FIX_FLOAT_ROUND(name, p, fsz, ftype, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, ftype, isz, itype, \ - float_round_to_zero, _round_to_zero) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, ftype, isz, itype, \ - get_float_rounding_mode(fpst), ) - -#define VFP_CONV_FIX_A64(name, p, fsz, ftype, isz, itype) \ -VFP_CONV_FIX_FLOAT(name, p, fsz, ftype, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, ftype, isz, itype, \ - get_float_rounding_mode(fpst), ) - -VFP_CONV_FIX(sh, d, 64, float64, 64, int16) -VFP_CONV_FIX(sl, d, 64, float64, 64, int32) -VFP_CONV_FIX_A64(sq, d, 64, float64, 64, int64) -VFP_CONV_FIX(uh, d, 64, float64, 64, uint16) -VFP_CONV_FIX(ul, d, 64, float64, 64, uint32) -VFP_CONV_FIX_A64(uq, d, 64, float64, 64, uint64) -VFP_CONV_FIX(sh, s, 32, float32, 32, int16) -VFP_CONV_FIX(sl, s, 32, float32, 32, int32) -VFP_CONV_FIX_A64(sq, s, 32, float32, 64, int64) -VFP_CONV_FIX(uh, s, 32, float32, 32, uint16) -VFP_CONV_FIX(ul, s, 32, float32, 32, uint32) -VFP_CONV_FIX_A64(uq, s, 32, float32, 64, uint64) -VFP_CONV_FIX(sh, h, 16, dh_ctype_f16, 32, int16) -VFP_CONV_FIX(sl, h, 16, dh_ctype_f16, 32, int32) -VFP_CONV_FIX_A64(sq, h, 16, dh_ctype_f16, 64, int64) -VFP_CONV_FIX(uh, h, 16, dh_ctype_f16, 32, uint16) -VFP_CONV_FIX(ul, h, 16, dh_ctype_f16, 32, uint32) -VFP_CONV_FIX_A64(uq, h, 16, dh_ctype_f16, 64, uint64) -VFP_CONV_FLOAT_FIX_ROUND(sq, d, 64, float64, 64, int64, - float_round_to_zero, _round_to_zero) -VFP_CONV_FLOAT_FIX_ROUND(uq, d, 64, float64, 64, uint64, - float_round_to_zero, _round_to_zero) - -#undef VFP_CONV_FIX -#undef VFP_CONV_FIX_FLOAT -#undef VFP_CONV_FLOAT_FIX_ROUND -#undef VFP_CONV_FIX_A64 - -/* Set the current fp rounding mode and return the old one. - * The argument is a softfloat float_round_ value. - */ -uint32_t HELPER(set_rmode)(uint32_t rmode, float_status *fp_status) -{ - uint32_t prev_rmode = get_float_rounding_mode(fp_status); - set_float_rounding_mode(rmode, fp_status); - - return prev_rmode; -} - -/* Half precision conversions. */ -float32 HELPER(vfp_fcvt_f16_to_f32)(uint32_t a, float_status *fpst, - uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing input denormals. - */ - bool save = get_flush_inputs_to_zero(fpst); - set_flush_inputs_to_zero(false, fpst); - float32 r = float16_to_float32(a, !ahp_mode, fpst); - set_flush_inputs_to_zero(save, fpst); - return r; -} - -uint32_t HELPER(vfp_fcvt_f32_to_f16)(float32 a, float_status *fpst, - uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing output denormals. - */ - bool save = get_flush_to_zero(fpst); - set_flush_to_zero(false, fpst); - float16 r = float32_to_float16(a, !ahp_mode, fpst); - set_flush_to_zero(save, fpst); - return r; -} - -float64 HELPER(vfp_fcvt_f16_to_f64)(uint32_t a, float_status *fpst, - uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing input denormals. - */ - bool save = get_flush_inputs_to_zero(fpst); - set_flush_inputs_to_zero(false, fpst); - float64 r = float16_to_float64(a, !ahp_mode, fpst); - set_flush_inputs_to_zero(save, fpst); - return r; -} - -uint32_t HELPER(vfp_fcvt_f64_to_f16)(float64 a, float_status *fpst, - uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing output denormals. - */ - bool save = get_flush_to_zero(fpst); - set_flush_to_zero(false, fpst); - float16 r = float64_to_float16(a, !ahp_mode, fpst); - set_flush_to_zero(save, fpst); - return r; -} - -/* NEON helpers. */ - -/* Constants 256 and 512 are used in some helpers; we avoid relying on - * int->float conversions at run-time. */ -#define float64_256 make_float64(0x4070000000000000LL) -#define float64_512 make_float64(0x4080000000000000LL) -#define float16_maxnorm make_float16(0x7bff) -#define float32_maxnorm make_float32(0x7f7fffff) -#define float64_maxnorm make_float64(0x7fefffffffffffffLL) - -/* Reciprocal functions - * - * The algorithm that must be used to calculate the estimate - * is specified by the ARM ARM, see FPRecipEstimate()/RecipEstimate - */ - -/* See RecipEstimate() - * - * input is a 9 bit fixed point number - * input range 256 .. 511 for a number from 0.5 <= x < 1.0. - * result range 256 .. 511 for a number from 1.0 to 511/256. - */ - -static int recip_estimate(int input) -{ - int a, b, r; - assert(256 <= input && input < 512); - a = (input * 2) + 1; - b = (1 << 19) / a; - r = (b + 1) >> 1; - assert(256 <= r && r < 512); - return r; -} - -/* - * Increased precision version: - * input is a 13 bit fixed point number - * input range 2048 .. 4095 for a number from 0.5 <= x < 1.0. - * result range 4096 .. 8191 for a number from 1.0 to 2.0 - */ -static int recip_estimate_incprec(int input) -{ - int a, b, r; - assert(2048 <= input && input < 4096); - a = (input * 2) + 1; - /* - * The pseudocode expresses this as an operation on infinite - * precision reals where it calculates 2^25 / a and then looks - * at the error between that and the rounded-down-to-integer - * value to see if it should instead round up. We instead - * follow the same approach as the pseudocode for the 8-bit - * precision version, and calculate (2 * (2^25 / a)) as an - * integer so we can do the "add one and halve" to round it. - * So the 1 << 26 here is correct. - */ - b = (1 << 26) / a; - r = (b + 1) >> 1; - assert(4096 <= r && r < 8192); - return r; -} - -/* - * Common wrapper to call recip_estimate - * - * The parameters are exponent and 64 bit fraction (without implicit - * bit) where the binary point is nominally at bit 52. Returns a - * float64 which can then be rounded to the appropriate size by the - * callee. - */ - -static uint64_t call_recip_estimate(int *exp, int exp_off, uint64_t frac, - bool increasedprecision) -{ - uint32_t scaled, estimate; - uint64_t result_frac; - int result_exp; - - /* Handle sub-normals */ - if (*exp == 0) { - if (extract64(frac, 51, 1) == 0) { - *exp = -1; - frac <<= 2; - } else { - frac <<= 1; - } - } - - if (increasedprecision) { - /* scaled = UInt('1':fraction<51:41>) */ - scaled = deposit32(1 << 11, 0, 11, extract64(frac, 41, 11)); - estimate = recip_estimate_incprec(scaled); - } else { - /* scaled = UInt('1':fraction<51:44>) */ - scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); - estimate = recip_estimate(scaled); - } - - result_exp = exp_off - *exp; - if (increasedprecision) { - result_frac = deposit64(0, 40, 12, estimate); - } else { - result_frac = deposit64(0, 44, 8, estimate); - } - if (result_exp == 0) { - result_frac = deposit64(result_frac >> 1, 51, 1, 1); - } else if (result_exp == -1) { - result_frac = deposit64(result_frac >> 2, 50, 2, 1); - result_exp = 0; - } - - *exp = result_exp; - - return result_frac; -} - -static bool round_to_inf(float_status *fpst, bool sign_bit) -{ - switch (fpst->float_rounding_mode) { - case float_round_nearest_even: /* Round to Nearest */ - return true; - case float_round_up: /* Round to +Inf */ - return !sign_bit; - case float_round_down: /* Round to -Inf */ - return sign_bit; - case float_round_to_zero: /* Round to Zero */ - return false; - default: - g_assert_not_reached(); - } -} - -uint32_t HELPER(recpe_f16)(uint32_t input, float_status *fpst) -{ - float16 f16 = float16_squash_input_denormal(input, fpst); - uint32_t f16_val = float16_val(f16); - uint32_t f16_sign = float16_is_neg(f16); - int f16_exp = extract32(f16_val, 10, 5); - uint32_t f16_frac = extract32(f16_val, 0, 10); - uint64_t f64_frac; - - if (float16_is_any_nan(f16)) { - float16 nan = f16; - if (float16_is_signaling_nan(f16, fpst)) { - float_raise(float_flag_invalid, fpst); - if (!fpst->default_nan_mode) { - nan = float16_silence_nan(f16, fpst); - } - } - if (fpst->default_nan_mode) { - nan = float16_default_nan(fpst); - } - return nan; - } else if (float16_is_infinity(f16)) { - return float16_set_sign(float16_zero, float16_is_neg(f16)); - } else if (float16_is_zero(f16)) { - float_raise(float_flag_divbyzero, fpst); - return float16_set_sign(float16_infinity, float16_is_neg(f16)); - } else if (float16_abs(f16) < (1 << 8)) { - /* Abs(value) < 2.0^-16 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f16_sign)) { - return float16_set_sign(float16_infinity, f16_sign); - } else { - return float16_set_sign(float16_maxnorm, f16_sign); - } - } else if (f16_exp >= 29 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float16_set_sign(float16_zero, float16_is_neg(f16)); - } - - f64_frac = call_recip_estimate(&f16_exp, 29, - ((uint64_t) f16_frac) << (52 - 10), false); - - /* result = sign : result_exp<4:0> : fraction<51:42> */ - f16_val = deposit32(0, 15, 1, f16_sign); - f16_val = deposit32(f16_val, 10, 5, f16_exp); - f16_val = deposit32(f16_val, 0, 10, extract64(f64_frac, 52 - 10, 10)); - return make_float16(f16_val); -} - -/* - * FEAT_RPRES means the f32 FRECPE has an "increased precision" variant - * which is used when FPCR.AH == 1. - */ -static float32 do_recpe_f32(float32 input, float_status *fpst, bool rpres) -{ - float32 f32 = float32_squash_input_denormal(input, fpst); - uint32_t f32_val = float32_val(f32); - bool f32_sign = float32_is_neg(f32); - int f32_exp = extract32(f32_val, 23, 8); - uint32_t f32_frac = extract32(f32_val, 0, 23); - uint64_t f64_frac; - - if (float32_is_any_nan(f32)) { - float32 nan = f32; - if (float32_is_signaling_nan(f32, fpst)) { - float_raise(float_flag_invalid, fpst); - if (!fpst->default_nan_mode) { - nan = float32_silence_nan(f32, fpst); - } - } - if (fpst->default_nan_mode) { - nan = float32_default_nan(fpst); - } - return nan; - } else if (float32_is_infinity(f32)) { - return float32_set_sign(float32_zero, float32_is_neg(f32)); - } else if (float32_is_zero(f32)) { - float_raise(float_flag_divbyzero, fpst); - return float32_set_sign(float32_infinity, float32_is_neg(f32)); - } else if (float32_abs(f32) < (1ULL << 21)) { - /* Abs(value) < 2.0^-128 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f32_sign)) { - return float32_set_sign(float32_infinity, f32_sign); - } else { - return float32_set_sign(float32_maxnorm, f32_sign); - } - } else if (f32_exp >= 253 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float32_set_sign(float32_zero, float32_is_neg(f32)); - } - - f64_frac = call_recip_estimate(&f32_exp, 253, - ((uint64_t) f32_frac) << (52 - 23), rpres); - - /* result = sign : result_exp<7:0> : fraction<51:29> */ - f32_val = deposit32(0, 31, 1, f32_sign); - f32_val = deposit32(f32_val, 23, 8, f32_exp); - f32_val = deposit32(f32_val, 0, 23, extract64(f64_frac, 52 - 23, 23)); - return make_float32(f32_val); -} - -float32 HELPER(recpe_f32)(float32 input, float_status *fpst) -{ - return do_recpe_f32(input, fpst, false); -} - -float32 HELPER(recpe_rpres_f32)(float32 input, float_status *fpst) -{ - return do_recpe_f32(input, fpst, true); -} - -float64 HELPER(recpe_f64)(float64 input, float_status *fpst) -{ - float64 f64 = float64_squash_input_denormal(input, fpst); - uint64_t f64_val = float64_val(f64); - bool f64_sign = float64_is_neg(f64); - int f64_exp = extract64(f64_val, 52, 11); - uint64_t f64_frac = extract64(f64_val, 0, 52); - - /* Deal with any special cases */ - if (float64_is_any_nan(f64)) { - float64 nan = f64; - if (float64_is_signaling_nan(f64, fpst)) { - float_raise(float_flag_invalid, fpst); - if (!fpst->default_nan_mode) { - nan = float64_silence_nan(f64, fpst); - } - } - if (fpst->default_nan_mode) { - nan = float64_default_nan(fpst); - } - return nan; - } else if (float64_is_infinity(f64)) { - return float64_set_sign(float64_zero, float64_is_neg(f64)); - } else if (float64_is_zero(f64)) { - float_raise(float_flag_divbyzero, fpst); - return float64_set_sign(float64_infinity, float64_is_neg(f64)); - } else if ((f64_val & ~(1ULL << 63)) < (1ULL << 50)) { - /* Abs(value) < 2.0^-1024 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f64_sign)) { - return float64_set_sign(float64_infinity, f64_sign); - } else { - return float64_set_sign(float64_maxnorm, f64_sign); - } - } else if (f64_exp >= 2045 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float64_set_sign(float64_zero, float64_is_neg(f64)); - } - - f64_frac = call_recip_estimate(&f64_exp, 2045, f64_frac, false); - - /* result = sign : result_exp<10:0> : fraction<51:0>; */ - f64_val = deposit64(0, 63, 1, f64_sign); - f64_val = deposit64(f64_val, 52, 11, f64_exp); - f64_val = deposit64(f64_val, 0, 52, f64_frac); - return make_float64(f64_val); -} - -/* The algorithm that must be used to calculate the estimate - * is specified by the ARM ARM. - */ - -static int do_recip_sqrt_estimate(int a) -{ - int b, estimate; - - assert(128 <= a && a < 512); - if (a < 256) { - a = a * 2 + 1; - } else { - a = (a >> 1) << 1; - a = (a + 1) * 2; - } - b = 512; - while (a * (b + 1) * (b + 1) < (1 << 28)) { - b += 1; - } - estimate = (b + 1) / 2; - assert(256 <= estimate && estimate < 512); - - return estimate; -} - -static int do_recip_sqrt_estimate_incprec(int a) -{ - /* - * The Arm ARM describes the 12-bit precision version of RecipSqrtEstimate - * in terms of an infinite-precision floating point calculation of a - * square root. We implement this using the same kind of pure integer - * algorithm as the 8-bit mantissa, to get the same bit-for-bit result. - */ - int64_t b, estimate; - - assert(1024 <= a && a < 4096); - if (a < 2048) { - a = a * 2 + 1; - } else { - a = (a >> 1) << 1; - a = (a + 1) * 2; - } - b = 8192; - while (a * (b + 1) * (b + 1) < (1ULL << 39)) { - b += 1; - } - estimate = (b + 1) / 2; - - assert(4096 <= estimate && estimate < 8192); - - return estimate; -} - -static uint64_t recip_sqrt_estimate(int *exp , int exp_off, uint64_t frac, - bool increasedprecision) -{ - int estimate; - uint32_t scaled; - - if (*exp == 0) { - while (extract64(frac, 51, 1) == 0) { - frac = frac << 1; - *exp -= 1; - } - frac = extract64(frac, 0, 51) << 1; - } - - if (increasedprecision) { - if (*exp & 1) { - /* scaled = UInt('01':fraction<51:42>) */ - scaled = deposit32(1 << 10, 0, 10, extract64(frac, 42, 10)); - } else { - /* scaled = UInt('1':fraction<51:41>) */ - scaled = deposit32(1 << 11, 0, 11, extract64(frac, 41, 11)); - } - estimate = do_recip_sqrt_estimate_incprec(scaled); - } else { - if (*exp & 1) { - /* scaled = UInt('01':fraction<51:45>) */ - scaled = deposit32(1 << 7, 0, 7, extract64(frac, 45, 7)); - } else { - /* scaled = UInt('1':fraction<51:44>) */ - scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); - } - estimate = do_recip_sqrt_estimate(scaled); - } - - *exp = (exp_off - *exp) / 2; - if (increasedprecision) { - return extract64(estimate, 0, 12) << 40; - } else { - return extract64(estimate, 0, 8) << 44; - } -} - -uint32_t HELPER(rsqrte_f16)(uint32_t input, float_status *s) -{ - float16 f16 = float16_squash_input_denormal(input, s); - uint16_t val = float16_val(f16); - bool f16_sign = float16_is_neg(f16); - int f16_exp = extract32(val, 10, 5); - uint16_t f16_frac = extract32(val, 0, 10); - uint64_t f64_frac; - - if (float16_is_any_nan(f16)) { - float16 nan = f16; - if (float16_is_signaling_nan(f16, s)) { - float_raise(float_flag_invalid, s); - if (!s->default_nan_mode) { - nan = float16_silence_nan(f16, s); - } - } - if (s->default_nan_mode) { - nan = float16_default_nan(s); - } - return nan; - } else if (float16_is_zero(f16)) { - float_raise(float_flag_divbyzero, s); - return float16_set_sign(float16_infinity, f16_sign); - } else if (f16_sign) { - float_raise(float_flag_invalid, s); - return float16_default_nan(s); - } else if (float16_is_infinity(f16)) { - return float16_zero; - } - - /* Scale and normalize to a double-precision value between 0.25 and 1.0, - * preserving the parity of the exponent. */ - - f64_frac = ((uint64_t) f16_frac) << (52 - 10); - - f64_frac = recip_sqrt_estimate(&f16_exp, 44, f64_frac, false); - - /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(2) */ - val = deposit32(0, 15, 1, f16_sign); - val = deposit32(val, 10, 5, f16_exp); - val = deposit32(val, 2, 8, extract64(f64_frac, 52 - 8, 8)); - return make_float16(val); -} - -/* - * FEAT_RPRES means the f32 FRSQRTE has an "increased precision" variant - * which is used when FPCR.AH == 1. - */ -static float32 do_rsqrte_f32(float32 input, float_status *s, bool rpres) -{ - float32 f32 = float32_squash_input_denormal(input, s); - uint32_t val = float32_val(f32); - uint32_t f32_sign = float32_is_neg(f32); - int f32_exp = extract32(val, 23, 8); - uint32_t f32_frac = extract32(val, 0, 23); - uint64_t f64_frac; - - if (float32_is_any_nan(f32)) { - float32 nan = f32; - if (float32_is_signaling_nan(f32, s)) { - float_raise(float_flag_invalid, s); - if (!s->default_nan_mode) { - nan = float32_silence_nan(f32, s); - } - } - if (s->default_nan_mode) { - nan = float32_default_nan(s); - } - return nan; - } else if (float32_is_zero(f32)) { - float_raise(float_flag_divbyzero, s); - return float32_set_sign(float32_infinity, float32_is_neg(f32)); - } else if (float32_is_neg(f32)) { - float_raise(float_flag_invalid, s); - return float32_default_nan(s); - } else if (float32_is_infinity(f32)) { - return float32_zero; - } - - /* Scale and normalize to a double-precision value between 0.25 and 1.0, - * preserving the parity of the exponent. */ - - f64_frac = ((uint64_t) f32_frac) << 29; - - f64_frac = recip_sqrt_estimate(&f32_exp, 380, f64_frac, rpres); - - /* - * result = sign : result_exp<7:0> : estimate<7:0> : Zeros(15) - * or for increased precision - * result = sign : result_exp<7:0> : estimate<11:0> : Zeros(11) - */ - val = deposit32(0, 31, 1, f32_sign); - val = deposit32(val, 23, 8, f32_exp); - if (rpres) { - val = deposit32(val, 11, 12, extract64(f64_frac, 52 - 12, 12)); - } else { - val = deposit32(val, 15, 8, extract64(f64_frac, 52 - 8, 8)); - } - return make_float32(val); -} - -float32 HELPER(rsqrte_f32)(float32 input, float_status *s) -{ - return do_rsqrte_f32(input, s, false); -} - -float32 HELPER(rsqrte_rpres_f32)(float32 input, float_status *s) -{ - return do_rsqrte_f32(input, s, true); -} - -float64 HELPER(rsqrte_f64)(float64 input, float_status *s) -{ - float64 f64 = float64_squash_input_denormal(input, s); - uint64_t val = float64_val(f64); - bool f64_sign = float64_is_neg(f64); - int f64_exp = extract64(val, 52, 11); - uint64_t f64_frac = extract64(val, 0, 52); - - if (float64_is_any_nan(f64)) { - float64 nan = f64; - if (float64_is_signaling_nan(f64, s)) { - float_raise(float_flag_invalid, s); - if (!s->default_nan_mode) { - nan = float64_silence_nan(f64, s); - } - } - if (s->default_nan_mode) { - nan = float64_default_nan(s); - } - return nan; - } else if (float64_is_zero(f64)) { - float_raise(float_flag_divbyzero, s); - return float64_set_sign(float64_infinity, float64_is_neg(f64)); - } else if (float64_is_neg(f64)) { - float_raise(float_flag_invalid, s); - return float64_default_nan(s); - } else if (float64_is_infinity(f64)) { - return float64_zero; - } - - f64_frac = recip_sqrt_estimate(&f64_exp, 3068, f64_frac, false); - - /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(44) */ - val = deposit64(0, 61, 1, f64_sign); - val = deposit64(val, 52, 11, f64_exp); - val = deposit64(val, 44, 8, extract64(f64_frac, 52 - 8, 8)); - return make_float64(val); -} - -uint32_t HELPER(recpe_u32)(uint32_t a) -{ - int input, estimate; - - if ((a & 0x80000000) == 0) { - return 0xffffffff; - } - - input = extract32(a, 23, 9); - estimate = recip_estimate(input); - - return deposit32(0, (32 - 9), 9, estimate); -} - -uint32_t HELPER(rsqrte_u32)(uint32_t a) -{ - int estimate; - - if ((a & 0xc0000000) == 0) { - return 0xffffffff; - } - - estimate = do_recip_sqrt_estimate(extract32(a, 23, 9)); - - return deposit32(0, 23, 9, estimate); -} - -/* VFPv4 fused multiply-accumulate */ -dh_ctype_f16 VFP_HELPER(muladd, h)(dh_ctype_f16 a, dh_ctype_f16 b, - dh_ctype_f16 c, float_status *fpst) -{ - return float16_muladd(a, b, c, 0, fpst); -} - -float32 VFP_HELPER(muladd, s)(float32 a, float32 b, float32 c, - float_status *fpst) -{ - return float32_muladd(a, b, c, 0, fpst); -} - -float64 VFP_HELPER(muladd, d)(float64 a, float64 b, float64 c, - float_status *fpst) -{ - return float64_muladd(a, b, c, 0, fpst); -} - -/* ARMv8 round to integral */ -dh_ctype_f16 HELPER(rinth_exact)(dh_ctype_f16 x, float_status *fp_status) -{ - return float16_round_to_int(x, fp_status); -} - -float32 HELPER(rints_exact)(float32 x, float_status *fp_status) -{ - return float32_round_to_int(x, fp_status); -} - -float64 HELPER(rintd_exact)(float64 x, float_status *fp_status) -{ - return float64_round_to_int(x, fp_status); -} - -dh_ctype_f16 HELPER(rinth)(dh_ctype_f16 x, float_status *fp_status) -{ - int old_flags = get_float_exception_flags(fp_status), new_flags; - float16 ret; - - ret = float16_round_to_int(x, fp_status); - - /* Suppress any inexact exceptions the conversion produced */ - if (!(old_flags & float_flag_inexact)) { - new_flags = get_float_exception_flags(fp_status); - set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); - } - - return ret; -} - -float32 HELPER(rints)(float32 x, float_status *fp_status) -{ - int old_flags = get_float_exception_flags(fp_status), new_flags; - float32 ret; - - ret = float32_round_to_int(x, fp_status); - - /* Suppress any inexact exceptions the conversion produced */ - if (!(old_flags & float_flag_inexact)) { - new_flags = get_float_exception_flags(fp_status); - set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); - } - - return ret; -} - -float64 HELPER(rintd)(float64 x, float_status *fp_status) -{ - int old_flags = get_float_exception_flags(fp_status), new_flags; - float64 ret; - - ret = float64_round_to_int(x, fp_status); - - /* Suppress any inexact exceptions the conversion produced */ - if (!(old_flags & float_flag_inexact)) { - new_flags = get_float_exception_flags(fp_status); - set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); - } - - return ret; -} - -/* Convert ARM rounding mode to softfloat */ -const FloatRoundMode arm_rmode_to_sf_map[] = { - [FPROUNDING_TIEEVEN] = float_round_nearest_even, - [FPROUNDING_POSINF] = float_round_up, - [FPROUNDING_NEGINF] = float_round_down, - [FPROUNDING_ZERO] = float_round_to_zero, - [FPROUNDING_TIEAWAY] = float_round_ties_away, - [FPROUNDING_ODD] = float_round_to_odd, -}; - -/* - * Implement float64 to int32_t conversion without saturation; - * the result is supplied modulo 2^32. - */ -uint64_t HELPER(fjcvtzs)(float64 value, float_status *status) -{ - uint32_t frac, e_old, e_new; - bool inexact; - - e_old = get_float_exception_flags(status); - set_float_exception_flags(0, status); - frac = float64_to_int32_modulo(value, float_round_to_zero, status); - e_new = get_float_exception_flags(status); - set_float_exception_flags(e_old | e_new, status); - - /* Normal inexact, denormal with flush-to-zero, or overflow or NaN */ - inexact = e_new & (float_flag_inexact | - float_flag_input_denormal_flushed | - float_flag_invalid); - - /* While not inexact for IEEE FP, -0.0 is inexact for JavaScript. */ - inexact |= value == float64_chs(float64_zero); - - /* Pack the result and the env->ZF representation of Z together. */ - return deposit64(frac, 32, 32, inexact); -} - -uint32_t HELPER(vjcvt)(float64 value, CPUARMState *env) -{ - uint64_t pair = HELPER(fjcvtzs)(value, &env->vfp.fp_status[FPST_A32]); - uint32_t result = pair; - uint32_t z = (pair >> 32) == 0; - - /* Store Z, clear NCV, in FPSCR.NZCV. */ - env->vfp.fpsr = (env->vfp.fpsr & ~FPSR_NZCV_MASK) | (z * FPSR_Z); - - return result; -} - -/* Round a float32 to an integer that fits in int32_t or int64_t. */ -static float32 frint_s(float32 f, float_status *fpst, int intsize) -{ - int old_flags = get_float_exception_flags(fpst); - uint32_t exp = extract32(f, 23, 8); - - if (unlikely(exp == 0xff)) { - /* NaN or Inf. */ - goto overflow; - } - - /* Round and re-extract the exponent. */ - f = float32_round_to_int(f, fpst); - exp = extract32(f, 23, 8); - - /* Validate the range of the result. */ - if (exp < 126 + intsize) { - /* abs(F) <= INT{N}_MAX */ - return f; - } - if (exp == 126 + intsize) { - uint32_t sign = extract32(f, 31, 1); - uint32_t frac = extract32(f, 0, 23); - if (sign && frac == 0) { - /* F == INT{N}_MIN */ - return f; - } - } - - overflow: - /* - * Raise Invalid and return INT{N}_MIN as a float. Revert any - * inexact exception float32_round_to_int may have raised. - */ - set_float_exception_flags(old_flags | float_flag_invalid, fpst); - return (0x100u + 126u + intsize) << 23; -} - -float32 HELPER(frint32_s)(float32 f, float_status *fpst) -{ - return frint_s(f, fpst, 32); -} - -float32 HELPER(frint64_s)(float32 f, float_status *fpst) -{ - return frint_s(f, fpst, 64); -} - -/* Round a float64 to an integer that fits in int32_t or int64_t. */ -static float64 frint_d(float64 f, float_status *fpst, int intsize) -{ - int old_flags = get_float_exception_flags(fpst); - uint32_t exp = extract64(f, 52, 11); - - if (unlikely(exp == 0x7ff)) { - /* NaN or Inf. */ - goto overflow; - } - - /* Round and re-extract the exponent. */ - f = float64_round_to_int(f, fpst); - exp = extract64(f, 52, 11); - - /* Validate the range of the result. */ - if (exp < 1022 + intsize) { - /* abs(F) <= INT{N}_MAX */ - return f; - } - if (exp == 1022 + intsize) { - uint64_t sign = extract64(f, 63, 1); - uint64_t frac = extract64(f, 0, 52); - if (sign && frac == 0) { - /* F == INT{N}_MIN */ - return f; - } - } - - overflow: - /* - * Raise Invalid and return INT{N}_MIN as a float. Revert any - * inexact exception float64_round_to_int may have raised. - */ - set_float_exception_flags(old_flags | float_flag_invalid, fpst); - return (uint64_t)(0x800 + 1022 + intsize) << 52; -} - -float64 HELPER(frint32_d)(float64 f, float_status *fpst) -{ - return frint_d(f, fpst, 32); -} - -float64 HELPER(frint64_d)(float64 f, float_status *fpst) -{ - return frint_d(f, fpst, 64); -} - -void HELPER(check_hcr_el2_trap)(CPUARMState *env, uint32_t rt, uint32_t reg) -{ - uint32_t syndrome; - - switch (reg) { - case ARM_VFP_MVFR0: - case ARM_VFP_MVFR1: - case ARM_VFP_MVFR2: - if (!(arm_hcr_el2_eff(env) & HCR_TID3)) { - return; - } - break; - case ARM_VFP_FPSID: - if (!(arm_hcr_el2_eff(env) & HCR_TID0)) { - return; - } - break; - default: - g_assert_not_reached(); - } - - syndrome = ((EC_FPIDTRAP << ARM_EL_EC_SHIFT) - | ARM_EL_IL - | (1 << 24) | (0xe << 20) | (7 << 14) - | (reg << 10) | (rt << 5) | 1); - - raise_exception(env, EXCP_HYP_TRAP, syndrome, 2); -} - -#endif diff --git a/target/arm/tcg/meson.build b/target/arm/tcg/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/meson.build +++ b/target/arm/tcg/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(files( 'vec_helper.c', 'tlb-insns.c', 'arith_helper.c', + 'vfp_helper.c', )) arm_ss.add(when: 'TARGET_AARCH64', if_true: files( -- 2.43.0 Currently the helper_vfp_get_fpscr() and helper_vfp_set_fpscr() functions do the actual work of updating the FPSCR, and we have wrappers vfp_get_fpscr() and vfp_set_fpscr() which we use for calls from other QEMU C code. Flip these around so that it is vfp_get_fpscr() and vfp_set_fpscr() which do the actual work, and helper_vfp_get_fpscr() and helper_vfp_set_fpscr() which are the wrappers; this allows us to move them to tcg/vfp_helper.c. Since this is the last HELPER() we had in arm/vfp_helper.c, we can drop the include of helper-proto.h. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250221190957.811948-3-peter.maydell@linaro.org --- target/arm/tcg/vfp_helper.c | 10 ++++++++++ target/arm/vfp_helper.c | 15 ++------------- 2 files changed, 12 insertions(+), 13 deletions(-) diff --git a/target/arm/tcg/vfp_helper.c b/target/arm/tcg/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/vfp_helper.c +++ b/target/arm/tcg/vfp_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(check_hcr_el2_trap)(CPUARMState *env, uint32_t rt, uint32_t reg) raise_exception(env, EXCP_HYP_TRAP, syndrome, 2); } + +uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) +{ + return vfp_get_fpscr(env); +} + +void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) +{ + vfp_set_fpscr(env, val); +} diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "exec/helper-proto.h" #include "internals.h" #include "cpu-features.h" #include "fpu/softfloat.h" @@ -XXX,XX +XXX,XX @@ uint32_t vfp_get_fpsr(CPUARMState *env) return fpsr; } -uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) +uint32_t vfp_get_fpscr(CPUARMState *env) { return (vfp_get_fpcr(env) & FPSCR_FPCR_MASK) | (vfp_get_fpsr(env) & FPSCR_FPSR_MASK); } -uint32_t vfp_get_fpscr(CPUARMState *env) -{ - return HELPER(vfp_get_fpscr)(env); -} - void vfp_set_fpsr(CPUARMState *env, uint32_t val) { ARMCPU *cpu = env_archcpu(env); @@ -XXX,XX +XXX,XX @@ void vfp_set_fpcr(CPUARMState *env, uint32_t val) vfp_set_fpcr_masked(env, val, MAKE_64BIT_MASK(0, 32)); } -void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) +void vfp_set_fpscr(CPUARMState *env, uint32_t val) { vfp_set_fpcr_masked(env, val, FPSCR_FPCR_MASK); vfp_set_fpsr(env, val & FPSCR_FPSR_MASK); } - -void vfp_set_fpscr(CPUARMState *env, uint32_t val) -{ - HELPER(vfp_set_fpscr)(env, val); -} -- 2.43.0 The softfloat (i.e. TCG) specific handling for the FPCR and FPSR is abstracted behind five functions: arm_set_default_fp_behaviours arm_set_ah_fp_behaviours vfp_get_fpsr_from_host vfp_clear_float_status_exc_flags vfp_set_fpsr_to_host Currently we rely on the first two calling softfloat functions that work even in a KVM-only compile because they're defined as inline in the softfloat header file, and we provide stub versions of the last three in arm/vfp_helper.c if CONFIG_TCG isn't defined. Move the softfloat-specific versions of these functions to tcg/vfp_helper.c, and provide the non-TCG stub versions in tcg-stubs.c. This lets us drop the softfloat header include and the last set of CONFIG_TCG ifdefs from arm/vfp_helper.c. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250221190957.811948-4-peter.maydell@linaro.org --- target/arm/internals.h | 9 ++ target/arm/tcg-stubs.c | 22 ++++ target/arm/tcg/vfp_helper.c | 228 +++++++++++++++++++++++++++++++++ target/arm/vfp_helper.c | 248 ------------------------------------ 4 files changed, 259 insertions(+), 248 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ int alle1_tlbmask(CPUARMState *env); void arm_set_default_fp_behaviours(float_status *s); /* Set the float_status behaviour to match Arm FPCR.AH=1 behaviour */ void arm_set_ah_fp_behaviours(float_status *s); +/* Read the float_status info and return the appropriate FPSR value */ +uint32_t vfp_get_fpsr_from_host(CPUARMState *env); +/* Clear the exception status flags from all float_status fields */ +void vfp_clear_float_status_exc_flags(CPUARMState *env); +/* + * Update float_status fields to handle the bits of the FPCR + * specified by mask changing to the values in val. + */ +void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask); #endif diff --git a/target/arm/tcg-stubs.c b/target/arm/tcg-stubs.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg-stubs.c +++ b/target/arm/tcg-stubs.c @@ -XXX,XX +XXX,XX @@ void assert_hflags_rebuild_correctly(CPUARMState *env) void define_tlb_insn_regs(ARMCPU *cpu) { } + +/* With KVM, we never use float_status, so these can be no-ops */ +void arm_set_default_fp_behaviours(float_status *s) +{ +} + +void arm_set_ah_fp_behaviours(float_status *s) +{ +} + +uint32_t vfp_get_fpsr_from_host(CPUARMState *env) +{ + return 0; +} + +void vfp_clear_float_status_exc_flags(CPUARMState *env) +{ +} + +void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) +{ +} diff --git a/target/arm/tcg/vfp_helper.c b/target/arm/tcg/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/vfp_helper.c +++ b/target/arm/tcg/vfp_helper.c @@ -XXX,XX +XXX,XX @@ #include "fpu/softfloat.h" #include "qemu/log.h" +/* + * Set the float_status behaviour to match the Arm defaults: + * * tininess-before-rounding + * * 2-input NaN propagation prefers SNaN over QNaN, and then + * operand A over operand B (see FPProcessNaNs() pseudocode) + * * 3-input NaN propagation prefers SNaN over QNaN, and then + * operand C over A over B (see FPProcessNaNs3() pseudocode, + * but note that for QEMU muladd is a * b + c, whereas for + * the pseudocode function the arguments are in the order c, a, b. + * * 0 * Inf + NaN returns the default NaN if the input NaN is quiet, + * and the input NaN if it is signalling + * * Default NaN has sign bit clear, msb frac bit set + */ +void arm_set_default_fp_behaviours(float_status *s) +{ + set_float_detect_tininess(float_tininess_before_rounding, s); + set_float_ftz_detection(float_ftz_before_rounding, s); + set_float_2nan_prop_rule(float_2nan_prop_s_ab, s); + set_float_3nan_prop_rule(float_3nan_prop_s_cab, s); + set_float_infzeronan_rule(float_infzeronan_dnan_if_qnan, s); + set_float_default_nan_pattern(0b01000000, s); +} + +/* + * Set the float_status behaviour to match the FEAT_AFP + * FPCR.AH=1 requirements: + * * tininess-after-rounding + * * 2-input NaN propagation prefers the first NaN + * * 3-input NaN propagation prefers a over b over c + * * 0 * Inf + NaN always returns the input NaN and doesn't + * set Invalid for a QNaN + * * default NaN has sign bit set, msb frac bit set + */ +void arm_set_ah_fp_behaviours(float_status *s) +{ + set_float_detect_tininess(float_tininess_after_rounding, s); + set_float_ftz_detection(float_ftz_after_rounding, s); + set_float_2nan_prop_rule(float_2nan_prop_ab, s); + set_float_3nan_prop_rule(float_3nan_prop_abc, s); + set_float_infzeronan_rule(float_infzeronan_dnan_never | + float_infzeronan_suppress_invalid, s); + set_float_default_nan_pattern(0b11000000, s); +} + +/* Convert host exception flags to vfp form. */ +static inline uint32_t vfp_exceptbits_from_host(int host_bits, bool ah) +{ + uint32_t target_bits = 0; + + if (host_bits & float_flag_invalid) { + target_bits |= FPSR_IOC; + } + if (host_bits & float_flag_divbyzero) { + target_bits |= FPSR_DZC; + } + if (host_bits & float_flag_overflow) { + target_bits |= FPSR_OFC; + } + if (host_bits & (float_flag_underflow | float_flag_output_denormal_flushed)) { + target_bits |= FPSR_UFC; + } + if (host_bits & float_flag_inexact) { + target_bits |= FPSR_IXC; + } + if (host_bits & float_flag_input_denormal_flushed) { + target_bits |= FPSR_IDC; + } + /* + * With FPCR.AH, IDC is set when an input denormal is used, + * and flushing an output denormal to zero sets both IXC and UFC. + */ + if (ah && (host_bits & float_flag_input_denormal_used)) { + target_bits |= FPSR_IDC; + } + if (ah && (host_bits & float_flag_output_denormal_flushed)) { + target_bits |= FPSR_IXC; + } + return target_bits; +} + +uint32_t vfp_get_fpsr_from_host(CPUARMState *env) +{ + uint32_t a32_flags = 0, a64_flags = 0; + + a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A32]); + a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_STD]); + /* FZ16 does not generate an input denormal exception. */ + a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A32_F16]) + & ~float_flag_input_denormal_flushed); + a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_STD_F16]) + & ~float_flag_input_denormal_flushed); + + a64_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A64]); + a64_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A64_F16]) + & ~(float_flag_input_denormal_flushed | float_flag_input_denormal_used)); + /* + * We do not merge in flags from FPST_AH or FPST_AH_F16, because + * they are used for insns that must not set the cumulative exception bits. + */ + + /* + * Flushing an input denormal *only* because FPCR.FIZ == 1 does + * not set FPSR.IDC; if FPCR.FZ is also set then this takes + * precedence and IDC is set (see the FPUnpackBase pseudocode). + * So squash it unless (FPCR.AH == 0 && FPCR.FZ == 1). + * We only do this for the a64 flags because FIZ has no effect + * on AArch32 even if it is set. + */ + if ((env->vfp.fpcr & (FPCR_FZ | FPCR_AH)) != FPCR_FZ) { + a64_flags &= ~float_flag_input_denormal_flushed; + } + return vfp_exceptbits_from_host(a64_flags, env->vfp.fpcr & FPCR_AH) | + vfp_exceptbits_from_host(a32_flags, false); +} + +void vfp_clear_float_status_exc_flags(CPUARMState *env) +{ + /* + * Clear out all the exception-flag information in the float_status + * values. The caller should have arranged for env->vfp.fpsr to + * be the architecturally up-to-date exception flag information first. + */ + set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32_F16]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64_F16]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD_F16]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH]); + set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH_F16]); +} + +static void vfp_sync_and_clear_float_status_exc_flags(CPUARMState *env) +{ + /* + * Synchronize any pending exception-flag information in the + * float_status values into env->vfp.fpsr, and then clear out + * the float_status data. + */ + env->vfp.fpsr |= vfp_get_fpsr_from_host(env); + vfp_clear_float_status_exc_flags(env); +} + +void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) +{ + uint64_t changed = env->vfp.fpcr; + + changed ^= val; + changed &= mask; + if (changed & (3 << 22)) { + int i = (val >> 22) & 3; + switch (i) { + case FPROUNDING_TIEEVEN: + i = float_round_nearest_even; + break; + case FPROUNDING_POSINF: + i = float_round_up; + break; + case FPROUNDING_NEGINF: + i = float_round_down; + break; + case FPROUNDING_ZERO: + i = float_round_to_zero; + break; + } + set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32]); + set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64]); + set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32_F16]); + set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64_F16]); + } + if (changed & FPCR_FZ16) { + bool ftz_enabled = val & FPCR_FZ16; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); + } + if (changed & FPCR_FZ) { + bool ftz_enabled = val & FPCR_FZ; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64]); + /* FIZ is A64 only so FZ always makes A32 code flush inputs to zero */ + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); + } + if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { + /* + * A64: Flush denormalized inputs to zero if FPCR.FIZ = 1, or + * both FPCR.AH = 0 and FPCR.FZ = 1. + */ + bool fitz_enabled = (val & FPCR_FIZ) || + (val & (FPCR_FZ | FPCR_AH)) == FPCR_FZ; + set_flush_inputs_to_zero(fitz_enabled, &env->vfp.fp_status[FPST_A64]); + } + if (changed & FPCR_DN) { + bool dnan_enabled = val & FPCR_DN; + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32]); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64]); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32_F16]); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64_F16]); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH]); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH_F16]); + } + if (changed & FPCR_AH) { + bool ah_enabled = val & FPCR_AH; + + if (ah_enabled) { + /* Change behaviours for A64 FP operations */ + arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64]); + arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); + } else { + arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64]); + arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); + } + } + /* + * If any bits changed that we look at in vfp_get_fpsr_from_host(), + * we must sync the float_status flags into vfp.fpsr now (under the + * old regime) before we update vfp.fpcr. + */ + if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { + vfp_sync_and_clear_float_status_exc_flags(env); + } +} + /* * VFP support. We follow the convention used for VFP instructions: * Single precision routines have a "s" suffix, double precision a diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "internals.h" #include "cpu-features.h" -#include "fpu/softfloat.h" - -/* - * Set the float_status behaviour to match the Arm defaults: - * * tininess-before-rounding - * * 2-input NaN propagation prefers SNaN over QNaN, and then - * operand A over operand B (see FPProcessNaNs() pseudocode) - * * 3-input NaN propagation prefers SNaN over QNaN, and then - * operand C over A over B (see FPProcessNaNs3() pseudocode, - * but note that for QEMU muladd is a * b + c, whereas for - * the pseudocode function the arguments are in the order c, a, b. - * * 0 * Inf + NaN returns the default NaN if the input NaN is quiet, - * and the input NaN if it is signalling - * * Default NaN has sign bit clear, msb frac bit set - */ -void arm_set_default_fp_behaviours(float_status *s) -{ - set_float_detect_tininess(float_tininess_before_rounding, s); - set_float_ftz_detection(float_ftz_before_rounding, s); - set_float_2nan_prop_rule(float_2nan_prop_s_ab, s); - set_float_3nan_prop_rule(float_3nan_prop_s_cab, s); - set_float_infzeronan_rule(float_infzeronan_dnan_if_qnan, s); - set_float_default_nan_pattern(0b01000000, s); -} - -/* - * Set the float_status behaviour to match the FEAT_AFP - * FPCR.AH=1 requirements: - * * tininess-after-rounding - * * 2-input NaN propagation prefers the first NaN - * * 3-input NaN propagation prefers a over b over c - * * 0 * Inf + NaN always returns the input NaN and doesn't - * set Invalid for a QNaN - * * default NaN has sign bit set, msb frac bit set - */ -void arm_set_ah_fp_behaviours(float_status *s) -{ - set_float_detect_tininess(float_tininess_after_rounding, s); - set_float_ftz_detection(float_ftz_after_rounding, s); - set_float_2nan_prop_rule(float_2nan_prop_ab, s); - set_float_3nan_prop_rule(float_3nan_prop_abc, s); - set_float_infzeronan_rule(float_infzeronan_dnan_never | - float_infzeronan_suppress_invalid, s); - set_float_default_nan_pattern(0b11000000, s); -} - -#ifdef CONFIG_TCG - -/* Convert host exception flags to vfp form. */ -static inline uint32_t vfp_exceptbits_from_host(int host_bits, bool ah) -{ - uint32_t target_bits = 0; - - if (host_bits & float_flag_invalid) { - target_bits |= FPSR_IOC; - } - if (host_bits & float_flag_divbyzero) { - target_bits |= FPSR_DZC; - } - if (host_bits & float_flag_overflow) { - target_bits |= FPSR_OFC; - } - if (host_bits & (float_flag_underflow | float_flag_output_denormal_flushed)) { - target_bits |= FPSR_UFC; - } - if (host_bits & float_flag_inexact) { - target_bits |= FPSR_IXC; - } - if (host_bits & float_flag_input_denormal_flushed) { - target_bits |= FPSR_IDC; - } - /* - * With FPCR.AH, IDC is set when an input denormal is used, - * and flushing an output denormal to zero sets both IXC and UFC. - */ - if (ah && (host_bits & float_flag_input_denormal_used)) { - target_bits |= FPSR_IDC; - } - if (ah && (host_bits & float_flag_output_denormal_flushed)) { - target_bits |= FPSR_IXC; - } - return target_bits; -} - -static uint32_t vfp_get_fpsr_from_host(CPUARMState *env) -{ - uint32_t a32_flags = 0, a64_flags = 0; - - a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A32]); - a32_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_STD]); - /* FZ16 does not generate an input denormal exception. */ - a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A32_F16]) - & ~float_flag_input_denormal_flushed); - a32_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_STD_F16]) - & ~float_flag_input_denormal_flushed); - - a64_flags |= get_float_exception_flags(&env->vfp.fp_status[FPST_A64]); - a64_flags |= (get_float_exception_flags(&env->vfp.fp_status[FPST_A64_F16]) - & ~(float_flag_input_denormal_flushed | float_flag_input_denormal_used)); - /* - * We do not merge in flags from FPST_AH or FPST_AH_F16, because - * they are used for insns that must not set the cumulative exception bits. - */ - - /* - * Flushing an input denormal *only* because FPCR.FIZ == 1 does - * not set FPSR.IDC; if FPCR.FZ is also set then this takes - * precedence and IDC is set (see the FPUnpackBase pseudocode). - * So squash it unless (FPCR.AH == 0 && FPCR.FZ == 1). - * We only do this for the a64 flags because FIZ has no effect - * on AArch32 even if it is set. - */ - if ((env->vfp.fpcr & (FPCR_FZ | FPCR_AH)) != FPCR_FZ) { - a64_flags &= ~float_flag_input_denormal_flushed; - } - return vfp_exceptbits_from_host(a64_flags, env->vfp.fpcr & FPCR_AH) | - vfp_exceptbits_from_host(a32_flags, false); -} - -static void vfp_clear_float_status_exc_flags(CPUARMState *env) -{ - /* - * Clear out all the exception-flag information in the float_status - * values. The caller should have arranged for env->vfp.fpsr to - * be the architecturally up-to-date exception flag information first. - */ - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A32_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_A64_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_STD_F16]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH]); - set_float_exception_flags(0, &env->vfp.fp_status[FPST_AH_F16]); -} - -static void vfp_sync_and_clear_float_status_exc_flags(CPUARMState *env) -{ - /* - * Synchronize any pending exception-flag information in the - * float_status values into env->vfp.fpsr, and then clear out - * the float_status data. - */ - env->vfp.fpsr |= vfp_get_fpsr_from_host(env); - vfp_clear_float_status_exc_flags(env); -} - -static void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) -{ - uint64_t changed = env->vfp.fpcr; - - changed ^= val; - changed &= mask; - if (changed & (3 << 22)) { - int i = (val >> 22) & 3; - switch (i) { - case FPROUNDING_TIEEVEN: - i = float_round_nearest_even; - break; - case FPROUNDING_POSINF: - i = float_round_up; - break; - case FPROUNDING_NEGINF: - i = float_round_down; - break; - case FPROUNDING_ZERO: - i = float_round_to_zero; - break; - } - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A32_F16]); - set_float_rounding_mode(i, &env->vfp.fp_status[FPST_A64_F16]); - } - if (changed & FPCR_FZ16) { - bool ftz_enabled = val & FPCR_FZ16; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_STD_F16]); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_AH_F16]); - } - if (changed & FPCR_FZ) { - bool ftz_enabled = val & FPCR_FZ; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A64]); - /* FIZ is A64 only so FZ always makes A32 code flush inputs to zero */ - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status[FPST_A32]); - } - if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { - /* - * A64: Flush denormalized inputs to zero if FPCR.FIZ = 1, or - * both FPCR.AH = 0 and FPCR.FZ = 1. - */ - bool fitz_enabled = (val & FPCR_FIZ) || - (val & (FPCR_FZ | FPCR_AH)) == FPCR_FZ; - set_flush_inputs_to_zero(fitz_enabled, &env->vfp.fp_status[FPST_A64]); - } - if (changed & FPCR_DN) { - bool dnan_enabled = val & FPCR_DN; - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A32_F16]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_A64_F16]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH]); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status[FPST_AH_F16]); - } - if (changed & FPCR_AH) { - bool ah_enabled = val & FPCR_AH; - - if (ah_enabled) { - /* Change behaviours for A64 FP operations */ - arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64]); - arm_set_ah_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); - } else { - arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64]); - arm_set_default_fp_behaviours(&env->vfp.fp_status[FPST_A64_F16]); - } - } - /* - * If any bits changed that we look at in vfp_get_fpsr_from_host(), - * we must sync the float_status flags into vfp.fpsr now (under the - * old regime) before we update vfp.fpcr. - */ - if (changed & (FPCR_FZ | FPCR_AH | FPCR_FIZ)) { - vfp_sync_and_clear_float_status_exc_flags(env); - } -} - -#else - -static uint32_t vfp_get_fpsr_from_host(CPUARMState *env) -{ - return 0; -} - -static void vfp_clear_float_status_exc_flags(CPUARMState *env) -{ -} - -static void vfp_set_fpcr_to_host(CPUARMState *env, uint32_t val, uint32_t mask) -{ -} - -#endif uint32_t vfp_get_fpcr(CPUARMState *env) { -- 2.43.0 The vfp_helper.c in the target/arm directory now only has code for handling FPSCR/FPCR/FPSR in it, and no helper functions. Rename it to vfp_fpscr.c; this helps keep it distinct from tcg/vfp_helper.c. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20250221190957.811948-5-peter.maydell@linaro.org --- target/arm/{vfp_helper.c => vfp_fpscr.c} | 2 +- target/arm/meson.build | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) rename target/arm/{vfp_helper.c => vfp_fpscr.c} (98%) diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_fpscr.c similarity index 98% rename from target/arm/vfp_helper.c rename to target/arm/vfp_fpscr.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_fpscr.c @@ -XXX,XX +XXX,XX @@ /* - * ARM VFP floating-point operations + * ARM VFP floating-point: handling of FPSCR/FPCR/FPSR * * Copyright (c) 2003 Fabrice Bellard * diff --git a/target/arm/meson.build b/target/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/meson.build +++ b/target/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(files( 'debug_helper.c', 'gdbstub.c', 'helper.c', - 'vfp_helper.c', + 'vfp_fpscr.c', )) arm_ss.add(zlib) -- 2.43.0 From: Joelle van Dyne <j@getutm.app> macOS 15.2's Hypervisor.framework exposes SME feature on M4 Macs. However, QEMU's hvf accelerator code does not properly support it yet, causing QEMU to fail to start when hvf accelerator is used on these systems, with the error message: qemu-aarch64-softmmu: cannot disable sme4224 All SME vector lengths are disabled. With SME enabled, at least one vector length must be enabled. Ideally we would have SME support on these hosts; however, until that point, we must suppress the SME feature in the ID registers, so that users can at least run non-SME guests. Cc: qemu-stable@nongnu.org Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2665 Signed-off-by: Joelle van Dyne <j@getutm.app> Message-id: 20250224165735.36792-1-j@getutm.app Reviewed-by: Peter Maydell <peter.maydell@linaro.org> [PMM: expanded commit message, comment] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/hvf/hvf.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/hvf/hvf.c +++ b/target/arm/hvf/hvf.c @@ -XXX,XX +XXX,XX @@ static bool hvf_arm_get_host_cpu_features(ARMHostCPUFeatures *ahcf) clamp_id_aa64mmfr0_parange_to_ipa_size(&host_isar.id_aa64mmfr0); + /* + * Disable SME, which is not properly handled by QEMU hvf yet. + * To allow this through we would need to: + * - make sure that the SME state is correctly handled in the + * get_registers/put_registers functions + * - get the SME-specific CPU properties to work with accelerators + * other than TCG + * - fix any assumptions we made that SME implies SVE (since + * on the M4 there is SME but not SVE) + */ + host_isar.id_aa64pfr1 &= ~R_ID_AA64PFR1_SME_MASK; + ahcf->isar = host_isar; /* -- 2.43.0 From: Joelle van Dyne <j@getutm.app> In the syndrome value for a data abort, bit 21 is SSE, which is set to indicate that the abort was on a sign-extending load. When we handle the data abort from the guest via address_space_read(), we forgot to handle this and so would return the wrong value if the guest did a sign-extending load to an MMIO region. Add the sign-extension of the returned data. Cc: qemu-stable@nongnu.org Signed-off-by: Joelle van Dyne <j@getutm.app> Message-id: 20250224184123.50780-1-j@getutm.app [PMM: Drop an unnecessary check on 'len'; expand commit message] Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/hvf/hvf.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/hvf/hvf.c +++ b/target/arm/hvf/hvf.c @@ -XXX,XX +XXX,XX @@ int hvf_vcpu_exec(CPUState *cpu) bool isv = syndrome & ARM_EL_ISV; bool iswrite = (syndrome >> 6) & 1; bool s1ptw = (syndrome >> 7) & 1; + bool sse = (syndrome >> 21) & 1; uint32_t sas = (syndrome >> 22) & 3; uint32_t len = 1 << sas; uint32_t srt = (syndrome >> 16) & 0x1f; @@ -XXX,XX +XXX,XX @@ int hvf_vcpu_exec(CPUState *cpu) address_space_read(&address_space_memory, hvf_exit->exception.physical_address, MEMTXATTRS_UNSPECIFIED, &val, len); + if (sse) { + val = sextract64(val, 0, len * 8); + } hvf_set_reg(cpu, srt, val); } -- 2.43.0 From: Pierrick Bouvier <pierrick.bouvier@linaro.org> Regression introduced by cf76c4 (hw/misc: Add nr_regs and cold_reset_values to NPCM CLK) cold_reset_values has a different size, depending on device used (NPCM7xx vs NPCM8xx). However, s->regs has a fixed size, which matches NPCM8xx. Thus, when initializing a NPCM7xx, we go past cold_reset_values ending. Report by asan: ==2066==ERROR: AddressSanitizer: global-buffer-overflow on address 0x55d68a3e97f0 at pc 0x7fcaf2b2d14b bp 0x7ffff0cc3890 sp 0x7ffff0cc3040 READ of size 196 at 0x55d68a3e97f0 thread T0 #0 0x7fcaf2b2d14a in __interceptor_memcpy ../../../../src/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:827 #1 0x55d688447e0d in memcpy /usr/include/x86_64-linux-gnu/bits/string_fortified.h:29 #2 0x55d688447e0d in npcm_clk_enter_reset ../hw/misc/npcm_clk.c:968 #3 0x55d6899b7213 in resettable_phase_enter ../hw/core/resettable.c:136 #4 0x55d6899a1ef7 in bus_reset_child_foreach ../hw/core/bus.c:97 #5 0x55d6899b717d in resettable_child_foreach ../hw/core/resettable.c:92 #6 0x55d6899b717d in resettable_phase_enter ../hw/core/resettable.c:129 #7 0x55d6899b4ead in resettable_container_child_foreach ../hw/core/resetcontainer.c:54 #8 0x55d6899b717d in resettable_child_foreach ../hw/core/resettable.c:92 #9 0x55d6899b717d in resettable_phase_enter ../hw/core/resettable.c:129 #10 0x55d6899b7bfa in resettable_assert_reset ../hw/core/resettable.c:55 #11 0x55d6899b8666 in resettable_reset ../hw/core/resettable.c:45 #12 0x55d688d15cd2 in qemu_system_reset ../system/runstate.c:527 #13 0x55d687fc5edd in qdev_machine_creation_done ../hw/core/machine.c:1738 #14 0x55d688d209bd in qemu_machine_creation_done ../system/vl.c:2779 #15 0x55d688d209bd in qmp_x_exit_preconfig ../system/vl.c:2807 #16 0x55d688d281fb in qemu_init ../system/vl.c:3838 #17 0x55d687ceab12 in main ../system/main.c:68 #18 0x7fcaef006249 (/lib/x86_64-linux-gnu/libc.so.6+0x27249) #19 0x7fcaef006304 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x27304) #20 0x55d687cf0010 in _start (/home/runner/work/qemu-ci/qemu-ci/build/qemu-system-arm+0x371c010) 0x55d68a3e97f0 is located 0 bytes to the right of global variable 'npcm7xx_cold_reset_values' defined in '../hw/misc/npcm_clk.c:134:23' (0x55d68a3e9780) of size 112 Impacted tests: Summary of Failures: check: 2/747 qemu:qtest+qtest-aarch64 / qtest-aarch64/qom-test ERROR 9.28s killed by signal 6 SIGABRT 4/747 qemu:qtest+qtest-arm / qtest-arm/qom-test ERROR 7.82s killed by signal 6 SIGABRT 32/747 qemu:qtest+qtest-aarch64 / qtest-aarch64/device-introspect-test ERROR 10.91s killed by signal 6 SIGABRT 35/747 qemu:qtest+qtest-arm / qtest-arm/device-introspect-test ERROR 11.33s killed by signal 6 SIGABRT 114/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_pwm-test ERROR 0.98s killed by signal 6 SIGABRT 115/747 qemu:qtest+qtest-aarch64 / qtest-aarch64/test-hmp ERROR 2.95s killed by signal 6 SIGABRT 117/747 qemu:qtest+qtest-arm / qtest-arm/test-hmp ERROR 2.54s killed by signal 6 SIGABRT 151/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_watchdog_timer-test ERROR 0.96s killed by signal 6 SIGABRT 247/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_adc-test ERROR 0.96s killed by signal 6 SIGABRT 248/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_gpio-test ERROR 1.05s killed by signal 6 SIGABRT 249/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_rng-test ERROR 0.97s killed by signal 6 SIGABRT 250/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_sdhci-test ERROR 0.97s killed by signal 6 SIGABRT 251/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_smbus-test ERROR 0.89s killed by signal 6 SIGABRT 252/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_timer-test ERROR 1.09s killed by signal 6 SIGABRT 253/747 qemu:qtest+qtest-arm / qtest-arm/npcm_gmac-test ERROR 1.12s killed by signal 6 SIGABRT 255/747 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_emc-test ERROR 1.05s killed by signal 6 SIGABRT check-functional: 22/203 qemu:func-thorough+func-arm-thorough+thorough / func-arm-arm_quanta_gsj ERROR 0.79s exit status 1 38/203 qemu:func-quick+func-aarch64 / func-aarch64-migration ERROR 1.97s exit status 1 45/203 qemu:func-quick+func-arm / func-arm-migration ERROR 1.90s exit status 1 Fixes: cf76c4e174e1 ("hw/misc: Add nr_regs and cold_reset_values to NPCM CLK") Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Reviewed-by: Hao Wu <wuhaotsh@google.com> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/npcm_clk.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/hw/misc/npcm_clk.c b/hw/misc/npcm_clk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/npcm_clk.c +++ b/hw/misc/npcm_clk.c @@ -XXX,XX +XXX,XX @@ static void npcm_clk_enter_reset(Object *obj, ResetType type) NPCMCLKState *s = NPCM_CLK(obj); NPCMCLKClass *c = NPCM_CLK_GET_CLASS(s); - g_assert(sizeof(s->regs) >= c->nr_regs * sizeof(uint32_t)); - memcpy(s->regs, c->cold_reset_values, sizeof(s->regs)); + size_t sizeof_regs = c->nr_regs * sizeof(uint32_t); + g_assert(sizeof(s->regs) >= sizeof_regs); + memcpy(s->regs, c->cold_reset_values, sizeof_regs); s->ref_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); npcm7xx_clk_update_all_clocks(s); /* -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> While at it add missing GUSB2RHBCTL register as found in i.MX 8M Plus reference manual. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-2-shentey@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/usb/hcd-dwc3.h | 2 +- hw/usb/hcd-dwc3.c | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/include/hw/usb/hcd-dwc3.h b/include/hw/usb/hcd-dwc3.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/usb/hcd-dwc3.h +++ b/include/hw/usb/hcd-dwc3.h @@ -XXX,XX +XXX,XX @@ #define USB_DWC3(obj) \ OBJECT_CHECK(USBDWC3, (obj), TYPE_USB_DWC3) -#define USB_DWC3_R_MAX ((0x530 / 4) + 1) +#define USB_DWC3_R_MAX (0x600 / 4) #define DWC3_SIZE 0x10000 typedef struct USBDWC3 { diff --git a/hw/usb/hcd-dwc3.c b/hw/usb/hcd-dwc3.c index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/hcd-dwc3.c +++ b/hw/usb/hcd-dwc3.c @@ -XXX,XX +XXX,XX @@ REG32(GFLADJ, 0x530) FIELD(GFLADJ, GFLADJ_REFCLK_FLADJ, 8, 14) FIELD(GFLADJ, GFLADJ_30MHZ_SDBND_SEL, 7, 1) FIELD(GFLADJ, GFLADJ_30MHZ, 0, 6) +REG32(GUSB2RHBCTL, 0x540) + FIELD(GUSB2RHBCTL, OVRD_L1TIMEOUT, 0, 4) #define DWC3_GLOBAL_OFFSET 0xC100 static void reset_csr(USBDWC3 * s) @@ -XXX,XX +XXX,XX @@ static const RegisterAccessInfo usb_dwc3_regs_info[] = { .rsvd = 0x40, .ro = 0x400040, .unimp = 0xffffffff, + },{ .name = "GUSB2RHBCTL", .addr = A_GUSB2RHBCTL, + .rsvd = 0xfffffff0, + .unimp = 0xffffffff, } }; -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> On the real device, the PCIe root bus is only connected to a PCIe bridge and does not allow for direct attachment of devices. Doing so in QEMU results in no PCI devices being detected by Linux. Instead, PCI devices should plug into the secondary PCIe bus spawned by the internal PCIe bridge. Unfortunately, QEMU defaults to plugging devices into the PCIe root bus. To work around this, every PCI device created on the command line needs an extra `bus=dw-pcie` option which is error prone. Fix that by marking the PCIe root bus as full which makes QEMU decend into the child PCIe bus. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-3-shentey@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/pci-host/designware.h | 7 +++++++ hw/pci-host/designware.c | 18 +++++++++++++++++- 2 files changed, 24 insertions(+), 1 deletion(-) diff --git a/include/hw/pci-host/designware.h b/include/hw/pci-host/designware.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/pci-host/designware.h +++ b/include/hw/pci-host/designware.h @@ -XXX,XX +XXX,XX @@ #include "hw/pci/pci_bridge.h" #include "qom/object.h" +#define TYPE_DESIGNWARE_PCIE_ROOT_BUS "designware-pcie-root-BUS" +OBJECT_DECLARE_SIMPLE_TYPE(DesignwarePCIERootBus, DESIGNWARE_PCIE_ROOT_BUS) + #define TYPE_DESIGNWARE_PCIE_HOST "designware-pcie-host" OBJECT_DECLARE_SIMPLE_TYPE(DesignwarePCIEHost, DESIGNWARE_PCIE_HOST) #define TYPE_DESIGNWARE_PCIE_ROOT "designware-pcie-root" OBJECT_DECLARE_SIMPLE_TYPE(DesignwarePCIERoot, DESIGNWARE_PCIE_ROOT) +struct DesignwarePCIERootBus { + PCIBus parent; +}; + typedef struct DesignwarePCIEViewport { DesignwarePCIERoot *root; diff --git a/hw/pci-host/designware.c b/hw/pci-host/designware.c index XXXXXXX..XXXXXXX 100644 --- a/hw/pci-host/designware.c +++ b/hw/pci-host/designware.c @@ -XXX,XX +XXX,XX @@ #define DESIGNWARE_PCIE_ATU_DEVFN(x) (((x) >> 16) & 0xff) #define DESIGNWARE_PCIE_ATU_UPPER_TARGET 0x91C +static void designware_pcie_root_bus_class_init(ObjectClass *klass, void *data) +{ + BusClass *k = BUS_CLASS(klass); + + /* + * Designware has only a single root complex. Enforce the limit on the + * parent bus + */ + k->max_dev = 1; +} + static DesignwarePCIEHost * designware_pcie_root_to_host(DesignwarePCIERoot *root) { @@ -XXX,XX +XXX,XX @@ static void designware_pcie_host_realize(DeviceState *dev, Error **errp) &s->pci.memory, &s->pci.io, 0, 4, - TYPE_PCIE_BUS); + TYPE_DESIGNWARE_PCIE_ROOT_BUS); pci->bus->flags |= PCI_BUS_EXTENDED_CONFIG_SPACE; memory_region_init(&s->pci.address_space_root, @@ -XXX,XX +XXX,XX @@ static void designware_pcie_host_init(Object *obj) static const TypeInfo designware_pcie_types[] = { { + .name = TYPE_DESIGNWARE_PCIE_ROOT_BUS, + .parent = TYPE_PCIE_BUS, + .instance_size = sizeof(DesignwarePCIERootBus), + .class_init = designware_pcie_root_bus_class_init, + }, { .name = TYPE_DESIGNWARE_PCIE_HOST, .parent = TYPE_PCI_HOST_BRIDGE, .instance_size = sizeof(DesignwarePCIEHost), -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> The move of the Kconfig bits to hw/gpio is fixing a bug in 6328d8ffa6cb9d ("misc/pca955*: Move models under hw/gpio"), which moved the code but forgot to move the Kconfig sections. Fixes: 6328d8ffa6cb9d "misc/pca955*: Move models under hw/gpio" Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-4-shentey@gmail.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/gpio/Kconfig | 8 ++++++++ hw/misc/Kconfig | 8 -------- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/hw/gpio/Kconfig b/hw/gpio/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/gpio/Kconfig +++ b/hw/gpio/Kconfig @@ -XXX,XX +XXX,XX @@ config SIFIVE_GPIO config STM32L4X5_GPIO bool +config PCA9552 + bool + depends on I2C + +config PCA9554 + bool + depends on I2C + config PCF8574 bool depends on I2C diff --git a/hw/misc/Kconfig b/hw/misc/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Kconfig +++ b/hw/misc/Kconfig @@ -XXX,XX +XXX,XX @@ config EDU default y if TEST_DEVICES depends on PCI && MSI_NONBROKEN -config PCA9552 - bool - depends on I2C - -config PCA9554 - bool - depends on I2C - config I2C_ECHO bool default y if TEST_DEVICES -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> As a first step, implement the bare minimum: CPUs, RAM, interrupt controller, serial. All other devices of the A53 memory map are represented as TYPE_UNIMPLEMENTED_DEVICE, i.e. the whole memory map is provided. This allows for running Linux without it crashing due to invalid memory accesses. Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-5-shentey@gmail.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> [PMM: drop 'static const' from serial_table[] definition to avoid compile failure on GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 9 + docs/system/arm/imx8mp-evk.rst | 54 +++++ docs/system/target-arm.rst | 1 + include/hw/arm/fsl-imx8mp.h | 189 +++++++++++++++++ hw/arm/fsl-imx8mp.c | 367 +++++++++++++++++++++++++++++++++ hw/arm/imx8mp-evk.c | 55 +++++ hw/arm/Kconfig | 12 ++ hw/arm/meson.build | 2 + 8 files changed, 689 insertions(+) create mode 100644 docs/system/arm/imx8mp-evk.rst create mode 100644 include/hw/arm/fsl-imx8mp.h create mode 100644 hw/arm/fsl-imx8mp.c create mode 100644 hw/arm/imx8mp-evk.c diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/pci-host/designware.c F: include/hw/pci-host/designware.h F: docs/system/arm/mcimx7d-sabre.rst +MCIMX8MP-EVK / i.MX8MP +M: Bernhard Beschow <shentey@gmail.com> +L: qemu-arm@nongnu.org +S: Maintained +F: hw/arm/imx8mp-evk.c +F: hw/arm/fsl-imx8mp.c +F: include/hw/arm/fsl-imx8mp.h +F: docs/system/arm/imx8mp-evk.rst + MPS2 / MPS3 M: Peter Maydell <peter.maydell@linaro.org> L: qemu-arm@nongnu.org diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ +NXP i.MX 8M Plus Evaluation Kit (``imx8mp-evk``) +================================================ + +The ``imx8mp-evk`` machine models the i.MX 8M Plus Evaluation Kit, based on an +i.MX 8M Plus SoC. + +Supported devices +----------------- + +The ``imx8mp-evk`` machine implements the following devices: + + * Up to 4 Cortex-A53 cores + * Generic Interrupt Controller (GICv3) + * 4 UARTs + +Boot options +------------ + +The ``imx8mp-evk`` machine can start a Linux kernel directly using the standard +``-kernel`` functionality. + +Direct Linux Kernel Boot +'''''''''''''''''''''''' + +Probably the easiest way to get started with a whole Linux system on the machine +is to generate an image with Buildroot. Version 2024.11.1 is tested at the time +of writing and involves two steps. First run the following commands in the +toplevel directory of the Buildroot source tree: + +.. code-block:: bash + + $ echo "BR2_TARGET_ROOTFS_CPIO=y" >> configs/freescale_imx8mpevk_defconfig + $ make freescale_imx8mpevk_defconfig + $ make + +Once finished successfully there is an ``output/image`` subfolder. Navigate into +it and patch the device tree with the following commands which will remove the +``cpu-idle-states`` properties from CPU nodes: + +.. code-block:: bash + + $ dtc imx8mp-evk.dtb | sed '/cpu-idle-states/d' > imx8mp-evk-patched.dts + $ dtc imx8mp-evk-patched.dts -o imx8mp-evk-patched.dtb + +Now that everything is prepared the machine can be started as follows: + +.. code-block:: bash + + $ qemu-system-aarch64 -M imx8mp-evk -smp 4 -m 3G \ + -display none -serial null -serial stdio \ + -kernel Image \ + -dtb imx8mp-evk-patched.dtb \ + -initrd rootfs.cpio \ + -append "root=/dev/ram" diff --git a/docs/system/target-arm.rst b/docs/system/target-arm.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/target-arm.rst +++ b/docs/system/target-arm.rst @@ -XXX,XX +XXX,XX @@ Board-specific documentation arm/imx25-pdk arm/mcimx6ul-evk arm/mcimx7d-sabre + arm/imx8mp-evk arm/orangepi arm/raspi arm/collie diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ +/* + * i.MX 8M Plus SoC Definitions + * + * Copyright (c) 2024, Bernhard Beschow <shentey@gmail.com> + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#ifndef FSL_IMX8MP_H +#define FSL_IMX8MP_H + +#include "cpu.h" +#include "hw/char/imx_serial.h" +#include "hw/intc/arm_gicv3_common.h" +#include "qom/object.h" +#include "qemu/units.h" + +#define TYPE_FSL_IMX8MP "fsl-imx8mp" +OBJECT_DECLARE_SIMPLE_TYPE(FslImx8mpState, FSL_IMX8MP) + +#define FSL_IMX8MP_RAM_START 0x40000000 +#define FSL_IMX8MP_RAM_SIZE_MAX (8 * GiB) + +enum FslImx8mpConfiguration { + FSL_IMX8MP_NUM_CPUS = 4, + FSL_IMX8MP_NUM_IRQS = 160, + FSL_IMX8MP_NUM_UARTS = 4, +}; + +struct FslImx8mpState { + DeviceState parent_obj; + + ARMCPU cpu[FSL_IMX8MP_NUM_CPUS]; + GICv3State gic; + IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; +}; + +enum FslImx8mpMemoryRegions { + FSL_IMX8MP_A53_DAP, + FSL_IMX8MP_AIPS1_CONFIGURATION, + FSL_IMX8MP_AIPS2_CONFIGURATION, + FSL_IMX8MP_AIPS3_CONFIGURATION, + FSL_IMX8MP_AIPS4_CONFIGURATION, + FSL_IMX8MP_AIPS5_CONFIGURATION, + FSL_IMX8MP_ANA_OSC, + FSL_IMX8MP_ANA_PLL, + FSL_IMX8MP_ANA_TSENSOR, + FSL_IMX8MP_APBH_DMA, + FSL_IMX8MP_ASRC, + FSL_IMX8MP_AUDIO_BLK_CTRL, + FSL_IMX8MP_AUDIO_DSP, + FSL_IMX8MP_AUDIO_XCVR_RX, + FSL_IMX8MP_AUD_IRQ_STEER, + FSL_IMX8MP_BOOT_ROM, + FSL_IMX8MP_BOOT_ROM_PROTECTED, + FSL_IMX8MP_CAAM, + FSL_IMX8MP_CAAM_MEM, + FSL_IMX8MP_CCM, + FSL_IMX8MP_CSU, + FSL_IMX8MP_DDR_BLK_CTRL, + FSL_IMX8MP_DDR_CTL, + FSL_IMX8MP_DDR_PERF_MON, + FSL_IMX8MP_DDR_PHY, + FSL_IMX8MP_DDR_PHY_BROADCAST, + FSL_IMX8MP_ECSPI1, + FSL_IMX8MP_ECSPI2, + FSL_IMX8MP_ECSPI3, + FSL_IMX8MP_EDMA_CHANNELS, + FSL_IMX8MP_EDMA_MANAGEMENT_PAGE, + FSL_IMX8MP_ENET1, + FSL_IMX8MP_ENET2_TSN, + FSL_IMX8MP_FLEXCAN1, + FSL_IMX8MP_FLEXCAN2, + FSL_IMX8MP_GIC_DIST, + FSL_IMX8MP_GIC_REDIST, + FSL_IMX8MP_GPC, + FSL_IMX8MP_GPIO1, + FSL_IMX8MP_GPIO2, + FSL_IMX8MP_GPIO3, + FSL_IMX8MP_GPIO4, + FSL_IMX8MP_GPIO5, + FSL_IMX8MP_GPT1, + FSL_IMX8MP_GPT2, + FSL_IMX8MP_GPT3, + FSL_IMX8MP_GPT4, + FSL_IMX8MP_GPT5, + FSL_IMX8MP_GPT6, + FSL_IMX8MP_GPU2D, + FSL_IMX8MP_GPU3D, + FSL_IMX8MP_HDMI_TX, + FSL_IMX8MP_HDMI_TX_AUDLNK_MSTR, + FSL_IMX8MP_HSIO_BLK_CTL, + FSL_IMX8MP_I2C1, + FSL_IMX8MP_I2C2, + FSL_IMX8MP_I2C3, + FSL_IMX8MP_I2C4, + FSL_IMX8MP_I2C5, + FSL_IMX8MP_I2C6, + FSL_IMX8MP_INTERCONNECT, + FSL_IMX8MP_IOMUXC, + FSL_IMX8MP_IOMUXC_GPR, + FSL_IMX8MP_IPS_DEWARP, + FSL_IMX8MP_ISI, + FSL_IMX8MP_ISP1, + FSL_IMX8MP_ISP2, + FSL_IMX8MP_LCDIF1, + FSL_IMX8MP_LCDIF2, + FSL_IMX8MP_MEDIA_BLK_CTL, + FSL_IMX8MP_MIPI_CSI1, + FSL_IMX8MP_MIPI_CSI2, + FSL_IMX8MP_MIPI_DSI1, + FSL_IMX8MP_MU_1_A, + FSL_IMX8MP_MU_1_B, + FSL_IMX8MP_MU_2_A, + FSL_IMX8MP_MU_2_B, + FSL_IMX8MP_MU_3_A, + FSL_IMX8MP_MU_3_B, + FSL_IMX8MP_NPU, + FSL_IMX8MP_OCOTP_CTRL, + FSL_IMX8MP_OCRAM, + FSL_IMX8MP_OCRAM_S, + FSL_IMX8MP_PCIE1, + FSL_IMX8MP_PCIE1_MEM, + FSL_IMX8MP_PCIE_PHY1, + FSL_IMX8MP_PDM, + FSL_IMX8MP_PERFMON1, + FSL_IMX8MP_PERFMON2, + FSL_IMX8MP_PWM1, + FSL_IMX8MP_PWM2, + FSL_IMX8MP_PWM3, + FSL_IMX8MP_PWM4, + FSL_IMX8MP_QOSC, + FSL_IMX8MP_QSPI, + FSL_IMX8MP_QSPI1_RX_BUFFER, + FSL_IMX8MP_QSPI1_TX_BUFFER, + FSL_IMX8MP_QSPI_MEM, + FSL_IMX8MP_RAM, + FSL_IMX8MP_RDC, + FSL_IMX8MP_SAI1, + FSL_IMX8MP_SAI2, + FSL_IMX8MP_SAI3, + FSL_IMX8MP_SAI5, + FSL_IMX8MP_SAI6, + FSL_IMX8MP_SAI7, + FSL_IMX8MP_SDMA1, + FSL_IMX8MP_SDMA2, + FSL_IMX8MP_SDMA3, + FSL_IMX8MP_SEMAPHORE1, + FSL_IMX8MP_SEMAPHORE2, + FSL_IMX8MP_SEMAPHORE_HS, + FSL_IMX8MP_SNVS_HP, + FSL_IMX8MP_SPBA1, + FSL_IMX8MP_SPBA2, + FSL_IMX8MP_SRC, + FSL_IMX8MP_SYSCNT_CMP, + FSL_IMX8MP_SYSCNT_CTRL, + FSL_IMX8MP_SYSCNT_RD, + FSL_IMX8MP_TCM_DTCM, + FSL_IMX8MP_TCM_ITCM, + FSL_IMX8MP_TZASC, + FSL_IMX8MP_UART1, + FSL_IMX8MP_UART2, + FSL_IMX8MP_UART3, + FSL_IMX8MP_UART4, + FSL_IMX8MP_USB1, + FSL_IMX8MP_USB2, + FSL_IMX8MP_USDHC1, + FSL_IMX8MP_USDHC2, + FSL_IMX8MP_USDHC3, + FSL_IMX8MP_VPU, + FSL_IMX8MP_VPU_BLK_CTRL, + FSL_IMX8MP_VPU_G1_DECODER, + FSL_IMX8MP_VPU_G2_DECODER, + FSL_IMX8MP_VPU_VC8000E_ENCODER, + FSL_IMX8MP_WDOG1, + FSL_IMX8MP_WDOG2, + FSL_IMX8MP_WDOG3, +}; + +enum FslImx8mpIrqs { + FSL_IMX8MP_UART1_IRQ = 26, + FSL_IMX8MP_UART2_IRQ = 27, + FSL_IMX8MP_UART3_IRQ = 28, + FSL_IMX8MP_UART4_IRQ = 29, + FSL_IMX8MP_UART5_IRQ = 30, + FSL_IMX8MP_UART6_IRQ = 16, +}; + +#endif /* FSL_IMX8MP_H */ diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ +/* + * i.MX 8M Plus SoC Implementation + * + * Based on hw/arm/fsl-imx6.c + * + * Copyright (c) 2024, Bernhard Beschow <shentey@gmail.com> + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "exec/address-spaces.h" +#include "hw/arm/bsa.h" +#include "hw/arm/fsl-imx8mp.h" +#include "hw/intc/arm_gicv3.h" +#include "hw/misc/unimp.h" +#include "hw/boards.h" +#include "system/system.h" +#include "target/arm/cpu-qom.h" +#include "qapi/error.h" +#include "qobject/qlist.h" + +static const struct { + hwaddr addr; + size_t size; + const char *name; +} fsl_imx8mp_memmap[] = { + [FSL_IMX8MP_RAM] = { FSL_IMX8MP_RAM_START, FSL_IMX8MP_RAM_SIZE_MAX, "ram" }, + [FSL_IMX8MP_DDR_PHY_BROADCAST] = { 0x3dc00000, 4 * MiB, "ddr_phy_broadcast" }, + [FSL_IMX8MP_DDR_PERF_MON] = { 0x3d800000, 4 * MiB, "ddr_perf_mon" }, + [FSL_IMX8MP_DDR_CTL] = { 0x3d400000, 4 * MiB, "ddr_ctl" }, + [FSL_IMX8MP_DDR_BLK_CTRL] = { 0x3d000000, 1 * MiB, "ddr_blk_ctrl" }, + [FSL_IMX8MP_DDR_PHY] = { 0x3c000000, 16 * MiB, "ddr_phy" }, + [FSL_IMX8MP_AUDIO_DSP] = { 0x3b000000, 16 * MiB, "audio_dsp" }, + [FSL_IMX8MP_GIC_DIST] = { 0x38800000, 512 * KiB, "gic_dist" }, + [FSL_IMX8MP_GIC_REDIST] = { 0x38880000, 512 * KiB, "gic_redist" }, + [FSL_IMX8MP_NPU] = { 0x38500000, 2 * MiB, "npu" }, + [FSL_IMX8MP_VPU] = { 0x38340000, 2 * MiB, "vpu" }, + [FSL_IMX8MP_VPU_BLK_CTRL] = { 0x38330000, 2 * MiB, "vpu_blk_ctrl" }, + [FSL_IMX8MP_VPU_VC8000E_ENCODER] = { 0x38320000, 2 * MiB, "vpu_vc8000e_encoder" }, + [FSL_IMX8MP_VPU_G2_DECODER] = { 0x38310000, 2 * MiB, "vpu_g2_decoder" }, + [FSL_IMX8MP_VPU_G1_DECODER] = { 0x38300000, 2 * MiB, "vpu_g1_decoder" }, + [FSL_IMX8MP_USB2] = { 0x38200000, 1 * MiB, "usb2" }, + [FSL_IMX8MP_USB1] = { 0x38100000, 1 * MiB, "usb1" }, + [FSL_IMX8MP_GPU2D] = { 0x38008000, 32 * KiB, "gpu2d" }, + [FSL_IMX8MP_GPU3D] = { 0x38000000, 32 * KiB, "gpu3d" }, + [FSL_IMX8MP_QSPI1_RX_BUFFER] = { 0x34000000, 32 * MiB, "qspi1_rx_buffer" }, + [FSL_IMX8MP_PCIE1] = { 0x33800000, 4 * MiB, "pcie1" }, + [FSL_IMX8MP_QSPI1_TX_BUFFER] = { 0x33008000, 32 * KiB, "qspi1_tx_buffer" }, + [FSL_IMX8MP_APBH_DMA] = { 0x33000000, 32 * KiB, "apbh_dma" }, + + /* AIPS-5 Begin */ + [FSL_IMX8MP_MU_3_B] = { 0x30e90000, 64 * KiB, "mu_3_b" }, + [FSL_IMX8MP_MU_3_A] = { 0x30e80000, 64 * KiB, "mu_3_a" }, + [FSL_IMX8MP_MU_2_B] = { 0x30e70000, 64 * KiB, "mu_2_b" }, + [FSL_IMX8MP_MU_2_A] = { 0x30e60000, 64 * KiB, "mu_2_a" }, + [FSL_IMX8MP_EDMA_CHANNELS] = { 0x30e40000, 128 * KiB, "edma_channels" }, + [FSL_IMX8MP_EDMA_MANAGEMENT_PAGE] = { 0x30e30000, 64 * KiB, "edma_management_page" }, + [FSL_IMX8MP_AUDIO_BLK_CTRL] = { 0x30e20000, 64 * KiB, "audio_blk_ctrl" }, + [FSL_IMX8MP_SDMA2] = { 0x30e10000, 64 * KiB, "sdma2" }, + [FSL_IMX8MP_SDMA3] = { 0x30e00000, 64 * KiB, "sdma3" }, + [FSL_IMX8MP_AIPS5_CONFIGURATION] = { 0x30df0000, 64 * KiB, "aips5_configuration" }, + [FSL_IMX8MP_SPBA2] = { 0x30cf0000, 64 * KiB, "spba2" }, + [FSL_IMX8MP_AUDIO_XCVR_RX] = { 0x30cc0000, 64 * KiB, "audio_xcvr_rx" }, + [FSL_IMX8MP_HDMI_TX_AUDLNK_MSTR] = { 0x30cb0000, 64 * KiB, "hdmi_tx_audlnk_mstr" }, + [FSL_IMX8MP_PDM] = { 0x30ca0000, 64 * KiB, "pdm" }, + [FSL_IMX8MP_ASRC] = { 0x30c90000, 64 * KiB, "asrc" }, + [FSL_IMX8MP_SAI7] = { 0x30c80000, 64 * KiB, "sai7" }, + [FSL_IMX8MP_SAI6] = { 0x30c60000, 64 * KiB, "sai6" }, + [FSL_IMX8MP_SAI5] = { 0x30c50000, 64 * KiB, "sai5" }, + [FSL_IMX8MP_SAI3] = { 0x30c30000, 64 * KiB, "sai3" }, + [FSL_IMX8MP_SAI2] = { 0x30c20000, 64 * KiB, "sai2" }, + [FSL_IMX8MP_SAI1] = { 0x30c10000, 64 * KiB, "sai1" }, + /* AIPS-5 End */ + + /* AIPS-4 Begin */ + [FSL_IMX8MP_HDMI_TX] = { 0x32fc0000, 128 * KiB, "hdmi_tx" }, + [FSL_IMX8MP_TZASC] = { 0x32f80000, 64 * KiB, "tzasc" }, + [FSL_IMX8MP_HSIO_BLK_CTL] = { 0x32f10000, 64 * KiB, "hsio_blk_ctl" }, + [FSL_IMX8MP_PCIE_PHY1] = { 0x32f00000, 64 * KiB, "pcie_phy1" }, + [FSL_IMX8MP_MEDIA_BLK_CTL] = { 0x32ec0000, 64 * KiB, "media_blk_ctl" }, + [FSL_IMX8MP_LCDIF2] = { 0x32e90000, 64 * KiB, "lcdif2" }, + [FSL_IMX8MP_LCDIF1] = { 0x32e80000, 64 * KiB, "lcdif1" }, + [FSL_IMX8MP_MIPI_DSI1] = { 0x32e60000, 64 * KiB, "mipi_dsi1" }, + [FSL_IMX8MP_MIPI_CSI2] = { 0x32e50000, 64 * KiB, "mipi_csi2" }, + [FSL_IMX8MP_MIPI_CSI1] = { 0x32e40000, 64 * KiB, "mipi_csi1" }, + [FSL_IMX8MP_IPS_DEWARP] = { 0x32e30000, 64 * KiB, "ips_dewarp" }, + [FSL_IMX8MP_ISP2] = { 0x32e20000, 64 * KiB, "isp2" }, + [FSL_IMX8MP_ISP1] = { 0x32e10000, 64 * KiB, "isp1" }, + [FSL_IMX8MP_ISI] = { 0x32e00000, 64 * KiB, "isi" }, + [FSL_IMX8MP_AIPS4_CONFIGURATION] = { 0x32df0000, 64 * KiB, "aips4_configuration" }, + /* AIPS-4 End */ + + [FSL_IMX8MP_INTERCONNECT] = { 0x32700000, 1 * MiB, "interconnect" }, + + /* AIPS-3 Begin */ + [FSL_IMX8MP_ENET2_TSN] = { 0x30bf0000, 64 * KiB, "enet2_tsn" }, + [FSL_IMX8MP_ENET1] = { 0x30be0000, 64 * KiB, "enet1" }, + [FSL_IMX8MP_SDMA1] = { 0x30bd0000, 64 * KiB, "sdma1" }, + [FSL_IMX8MP_QSPI] = { 0x30bb0000, 64 * KiB, "qspi" }, + [FSL_IMX8MP_USDHC3] = { 0x30b60000, 64 * KiB, "usdhc3" }, + [FSL_IMX8MP_USDHC2] = { 0x30b50000, 64 * KiB, "usdhc2" }, + [FSL_IMX8MP_USDHC1] = { 0x30b40000, 64 * KiB, "usdhc1" }, + [FSL_IMX8MP_I2C6] = { 0x30ae0000, 64 * KiB, "i2c6" }, + [FSL_IMX8MP_I2C5] = { 0x30ad0000, 64 * KiB, "i2c5" }, + [FSL_IMX8MP_SEMAPHORE_HS] = { 0x30ac0000, 64 * KiB, "semaphore_hs" }, + [FSL_IMX8MP_MU_1_B] = { 0x30ab0000, 64 * KiB, "mu_1_b" }, + [FSL_IMX8MP_MU_1_A] = { 0x30aa0000, 64 * KiB, "mu_1_a" }, + [FSL_IMX8MP_AUD_IRQ_STEER] = { 0x30a80000, 64 * KiB, "aud_irq_steer" }, + [FSL_IMX8MP_UART4] = { 0x30a60000, 64 * KiB, "uart4" }, + [FSL_IMX8MP_I2C4] = { 0x30a50000, 64 * KiB, "i2c4" }, + [FSL_IMX8MP_I2C3] = { 0x30a40000, 64 * KiB, "i2c3" }, + [FSL_IMX8MP_I2C2] = { 0x30a30000, 64 * KiB, "i2c2" }, + [FSL_IMX8MP_I2C1] = { 0x30a20000, 64 * KiB, "i2c1" }, + [FSL_IMX8MP_AIPS3_CONFIGURATION] = { 0x309f0000, 64 * KiB, "aips3_configuration" }, + [FSL_IMX8MP_CAAM] = { 0x30900000, 256 * KiB, "caam" }, + [FSL_IMX8MP_SPBA1] = { 0x308f0000, 64 * KiB, "spba1" }, + [FSL_IMX8MP_FLEXCAN2] = { 0x308d0000, 64 * KiB, "flexcan2" }, + [FSL_IMX8MP_FLEXCAN1] = { 0x308c0000, 64 * KiB, "flexcan1" }, + [FSL_IMX8MP_UART2] = { 0x30890000, 64 * KiB, "uart2" }, + [FSL_IMX8MP_UART3] = { 0x30880000, 64 * KiB, "uart3" }, + [FSL_IMX8MP_UART1] = { 0x30860000, 64 * KiB, "uart1" }, + [FSL_IMX8MP_ECSPI3] = { 0x30840000, 64 * KiB, "ecspi3" }, + [FSL_IMX8MP_ECSPI2] = { 0x30830000, 64 * KiB, "ecspi2" }, + [FSL_IMX8MP_ECSPI1] = { 0x30820000, 64 * KiB, "ecspi1" }, + /* AIPS-3 End */ + + /* AIPS-2 Begin */ + [FSL_IMX8MP_QOSC] = { 0x307f0000, 64 * KiB, "qosc" }, + [FSL_IMX8MP_PERFMON2] = { 0x307d0000, 64 * KiB, "perfmon2" }, + [FSL_IMX8MP_PERFMON1] = { 0x307c0000, 64 * KiB, "perfmon1" }, + [FSL_IMX8MP_GPT4] = { 0x30700000, 64 * KiB, "gpt4" }, + [FSL_IMX8MP_GPT5] = { 0x306f0000, 64 * KiB, "gpt5" }, + [FSL_IMX8MP_GPT6] = { 0x306e0000, 64 * KiB, "gpt6" }, + [FSL_IMX8MP_SYSCNT_CTRL] = { 0x306c0000, 64 * KiB, "syscnt_ctrl" }, + [FSL_IMX8MP_SYSCNT_CMP] = { 0x306b0000, 64 * KiB, "syscnt_cmp" }, + [FSL_IMX8MP_SYSCNT_RD] = { 0x306a0000, 64 * KiB, "syscnt_rd" }, + [FSL_IMX8MP_PWM4] = { 0x30690000, 64 * KiB, "pwm4" }, + [FSL_IMX8MP_PWM3] = { 0x30680000, 64 * KiB, "pwm3" }, + [FSL_IMX8MP_PWM2] = { 0x30670000, 64 * KiB, "pwm2" }, + [FSL_IMX8MP_PWM1] = { 0x30660000, 64 * KiB, "pwm1" }, + [FSL_IMX8MP_AIPS2_CONFIGURATION] = { 0x305f0000, 64 * KiB, "aips2_configuration" }, + /* AIPS-2 End */ + + /* AIPS-1 Begin */ + [FSL_IMX8MP_CSU] = { 0x303e0000, 64 * KiB, "csu" }, + [FSL_IMX8MP_RDC] = { 0x303d0000, 64 * KiB, "rdc" }, + [FSL_IMX8MP_SEMAPHORE2] = { 0x303c0000, 64 * KiB, "semaphore2" }, + [FSL_IMX8MP_SEMAPHORE1] = { 0x303b0000, 64 * KiB, "semaphore1" }, + [FSL_IMX8MP_GPC] = { 0x303a0000, 64 * KiB, "gpc" }, + [FSL_IMX8MP_SRC] = { 0x30390000, 64 * KiB, "src" }, + [FSL_IMX8MP_CCM] = { 0x30380000, 64 * KiB, "ccm" }, + [FSL_IMX8MP_SNVS_HP] = { 0x30370000, 64 * KiB, "snvs_hp" }, + [FSL_IMX8MP_ANA_PLL] = { 0x30360000, 64 * KiB, "ana_pll" }, + [FSL_IMX8MP_OCOTP_CTRL] = { 0x30350000, 64 * KiB, "ocotp_ctrl" }, + [FSL_IMX8MP_IOMUXC_GPR] = { 0x30340000, 64 * KiB, "iomuxc_gpr" }, + [FSL_IMX8MP_IOMUXC] = { 0x30330000, 64 * KiB, "iomuxc" }, + [FSL_IMX8MP_GPT3] = { 0x302f0000, 64 * KiB, "gpt3" }, + [FSL_IMX8MP_GPT2] = { 0x302e0000, 64 * KiB, "gpt2" }, + [FSL_IMX8MP_GPT1] = { 0x302d0000, 64 * KiB, "gpt1" }, + [FSL_IMX8MP_WDOG3] = { 0x302a0000, 64 * KiB, "wdog3" }, + [FSL_IMX8MP_WDOG2] = { 0x30290000, 64 * KiB, "wdog2" }, + [FSL_IMX8MP_WDOG1] = { 0x30280000, 64 * KiB, "wdog1" }, + [FSL_IMX8MP_ANA_OSC] = { 0x30270000, 64 * KiB, "ana_osc" }, + [FSL_IMX8MP_ANA_TSENSOR] = { 0x30260000, 64 * KiB, "ana_tsensor" }, + [FSL_IMX8MP_GPIO5] = { 0x30240000, 64 * KiB, "gpio5" }, + [FSL_IMX8MP_GPIO4] = { 0x30230000, 64 * KiB, "gpio4" }, + [FSL_IMX8MP_GPIO3] = { 0x30220000, 64 * KiB, "gpio3" }, + [FSL_IMX8MP_GPIO2] = { 0x30210000, 64 * KiB, "gpio2" }, + [FSL_IMX8MP_GPIO1] = { 0x30200000, 64 * KiB, "gpio1" }, + [FSL_IMX8MP_AIPS1_CONFIGURATION] = { 0x301f0000, 64 * KiB, "aips1_configuration" }, + /* AIPS-1 End */ + + [FSL_IMX8MP_A53_DAP] = { 0x28000000, 16 * MiB, "a53_dap" }, + [FSL_IMX8MP_PCIE1_MEM] = { 0x18000000, 128 * MiB, "pcie1_mem" }, + [FSL_IMX8MP_QSPI_MEM] = { 0x08000000, 256 * MiB, "qspi_mem" }, + [FSL_IMX8MP_OCRAM] = { 0x00900000, 576 * KiB, "ocram" }, + [FSL_IMX8MP_TCM_DTCM] = { 0x00800000, 128 * KiB, "tcm_dtcm" }, + [FSL_IMX8MP_TCM_ITCM] = { 0x007e0000, 128 * KiB, "tcm_itcm" }, + [FSL_IMX8MP_OCRAM_S] = { 0x00180000, 36 * KiB, "ocram_s" }, + [FSL_IMX8MP_CAAM_MEM] = { 0x00100000, 32 * KiB, "caam_mem" }, + [FSL_IMX8MP_BOOT_ROM_PROTECTED] = { 0x0003f000, 4 * KiB, "boot_rom_protected" }, + [FSL_IMX8MP_BOOT_ROM] = { 0x00000000, 252 * KiB, "boot_rom" }, +}; + +static void fsl_imx8mp_init(Object *obj) +{ + MachineState *ms = MACHINE(qdev_get_machine()); + FslImx8mpState *s = FSL_IMX8MP(obj); + int i; + + for (i = 0; i < MIN(ms->smp.cpus, FSL_IMX8MP_NUM_CPUS); i++) { + g_autofree char *name = g_strdup_printf("cpu%d", i); + object_initialize_child(obj, name, &s->cpu[i], + ARM_CPU_TYPE_NAME("cortex-a53")); + } + + object_initialize_child(obj, "gic", &s->gic, TYPE_ARM_GICV3); + + for (i = 0; i < FSL_IMX8MP_NUM_UARTS; i++) { + g_autofree char *name = g_strdup_printf("uart%d", i + 1); + object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); + } +} + +static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) +{ + MachineState *ms = MACHINE(qdev_get_machine()); + FslImx8mpState *s = FSL_IMX8MP(dev); + DeviceState *gicdev = DEVICE(&s->gic); + int i; + + if (ms->smp.cpus > FSL_IMX8MP_NUM_CPUS) { + error_setg(errp, "%s: Only %d CPUs are supported (%d requested)", + TYPE_FSL_IMX8MP, FSL_IMX8MP_NUM_CPUS, ms->smp.cpus); + return; + } + + /* CPUs */ + for (i = 0; i < ms->smp.cpus; i++) { + /* On uniprocessor, the CBAR is set to 0 */ + if (ms->smp.cpus > 1) { + object_property_set_int(OBJECT(&s->cpu[i]), "reset-cbar", + fsl_imx8mp_memmap[FSL_IMX8MP_GIC_DIST].addr, + &error_abort); + } + + /* + * CNTFID0 base frequency in Hz of system counter + */ + object_property_set_int(OBJECT(&s->cpu[i]), "cntfrq", 8000000, + &error_abort); + + if (i) { + /* + * Secondary CPUs start in powered-down state (and can be + * powered up via the SRC system reset controller) + */ + object_property_set_bool(OBJECT(&s->cpu[i]), "start-powered-off", + true, &error_abort); + } + + if (!qdev_realize(DEVICE(&s->cpu[i]), NULL, errp)) { + return; + } + } + + /* GIC */ + { + SysBusDevice *gicsbd = SYS_BUS_DEVICE(&s->gic); + QList *redist_region_count; + + qdev_prop_set_uint32(gicdev, "num-cpu", ms->smp.cpus); + qdev_prop_set_uint32(gicdev, "num-irq", + FSL_IMX8MP_NUM_IRQS + GIC_INTERNAL); + redist_region_count = qlist_new(); + qlist_append_int(redist_region_count, ms->smp.cpus); + qdev_prop_set_array(gicdev, "redist-region-count", redist_region_count); + object_property_set_link(OBJECT(&s->gic), "sysmem", + OBJECT(get_system_memory()), &error_fatal); + if (!sysbus_realize(gicsbd, errp)) { + return; + } + sysbus_mmio_map(gicsbd, 0, fsl_imx8mp_memmap[FSL_IMX8MP_GIC_DIST].addr); + sysbus_mmio_map(gicsbd, 1, fsl_imx8mp_memmap[FSL_IMX8MP_GIC_REDIST].addr); + + /* + * Wire the outputs from each CPU's generic timer and the GICv3 + * maintenance interrupt signal to the appropriate GIC PPI inputs, and + * the GIC's IRQ/FIQ interrupt outputs to the CPU's inputs. + */ + for (i = 0; i < ms->smp.cpus; i++) { + DeviceState *cpudev = DEVICE(&s->cpu[i]); + int intidbase = FSL_IMX8MP_NUM_IRQS + i * GIC_INTERNAL; + qemu_irq irq; + + /* + * Mapping from the output timer irq lines from the CPU to the + * GIC PPI inputs. + */ + static const int timer_irqs[] = { + [GTIMER_PHYS] = ARCH_TIMER_NS_EL1_IRQ, + [GTIMER_VIRT] = ARCH_TIMER_VIRT_IRQ, + [GTIMER_HYP] = ARCH_TIMER_NS_EL2_IRQ, + [GTIMER_SEC] = ARCH_TIMER_S_EL1_IRQ, + }; + + for (int j = 0; j < ARRAY_SIZE(timer_irqs); j++) { + irq = qdev_get_gpio_in(gicdev, intidbase + timer_irqs[j]); + qdev_connect_gpio_out(cpudev, j, irq); + } + + irq = qdev_get_gpio_in(gicdev, intidbase + ARCH_GIC_MAINT_IRQ); + qdev_connect_gpio_out_named(cpudev, "gicv3-maintenance-interrupt", + 0, irq); + + irq = qdev_get_gpio_in(gicdev, intidbase + VIRTUAL_PMU_IRQ); + qdev_connect_gpio_out_named(cpudev, "pmu-interrupt", 0, irq); + + sysbus_connect_irq(gicsbd, i, + qdev_get_gpio_in(cpudev, ARM_CPU_IRQ)); + sysbus_connect_irq(gicsbd, i + ms->smp.cpus, + qdev_get_gpio_in(cpudev, ARM_CPU_FIQ)); + } + } + + /* UARTs */ + for (i = 0; i < FSL_IMX8MP_NUM_UARTS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } serial_table[FSL_IMX8MP_NUM_UARTS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_UART1].addr, FSL_IMX8MP_UART1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_UART2].addr, FSL_IMX8MP_UART2_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_UART3].addr, FSL_IMX8MP_UART3_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_UART4].addr, FSL_IMX8MP_UART4_IRQ }, + }; + + qdev_prop_set_chr(DEVICE(&s->uart[i]), "chardev", serial_hd(i)); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->uart[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->uart[i]), 0, serial_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->uart[i]), 0, + qdev_get_gpio_in(gicdev, serial_table[i].irq)); + } + + /* Unimplemented devices */ + for (i = 0; i < ARRAY_SIZE(fsl_imx8mp_memmap); i++) { + switch (i) { + case FSL_IMX8MP_GIC_DIST: + case FSL_IMX8MP_GIC_REDIST: + case FSL_IMX8MP_RAM: + case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: + /* device implemented and treated above */ + break; + + default: + create_unimplemented_device(fsl_imx8mp_memmap[i].name, + fsl_imx8mp_memmap[i].addr, + fsl_imx8mp_memmap[i].size); + break; + } + } +} + +static void fsl_imx8mp_class_init(ObjectClass *oc, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(oc); + + dc->realize = fsl_imx8mp_realize; + + dc->desc = "i.MX 8M Plus SoC"; +} + +static const TypeInfo fsl_imx8mp_types[] = { + { + .name = TYPE_FSL_IMX8MP, + .parent = TYPE_DEVICE, + .instance_size = sizeof(FslImx8mpState), + .instance_init = fsl_imx8mp_init, + .class_init = fsl_imx8mp_class_init, + }, +}; + +DEFINE_TYPES(fsl_imx8mp_types) diff --git a/hw/arm/imx8mp-evk.c b/hw/arm/imx8mp-evk.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/imx8mp-evk.c @@ -XXX,XX +XXX,XX @@ +/* + * NXP i.MX 8M Plus Evaluation Kit System Emulation + * + * Copyright (c) 2024, Bernhard Beschow <shentey@gmail.com> + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "exec/address-spaces.h" +#include "hw/arm/boot.h" +#include "hw/arm/fsl-imx8mp.h" +#include "hw/boards.h" +#include "system/qtest.h" +#include "qemu/error-report.h" +#include "qapi/error.h" + +static void imx8mp_evk_init(MachineState *machine) +{ + static struct arm_boot_info boot_info; + FslImx8mpState *s; + + if (machine->ram_size > FSL_IMX8MP_RAM_SIZE_MAX) { + error_report("RAM size " RAM_ADDR_FMT " above max supported (%08" PRIx64 ")", + machine->ram_size, FSL_IMX8MP_RAM_SIZE_MAX); + exit(1); + } + + boot_info = (struct arm_boot_info) { + .loader_start = FSL_IMX8MP_RAM_START, + .board_id = -1, + .ram_size = machine->ram_size, + .psci_conduit = QEMU_PSCI_CONDUIT_SMC, + }; + + s = FSL_IMX8MP(object_new(TYPE_FSL_IMX8MP)); + object_property_add_child(OBJECT(machine), "soc", OBJECT(s)); + qdev_realize(DEVICE(s), NULL, &error_fatal); + + memory_region_add_subregion(get_system_memory(), FSL_IMX8MP_RAM_START, + machine->ram); + + if (!qtest_enabled()) { + arm_load_kernel(&s->cpu[0], machine, &boot_info); + } +} + +static void imx8mp_evk_machine_init(MachineClass *mc) +{ + mc->desc = "NXP i.MX 8M Plus EVK Board"; + mc->init = imx8mp_evk_init; + mc->max_cpus = FSL_IMX8MP_NUM_CPUS; + mc->default_ram_id = "imx8mp-evk.ram"; +} +DEFINE_MACHINE("imx8mp-evk", imx8mp_evk_machine_init) diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX7 select UNIMP select USB_CHIPIDEA +config FSL_IMX8MP + bool + select ARM_GIC + select IMX + select UNIMP + +config FSL_IMX8MP_EVK + bool + default y + depends on TCG && AARCH64 + select FSL_IMX8MP + config ARM_SMMUV3 bool diff --git a/hw/arm/meson.build b/hw/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/meson.build +++ b/hw/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'CONFIG_MSF2', if_true: files('msf2-soc.c')) arm_ss.add(when: 'CONFIG_MUSCA', if_true: files('musca.c')) arm_ss.add(when: 'CONFIG_ARMSSE', if_true: files('armsse.c')) arm_ss.add(when: 'CONFIG_FSL_IMX7', if_true: files('fsl-imx7.c', 'mcimx7d-sabre.c')) +arm_ss.add(when: 'CONFIG_FSL_IMX8MP', if_true: files('fsl-imx8mp.c')) +arm_ss.add(when: 'CONFIG_FSL_IMX8MP_EVK', if_true: files('imx8mp-evk.c')) arm_ss.add(when: 'CONFIG_ARM_SMMUV3', if_true: files('smmuv3.c')) arm_ss.add(when: 'CONFIG_FSL_IMX6UL', if_true: files('fsl-imx6ul.c', 'mcimx6ul-evk.c')) arm_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_soc.c')) -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Fixes quite a few stack traces during the Linux boot process. Also provides the clocks for devices added later, e.g. enet1. Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-6-shentey@gmail.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 2 + docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 4 + include/hw/misc/imx8mp_analog.h | 81 +++++++++++++++ include/hw/misc/imx8mp_ccm.h | 30 ++++++ hw/arm/fsl-imx8mp.c | 20 ++++ hw/misc/imx8mp_analog.c | 160 +++++++++++++++++++++++++++++ hw/misc/imx8mp_ccm.c | 175 ++++++++++++++++++++++++++++++++ hw/arm/Kconfig | 2 + hw/misc/Kconfig | 6 ++ hw/misc/meson.build | 2 + 11 files changed, 483 insertions(+) create mode 100644 include/hw/misc/imx8mp_analog.h create mode 100644 include/hw/misc/imx8mp_ccm.h create mode 100644 hw/misc/imx8mp_analog.c create mode 100644 hw/misc/imx8mp_ccm.c diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ L: qemu-arm@nongnu.org S: Maintained F: hw/arm/imx8mp-evk.c F: hw/arm/fsl-imx8mp.c +F: hw/misc/imx8mp_*.c F: include/hw/arm/fsl-imx8mp.h +F: include/hw/misc/imx8mp_*.h F: docs/system/arm/imx8mp-evk.rst MPS2 / MPS3 diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * Up to 4 Cortex-A53 cores * Generic Interrupt Controller (GICv3) * 4 UARTs + * Clock Tree Boot options ------------ diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "hw/char/imx_serial.h" #include "hw/intc/arm_gicv3_common.h" +#include "hw/misc/imx8mp_analog.h" +#include "hw/misc/imx8mp_ccm.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { ARMCPU cpu[FSL_IMX8MP_NUM_CPUS]; GICv3State gic; + IMX8MPCCMState ccm; + IMX8MPAnalogState analog; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; }; diff --git a/include/hw/misc/imx8mp_analog.h b/include/hw/misc/imx8mp_analog.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/imx8mp_analog.h @@ -XXX,XX +XXX,XX @@ +/* + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * i.MX8MP ANALOG IP block emulation code + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#ifndef IMX8MP_ANALOG_H +#define IMX8MP_ANALOG_H + +#include "qom/object.h" +#include "hw/sysbus.h" + +enum IMX8MPAnalogRegisters { + ANALOG_AUDIO_PLL1_GEN_CTRL = 0x000 / 4, + ANALOG_AUDIO_PLL1_FDIV_CTL0 = 0x004 / 4, + ANALOG_AUDIO_PLL1_FDIV_CTL1 = 0x008 / 4, + ANALOG_AUDIO_PLL1_SSCG_CTRL = 0x00c / 4, + ANALOG_AUDIO_PLL1_MNIT_CTRL = 0x010 / 4, + ANALOG_AUDIO_PLL2_GEN_CTRL = 0x014 / 4, + ANALOG_AUDIO_PLL2_FDIV_CTL0 = 0x018 / 4, + ANALOG_AUDIO_PLL2_FDIV_CTL1 = 0x01c / 4, + ANALOG_AUDIO_PLL2_SSCG_CTRL = 0x020 / 4, + ANALOG_AUDIO_PLL2_MNIT_CTRL = 0x024 / 4, + ANALOG_VIDEO_PLL1_GEN_CTRL = 0x028 / 4, + ANALOG_VIDEO_PLL1_FDIV_CTL0 = 0x02c / 4, + ANALOG_VIDEO_PLL1_FDIV_CTL1 = 0x030 / 4, + ANALOG_VIDEO_PLL1_SSCG_CTRL = 0x034 / 4, + ANALOG_VIDEO_PLL1_MNIT_CTRL = 0x038 / 4, + ANALOG_DRAM_PLL_GEN_CTRL = 0x050 / 4, + ANALOG_DRAM_PLL_FDIV_CTL0 = 0x054 / 4, + ANALOG_DRAM_PLL_FDIV_CTL1 = 0x058 / 4, + ANALOG_DRAM_PLL_SSCG_CTRL = 0x05c / 4, + ANALOG_DRAM_PLL_MNIT_CTRL = 0x060 / 4, + ANALOG_GPU_PLL_GEN_CTRL = 0x064 / 4, + ANALOG_GPU_PLL_FDIV_CTL0 = 0x068 / 4, + ANALOG_GPU_PLL_LOCKD_CTRL = 0x06c / 4, + ANALOG_GPU_PLL_MNIT_CTRL = 0x070 / 4, + ANALOG_VPU_PLL_GEN_CTRL = 0x074 / 4, + ANALOG_VPU_PLL_FDIV_CTL0 = 0x078 / 4, + ANALOG_VPU_PLL_LOCKD_CTRL = 0x07c / 4, + ANALOG_VPU_PLL_MNIT_CTRL = 0x080 / 4, + ANALOG_ARM_PLL_GEN_CTRL = 0x084 / 4, + ANALOG_ARM_PLL_FDIV_CTL0 = 0x088 / 4, + ANALOG_ARM_PLL_LOCKD_CTRL = 0x08c / 4, + ANALOG_ARM_PLL_MNIT_CTRL = 0x090 / 4, + ANALOG_SYS_PLL1_GEN_CTRL = 0x094 / 4, + ANALOG_SYS_PLL1_FDIV_CTL0 = 0x098 / 4, + ANALOG_SYS_PLL1_LOCKD_CTRL = 0x09c / 4, + ANALOG_SYS_PLL1_MNIT_CTRL = 0x100 / 4, + ANALOG_SYS_PLL2_GEN_CTRL = 0x104 / 4, + ANALOG_SYS_PLL2_FDIV_CTL0 = 0x108 / 4, + ANALOG_SYS_PLL2_LOCKD_CTRL = 0x10c / 4, + ANALOG_SYS_PLL2_MNIT_CTRL = 0x110 / 4, + ANALOG_SYS_PLL3_GEN_CTRL = 0x114 / 4, + ANALOG_SYS_PLL3_FDIV_CTL0 = 0x118 / 4, + ANALOG_SYS_PLL3_LOCKD_CTRL = 0x11c / 4, + ANALOG_SYS_PLL3_MNIT_CTRL = 0x120 / 4, + ANALOG_OSC_MISC_CFG = 0x124 / 4, + ANALOG_ANAMIX_PLL_MNIT_CTL = 0x128 / 4, + + ANALOG_DIGPROG = 0x800 / 4, + ANALOG_MAX, +}; + +#define TYPE_IMX8MP_ANALOG "imx8mp.analog" +OBJECT_DECLARE_SIMPLE_TYPE(IMX8MPAnalogState, IMX8MP_ANALOG) + +struct IMX8MPAnalogState { + SysBusDevice parent_obj; + + struct { + MemoryRegion container; + MemoryRegion analog; + } mmio; + + uint32_t analog[ANALOG_MAX]; +}; + +#endif /* IMX8MP_ANALOG_H */ diff --git a/include/hw/misc/imx8mp_ccm.h b/include/hw/misc/imx8mp_ccm.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/imx8mp_ccm.h @@ -XXX,XX +XXX,XX @@ +/* + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * i.MX 8M Plus CCM IP block emulation code + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#ifndef IMX8MP_CCM_H +#define IMX8MP_CCM_H + +#include "hw/misc/imx_ccm.h" +#include "qom/object.h" + +enum IMX8MPCCMRegisters { + CCM_MAX = 0xc6fc / sizeof(uint32_t) + 1, +}; + +#define TYPE_IMX8MP_CCM "imx8mp.ccm" +OBJECT_DECLARE_SIMPLE_TYPE(IMX8MPCCMState, IMX8MP_CCM) + +struct IMX8MPCCMState { + IMXCCMState parent_obj; + + MemoryRegion iomem; + + uint32_t ccm[CCM_MAX]; +}; + +#endif /* IMX8MP_CCM_H */ diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, "gic", &s->gic, TYPE_ARM_GICV3); + object_initialize_child(obj, "ccm", &s->ccm, TYPE_IMX8MP_CCM); + + object_initialize_child(obj, "analog", &s->analog, TYPE_IMX8MP_ANALOG); + for (i = 0; i < FSL_IMX8MP_NUM_UARTS; i++) { g_autofree char *name = g_strdup_printf("uart%d", i + 1); object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) } } + /* CCM */ + if (!sysbus_realize(SYS_BUS_DEVICE(&s->ccm), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->ccm), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_CCM].addr); + + /* Analog */ + if (!sysbus_realize(SYS_BUS_DEVICE(&s->analog), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->analog), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_ANA_PLL].addr); + /* UARTs */ for (i = 0; i < FSL_IMX8MP_NUM_UARTS; i++) { struct { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) /* Unimplemented devices */ for (i = 0; i < ARRAY_SIZE(fsl_imx8mp_memmap); i++) { switch (i) { + case FSL_IMX8MP_ANA_PLL: + case FSL_IMX8MP_CCM: case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: case FSL_IMX8MP_RAM: diff --git a/hw/misc/imx8mp_analog.c b/hw/misc/imx8mp_analog.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/imx8mp_analog.c @@ -XXX,XX +XXX,XX @@ +/* + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * i.MX 8M Plus ANALOG IP block emulation code + * + * Based on hw/misc/imx7_ccm.c + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" + +#include "hw/misc/imx8mp_analog.h" +#include "migration/vmstate.h" + +#define ANALOG_PLL_LOCK BIT(31) + +static void imx8mp_analog_reset(DeviceState *dev) +{ + IMX8MPAnalogState *s = IMX8MP_ANALOG(dev); + + memset(s->analog, 0, sizeof(s->analog)); + + s->analog[ANALOG_AUDIO_PLL1_GEN_CTRL] = 0x00002010; + s->analog[ANALOG_AUDIO_PLL1_FDIV_CTL0] = 0x00145032; + s->analog[ANALOG_AUDIO_PLL1_FDIV_CTL1] = 0x00000000; + s->analog[ANALOG_AUDIO_PLL1_SSCG_CTRL] = 0x00000000; + s->analog[ANALOG_AUDIO_PLL1_MNIT_CTRL] = 0x00100103; + s->analog[ANALOG_AUDIO_PLL2_GEN_CTRL] = 0x00002010; + s->analog[ANALOG_AUDIO_PLL2_FDIV_CTL0] = 0x00145032; + s->analog[ANALOG_AUDIO_PLL2_FDIV_CTL1] = 0x00000000; + s->analog[ANALOG_AUDIO_PLL2_SSCG_CTRL] = 0x00000000; + s->analog[ANALOG_AUDIO_PLL2_MNIT_CTRL] = 0x00100103; + s->analog[ANALOG_VIDEO_PLL1_GEN_CTRL] = 0x00002010; + s->analog[ANALOG_VIDEO_PLL1_FDIV_CTL0] = 0x00145032; + s->analog[ANALOG_VIDEO_PLL1_FDIV_CTL1] = 0x00000000; + s->analog[ANALOG_VIDEO_PLL1_SSCG_CTRL] = 0x00000000; + s->analog[ANALOG_VIDEO_PLL1_MNIT_CTRL] = 0x00100103; + s->analog[ANALOG_DRAM_PLL_GEN_CTRL] = 0x00002010; + s->analog[ANALOG_DRAM_PLL_FDIV_CTL0] = 0x0012c032; + s->analog[ANALOG_DRAM_PLL_FDIV_CTL1] = 0x00000000; + s->analog[ANALOG_DRAM_PLL_SSCG_CTRL] = 0x00000000; + s->analog[ANALOG_DRAM_PLL_MNIT_CTRL] = 0x00100103; + s->analog[ANALOG_GPU_PLL_GEN_CTRL] = 0x00000810; + s->analog[ANALOG_GPU_PLL_FDIV_CTL0] = 0x000c8031; + s->analog[ANALOG_GPU_PLL_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_GPU_PLL_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_VPU_PLL_GEN_CTRL] = 0x00000810; + s->analog[ANALOG_VPU_PLL_FDIV_CTL0] = 0x0012c032; + s->analog[ANALOG_VPU_PLL_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_VPU_PLL_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_ARM_PLL_GEN_CTRL] = 0x00000810; + s->analog[ANALOG_ARM_PLL_FDIV_CTL0] = 0x000fa031; + s->analog[ANALOG_ARM_PLL_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_ARM_PLL_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_SYS_PLL1_GEN_CTRL] = 0x0aaaa810; + s->analog[ANALOG_SYS_PLL1_FDIV_CTL0] = 0x00190032; + s->analog[ANALOG_SYS_PLL1_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_SYS_PLL1_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_SYS_PLL2_GEN_CTRL] = 0x0aaaa810; + s->analog[ANALOG_SYS_PLL2_FDIV_CTL0] = 0x000fa031; + s->analog[ANALOG_SYS_PLL2_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_SYS_PLL2_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_SYS_PLL3_GEN_CTRL] = 0x00000810; + s->analog[ANALOG_SYS_PLL3_FDIV_CTL0] = 0x000fa031; + s->analog[ANALOG_SYS_PLL3_LOCKD_CTRL] = 0x0010003f; + s->analog[ANALOG_SYS_PLL3_MNIT_CTRL] = 0x00280081; + s->analog[ANALOG_OSC_MISC_CFG] = 0x00000000; + s->analog[ANALOG_ANAMIX_PLL_MNIT_CTL] = 0x00000000; + s->analog[ANALOG_DIGPROG] = 0x00824010; + + /* all PLLs need to be locked */ + s->analog[ANALOG_AUDIO_PLL1_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_AUDIO_PLL2_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_VIDEO_PLL1_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_DRAM_PLL_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_GPU_PLL_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_VPU_PLL_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_ARM_PLL_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_SYS_PLL1_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_SYS_PLL2_GEN_CTRL] |= ANALOG_PLL_LOCK; + s->analog[ANALOG_SYS_PLL3_GEN_CTRL] |= ANALOG_PLL_LOCK; +} + +static uint64_t imx8mp_analog_read(void *opaque, hwaddr offset, unsigned size) +{ + IMX8MPAnalogState *s = opaque; + + return s->analog[offset >> 2]; +} + +static void imx8mp_analog_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + IMX8MPAnalogState *s = opaque; + + if (offset >> 2 == ANALOG_DIGPROG) { + qemu_log_mask(LOG_GUEST_ERROR, + "Guest write to read-only ANALOG_DIGPROG register\n"); + } else { + s->analog[offset >> 2] = value; + } +} + +static const struct MemoryRegionOps imx8mp_analog_ops = { + .read = imx8mp_analog_read, + .write = imx8mp_analog_write, + .endianness = DEVICE_NATIVE_ENDIAN, + .impl = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void imx8mp_analog_init(Object *obj) +{ + IMX8MPAnalogState *s = IMX8MP_ANALOG(obj); + SysBusDevice *sd = SYS_BUS_DEVICE(obj); + + memory_region_init(&s->mmio.container, obj, TYPE_IMX8MP_ANALOG, 0x10000); + + memory_region_init_io(&s->mmio.analog, obj, &imx8mp_analog_ops, s, + TYPE_IMX8MP_ANALOG, sizeof(s->analog)); + memory_region_add_subregion(&s->mmio.container, 0, &s->mmio.analog); + + sysbus_init_mmio(sd, &s->mmio.container); +} + +static const VMStateDescription imx8mp_analog_vmstate = { + .name = TYPE_IMX8MP_ANALOG, + .version_id = 1, + .minimum_version_id = 1, + .fields = (const VMStateField[]) { + VMSTATE_UINT32_ARRAY(analog, IMX8MPAnalogState, ANALOG_MAX), + VMSTATE_END_OF_LIST() + }, +}; + +static void imx8mp_analog_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + device_class_set_legacy_reset(dc, imx8mp_analog_reset); + dc->vmsd = &imx8mp_analog_vmstate; + dc->desc = "i.MX 8M Plus Analog Module"; +} + +static const TypeInfo imx8mp_analog_types[] = { + { + .name = TYPE_IMX8MP_ANALOG, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(IMX8MPAnalogState), + .instance_init = imx8mp_analog_init, + .class_init = imx8mp_analog_class_init, + } +}; + +DEFINE_TYPES(imx8mp_analog_types); diff --git a/hw/misc/imx8mp_ccm.c b/hw/misc/imx8mp_ccm.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/imx8mp_ccm.c @@ -XXX,XX +XXX,XX @@ +/* + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * i.MX 8M Plus CCM IP block emulation code + * + * Based on hw/misc/imx7_ccm.c + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" + +#include "hw/misc/imx8mp_ccm.h" +#include "migration/vmstate.h" + +#include "trace.h" + +#define CKIH_FREQ 16000000 /* 16MHz crystal input */ + +static void imx8mp_ccm_reset(DeviceState *dev) +{ + IMX8MPCCMState *s = IMX8MP_CCM(dev); + + memset(s->ccm, 0, sizeof(s->ccm)); +} + +#define CCM_INDEX(offset) (((offset) & ~(hwaddr)0xF) / sizeof(uint32_t)) +#define CCM_BITOP(offset) ((offset) & (hwaddr)0xF) + +enum { + CCM_BITOP_NONE = 0x00, + CCM_BITOP_SET = 0x04, + CCM_BITOP_CLR = 0x08, + CCM_BITOP_TOG = 0x0C, +}; + +static uint64_t imx8mp_set_clr_tog_read(void *opaque, hwaddr offset, + unsigned size) +{ + const uint32_t *mmio = opaque; + + return mmio[CCM_INDEX(offset)]; +} + +static void imx8mp_set_clr_tog_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + const uint8_t bitop = CCM_BITOP(offset); + const uint32_t index = CCM_INDEX(offset); + uint32_t *mmio = opaque; + + switch (bitop) { + case CCM_BITOP_NONE: + mmio[index] = value; + break; + case CCM_BITOP_SET: + mmio[index] |= value; + break; + case CCM_BITOP_CLR: + mmio[index] &= ~value; + break; + case CCM_BITOP_TOG: + mmio[index] ^= value; + break; + }; +} + +static const struct MemoryRegionOps imx8mp_set_clr_tog_ops = { + .read = imx8mp_set_clr_tog_read, + .write = imx8mp_set_clr_tog_write, + .endianness = DEVICE_NATIVE_ENDIAN, + .impl = { + /* + * Our device would not work correctly if the guest was doing + * unaligned access. This might not be a limitation on the real + * device but in practice there is no reason for a guest to access + * this device unaligned. + */ + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void imx8mp_ccm_init(Object *obj) +{ + SysBusDevice *sd = SYS_BUS_DEVICE(obj); + IMX8MPCCMState *s = IMX8MP_CCM(obj); + + memory_region_init_io(&s->iomem, + obj, + &imx8mp_set_clr_tog_ops, + s->ccm, + TYPE_IMX8MP_CCM ".ccm", + sizeof(s->ccm)); + + sysbus_init_mmio(sd, &s->iomem); +} + +static const VMStateDescription imx8mp_ccm_vmstate = { + .name = TYPE_IMX8MP_CCM, + .version_id = 1, + .minimum_version_id = 1, + .fields = (const VMStateField[]) { + VMSTATE_UINT32_ARRAY(ccm, IMX8MPCCMState, CCM_MAX), + VMSTATE_END_OF_LIST() + }, +}; + +static uint32_t imx8mp_ccm_get_clock_frequency(IMXCCMState *dev, IMXClk clock) +{ + /* + * This function is "consumed" by GPT emulation code. Some clocks + * have fixed frequencies and we can provide requested frequency + * easily. However for CCM provided clocks (like IPG) each GPT + * timer can have its own clock root. + * This means we need additional information when calling this + * function to know the requester's identity. + */ + uint32_t freq = 0; + + switch (clock) { + case CLK_NONE: + break; + case CLK_32k: + freq = CKIL_FREQ; + break; + case CLK_HIGH: + freq = CKIH_FREQ; + break; + case CLK_IPG: + case CLK_IPG_HIGH: + /* + * For now we don't have a way to figure out the device this + * function is called for. Until then the IPG derived clocks + * are left unimplemented. + */ + qemu_log_mask(LOG_GUEST_ERROR, "[%s]%s: Clock %d Not implemented\n", + TYPE_IMX8MP_CCM, __func__, clock); + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, "[%s]%s: unsupported clock %d\n", + TYPE_IMX8MP_CCM, __func__, clock); + break; + } + + trace_ccm_clock_freq(clock, freq); + + return freq; +} + +static void imx8mp_ccm_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + IMXCCMClass *ccm = IMX_CCM_CLASS(klass); + + device_class_set_legacy_reset(dc, imx8mp_ccm_reset); + dc->vmsd = &imx8mp_ccm_vmstate; + dc->desc = "i.MX 8M Plus Clock Control Module"; + + ccm->get_clock_frequency = imx8mp_ccm_get_clock_frequency; +} + +static const TypeInfo imx8mp_ccm_types[] = { + { + .name = TYPE_IMX8MP_CCM, + .parent = TYPE_IMX_CCM, + .instance_size = sizeof(IMX8MPCCMState), + .instance_init = imx8mp_ccm_init, + .class_init = imx8mp_ccm_class_init, + }, +}; + +DEFINE_TYPES(imx8mp_ccm_types); diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX7 config FSL_IMX8MP bool select ARM_GIC + select FSL_IMX8MP_ANALOG + select FSL_IMX8MP_CCM select IMX select UNIMP diff --git a/hw/misc/Kconfig b/hw/misc/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Kconfig +++ b/hw/misc/Kconfig @@ -XXX,XX +XXX,XX @@ config IMX select SSI select USB_EHCI_SYSBUS +config FSL_IMX8MP_ANALOG + bool + +config FSL_IMX8MP_CCM + bool + config STM32_RCC bool diff --git a/hw/misc/meson.build b/hw/misc/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/meson.build +++ b/hw/misc/meson.build @@ -XXX,XX +XXX,XX @@ system_ss.add(when: 'CONFIG_AXP2XX_PMU', if_true: files('axp2xx.c')) system_ss.add(when: 'CONFIG_REALVIEW', if_true: files('arm_sysctl.c')) system_ss.add(when: 'CONFIG_ECCMEMCTL', if_true: files('eccmemctl.c')) system_ss.add(when: 'CONFIG_EXYNOS4', if_true: files('exynos4210_pmu.c', 'exynos4210_clk.c', 'exynos4210_rng.c')) +system_ss.add(when: 'CONFIG_FSL_IMX8MP_ANALOG', if_true: files('imx8mp_analog.c')) +system_ss.add(when: 'CONFIG_FSL_IMX8MP_CCM', if_true: files('imx8mp_ccm.c')) system_ss.add(when: 'CONFIG_IMX', if_true: files( 'imx25_ccm.c', 'imx31_ccm.c', -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> SNVS contains an RTC which allows Linux to deal correctly with time. This is particularly useful when handling persistent storage which will be done in the next patch. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-7-shentey@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 2 ++ hw/arm/fsl-imx8mp.c | 10 ++++++++++ 3 files changed, 13 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * Up to 4 Cortex-A53 cores * Generic Interrupt Controller (GICv3) * 4 UARTs + * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree Boot options diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "hw/char/imx_serial.h" #include "hw/intc/arm_gicv3_common.h" +#include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" #include "hw/misc/imx8mp_ccm.h" #include "qom/object.h" @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { GICv3State gic; IMX8MPCCMState ccm; IMX8MPAnalogState analog; + IMX7SNVSState snvs; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; }; diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, "analog", &s->analog, TYPE_IMX8MP_ANALOG); + object_initialize_child(obj, "snvs", &s->snvs, TYPE_IMX7_SNVS); + for (i = 0; i < FSL_IMX8MP_NUM_UARTS; i++) { g_autofree char *name = g_strdup_printf("uart%d", i + 1); object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, serial_table[i].irq)); } + /* SNVS */ + if (!sysbus_realize(SYS_BUS_DEVICE(&s->snvs), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->snvs), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_SNVS_HP].addr); + /* Unimplemented devices */ for (i = 0; i < ARRAY_SIZE(fsl_imx8mp_memmap); i++) { switch (i) { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: case FSL_IMX8MP_RAM: + case FSL_IMX8MP_SNVS_HP: case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: /* device implemented and treated above */ break; -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> The USDHC emulation allows for running real-world images such as those generated by Buildroot. Convert the board documentation accordingly instead of running a Linux kernel with ephemeral storage. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-8-shentey@gmail.com [PMM: drop 'static const' from usdhc_table[] for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 18 ++++++++++++------ include/hw/arm/fsl-imx8mp.h | 7 +++++++ hw/arm/fsl-imx8mp.c | 28 ++++++++++++++++++++++++++++ hw/arm/imx8mp-evk.c | 18 ++++++++++++++++++ hw/arm/Kconfig | 1 + 5 files changed, 66 insertions(+), 6 deletions(-) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * Up to 4 Cortex-A53 cores * Generic Interrupt Controller (GICv3) * 4 UARTs + * 3 USDHC Storage Controllers * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree @@ -XXX,XX +XXX,XX @@ Direct Linux Kernel Boot Probably the easiest way to get started with a whole Linux system on the machine is to generate an image with Buildroot. Version 2024.11.1 is tested at the time -of writing and involves two steps. First run the following commands in the +of writing and involves three steps. First run the following commands in the toplevel directory of the Buildroot source tree: .. code-block:: bash - $ echo "BR2_TARGET_ROOTFS_CPIO=y" >> configs/freescale_imx8mpevk_defconfig $ make freescale_imx8mpevk_defconfig $ make Once finished successfully there is an ``output/image`` subfolder. Navigate into -it and patch the device tree with the following commands which will remove the -``cpu-idle-states`` properties from CPU nodes: +it and resize the SD card image to a power of two: + +.. code-block:: bash + + $ qemu-img resize sdcard.img 256M + +Finally, the device tree needs to be patched with the following commands which +will remove the ``cpu-idle-states`` properties from CPU nodes: .. code-block:: bash @@ -XXX,XX +XXX,XX @@ Now that everything is prepared the machine can be started as follows: -display none -serial null -serial stdio \ -kernel Image \ -dtb imx8mp-evk-patched.dtb \ - -initrd rootfs.cpio \ - -append "root=/dev/ram" + -append "root=/dev/mmcblk2p2" \ + -drive file=sdcard.img,if=sd,bus=2,format=raw,id=mmcblk2 diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" #include "hw/misc/imx8mp_ccm.h" +#include "hw/sd/sdhci.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_CPUS = 4, FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, + FSL_IMX8MP_NUM_USDHCS = 3, }; struct FslImx8mpState { @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMX8MPAnalogState analog; IMX7SNVSState snvs; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; + SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; }; enum FslImx8mpMemoryRegions { @@ -XXX,XX +XXX,XX @@ enum FslImx8mpMemoryRegions { }; enum FslImx8mpIrqs { + FSL_IMX8MP_USDHC1_IRQ = 22, + FSL_IMX8MP_USDHC2_IRQ = 23, + FSL_IMX8MP_USDHC3_IRQ = 24, + FSL_IMX8MP_UART1_IRQ = 26, FSL_IMX8MP_UART2_IRQ = 27, FSL_IMX8MP_UART3_IRQ = 28, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) g_autofree char *name = g_strdup_printf("uart%d", i + 1); object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); } + + for (i = 0; i < FSL_IMX8MP_NUM_USDHCS; i++) { + g_autofree char *name = g_strdup_printf("usdhc%d", i + 1); + object_initialize_child(obj, name, &s->usdhc[i], TYPE_IMX_USDHC); + } } static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, serial_table[i].irq)); } + /* USDHCs */ + for (i = 0; i < FSL_IMX8MP_NUM_USDHCS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } usdhc_table[FSL_IMX8MP_NUM_USDHCS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_USDHC1].addr, FSL_IMX8MP_USDHC1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_USDHC2].addr, FSL_IMX8MP_USDHC2_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_USDHC3].addr, FSL_IMX8MP_USDHC3_IRQ }, + }; + + object_property_set_uint(OBJECT(&s->usdhc[i]), "vendor", + SDHCI_VENDOR_IMX, &error_abort); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->usdhc[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->usdhc[i]), 0, usdhc_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->usdhc[i]), 0, + qdev_get_gpio_in(gicdev, usdhc_table[i].irq)); + } + /* SNVS */ if (!sysbus_realize(SYS_BUS_DEVICE(&s->snvs), errp)) { return; @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_RAM: case FSL_IMX8MP_SNVS_HP: case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: + case FSL_IMX8MP_USDHC1 ... FSL_IMX8MP_USDHC3: /* device implemented and treated above */ break; diff --git a/hw/arm/imx8mp-evk.c b/hw/arm/imx8mp-evk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/imx8mp-evk.c +++ b/hw/arm/imx8mp-evk.c @@ -XXX,XX +XXX,XX @@ #include "hw/arm/boot.h" #include "hw/arm/fsl-imx8mp.h" #include "hw/boards.h" +#include "hw/qdev-properties.h" #include "system/qtest.h" #include "qemu/error-report.h" #include "qapi/error.h" @@ -XXX,XX +XXX,XX @@ static void imx8mp_evk_init(MachineState *machine) memory_region_add_subregion(get_system_memory(), FSL_IMX8MP_RAM_START, machine->ram); + for (int i = 0; i < FSL_IMX8MP_NUM_USDHCS; i++) { + BusState *bus; + DeviceState *carddev; + BlockBackend *blk; + DriveInfo *di = drive_get(IF_SD, i, 0); + + if (!di) { + continue; + } + + blk = blk_by_legacy_dinfo(di); + bus = qdev_get_child_bus(DEVICE(&s->usdhc[i]), "sd-bus"); + carddev = qdev_new(TYPE_SD_CARD); + qdev_prop_set_drive_err(carddev, "drive", blk, &error_fatal); + qdev_realize_and_unref(carddev, bus, &error_fatal); + } + if (!qtest_enabled()) { arm_load_kernel(&s->cpu[0], machine, &boot_info); } diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX8MP select FSL_IMX8MP_ANALOG select FSL_IMX8MP_CCM select IMX + select SDHCI select UNIMP config FSL_IMX8MP_EVK -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Linux checks for the PLLs in the PHY to be locked, so implement a model emulating that. Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-9-shentey@gmail.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 2 + docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 10 +++ include/hw/pci-host/fsl_imx8m_phy.h | 28 +++++++++ hw/arm/fsl-imx8mp.c | 30 +++++++++ hw/pci-host/fsl_imx8m_phy.c | 98 +++++++++++++++++++++++++++++ hw/arm/Kconfig | 3 + hw/pci-host/Kconfig | 3 + hw/pci-host/meson.build | 1 + 9 files changed, 176 insertions(+) create mode 100644 include/hw/pci-host/fsl_imx8m_phy.h create mode 100644 hw/pci-host/fsl_imx8m_phy.c diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ S: Maintained F: hw/arm/imx8mp-evk.c F: hw/arm/fsl-imx8mp.c F: hw/misc/imx8mp_*.c +F: hw/pci-host/fsl_imx8m_phy.c F: include/hw/arm/fsl-imx8mp.h F: include/hw/misc/imx8mp_*.h +F: include/hw/pci-host/fsl_imx8m_phy.h F: docs/system/arm/imx8mp-evk.rst MPS2 / MPS3 diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * Generic Interrupt Controller (GICv3) * 4 UARTs * 3 USDHC Storage Controllers + * 1 Designware PCI Express Controller * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" #include "hw/misc/imx8mp_ccm.h" +#include "hw/pci-host/designware.h" +#include "hw/pci-host/fsl_imx8m_phy.h" #include "hw/sd/sdhci.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMX7SNVSState snvs; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; + DesignwarePCIEHost pcie; + FslImx8mPciePhyState pcie_phy; }; enum FslImx8mpMemoryRegions { @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_UART4_IRQ = 29, FSL_IMX8MP_UART5_IRQ = 30, FSL_IMX8MP_UART6_IRQ = 16, + + FSL_IMX8MP_PCI_INTA_IRQ = 126, + FSL_IMX8MP_PCI_INTB_IRQ = 125, + FSL_IMX8MP_PCI_INTC_IRQ = 124, + FSL_IMX8MP_PCI_INTD_IRQ = 123, + FSL_IMX8MP_PCI_MSI_IRQ = 140, }; #endif /* FSL_IMX8MP_H */ diff --git a/include/hw/pci-host/fsl_imx8m_phy.h b/include/hw/pci-host/fsl_imx8m_phy.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/pci-host/fsl_imx8m_phy.h @@ -XXX,XX +XXX,XX @@ +/* + * i.MX8 PCIe PHY emulation + * + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#ifndef HW_PCIHOST_FSLIMX8MPCIEPHY_H +#define HW_PCIHOST_FSLIMX8MPCIEPHY_H + +#include "hw/sysbus.h" +#include "qom/object.h" +#include "exec/memory.h" + +#define TYPE_FSL_IMX8M_PCIE_PHY "fsl-imx8m-pcie-phy" +OBJECT_DECLARE_SIMPLE_TYPE(FslImx8mPciePhyState, FSL_IMX8M_PCIE_PHY) + +#define FSL_IMX8M_PCIE_PHY_DATA_SIZE 0x800 + +struct FslImx8mPciePhyState { + SysBusDevice parent_obj; + + MemoryRegion iomem; + uint8_t data[FSL_IMX8M_PCIE_PHY_DATA_SIZE]; +}; + +#endif diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) g_autofree char *name = g_strdup_printf("usdhc%d", i + 1); object_initialize_child(obj, name, &s->usdhc[i], TYPE_IMX_USDHC); } + + object_initialize_child(obj, "pcie", &s->pcie, TYPE_DESIGNWARE_PCIE_HOST); + object_initialize_child(obj, "pcie_phy", &s->pcie_phy, + TYPE_FSL_IMX8M_PCIE_PHY); } static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) sysbus_mmio_map(SYS_BUS_DEVICE(&s->snvs), 0, fsl_imx8mp_memmap[FSL_IMX8MP_SNVS_HP].addr); + /* PCIe */ + if (!sysbus_realize(SYS_BUS_DEVICE(&s->pcie), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->pcie), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_PCIE1].addr); + + sysbus_connect_irq(SYS_BUS_DEVICE(&s->pcie), 0, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_PCI_INTA_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->pcie), 1, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_PCI_INTB_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->pcie), 2, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_PCI_INTC_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->pcie), 3, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_PCI_INTD_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->pcie), 4, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_PCI_MSI_IRQ)); + + if (!sysbus_realize(SYS_BUS_DEVICE(&s->pcie_phy), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->pcie_phy), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_PCIE_PHY1].addr); + /* Unimplemented devices */ for (i = 0; i < ARRAY_SIZE(fsl_imx8mp_memmap); i++) { switch (i) { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_CCM: case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: + case FSL_IMX8MP_PCIE1: + case FSL_IMX8MP_PCIE_PHY1: case FSL_IMX8MP_RAM: case FSL_IMX8MP_SNVS_HP: case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: diff --git a/hw/pci-host/fsl_imx8m_phy.c b/hw/pci-host/fsl_imx8m_phy.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/pci-host/fsl_imx8m_phy.c @@ -XXX,XX +XXX,XX @@ +/* + * i.MX8 PCIe PHY emulation + * + * Copyright (c) 2025 Bernhard Beschow <shentey@gmail.com> + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "hw/pci-host/fsl_imx8m_phy.h" +#include "hw/resettable.h" +#include "migration/vmstate.h" + +#define CMN_REG075 0x1d4 +#define ANA_PLL_LOCK_DONE BIT(1) +#define ANA_PLL_AFC_DONE BIT(0) + +static uint64_t fsl_imx8m_pcie_phy_read(void *opaque, hwaddr offset, + unsigned size) +{ + FslImx8mPciePhyState *s = opaque; + + if (offset == CMN_REG075) { + return s->data[offset] | ANA_PLL_LOCK_DONE | ANA_PLL_AFC_DONE; + } + + return s->data[offset]; +} + +static void fsl_imx8m_pcie_phy_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + FslImx8mPciePhyState *s = opaque; + + s->data[offset] = value; +} + +static const MemoryRegionOps fsl_imx8m_pcie_phy_ops = { + .read = fsl_imx8m_pcie_phy_read, + .write = fsl_imx8m_pcie_phy_write, + .impl = { + .min_access_size = 1, + .max_access_size = 1, + }, + .valid = { + .min_access_size = 1, + .max_access_size = 8, + }, + .endianness = DEVICE_LITTLE_ENDIAN, +}; + +static void fsl_imx8m_pcie_phy_realize(DeviceState *dev, Error **errp) +{ + FslImx8mPciePhyState *s = FSL_IMX8M_PCIE_PHY(dev); + + memory_region_init_io(&s->iomem, OBJECT(s), &fsl_imx8m_pcie_phy_ops, s, + TYPE_FSL_IMX8M_PCIE_PHY, ARRAY_SIZE(s->data)); + sysbus_init_mmio(SYS_BUS_DEVICE(s), &s->iomem); +} + +static void fsl_imx8m_pcie_phy_reset_hold(Object *obj, ResetType type) +{ + FslImx8mPciePhyState *s = FSL_IMX8M_PCIE_PHY(obj); + + memset(s->data, 0, sizeof(s->data)); +} + +static const VMStateDescription fsl_imx8m_pcie_phy_vmstate = { + .name = "fsl-imx8m-pcie-phy", + .version_id = 1, + .minimum_version_id = 1, + .fields = (const VMStateField[]) { + VMSTATE_UINT8_ARRAY(data, FslImx8mPciePhyState, + FSL_IMX8M_PCIE_PHY_DATA_SIZE), + VMSTATE_END_OF_LIST() + } +}; + +static void fsl_imx8m_pcie_phy_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + ResettableClass *rc = RESETTABLE_CLASS(klass); + + dc->realize = fsl_imx8m_pcie_phy_realize; + dc->vmsd = &fsl_imx8m_pcie_phy_vmstate; + rc->phases.hold = fsl_imx8m_pcie_phy_reset_hold; +} + +static const TypeInfo fsl_imx8m_pcie_phy_types[] = { + { + .name = TYPE_FSL_IMX8M_PCIE_PHY, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(FslImx8mPciePhyState), + .class_init = fsl_imx8m_pcie_phy_class_init, + } +}; + +DEFINE_TYPES(fsl_imx8m_pcie_phy_types) diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX7 config FSL_IMX8MP bool + imply PCI_DEVICES select ARM_GIC select FSL_IMX8MP_ANALOG select FSL_IMX8MP_CCM select IMX + select PCI_EXPRESS_DESIGNWARE + select PCI_EXPRESS_FSL_IMX8M_PHY select SDHCI select UNIMP diff --git a/hw/pci-host/Kconfig b/hw/pci-host/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/pci-host/Kconfig +++ b/hw/pci-host/Kconfig @@ -XXX,XX +XXX,XX @@ config ASTRO bool select PCI +config PCI_EXPRESS_FSL_IMX8M_PHY + bool + config GT64120 bool select PCI diff --git a/hw/pci-host/meson.build b/hw/pci-host/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/pci-host/meson.build +++ b/hw/pci-host/meson.build @@ -XXX,XX +XXX,XX @@ pci_ss.add(when: 'CONFIG_ARTICIA', if_true: files('articia.c')) pci_ss.add(when: 'CONFIG_MV64361', if_true: files('mv64361.c')) # ARM devices +pci_ss.add(when: 'CONFIG_PCI_EXPRESS_FSL_IMX8M_PHY', if_true: files('fsl_imx8m_phy.c')) pci_ss.add(when: 'CONFIG_VERSATILE_PCI', if_true: files('versatile.c')) # HPPA devices -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-10-shentey@gmail.com [PMM: drop static const from gpio_table for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 14 +++++++++ hw/arm/fsl-imx8mp.c | 55 ++++++++++++++++++++++++++++++++++ 3 files changed, 70 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 4 UARTs * 3 USDHC Storage Controllers * 1 Designware PCI Express Controller + * 5 GPIO Controllers * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "hw/char/imx_serial.h" +#include "hw/gpio/imx_gpio.h" #include "hw/intc/arm_gicv3_common.h" #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" @@ -XXX,XX +XXX,XX @@ OBJECT_DECLARE_SIMPLE_TYPE(FslImx8mpState, FSL_IMX8MP) enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_CPUS = 4, + FSL_IMX8MP_NUM_GPIOS = 5, FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, FSL_IMX8MP_NUM_USDHCS = 3, @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { ARMCPU cpu[FSL_IMX8MP_NUM_CPUS]; GICv3State gic; + IMXGPIOState gpio[FSL_IMX8MP_NUM_GPIOS]; IMX8MPCCMState ccm; IMX8MPAnalogState analog; IMX7SNVSState snvs; @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_UART5_IRQ = 30, FSL_IMX8MP_UART6_IRQ = 16, + FSL_IMX8MP_GPIO1_LOW_IRQ = 64, + FSL_IMX8MP_GPIO1_HIGH_IRQ = 65, + FSL_IMX8MP_GPIO2_LOW_IRQ = 66, + FSL_IMX8MP_GPIO2_HIGH_IRQ = 67, + FSL_IMX8MP_GPIO3_LOW_IRQ = 68, + FSL_IMX8MP_GPIO3_HIGH_IRQ = 69, + FSL_IMX8MP_GPIO4_LOW_IRQ = 70, + FSL_IMX8MP_GPIO4_HIGH_IRQ = 71, + FSL_IMX8MP_GPIO5_LOW_IRQ = 72, + FSL_IMX8MP_GPIO5_HIGH_IRQ = 73, + FSL_IMX8MP_PCI_INTA_IRQ = 126, FSL_IMX8MP_PCI_INTB_IRQ = 125, FSL_IMX8MP_PCI_INTC_IRQ = 124, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); } + for (i = 0; i < FSL_IMX8MP_NUM_GPIOS; i++) { + g_autofree char *name = g_strdup_printf("gpio%d", i + 1); + object_initialize_child(obj, name, &s->gpio[i], TYPE_IMX_GPIO); + } + for (i = 0; i < FSL_IMX8MP_NUM_USDHCS; i++) { g_autofree char *name = g_strdup_printf("usdhc%d", i + 1); object_initialize_child(obj, name, &s->usdhc[i], TYPE_IMX_USDHC); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, serial_table[i].irq)); } + /* GPIOs */ + for (i = 0; i < FSL_IMX8MP_NUM_GPIOS; i++) { + struct { + hwaddr addr; + unsigned int irq_low; + unsigned int irq_high; + } gpio_table[FSL_IMX8MP_NUM_GPIOS] = { + { + fsl_imx8mp_memmap[FSL_IMX8MP_GPIO1].addr, + FSL_IMX8MP_GPIO1_LOW_IRQ, + FSL_IMX8MP_GPIO1_HIGH_IRQ + }, + { + fsl_imx8mp_memmap[FSL_IMX8MP_GPIO2].addr, + FSL_IMX8MP_GPIO2_LOW_IRQ, + FSL_IMX8MP_GPIO2_HIGH_IRQ + }, + { + fsl_imx8mp_memmap[FSL_IMX8MP_GPIO3].addr, + FSL_IMX8MP_GPIO3_LOW_IRQ, + FSL_IMX8MP_GPIO3_HIGH_IRQ + }, + { + fsl_imx8mp_memmap[FSL_IMX8MP_GPIO4].addr, + FSL_IMX8MP_GPIO4_LOW_IRQ, + FSL_IMX8MP_GPIO4_HIGH_IRQ + }, + { + fsl_imx8mp_memmap[FSL_IMX8MP_GPIO5].addr, + FSL_IMX8MP_GPIO5_LOW_IRQ, + FSL_IMX8MP_GPIO5_HIGH_IRQ + }, + }; + + object_property_set_bool(OBJECT(&s->gpio[i]), "has-edge-sel", true, + &error_abort); + object_property_set_bool(OBJECT(&s->gpio[i]), "has-upper-pin-irq", + true, &error_abort); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->gpio[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->gpio[i]), 0, gpio_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->gpio[i]), 0, + qdev_get_gpio_in(gicdev, gpio_table[i].irq_low)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->gpio[i]), 1, + qdev_get_gpio_in(gicdev, gpio_table[i].irq_high)); + } + /* USDHCs */ for (i = 0; i < FSL_IMX8MP_NUM_USDHCS; i++) { struct { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_CCM: case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: + case FSL_IMX8MP_GPIO1 ... FSL_IMX8MP_GPIO5: case FSL_IMX8MP_PCIE1: case FSL_IMX8MP_PCIE_PHY1: case FSL_IMX8MP_RAM: -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-11-shentey@gmail.com [PMM: drop static const from i2c_table for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 11 +++++++++++ hw/arm/fsl-imx8mp.c | 29 +++++++++++++++++++++++++++++ hw/arm/Kconfig | 2 ++ 4 files changed, 43 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 3 USDHC Storage Controllers * 1 Designware PCI Express Controller * 5 GPIO Controllers + * 6 I2C Controllers * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "hw/char/imx_serial.h" #include "hw/gpio/imx_gpio.h" +#include "hw/i2c/imx_i2c.h" #include "hw/intc/arm_gicv3_common.h" #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" @@ -XXX,XX +XXX,XX @@ OBJECT_DECLARE_SIMPLE_TYPE(FslImx8mpState, FSL_IMX8MP) enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_CPUS = 4, FSL_IMX8MP_NUM_GPIOS = 5, + FSL_IMX8MP_NUM_I2CS = 6, FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, FSL_IMX8MP_NUM_USDHCS = 3, @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMX8MPCCMState ccm; IMX8MPAnalogState analog; IMX7SNVSState snvs; + IMXI2CState i2c[FSL_IMX8MP_NUM_I2CS]; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; DesignwarePCIEHost pcie; @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_UART5_IRQ = 30, FSL_IMX8MP_UART6_IRQ = 16, + FSL_IMX8MP_I2C1_IRQ = 35, + FSL_IMX8MP_I2C2_IRQ = 36, + FSL_IMX8MP_I2C3_IRQ = 37, + FSL_IMX8MP_I2C4_IRQ = 38, + FSL_IMX8MP_GPIO1_LOW_IRQ = 64, FSL_IMX8MP_GPIO1_HIGH_IRQ = 65, FSL_IMX8MP_GPIO2_LOW_IRQ = 66, @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_GPIO5_LOW_IRQ = 72, FSL_IMX8MP_GPIO5_HIGH_IRQ = 73, + FSL_IMX8MP_I2C5_IRQ = 76, + FSL_IMX8MP_I2C6_IRQ = 77, + FSL_IMX8MP_PCI_INTA_IRQ = 126, FSL_IMX8MP_PCI_INTB_IRQ = 125, FSL_IMX8MP_PCI_INTC_IRQ = 124, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); } + for (i = 0; i < FSL_IMX8MP_NUM_I2CS; i++) { + g_autofree char *name = g_strdup_printf("i2c%d", i + 1); + object_initialize_child(obj, name, &s->i2c[i], TYPE_IMX_I2C); + } + for (i = 0; i < FSL_IMX8MP_NUM_GPIOS; i++) { g_autofree char *name = g_strdup_printf("gpio%d", i + 1); object_initialize_child(obj, name, &s->gpio[i], TYPE_IMX_GPIO); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, serial_table[i].irq)); } + /* I2Cs */ + for (i = 0; i < FSL_IMX8MP_NUM_I2CS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } i2c_table[FSL_IMX8MP_NUM_I2CS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C1].addr, FSL_IMX8MP_I2C1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C2].addr, FSL_IMX8MP_I2C2_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C3].addr, FSL_IMX8MP_I2C3_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C4].addr, FSL_IMX8MP_I2C4_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C5].addr, FSL_IMX8MP_I2C5_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_I2C6].addr, FSL_IMX8MP_I2C6_IRQ }, + }; + + if (!sysbus_realize(SYS_BUS_DEVICE(&s->i2c[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->i2c[i]), 0, i2c_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->i2c[i]), 0, + qdev_get_gpio_in(gicdev, i2c_table[i].irq)); + } + /* GPIOs */ for (i = 0; i < FSL_IMX8MP_NUM_GPIOS; i++) { struct { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: case FSL_IMX8MP_GPIO1 ... FSL_IMX8MP_GPIO5: + case FSL_IMX8MP_I2C1 ... FSL_IMX8MP_I2C6: case FSL_IMX8MP_PCIE1: case FSL_IMX8MP_PCIE_PHY1: case FSL_IMX8MP_RAM: diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX7 config FSL_IMX8MP bool + imply I2C_DEVICES imply PCI_DEVICES select ARM_GIC select FSL_IMX8MP_ANALOG select FSL_IMX8MP_CCM select IMX + select IMX_I2C select PCI_EXPRESS_DESIGNWARE select PCI_EXPRESS_FSL_IMX8M_PHY select SDHCI -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-12-shentey@gmail.com [PMM: drop static const from spi_table for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 8 ++++++++ hw/arm/fsl-imx8mp.c | 26 ++++++++++++++++++++++++++ 3 files changed, 35 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 1 Designware PCI Express Controller * 5 GPIO Controllers * 6 I2C Controllers + * 3 SPI Controllers * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/pci-host/designware.h" #include "hw/pci-host/fsl_imx8m_phy.h" #include "hw/sd/sdhci.h" +#include "hw/ssi/imx_spi.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ OBJECT_DECLARE_SIMPLE_TYPE(FslImx8mpState, FSL_IMX8MP) enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_CPUS = 4, + FSL_IMX8MP_NUM_ECSPIS = 3, FSL_IMX8MP_NUM_GPIOS = 5, FSL_IMX8MP_NUM_I2CS = 6, FSL_IMX8MP_NUM_IRQS = 160, @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMX8MPCCMState ccm; IMX8MPAnalogState analog; IMX7SNVSState snvs; + IMXSPIState spi[FSL_IMX8MP_NUM_ECSPIS]; IMXI2CState i2c[FSL_IMX8MP_NUM_I2CS]; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_UART5_IRQ = 30, FSL_IMX8MP_UART6_IRQ = 16, + FSL_IMX8MP_ECSPI1_IRQ = 31, + FSL_IMX8MP_ECSPI2_IRQ = 32, + FSL_IMX8MP_ECSPI3_IRQ = 33, + FSL_IMX8MP_ECSPI4_IRQ = 34, + FSL_IMX8MP_I2C1_IRQ = 35, FSL_IMX8MP_I2C2_IRQ = 36, FSL_IMX8MP_I2C3_IRQ = 37, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->usdhc[i], TYPE_IMX_USDHC); } + for (i = 0; i < FSL_IMX8MP_NUM_ECSPIS; i++) { + g_autofree char *name = g_strdup_printf("spi%d", i + 1); + object_initialize_child(obj, name, &s->spi[i], TYPE_IMX_SPI); + } + object_initialize_child(obj, "pcie", &s->pcie, TYPE_DESIGNWARE_PCIE_HOST); object_initialize_child(obj, "pcie_phy", &s->pcie_phy, TYPE_FSL_IMX8M_PCIE_PHY); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, usdhc_table[i].irq)); } + /* ECSPIs */ + for (i = 0; i < FSL_IMX8MP_NUM_ECSPIS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } spi_table[FSL_IMX8MP_NUM_ECSPIS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_ECSPI1].addr, FSL_IMX8MP_ECSPI1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_ECSPI2].addr, FSL_IMX8MP_ECSPI2_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_ECSPI3].addr, FSL_IMX8MP_ECSPI3_IRQ }, + }; + + if (!sysbus_realize(SYS_BUS_DEVICE(&s->spi[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->spi[i]), 0, spi_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->spi[i]), 0, + qdev_get_gpio_in(gicdev, spi_table[i].irq)); + } + /* SNVS */ if (!sysbus_realize(SYS_BUS_DEVICE(&s->snvs), errp)) { return; @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_GIC_DIST: case FSL_IMX8MP_GIC_REDIST: case FSL_IMX8MP_GPIO1 ... FSL_IMX8MP_GPIO5: + case FSL_IMX8MP_ECSPI1 ... FSL_IMX8MP_ECSPI3: case FSL_IMX8MP_I2C1 ... FSL_IMX8MP_I2C6: case FSL_IMX8MP_PCIE1: case FSL_IMX8MP_PCIE_PHY1: -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-13-shentey@gmail.com [PMM: drop static const from wdog_table for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 7 +++++++ hw/arm/fsl-imx8mp.c | 28 ++++++++++++++++++++++++++++ hw/arm/Kconfig | 1 + 4 files changed, 37 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 5 GPIO Controllers * 6 I2C Controllers * 3 SPI Controllers + * 3 Watchdogs * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/pci-host/fsl_imx8m_phy.h" #include "hw/sd/sdhci.h" #include "hw/ssi/imx_spi.h" +#include "hw/watchdog/wdt_imx2.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, FSL_IMX8MP_NUM_USDHCS = 3, + FSL_IMX8MP_NUM_WDTS = 3, }; struct FslImx8mpState { @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMXI2CState i2c[FSL_IMX8MP_NUM_I2CS]; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; + IMX2WdtState wdt[FSL_IMX8MP_NUM_WDTS]; DesignwarePCIEHost pcie; FslImx8mPciePhyState pcie_phy; }; @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_I2C5_IRQ = 76, FSL_IMX8MP_I2C6_IRQ = 77, + FSL_IMX8MP_WDOG1_IRQ = 78, + FSL_IMX8MP_WDOG2_IRQ = 79, + FSL_IMX8MP_WDOG3_IRQ = 10, + FSL_IMX8MP_PCI_INTA_IRQ = 126, FSL_IMX8MP_PCI_INTB_IRQ = 125, FSL_IMX8MP_PCI_INTC_IRQ = 124, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->spi[i], TYPE_IMX_SPI); } + for (i = 0; i < FSL_IMX8MP_NUM_WDTS; i++) { + g_autofree char *name = g_strdup_printf("wdt%d", i); + object_initialize_child(obj, name, &s->wdt[i], TYPE_IMX2_WDT); + } + object_initialize_child(obj, "pcie", &s->pcie, TYPE_DESIGNWARE_PCIE_HOST); object_initialize_child(obj, "pcie_phy", &s->pcie_phy, TYPE_FSL_IMX8M_PCIE_PHY); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) sysbus_mmio_map(SYS_BUS_DEVICE(&s->snvs), 0, fsl_imx8mp_memmap[FSL_IMX8MP_SNVS_HP].addr); + /* Watchdogs */ + for (i = 0; i < FSL_IMX8MP_NUM_WDTS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } wdog_table[FSL_IMX8MP_NUM_WDTS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_WDOG1].addr, FSL_IMX8MP_WDOG1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_WDOG2].addr, FSL_IMX8MP_WDOG2_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_WDOG3].addr, FSL_IMX8MP_WDOG3_IRQ }, + }; + + object_property_set_bool(OBJECT(&s->wdt[i]), "pretimeout-support", + true, &error_abort); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->wdt[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->wdt[i]), 0, wdog_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->wdt[i]), 0, + qdev_get_gpio_in(gicdev, wdog_table[i].irq)); + } + /* PCIe */ if (!sysbus_realize(SYS_BUS_DEVICE(&s->pcie), errp)) { return; @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_SNVS_HP: case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: case FSL_IMX8MP_USDHC1 ... FSL_IMX8MP_USDHC3: + case FSL_IMX8MP_WDOG1 ... FSL_IMX8MP_WDOG3: /* device implemented and treated above */ break; diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX8MP select PCI_EXPRESS_FSL_IMX8M_PHY select SDHCI select UNIMP + select WDT_IMX2 config FSL_IMX8MP_EVK bool -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-14-shentey@gmail.com [PMM: drop static const from gpt_attrs for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 11 +++++++ include/hw/timer/imx_gpt.h | 1 + hw/arm/fsl-imx8mp.c | 53 ++++++++++++++++++++++++++++++++++ hw/timer/imx_gpt.c | 25 ++++++++++++++++ hw/arm/Kconfig | 1 + 6 files changed, 92 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 6 I2C Controllers * 3 SPI Controllers * 3 Watchdogs + * 6 General Purpose Timers * Secure Non-Volatile Storage (SNVS) including an RTC * Clock Tree diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" #include "hw/misc/imx8mp_ccm.h" +#include "hw/or-irq.h" #include "hw/pci-host/designware.h" #include "hw/pci-host/fsl_imx8m_phy.h" #include "hw/sd/sdhci.h" #include "hw/ssi/imx_spi.h" +#include "hw/timer/imx_gpt.h" #include "hw/watchdog/wdt_imx2.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_CPUS = 4, FSL_IMX8MP_NUM_ECSPIS = 3, FSL_IMX8MP_NUM_GPIOS = 5, + FSL_IMX8MP_NUM_GPTS = 6, FSL_IMX8MP_NUM_I2CS = 6, FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { ARMCPU cpu[FSL_IMX8MP_NUM_CPUS]; GICv3State gic; + IMXGPTState gpt[FSL_IMX8MP_NUM_GPTS]; IMXGPIOState gpio[FSL_IMX8MP_NUM_GPIOS]; IMX8MPCCMState ccm; IMX8MPAnalogState analog; @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMX2WdtState wdt[FSL_IMX8MP_NUM_WDTS]; DesignwarePCIEHost pcie; FslImx8mPciePhyState pcie_phy; + OrIRQState gpt5_gpt6_irq; }; enum FslImx8mpMemoryRegions { @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_I2C3_IRQ = 37, FSL_IMX8MP_I2C4_IRQ = 38, + FSL_IMX8MP_GPT1_IRQ = 55, + FSL_IMX8MP_GPT2_IRQ = 54, + FSL_IMX8MP_GPT3_IRQ = 53, + FSL_IMX8MP_GPT4_IRQ = 52, + FSL_IMX8MP_GPT5_GPT6_IRQ = 51, + FSL_IMX8MP_GPIO1_LOW_IRQ = 64, FSL_IMX8MP_GPIO1_HIGH_IRQ = 65, FSL_IMX8MP_GPIO2_LOW_IRQ = 66, diff --git a/include/hw/timer/imx_gpt.h b/include/hw/timer/imx_gpt.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/timer/imx_gpt.h +++ b/include/hw/timer/imx_gpt.h @@ -XXX,XX +XXX,XX @@ #define TYPE_IMX6_GPT "imx6.gpt" #define TYPE_IMX6UL_GPT "imx6ul.gpt" #define TYPE_IMX7_GPT "imx7.gpt" +#define TYPE_IMX8MP_GPT "imx8mp.gpt" #define TYPE_IMX_GPT TYPE_IMX25_GPT diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->uart[i], TYPE_IMX_SERIAL); } + for (i = 0; i < FSL_IMX8MP_NUM_GPTS; i++) { + g_autofree char *name = g_strdup_printf("gpt%d", i + 1); + object_initialize_child(obj, name, &s->gpt[i], TYPE_IMX8MP_GPT); + } + object_initialize_child(obj, "gpt5-gpt6-irq", &s->gpt5_gpt6_irq, + TYPE_OR_IRQ); + for (i = 0; i < FSL_IMX8MP_NUM_I2CS; i++) { g_autofree char *name = g_strdup_printf("i2c%d", i + 1); object_initialize_child(obj, name, &s->i2c[i], TYPE_IMX_I2C); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, serial_table[i].irq)); } + /* GPTs */ + object_property_set_int(OBJECT(&s->gpt5_gpt6_irq), "num-lines", 2, + &error_abort); + if (!qdev_realize(DEVICE(&s->gpt5_gpt6_irq), NULL, errp)) { + return; + } + + qdev_connect_gpio_out(DEVICE(&s->gpt5_gpt6_irq), 0, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_GPT5_GPT6_IRQ)); + + for (i = 0; i < FSL_IMX8MP_NUM_GPTS; i++) { + hwaddr gpt_addrs[FSL_IMX8MP_NUM_GPTS] = { + fsl_imx8mp_memmap[FSL_IMX8MP_GPT1].addr, + fsl_imx8mp_memmap[FSL_IMX8MP_GPT2].addr, + fsl_imx8mp_memmap[FSL_IMX8MP_GPT3].addr, + fsl_imx8mp_memmap[FSL_IMX8MP_GPT4].addr, + fsl_imx8mp_memmap[FSL_IMX8MP_GPT5].addr, + fsl_imx8mp_memmap[FSL_IMX8MP_GPT6].addr, + }; + + s->gpt[i].ccm = IMX_CCM(&s->ccm); + + if (!sysbus_realize(SYS_BUS_DEVICE(&s->gpt[i]), errp)) { + return; + } + + sysbus_mmio_map(SYS_BUS_DEVICE(&s->gpt[i]), 0, gpt_addrs[i]); + + if (i < FSL_IMX8MP_NUM_GPTS - 2) { + static const unsigned int gpt_irqs[FSL_IMX8MP_NUM_GPTS - 2] = { + FSL_IMX8MP_GPT1_IRQ, + FSL_IMX8MP_GPT2_IRQ, + FSL_IMX8MP_GPT3_IRQ, + FSL_IMX8MP_GPT4_IRQ, + }; + + sysbus_connect_irq(SYS_BUS_DEVICE(&s->gpt[i]), 0, + qdev_get_gpio_in(gicdev, gpt_irqs[i])); + } else { + int irq = i - FSL_IMX8MP_NUM_GPTS + 2; + + sysbus_connect_irq(SYS_BUS_DEVICE(&s->gpt[i]), 0, + qdev_get_gpio_in(DEVICE(&s->gpt5_gpt6_irq), irq)); + } + } + /* I2Cs */ for (i = 0; i < FSL_IMX8MP_NUM_I2CS; i++) { struct { diff --git a/hw/timer/imx_gpt.c b/hw/timer/imx_gpt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/imx_gpt.c +++ b/hw/timer/imx_gpt.c @@ -XXX,XX +XXX,XX @@ static const IMXClk imx7_gpt_clocks[] = { CLK_NONE, /* 111 not defined */ }; +static const IMXClk imx8mp_gpt_clocks[] = { + CLK_NONE, /* 000 No clock source */ + CLK_IPG, /* 001 ipg_clk, 532MHz */ + CLK_IPG_HIGH, /* 010 ipg_clk_highfreq */ + CLK_EXT, /* 011 External clock */ + CLK_32k, /* 100 ipg_clk_32k */ + CLK_HIGH, /* 101 ipg_clk_16M */ + CLK_NONE, /* 110 not defined */ + CLK_NONE, /* 111 not defined */ +}; + /* Must be called from within ptimer_transaction_begin/commit block */ static void imx_gpt_set_freq(IMXGPTState *s) { @@ -XXX,XX +XXX,XX @@ static void imx7_gpt_init(Object *obj) s->clocks = imx7_gpt_clocks; } +static void imx8mp_gpt_init(Object *obj) +{ + IMXGPTState *s = IMX_GPT(obj); + + s->clocks = imx8mp_gpt_clocks; +} + static const TypeInfo imx25_gpt_info = { .name = TYPE_IMX25_GPT, .parent = TYPE_SYS_BUS_DEVICE, @@ -XXX,XX +XXX,XX @@ static const TypeInfo imx7_gpt_info = { .instance_init = imx7_gpt_init, }; +static const TypeInfo imx8mp_gpt_info = { + .name = TYPE_IMX8MP_GPT, + .parent = TYPE_IMX25_GPT, + .instance_init = imx8mp_gpt_init, +}; + static void imx_gpt_register_types(void) { type_register_static(&imx25_gpt_info); @@ -XXX,XX +XXX,XX @@ static void imx_gpt_register_types(void) type_register_static(&imx6_gpt_info); type_register_static(&imx6ul_gpt_info); type_register_static(&imx7_gpt_info); + type_register_static(&imx8mp_gpt_info); } type_init(imx_gpt_register_types) diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX8MP select FSL_IMX8MP_CCM select IMX select IMX_I2C + select OR_IRQ select PCI_EXPRESS_DESIGNWARE select PCI_EXPRESS_FSL_IMX8M_PHY select SDHCI -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> The i.MX 8M Plus SoC actually has two ethernet controllers, the usual ENET one and a Designware one. There is no device model for the latter, so only add the ENET one. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-15-shentey@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 8 ++++++++ hw/arm/fsl-imx8mp.c | 24 ++++++++++++++++++++++++ hw/arm/imx8mp-evk.c | 1 + hw/arm/Kconfig | 1 + 5 files changed, 35 insertions(+) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 4 UARTs * 3 USDHC Storage Controllers * 1 Designware PCI Express Controller + * 1 Ethernet Controller * 5 GPIO Controllers * 6 I2C Controllers * 3 SPI Controllers diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/imx7_snvs.h" #include "hw/misc/imx8mp_analog.h" #include "hw/misc/imx8mp_ccm.h" +#include "hw/net/imx_fec.h" #include "hw/or-irq.h" #include "hw/pci-host/designware.h" #include "hw/pci-host/fsl_imx8m_phy.h" @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMXSPIState spi[FSL_IMX8MP_NUM_ECSPIS]; IMXI2CState i2c[FSL_IMX8MP_NUM_I2CS]; IMXSerialState uart[FSL_IMX8MP_NUM_UARTS]; + IMXFECState enet; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; IMX2WdtState wdt[FSL_IMX8MP_NUM_WDTS]; DesignwarePCIEHost pcie; FslImx8mPciePhyState pcie_phy; OrIRQState gpt5_gpt6_irq; + + uint32_t phy_num; + bool phy_connected; }; enum FslImx8mpMemoryRegions { @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_WDOG2_IRQ = 79, FSL_IMX8MP_WDOG3_IRQ = 10, + FSL_IMX8MP_ENET1_MAC_IRQ = 118, + FSL_IMX6_ENET1_MAC_1588_IRQ = 121, + FSL_IMX8MP_PCI_INTA_IRQ = 126, FSL_IMX8MP_PCI_INTB_IRQ = 125, FSL_IMX8MP_PCI_INTC_IRQ = 124, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->wdt[i], TYPE_IMX2_WDT); } + object_initialize_child(obj, "eth0", &s->enet, TYPE_IMX_ENET); + object_initialize_child(obj, "pcie", &s->pcie, TYPE_DESIGNWARE_PCIE_HOST); object_initialize_child(obj, "pcie_phy", &s->pcie_phy, TYPE_FSL_IMX8M_PCIE_PHY); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, spi_table[i].irq)); } + /* ENET1 */ + object_property_set_uint(OBJECT(&s->enet), "phy-num", s->phy_num, + &error_abort); + object_property_set_uint(OBJECT(&s->enet), "tx-ring-num", 3, &error_abort); + qemu_configure_nic_device(DEVICE(&s->enet), true, NULL); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->enet), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->enet), 0, + fsl_imx8mp_memmap[FSL_IMX8MP_ENET1].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->enet), 0, + qdev_get_gpio_in(gicdev, FSL_IMX8MP_ENET1_MAC_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->enet), 1, + qdev_get_gpio_in(gicdev, FSL_IMX6_ENET1_MAC_1588_IRQ)); + /* SNVS */ if (!sysbus_realize(SYS_BUS_DEVICE(&s->snvs), errp)) { return; @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_GIC_REDIST: case FSL_IMX8MP_GPIO1 ... FSL_IMX8MP_GPIO5: case FSL_IMX8MP_ECSPI1 ... FSL_IMX8MP_ECSPI3: + case FSL_IMX8MP_ENET1: case FSL_IMX8MP_I2C1 ... FSL_IMX8MP_I2C6: case FSL_IMX8MP_PCIE1: case FSL_IMX8MP_PCIE_PHY1: @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) } } +static const Property fsl_imx8mp_properties[] = { + DEFINE_PROP_UINT32("fec1-phy-num", FslImx8mpState, phy_num, 0), + DEFINE_PROP_BOOL("fec1-phy-connected", FslImx8mpState, phy_connected, true), +}; + static void fsl_imx8mp_class_init(ObjectClass *oc, void *data) { DeviceClass *dc = DEVICE_CLASS(oc); + device_class_set_props(dc, fsl_imx8mp_properties); dc->realize = fsl_imx8mp_realize; dc->desc = "i.MX 8M Plus SoC"; diff --git a/hw/arm/imx8mp-evk.c b/hw/arm/imx8mp-evk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/imx8mp-evk.c +++ b/hw/arm/imx8mp-evk.c @@ -XXX,XX +XXX,XX @@ static void imx8mp_evk_init(MachineState *machine) s = FSL_IMX8MP(object_new(TYPE_FSL_IMX8MP)); object_property_add_child(OBJECT(machine), "soc", OBJECT(s)); + object_property_set_uint(OBJECT(s), "fec1-phy-num", 1, &error_fatal); qdev_realize(DEVICE(s), NULL, &error_fatal); memory_region_add_subregion(get_system_memory(), FSL_IMX8MP_RAM_START, diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX8MP select FSL_IMX8MP_ANALOG select FSL_IMX8MP_CCM select IMX + select IMX_FEC select IMX_I2C select OR_IRQ select PCI_EXPRESS_DESIGNWARE -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Split the USB MMIO regions to better keep track of the implemented vs. unimplemented regions. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-16-shentey@gmail.com [PMM: drop "static const" from usb_table for GCC 7.5] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/imx8mp-evk.rst | 1 + include/hw/arm/fsl-imx8mp.h | 12 +++++++++++ hw/arm/fsl-imx8mp.c | 37 ++++++++++++++++++++++++++++++++-- hw/arm/Kconfig | 1 + 4 files changed, 49 insertions(+), 2 deletions(-) diff --git a/docs/system/arm/imx8mp-evk.rst b/docs/system/arm/imx8mp-evk.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/imx8mp-evk.rst +++ b/docs/system/arm/imx8mp-evk.rst @@ -XXX,XX +XXX,XX @@ The ``imx8mp-evk`` machine implements the following devices: * 3 USDHC Storage Controllers * 1 Designware PCI Express Controller * 1 Ethernet Controller + * 2 Designware USB 3 Controllers * 5 GPIO Controllers * 6 I2C Controllers * 3 SPI Controllers diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ #include "hw/sd/sdhci.h" #include "hw/ssi/imx_spi.h" #include "hw/timer/imx_gpt.h" +#include "hw/usb/hcd-dwc3.h" #include "hw/watchdog/wdt_imx2.h" #include "qom/object.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ enum FslImx8mpConfiguration { FSL_IMX8MP_NUM_I2CS = 6, FSL_IMX8MP_NUM_IRQS = 160, FSL_IMX8MP_NUM_UARTS = 4, + FSL_IMX8MP_NUM_USBS = 2, FSL_IMX8MP_NUM_USDHCS = 3, FSL_IMX8MP_NUM_WDTS = 3, }; @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { IMXFECState enet; SDHCIState usdhc[FSL_IMX8MP_NUM_USDHCS]; IMX2WdtState wdt[FSL_IMX8MP_NUM_WDTS]; + USBDWC3 usb[FSL_IMX8MP_NUM_USBS]; DesignwarePCIEHost pcie; FslImx8mPciePhyState pcie_phy; OrIRQState gpt5_gpt6_irq; @@ -XXX,XX +XXX,XX @@ enum FslImx8mpMemoryRegions { FSL_IMX8MP_UART4, FSL_IMX8MP_USB1, FSL_IMX8MP_USB2, + FSL_IMX8MP_USB1_DEV, + FSL_IMX8MP_USB2_DEV, + FSL_IMX8MP_USB1_OTG, + FSL_IMX8MP_USB2_OTG, + FSL_IMX8MP_USB1_GLUE, + FSL_IMX8MP_USB2_GLUE, FSL_IMX8MP_USDHC1, FSL_IMX8MP_USDHC2, FSL_IMX8MP_USDHC3, @@ -XXX,XX +XXX,XX @@ enum FslImx8mpIrqs { FSL_IMX8MP_I2C3_IRQ = 37, FSL_IMX8MP_I2C4_IRQ = 38, + FSL_IMX8MP_USB1_IRQ = 40, + FSL_IMX8MP_USB2_IRQ = 41, + FSL_IMX8MP_GPT1_IRQ = 55, FSL_IMX8MP_GPT2_IRQ = 54, FSL_IMX8MP_GPT3_IRQ = 53, diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static const struct { [FSL_IMX8MP_VPU_VC8000E_ENCODER] = { 0x38320000, 2 * MiB, "vpu_vc8000e_encoder" }, [FSL_IMX8MP_VPU_G2_DECODER] = { 0x38310000, 2 * MiB, "vpu_g2_decoder" }, [FSL_IMX8MP_VPU_G1_DECODER] = { 0x38300000, 2 * MiB, "vpu_g1_decoder" }, - [FSL_IMX8MP_USB2] = { 0x38200000, 1 * MiB, "usb2" }, - [FSL_IMX8MP_USB1] = { 0x38100000, 1 * MiB, "usb1" }, + [FSL_IMX8MP_USB2_GLUE] = { 0x382f0000, 0x100, "usb2_glue" }, + [FSL_IMX8MP_USB2_OTG] = { 0x3820cc00, 0x100, "usb2_otg" }, + [FSL_IMX8MP_USB2_DEV] = { 0x3820c700, 0x500, "usb2_dev" }, + [FSL_IMX8MP_USB2] = { 0x38200000, 0xc700, "usb2" }, + [FSL_IMX8MP_USB1_GLUE] = { 0x381f0000, 0x100, "usb1_glue" }, + [FSL_IMX8MP_USB1_OTG] = { 0x3810cc00, 0x100, "usb1_otg" }, + [FSL_IMX8MP_USB1_DEV] = { 0x3810c700, 0x500, "usb1_dev" }, + [FSL_IMX8MP_USB1] = { 0x38100000, 0xc700, "usb1" }, [FSL_IMX8MP_GPU2D] = { 0x38008000, 32 * KiB, "gpu2d" }, [FSL_IMX8MP_GPU3D] = { 0x38000000, 32 * KiB, "gpu3d" }, [FSL_IMX8MP_QSPI1_RX_BUFFER] = { 0x34000000, 32 * MiB, "qspi1_rx_buffer" }, @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_init(Object *obj) object_initialize_child(obj, name, &s->usdhc[i], TYPE_IMX_USDHC); } + for (i = 0; i < FSL_IMX8MP_NUM_USBS; i++) { + g_autofree char *name = g_strdup_printf("usb%d", i); + object_initialize_child(obj, name, &s->usb[i], TYPE_USB_DWC3); + } + for (i = 0; i < FSL_IMX8MP_NUM_ECSPIS; i++) { g_autofree char *name = g_strdup_printf("spi%d", i + 1); object_initialize_child(obj, name, &s->spi[i], TYPE_IMX_SPI); @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, usdhc_table[i].irq)); } + /* USBs */ + for (i = 0; i < FSL_IMX8MP_NUM_USBS; i++) { + struct { + hwaddr addr; + unsigned int irq; + } usb_table[FSL_IMX8MP_NUM_USBS] = { + { fsl_imx8mp_memmap[FSL_IMX8MP_USB1].addr, FSL_IMX8MP_USB1_IRQ }, + { fsl_imx8mp_memmap[FSL_IMX8MP_USB2].addr, FSL_IMX8MP_USB2_IRQ }, + }; + + qdev_prop_set_uint32(DEVICE(&s->usb[i].sysbus_xhci), "p2", 1); + qdev_prop_set_uint32(DEVICE(&s->usb[i].sysbus_xhci), "p3", 1); + qdev_prop_set_uint32(DEVICE(&s->usb[i].sysbus_xhci), "slots", 2); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->usb[i]), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->usb[i]), 0, usb_table[i].addr); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->usb[i].sysbus_xhci), 0, + qdev_get_gpio_in(gicdev, usb_table[i].irq)); + } + /* ECSPIs */ for (i = 0; i < FSL_IMX8MP_NUM_ECSPIS; i++) { struct { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_RAM: case FSL_IMX8MP_SNVS_HP: case FSL_IMX8MP_UART1 ... FSL_IMX8MP_UART4: + case FSL_IMX8MP_USB1 ... FSL_IMX8MP_USB2: case FSL_IMX8MP_USDHC1 ... FSL_IMX8MP_USDHC3: case FSL_IMX8MP_WDOG1 ... FSL_IMX8MP_WDOG3: /* device implemented and treated above */ diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config FSL_IMX8MP select PCI_EXPRESS_FSL_IMX8M_PHY select SDHCI select UNIMP + select USB_DWC3 select WDT_IMX2 config FSL_IMX8MP_EVK -- 2.43.0 From: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Message-id: 20250223114708.1780-18-shentey@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/fsl-imx8mp.h | 1 + hw/arm/fsl-imx8mp.c | 11 +++++++++++ 2 files changed, 12 insertions(+) diff --git a/include/hw/arm/fsl-imx8mp.h b/include/hw/arm/fsl-imx8mp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/fsl-imx8mp.h +++ b/include/hw/arm/fsl-imx8mp.h @@ -XXX,XX +XXX,XX @@ struct FslImx8mpState { DesignwarePCIEHost pcie; FslImx8mPciePhyState pcie_phy; OrIRQState gpt5_gpt6_irq; + MemoryRegion ocram; uint32_t phy_num; bool phy_connected; diff --git a/hw/arm/fsl-imx8mp.c b/hw/arm/fsl-imx8mp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx8mp.c +++ b/hw/arm/fsl-imx8mp.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) sysbus_mmio_map(SYS_BUS_DEVICE(&s->pcie_phy), 0, fsl_imx8mp_memmap[FSL_IMX8MP_PCIE_PHY1].addr); + /* On-Chip RAM */ + if (!memory_region_init_ram(&s->ocram, NULL, "imx8mp.ocram", + fsl_imx8mp_memmap[FSL_IMX8MP_OCRAM].size, + errp)) { + return; + } + memory_region_add_subregion(get_system_memory(), + fsl_imx8mp_memmap[FSL_IMX8MP_OCRAM].addr, + &s->ocram); + /* Unimplemented devices */ for (i = 0; i < ARRAY_SIZE(fsl_imx8mp_memmap); i++) { switch (i) { @@ -XXX,XX +XXX,XX @@ static void fsl_imx8mp_realize(DeviceState *dev, Error **errp) case FSL_IMX8MP_ECSPI1 ... FSL_IMX8MP_ECSPI3: case FSL_IMX8MP_ENET1: case FSL_IMX8MP_I2C1 ... FSL_IMX8MP_I2C6: + case FSL_IMX8MP_OCRAM: case FSL_IMX8MP_PCIE1: case FSL_IMX8MP_PCIE_PHY1: case FSL_IMX8MP_RAM: -- 2.43.0