From 87e9bf23236d3c9da84f2b6164e06be3ecfd45e0 Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Mon, 27 Dec 2021 07:01:22 -0800 Subject: linux-user: Split out do_prctl and subroutines MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since the prctl constants are supposed to be generic, supply any that are not provided by the host. Split out subroutines for PR_GET_FP_MODE, PR_SET_FP_MODE, PR_GET_VL, PR_SET_VL, PR_RESET_KEYS, PR_SET_TAGGED_ADDR_CTRL, PR_GET_TAGGED_ADDR_CTRL. Return EINVAL for guests that do not support these options rather than pass them on to the host. Reviewed-by: Laurent Vivier Reviewed-by: Philippe Mathieu-Daudé Signed-off-by: Richard Henderson Message-Id: <20211227150127.2659293-2-richard.henderson@linaro.org> Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 433 ++++++++++++++++++--------------------------------- 1 file changed, 148 insertions(+), 285 deletions(-) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index 56a3e17183..0f0f67d567 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -6294,9 +6294,155 @@ abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr) return ret; } #endif /* defined(TARGET_ABI32 */ - #endif /* defined(TARGET_I386) */ +/* + * These constants are generic. Supply any that are missing from the host. + */ +#ifndef PR_SET_NAME +# define PR_SET_NAME 15 +# define PR_GET_NAME 16 +#endif +#ifndef PR_SET_FP_MODE +# define PR_SET_FP_MODE 45 +# define PR_GET_FP_MODE 46 +# define PR_FP_MODE_FR (1 << 0) +# define PR_FP_MODE_FRE (1 << 1) +#endif +#ifndef PR_SVE_SET_VL +# define PR_SVE_SET_VL 50 +# define PR_SVE_GET_VL 51 +# define PR_SVE_VL_LEN_MASK 0xffff +# define PR_SVE_VL_INHERIT (1 << 17) +#endif +#ifndef PR_PAC_RESET_KEYS +# define PR_PAC_RESET_KEYS 54 +# define PR_PAC_APIAKEY (1 << 0) +# define PR_PAC_APIBKEY (1 << 1) +# define PR_PAC_APDAKEY (1 << 2) +# define PR_PAC_APDBKEY (1 << 3) +# define PR_PAC_APGAKEY (1 << 4) +#endif +#ifndef PR_SET_TAGGED_ADDR_CTRL +# define PR_SET_TAGGED_ADDR_CTRL 55 +# define PR_GET_TAGGED_ADDR_CTRL 56 +# define PR_TAGGED_ADDR_ENABLE (1UL << 0) +#endif +#ifndef PR_MTE_TCF_SHIFT +# define PR_MTE_TCF_SHIFT 1 +# define PR_MTE_TCF_NONE (0UL << PR_MTE_TCF_SHIFT) +# define PR_MTE_TCF_SYNC (1UL << PR_MTE_TCF_SHIFT) +# define PR_MTE_TCF_ASYNC (2UL << PR_MTE_TCF_SHIFT) +# define PR_MTE_TCF_MASK (3UL << PR_MTE_TCF_SHIFT) +# define PR_MTE_TAG_SHIFT 3 +# define PR_MTE_TAG_MASK (0xffffUL << PR_MTE_TAG_SHIFT) +#endif + +#include "target_prctl.h" + +static abi_long do_prctl_inval0(CPUArchState *env) +{ + return -TARGET_EINVAL; +} + +static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) +{ + return -TARGET_EINVAL; +} + +#ifndef do_prctl_get_fp_mode +#define do_prctl_get_fp_mode do_prctl_inval0 +#endif +#ifndef do_prctl_set_fp_mode +#define do_prctl_set_fp_mode do_prctl_inval1 +#endif +#ifndef do_prctl_get_vl +#define do_prctl_get_vl do_prctl_inval0 +#endif +#ifndef do_prctl_set_vl +#define do_prctl_set_vl do_prctl_inval1 +#endif +#ifndef do_prctl_reset_keys +#define do_prctl_reset_keys do_prctl_inval1 +#endif +#ifndef do_prctl_set_tagged_addr_ctrl +#define do_prctl_set_tagged_addr_ctrl do_prctl_inval1 +#endif +#ifndef do_prctl_get_tagged_addr_ctrl +#define do_prctl_get_tagged_addr_ctrl do_prctl_inval0 +#endif + +static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, + abi_long arg3, abi_long arg4, abi_long arg5) +{ + abi_long ret; + + switch (option) { + case PR_GET_PDEATHSIG: + { + int deathsig; + ret = get_errno(prctl(PR_GET_PDEATHSIG, &deathsig, + arg3, arg4, arg5)); + if (!is_error(ret) && arg2 && put_user_s32(deathsig, arg2)) { + return -TARGET_EFAULT; + } + return ret; + } + case PR_GET_NAME: + { + void *name = lock_user(VERIFY_WRITE, arg2, 16, 1); + if (!name) { + return -TARGET_EFAULT; + } + ret = get_errno(prctl(PR_GET_NAME, (uintptr_t)name, + arg3, arg4, arg5)); + unlock_user(name, arg2, 16); + return ret; + } + case PR_SET_NAME: + { + void *name = lock_user(VERIFY_READ, arg2, 16, 1); + if (!name) { + return -TARGET_EFAULT; + } + ret = get_errno(prctl(PR_SET_NAME, (uintptr_t)name, + arg3, arg4, arg5)); + unlock_user(name, arg2, 0); + return ret; + } + case PR_GET_FP_MODE: + return do_prctl_get_fp_mode(env); + case PR_SET_FP_MODE: + return do_prctl_set_fp_mode(env, arg2); + case PR_SVE_GET_VL: + return do_prctl_get_vl(env); + case PR_SVE_SET_VL: + return do_prctl_set_vl(env, arg2); + case PR_PAC_RESET_KEYS: + if (arg3 || arg4 || arg5) { + return -TARGET_EINVAL; + } + return do_prctl_reset_keys(env, arg2); + case PR_SET_TAGGED_ADDR_CTRL: + if (arg3 || arg4 || arg5) { + return -TARGET_EINVAL; + } + return do_prctl_set_tagged_addr_ctrl(env, arg2); + case PR_GET_TAGGED_ADDR_CTRL: + if (arg2 || arg3 || arg4 || arg5) { + return -TARGET_EINVAL; + } + return do_prctl_get_tagged_addr_ctrl(env); + case PR_GET_SECCOMP: + case PR_SET_SECCOMP: + /* Disable seccomp to prevent the target disabling syscalls we need. */ + return -TARGET_EINVAL; + default: + /* Most prctl options have no pointer arguments */ + return get_errno(prctl(option, arg2, arg3, arg4, arg5)); + } +} + #define NEW_STACK_SIZE 0x40000 @@ -10635,290 +10781,7 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1, return ret; #endif case TARGET_NR_prctl: - switch (arg1) { - case PR_GET_PDEATHSIG: - { - int deathsig; - ret = get_errno(prctl(arg1, &deathsig, arg3, arg4, arg5)); - if (!is_error(ret) && arg2 - && put_user_s32(deathsig, arg2)) { - return -TARGET_EFAULT; - } - return ret; - } -#ifdef PR_GET_NAME - case PR_GET_NAME: - { - void *name = lock_user(VERIFY_WRITE, arg2, 16, 1); - if (!name) { - return -TARGET_EFAULT; - } - ret = get_errno(prctl(arg1, (unsigned long)name, - arg3, arg4, arg5)); - unlock_user(name, arg2, 16); - return ret; - } - case PR_SET_NAME: - { - void *name = lock_user(VERIFY_READ, arg2, 16, 1); - if (!name) { - return -TARGET_EFAULT; - } - ret = get_errno(prctl(arg1, (unsigned long)name, - arg3, arg4, arg5)); - unlock_user(name, arg2, 0); - return ret; - } -#endif -#ifdef TARGET_MIPS - case TARGET_PR_GET_FP_MODE: - { - CPUMIPSState *env = ((CPUMIPSState *)cpu_env); - ret = 0; - if (env->CP0_Status & (1 << CP0St_FR)) { - ret |= TARGET_PR_FP_MODE_FR; - } - if (env->CP0_Config5 & (1 << CP0C5_FRE)) { - ret |= TARGET_PR_FP_MODE_FRE; - } - return ret; - } - case TARGET_PR_SET_FP_MODE: - { - CPUMIPSState *env = ((CPUMIPSState *)cpu_env); - bool old_fr = env->CP0_Status & (1 << CP0St_FR); - bool old_fre = env->CP0_Config5 & (1 << CP0C5_FRE); - bool new_fr = arg2 & TARGET_PR_FP_MODE_FR; - bool new_fre = arg2 & TARGET_PR_FP_MODE_FRE; - - const unsigned int known_bits = TARGET_PR_FP_MODE_FR | - TARGET_PR_FP_MODE_FRE; - - /* If nothing to change, return right away, successfully. */ - if (old_fr == new_fr && old_fre == new_fre) { - return 0; - } - /* Check the value is valid */ - if (arg2 & ~known_bits) { - return -TARGET_EOPNOTSUPP; - } - /* Setting FRE without FR is not supported. */ - if (new_fre && !new_fr) { - return -TARGET_EOPNOTSUPP; - } - if (new_fr && !(env->active_fpu.fcr0 & (1 << FCR0_F64))) { - /* FR1 is not supported */ - return -TARGET_EOPNOTSUPP; - } - if (!new_fr && (env->active_fpu.fcr0 & (1 << FCR0_F64)) - && !(env->CP0_Status_rw_bitmask & (1 << CP0St_FR))) { - /* cannot set FR=0 */ - return -TARGET_EOPNOTSUPP; - } - if (new_fre && !(env->active_fpu.fcr0 & (1 << FCR0_FREP))) { - /* Cannot set FRE=1 */ - return -TARGET_EOPNOTSUPP; - } - - int i; - fpr_t *fpr = env->active_fpu.fpr; - for (i = 0; i < 32 ; i += 2) { - if (!old_fr && new_fr) { - fpr[i].w[!FP_ENDIAN_IDX] = fpr[i + 1].w[FP_ENDIAN_IDX]; - } else if (old_fr && !new_fr) { - fpr[i + 1].w[FP_ENDIAN_IDX] = fpr[i].w[!FP_ENDIAN_IDX]; - } - } - - if (new_fr) { - env->CP0_Status |= (1 << CP0St_FR); - env->hflags |= MIPS_HFLAG_F64; - } else { - env->CP0_Status &= ~(1 << CP0St_FR); - env->hflags &= ~MIPS_HFLAG_F64; - } - if (new_fre) { - env->CP0_Config5 |= (1 << CP0C5_FRE); - if (env->active_fpu.fcr0 & (1 << FCR0_FREP)) { - env->hflags |= MIPS_HFLAG_FRE; - } - } else { - env->CP0_Config5 &= ~(1 << CP0C5_FRE); - env->hflags &= ~MIPS_HFLAG_FRE; - } - - return 0; - } -#endif /* MIPS */ -#ifdef TARGET_AARCH64 - case TARGET_PR_SVE_SET_VL: - /* - * We cannot support either PR_SVE_SET_VL_ONEXEC or - * PR_SVE_VL_INHERIT. Note the kernel definition - * of sve_vl_valid allows for VQ=512, i.e. VL=8192, - * even though the current architectural maximum is VQ=16. - */ - ret = -TARGET_EINVAL; - if (cpu_isar_feature(aa64_sve, env_archcpu(cpu_env)) - && arg2 >= 0 && arg2 <= 512 * 16 && !(arg2 & 15)) { - CPUARMState *env = cpu_env; - ARMCPU *cpu = env_archcpu(env); - uint32_t vq, old_vq; - - old_vq = (env->vfp.zcr_el[1] & 0xf) + 1; - vq = MAX(arg2 / 16, 1); - vq = MIN(vq, cpu->sve_max_vq); - - if (vq < old_vq) { - aarch64_sve_narrow_vq(env, vq); - } - env->vfp.zcr_el[1] = vq - 1; - arm_rebuild_hflags(env); - ret = vq * 16; - } - return ret; - case TARGET_PR_SVE_GET_VL: - ret = -TARGET_EINVAL; - { - ARMCPU *cpu = env_archcpu(cpu_env); - if (cpu_isar_feature(aa64_sve, cpu)) { - ret = ((cpu->env.vfp.zcr_el[1] & 0xf) + 1) * 16; - } - } - return ret; - case TARGET_PR_PAC_RESET_KEYS: - { - CPUARMState *env = cpu_env; - ARMCPU *cpu = env_archcpu(env); - - if (arg3 || arg4 || arg5) { - return -TARGET_EINVAL; - } - if (cpu_isar_feature(aa64_pauth, cpu)) { - int all = (TARGET_PR_PAC_APIAKEY | TARGET_PR_PAC_APIBKEY | - TARGET_PR_PAC_APDAKEY | TARGET_PR_PAC_APDBKEY | - TARGET_PR_PAC_APGAKEY); - int ret = 0; - Error *err = NULL; - - if (arg2 == 0) { - arg2 = all; - } else if (arg2 & ~all) { - return -TARGET_EINVAL; - } - if (arg2 & TARGET_PR_PAC_APIAKEY) { - ret |= qemu_guest_getrandom(&env->keys.apia, - sizeof(ARMPACKey), &err); - } - if (arg2 & TARGET_PR_PAC_APIBKEY) { - ret |= qemu_guest_getrandom(&env->keys.apib, - sizeof(ARMPACKey), &err); - } - if (arg2 & TARGET_PR_PAC_APDAKEY) { - ret |= qemu_guest_getrandom(&env->keys.apda, - sizeof(ARMPACKey), &err); - } - if (arg2 & TARGET_PR_PAC_APDBKEY) { - ret |= qemu_guest_getrandom(&env->keys.apdb, - sizeof(ARMPACKey), &err); - } - if (arg2 & TARGET_PR_PAC_APGAKEY) { - ret |= qemu_guest_getrandom(&env->keys.apga, - sizeof(ARMPACKey), &err); - } - if (ret != 0) { - /* - * Some unknown failure in the crypto. The best - * we can do is log it and fail the syscall. - * The real syscall cannot fail this way. - */ - qemu_log_mask(LOG_UNIMP, - "PR_PAC_RESET_KEYS: Crypto failure: %s", - error_get_pretty(err)); - error_free(err); - return -TARGET_EIO; - } - return 0; - } - } - return -TARGET_EINVAL; - case TARGET_PR_SET_TAGGED_ADDR_CTRL: - { - abi_ulong valid_mask = TARGET_PR_TAGGED_ADDR_ENABLE; - CPUARMState *env = cpu_env; - ARMCPU *cpu = env_archcpu(env); - - if (cpu_isar_feature(aa64_mte, cpu)) { - valid_mask |= TARGET_PR_MTE_TCF_MASK; - valid_mask |= TARGET_PR_MTE_TAG_MASK; - } - - if ((arg2 & ~valid_mask) || arg3 || arg4 || arg5) { - return -TARGET_EINVAL; - } - env->tagged_addr_enable = arg2 & TARGET_PR_TAGGED_ADDR_ENABLE; - - if (cpu_isar_feature(aa64_mte, cpu)) { - switch (arg2 & TARGET_PR_MTE_TCF_MASK) { - case TARGET_PR_MTE_TCF_NONE: - case TARGET_PR_MTE_TCF_SYNC: - case TARGET_PR_MTE_TCF_ASYNC: - break; - default: - return -EINVAL; - } - - /* - * Write PR_MTE_TCF to SCTLR_EL1[TCF0]. - * Note that the syscall values are consistent with hw. - */ - env->cp15.sctlr_el[1] = - deposit64(env->cp15.sctlr_el[1], 38, 2, - arg2 >> TARGET_PR_MTE_TCF_SHIFT); - - /* - * Write PR_MTE_TAG to GCR_EL1[Exclude]. - * Note that the syscall uses an include mask, - * and hardware uses an exclude mask -- invert. - */ - env->cp15.gcr_el1 = - deposit64(env->cp15.gcr_el1, 0, 16, - ~arg2 >> TARGET_PR_MTE_TAG_SHIFT); - arm_rebuild_hflags(env); - } - return 0; - } - case TARGET_PR_GET_TAGGED_ADDR_CTRL: - { - abi_long ret = 0; - CPUARMState *env = cpu_env; - ARMCPU *cpu = env_archcpu(env); - - if (arg2 || arg3 || arg4 || arg5) { - return -TARGET_EINVAL; - } - if (env->tagged_addr_enable) { - ret |= TARGET_PR_TAGGED_ADDR_ENABLE; - } - if (cpu_isar_feature(aa64_mte, cpu)) { - /* See above. */ - ret |= (extract64(env->cp15.sctlr_el[1], 38, 2) - << TARGET_PR_MTE_TCF_SHIFT); - ret = deposit64(ret, TARGET_PR_MTE_TAG_SHIFT, 16, - ~env->cp15.gcr_el1); - } - return ret; - } -#endif /* AARCH64 */ - case PR_GET_SECCOMP: - case PR_SET_SECCOMP: - /* Disable seccomp to prevent the target disabling syscalls we - * need. */ - return -TARGET_EINVAL; - default: - /* Most prctl options have no pointer arguments */ - return get_errno(prctl(arg1, arg2, arg3, arg4, arg5)); - } + return do_prctl(cpu_env, arg1, arg2, arg3, arg4, arg5); break; #ifdef TARGET_NR_arch_prctl case TARGET_NR_arch_prctl: -- cgit 1.4.1 From 220717a6f46a99031a5b1af964bbf4dec1310440 Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Mon, 27 Dec 2021 07:01:23 -0800 Subject: linux-user: Disable more prctl subcodes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Create a list of subcodes that we want to pass on, a list of subcodes that should not be passed on because they would affect the running qemu itself, and a list that probably could be implemented but require extra work. Do not pass on unknown subcodes. Reviewed-by: Laurent Vivier Reviewed-by: Philippe Mathieu-Daudé Signed-off-by: Richard Henderson Message-Id: <20211227150127.2659293-3-richard.henderson@linaro.org> Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 58 +++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 55 insertions(+), 3 deletions(-) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index 0f0f67d567..d868ef2910 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -6337,6 +6337,13 @@ abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr) # define PR_MTE_TAG_SHIFT 3 # define PR_MTE_TAG_MASK (0xffffUL << PR_MTE_TAG_SHIFT) #endif +#ifndef PR_SET_IO_FLUSHER +# define PR_SET_IO_FLUSHER 57 +# define PR_GET_IO_FLUSHER 58 +#endif +#ifndef PR_SET_SYSCALL_USER_DISPATCH +# define PR_SET_SYSCALL_USER_DISPATCH 59 +#endif #include "target_prctl.h" @@ -6433,13 +6440,58 @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, return -TARGET_EINVAL; } return do_prctl_get_tagged_addr_ctrl(env); + + case PR_GET_DUMPABLE: + case PR_SET_DUMPABLE: + case PR_GET_KEEPCAPS: + case PR_SET_KEEPCAPS: + case PR_GET_TIMING: + case PR_SET_TIMING: + case PR_GET_TIMERSLACK: + case PR_SET_TIMERSLACK: + case PR_MCE_KILL: + case PR_MCE_KILL_GET: + case PR_GET_NO_NEW_PRIVS: + case PR_SET_NO_NEW_PRIVS: + case PR_GET_IO_FLUSHER: + case PR_SET_IO_FLUSHER: + /* Some prctl options have no pointer arguments and we can pass on. */ + return get_errno(prctl(option, arg2, arg3, arg4, arg5)); + + case PR_GET_CHILD_SUBREAPER: + case PR_SET_CHILD_SUBREAPER: + case PR_GET_SPECULATION_CTRL: + case PR_SET_SPECULATION_CTRL: + case PR_GET_TID_ADDRESS: + /* TODO */ + return -TARGET_EINVAL; + + case PR_GET_FPEXC: + case PR_SET_FPEXC: + /* Was used for SPE on PowerPC. */ + return -TARGET_EINVAL; + + case PR_GET_ENDIAN: + case PR_SET_ENDIAN: + case PR_GET_FPEMU: + case PR_SET_FPEMU: + case PR_SET_MM: case PR_GET_SECCOMP: case PR_SET_SECCOMP: - /* Disable seccomp to prevent the target disabling syscalls we need. */ + case PR_SET_SYSCALL_USER_DISPATCH: + case PR_GET_THP_DISABLE: + case PR_SET_THP_DISABLE: + case PR_GET_TSC: + case PR_SET_TSC: + case PR_GET_UNALIGN: + case PR_SET_UNALIGN: + /* Disable to prevent the target disabling stuff we need. */ return -TARGET_EINVAL; + default: - /* Most prctl options have no pointer arguments */ - return get_errno(prctl(option, arg2, arg3, arg4, arg5)); + qemu_log_mask(LOG_UNIMP, "Unsupported prctl: " TARGET_ABI_FMT_ld "\n", + option); + return -TARGET_EINVAL; } } -- cgit 1.4.1 From 6e8dcacd0840b3c38e7ec664bb9adbccc2108fdd Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Mon, 27 Dec 2021 07:01:24 -0800 Subject: linux-user: Add code for PR_GET/SET_UNALIGN MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This requires extra work for each target, but adds the common syscall code, and the necessary flag in CPUState. Reviewed-by: Warner Losh Reviewed-by: Laurent Vivier Reviewed-by: Philippe Mathieu-Daudé Signed-off-by: Richard Henderson Message-Id: <20211227150127.2659293-4-richard.henderson@linaro.org> Signed-off-by: Laurent Vivier --- cpu.c | 20 +++++++++++++++----- include/hw/core/cpu.h | 3 +++ linux-user/generic/target_prctl_unalign.h | 27 +++++++++++++++++++++++++++ linux-user/syscall.c | 13 +++++++++++-- 4 files changed, 56 insertions(+), 7 deletions(-) create mode 100644 linux-user/generic/target_prctl_unalign.h (limited to 'linux-user/syscall.c') diff --git a/cpu.c b/cpu.c index 945dd3dded..016bf06a1a 100644 --- a/cpu.c +++ b/cpu.c @@ -174,13 +174,23 @@ void cpu_exec_unrealizefn(CPUState *cpu) cpu_list_remove(cpu); } +/* + * This can't go in hw/core/cpu.c because that file is compiled only + * once for both user-mode and system builds. + */ static Property cpu_common_props[] = { -#ifndef CONFIG_USER_ONLY +#ifdef CONFIG_USER_ONLY + /* + * Create a property for the user-only object, so users can + * adjust prctl(PR_SET_UNALIGN) from the command-line. + * Has no effect if the target does not support the feature. + */ + DEFINE_PROP_BOOL("prctl-unalign-sigbus", CPUState, + prctl_unalign_sigbus, false), +#else /* - * Create a memory property for softmmu CPU object, - * so users can wire up its memory. (This can't go in hw/core/cpu.c - * because that file is compiled only once for both user-mode - * and system builds.) The default if no link is set up is to use + * Create a memory property for softmmu CPU object, so users can + * wire up its memory. The default if no link is set up is to use * the system address space. */ DEFINE_PROP_LINK("memory", CPUState, memory, TYPE_MEMORY_REGION, diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h index e948e81f1a..76ab3b851c 100644 --- a/include/hw/core/cpu.h +++ b/include/hw/core/cpu.h @@ -413,6 +413,9 @@ struct CPUState { bool ignore_memory_transaction_failures; + /* Used for user-only emulation of prctl(PR_SET_UNALIGN). */ + bool prctl_unalign_sigbus; + struct hax_vcpu_state *hax_vcpu; struct hvf_vcpu_state *hvf; diff --git a/linux-user/generic/target_prctl_unalign.h b/linux-user/generic/target_prctl_unalign.h new file mode 100644 index 0000000000..bc3b83af2a --- /dev/null +++ b/linux-user/generic/target_prctl_unalign.h @@ -0,0 +1,27 @@ +/* + * Generic prctl unalign functions for linux-user + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ +#ifndef GENERIC_TARGET_PRCTL_UNALIGN_H +#define GENERIC_TARGET_PRCTL_UNALIGN_H + +static abi_long do_prctl_get_unalign(CPUArchState *env, target_long arg2) +{ + CPUState *cs = env_cpu(env); + uint32_t res = PR_UNALIGN_NOPRINT; + if (cs->prctl_unalign_sigbus) { + res |= PR_UNALIGN_SIGBUS; + } + return put_user_u32(res, arg2); +} +#define do_prctl_get_unalign do_prctl_get_unalign + +static abi_long do_prctl_set_unalign(CPUArchState *env, target_long arg2) +{ + env_cpu(env)->prctl_unalign_sigbus = arg2 & PR_UNALIGN_SIGBUS; + return 0; +} +#define do_prctl_set_unalign do_prctl_set_unalign + +#endif /* GENERIC_TARGET_PRCTL_UNALIGN_H */ diff --git a/linux-user/syscall.c b/linux-user/syscall.c index d868ef2910..b5112891b0 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -6378,6 +6378,12 @@ static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) #ifndef do_prctl_get_tagged_addr_ctrl #define do_prctl_get_tagged_addr_ctrl do_prctl_inval0 #endif +#ifndef do_prctl_get_unalign +#define do_prctl_get_unalign do_prctl_inval1 +#endif +#ifndef do_prctl_set_unalign +#define do_prctl_set_unalign do_prctl_inval1 +#endif static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, abi_long arg3, abi_long arg4, abi_long arg5) @@ -6441,6 +6447,11 @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, } return do_prctl_get_tagged_addr_ctrl(env); + case PR_GET_UNALIGN: + return do_prctl_get_unalign(env, arg2); + case PR_SET_UNALIGN: + return do_prctl_set_unalign(env, arg2); + case PR_GET_DUMPABLE: case PR_SET_DUMPABLE: case PR_GET_KEEPCAPS: @@ -6483,8 +6494,6 @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, case PR_SET_THP_DISABLE: case PR_GET_TSC: case PR_SET_TSC: - case PR_GET_UNALIGN: - case PR_SET_UNALIGN: /* Disable to prevent the target disabling stuff we need. */ return -TARGET_EINVAL; -- cgit 1.4.1 From 45ad761c27dfb571d9a1ab3af80e8034d36d1b59 Mon Sep 17 00:00:00 2001 From: Tonis Tiigi Date: Tue, 4 Jan 2022 20:18:18 -0800 Subject: linux-user: add sched_getattr support These syscalls are not exposed by glibc. The struct type need to be redefined as it can't be included directly before https://lkml.org/lkml/2020/5/28/810 . sched_attr type can grow in future kernel versions. When client sends values that QEMU does not understand it will return E2BIG with same semantics as old kernel would so client can retry with smaller inputs. Signed-off-by: Tonis Tiigi Message-Id: <20220105041819.24160-2-tonistiigi@gmail.com> Reviewed-by: Laurent Vivier Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 123 ++++++++++++++++++++++++++++++++++++++++++++++ linux-user/syscall_defs.h | 14 ++++++ 2 files changed, 137 insertions(+) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index b5112891b0..6de116eb90 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -340,6 +340,25 @@ _syscall3(int, sys_sched_getaffinity, pid_t, pid, unsigned int, len, #define __NR_sys_sched_setaffinity __NR_sched_setaffinity _syscall3(int, sys_sched_setaffinity, pid_t, pid, unsigned int, len, unsigned long *, user_mask_ptr); +/* sched_attr is not defined in glibc */ +struct sched_attr { + uint32_t size; + uint32_t sched_policy; + uint64_t sched_flags; + int32_t sched_nice; + uint32_t sched_priority; + uint64_t sched_runtime; + uint64_t sched_deadline; + uint64_t sched_period; + uint32_t sched_util_min; + uint32_t sched_util_max; +}; +#define __NR_sys_sched_getattr __NR_sched_getattr +_syscall4(int, sys_sched_getattr, pid_t, pid, struct sched_attr *, attr, + unsigned int, size, unsigned int, flags); +#define __NR_sys_sched_setattr __NR_sched_setattr +_syscall3(int, sys_sched_setattr, pid_t, pid, struct sched_attr *, attr, + unsigned int, flags); #define __NR_sys_getcpu __NR_getcpu _syscall3(int, sys_getcpu, unsigned *, cpu, unsigned *, node, void *, tcache); _syscall4(int, reboot, int, magic1, int, magic2, unsigned int, cmd, @@ -558,6 +577,24 @@ const char *target_strerror(int err) return strerror(target_to_host_errno(err)); } +static int check_zeroed_user(abi_long addr, size_t ksize, size_t usize) +{ + int i; + uint8_t b; + if (usize <= ksize) { + return 1; + } + for (i = ksize; i < usize; i++) { + if (get_user_u8(b, addr + i)) { + return -TARGET_EFAULT; + } + if (b != 0) { + return 0; + } + } + return 1; +} + #define safe_syscall0(type, name) \ static type safe_##name(void) \ { \ @@ -10801,6 +10838,92 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1, } case TARGET_NR_sched_getscheduler: return get_errno(sched_getscheduler(arg1)); + case TARGET_NR_sched_getattr: + { + struct target_sched_attr *target_scha; + struct sched_attr scha; + if (arg2 == 0) { + return -TARGET_EINVAL; + } + if (arg3 > sizeof(scha)) { + arg3 = sizeof(scha); + } + ret = get_errno(sys_sched_getattr(arg1, &scha, arg3, arg4)); + if (!is_error(ret)) { + target_scha = lock_user(VERIFY_WRITE, arg2, arg3, 0); + if (!target_scha) { + return -TARGET_EFAULT; + } + target_scha->size = tswap32(scha.size); + target_scha->sched_policy = tswap32(scha.sched_policy); + target_scha->sched_flags = tswap64(scha.sched_flags); + target_scha->sched_nice = tswap32(scha.sched_nice); + target_scha->sched_priority = tswap32(scha.sched_priority); + target_scha->sched_runtime = tswap64(scha.sched_runtime); + target_scha->sched_deadline = tswap64(scha.sched_deadline); + target_scha->sched_period = tswap64(scha.sched_period); + if (scha.size > offsetof(struct sched_attr, sched_util_min)) { + target_scha->sched_util_min = tswap32(scha.sched_util_min); + target_scha->sched_util_max = tswap32(scha.sched_util_max); + } + unlock_user(target_scha, arg2, arg3); + } + return ret; + } + case TARGET_NR_sched_setattr: + { + struct target_sched_attr *target_scha; + struct sched_attr scha; + uint32_t size; + int zeroed; + if (arg2 == 0) { + return -TARGET_EINVAL; + } + if (get_user_u32(size, arg2)) { + return -TARGET_EFAULT; + } + if (!size) { + size = offsetof(struct target_sched_attr, sched_util_min); + } + if (size < offsetof(struct target_sched_attr, sched_util_min)) { + if (put_user_u32(sizeof(struct target_sched_attr), arg2)) { + return -TARGET_EFAULT; + } + return -TARGET_E2BIG; + } + + zeroed = check_zeroed_user(arg2, sizeof(struct target_sched_attr), size); + if (zeroed < 0) { + return zeroed; + } else if (zeroed == 0) { + if (put_user_u32(sizeof(struct target_sched_attr), arg2)) { + return -TARGET_EFAULT; + } + return -TARGET_E2BIG; + } + if (size > sizeof(struct target_sched_attr)) { + size = sizeof(struct target_sched_attr); + } + + target_scha = lock_user(VERIFY_READ, arg2, size, 1); + if (!target_scha) { + return -TARGET_EFAULT; + } + scha.size = size; + scha.sched_policy = tswap32(target_scha->sched_policy); + scha.sched_flags = tswap64(target_scha->sched_flags); + scha.sched_nice = tswap32(target_scha->sched_nice); + scha.sched_priority = tswap32(target_scha->sched_priority); + scha.sched_runtime = tswap64(target_scha->sched_runtime); + scha.sched_deadline = tswap64(target_scha->sched_deadline); + scha.sched_period = tswap64(target_scha->sched_period); + if (size > offsetof(struct target_sched_attr, sched_util_min)) { + scha.sched_util_min = tswap32(target_scha->sched_util_min); + scha.sched_util_max = tswap32(target_scha->sched_util_max); + } + unlock_user(target_scha, arg2, 0); + return get_errno(sys_sched_setattr(arg1, &scha, arg3)); + } case TARGET_NR_sched_yield: return get_errno(sched_yield()); case TARGET_NR_sched_get_priority_max: diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h index 585e933140..66244589aa 100644 --- a/linux-user/syscall_defs.h +++ b/linux-user/syscall_defs.h @@ -2890,4 +2890,18 @@ struct target_statx { /* 0x100 */ }; +/* from kernel's include/linux/sched/types.h */ +struct target_sched_attr { + abi_uint size; + abi_uint sched_policy; + abi_ullong sched_flags; + abi_int sched_nice; + abi_uint sched_priority; + abi_ullong sched_runtime; + abi_ullong sched_deadline; + abi_ullong sched_period; + abi_uint sched_util_min; + abi_uint sched_util_max; +}; + #endif -- cgit 1.4.1 From 407a119bfd350101a8ef947a9edd58b01644ec89 Mon Sep 17 00:00:00 2001 From: Tonis Tiigi Date: Tue, 4 Jan 2022 20:18:19 -0800 Subject: linux-user: call set/getscheduler set/getparam directly There seems to be difference in syscall and libc definition of these methods and therefore musl does not implement them (1e21e78bf7). Call syscall directly to ensure the behavior of the libc of user application, not the libc that was used to build QEMU. Signed-off-by: Tonis Tiigi Message-Id: <20220105041819.24160-3-tonistiigi@gmail.com> Reviewed-by: Laurent Vivier Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 34 ++++++++++++++++++++++++---------- linux-user/syscall_defs.h | 4 ++++ 2 files changed, 28 insertions(+), 10 deletions(-) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index 6de116eb90..01cd59cdce 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -359,6 +359,17 @@ _syscall4(int, sys_sched_getattr, pid_t, pid, struct sched_attr *, attr, #define __NR_sys_sched_setattr __NR_sched_setattr _syscall3(int, sys_sched_setattr, pid_t, pid, struct sched_attr *, attr, unsigned int, flags); +#define __NR_sys_sched_getscheduler __NR_sched_getscheduler +_syscall1(int, sys_sched_getscheduler, pid_t, pid); +#define __NR_sys_sched_setscheduler __NR_sched_setscheduler +_syscall3(int, sys_sched_setscheduler, pid_t, pid, int, policy, + const struct sched_param *, param); +#define __NR_sys_sched_getparam __NR_sched_getparam +_syscall2(int, sys_sched_getparam, pid_t, pid, + struct sched_param *, param); +#define __NR_sys_sched_setparam __NR_sched_setparam +_syscall2(int, sys_sched_setparam, pid_t, pid, + const struct sched_param *, param); #define __NR_sys_getcpu __NR_getcpu _syscall3(int, sys_getcpu, unsigned *, cpu, unsigned *, node, void *, tcache); _syscall4(int, reboot, int, magic1, int, magic2, unsigned int, cmd, @@ -10794,30 +10805,32 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1, return ret; case TARGET_NR_sched_setparam: { - struct sched_param *target_schp; + struct target_sched_param *target_schp; struct sched_param schp; if (arg2 == 0) { return -TARGET_EINVAL; } - if (!lock_user_struct(VERIFY_READ, target_schp, arg2, 1)) + if (!lock_user_struct(VERIFY_READ, target_schp, arg2, 1)) { return -TARGET_EFAULT; + } schp.sched_priority = tswap32(target_schp->sched_priority); unlock_user_struct(target_schp, arg2, 0); - return get_errno(sched_setparam(arg1, &schp)); + return get_errno(sys_sched_setparam(arg1, &schp)); } case TARGET_NR_sched_getparam: { - struct sched_param *target_schp; + struct target_sched_param *target_schp; struct sched_param schp; if (arg2 == 0) { return -TARGET_EINVAL; } - ret = get_errno(sched_getparam(arg1, &schp)); + ret = get_errno(sys_sched_getparam(arg1, &schp)); if (!is_error(ret)) { - if (!lock_user_struct(VERIFY_WRITE, target_schp, arg2, 0)) + if (!lock_user_struct(VERIFY_WRITE, target_schp, arg2, 0)) { return -TARGET_EFAULT; + } target_schp->sched_priority = tswap32(schp.sched_priority); unlock_user_struct(target_schp, arg2, 1); } @@ -10825,19 +10838,20 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1, return ret; case TARGET_NR_sched_setscheduler: { - struct sched_param *target_schp; + struct target_sched_param *target_schp; struct sched_param schp; if (arg3 == 0) { return -TARGET_EINVAL; } - if (!lock_user_struct(VERIFY_READ, target_schp, arg3, 1)) + if (!lock_user_struct(VERIFY_READ, target_schp, arg3, 1)) { return -TARGET_EFAULT; + } schp.sched_priority = tswap32(target_schp->sched_priority); unlock_user_struct(target_schp, arg3, 0); - return get_errno(sched_setscheduler(arg1, arg2, &schp)); + return get_errno(sys_sched_setscheduler(arg1, arg2, &schp)); } case TARGET_NR_sched_getscheduler: - return get_errno(sched_getscheduler(arg1)); + return get_errno(sys_sched_getscheduler(arg1)); case TARGET_NR_sched_getattr: { struct target_sched_attr *target_scha; diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h index 66244589aa..cca561f622 100644 --- a/linux-user/syscall_defs.h +++ b/linux-user/syscall_defs.h @@ -2904,4 +2904,8 @@ struct target_sched_attr { abi_uint sched_util_max; }; +struct target_sched_param { + abi_int sched_priority; +}; + #endif -- cgit 1.4.1 From e13685a6e5c92612c6eeeb8a17c052539c0f9da5 Mon Sep 17 00:00:00 2001 From: Andrey Kazmin Date: Mon, 27 Dec 2021 15:50:48 +0300 Subject: linux-user/syscall.c: fix missed flag for shared memory in open_self_maps MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The possible variants for region type in /proc/self/maps are either private "p" or shared "s". In the current implementation, we mark shared regions as "-". It could break memory mapping parsers such as included into ASan/HWASan sanitizers. Fixes: 01ef6b9e4e4e ("linux-user: factor out reading of /proc/self/maps") Signed-off-by: Andrey Kazmin Reviewed-by: Laurent Vivier Acked-by: Alex Bennée Message-Id: <20211227125048.22610-1-a.kazmin@partner.samsung.com> Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index 01cd59cdce..3160d77154 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -8045,7 +8045,7 @@ static int open_self_maps(void *cpu_env, int fd) (flags & PAGE_READ) ? 'r' : '-', (flags & PAGE_WRITE_ORG) ? 'w' : '-', (flags & PAGE_EXEC) ? 'x' : '-', - e->is_priv ? 'p' : '-', + e->is_priv ? 'p' : 's', (uint64_t) e->offset, e->dev, e->inode); if (path) { dprintf(fd, "%*s%s\n", 73 - count, "", path); -- cgit 1.4.1 From 7a5626a1d8a193106dc709608273ed48729967d5 Mon Sep 17 00:00:00 2001 From: Ahmed Abouzied Date: Tue, 4 Jan 2022 16:38:41 +0200 Subject: linux-user/syscall.c: malloc to g_try_malloc Use g_try_malloc instead of malloc to alocate the target ifconfig. Also replace the corresponding free with g_free. Signed-off-by: Ahmed Abouzied Message-Id: <20220104143841.25116-1-email@aabouzied.com> Signed-off-by: Laurent Vivier --- linux-user/syscall.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'linux-user/syscall.c') diff --git a/linux-user/syscall.c b/linux-user/syscall.c index 3160d77154..ce9d64896c 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -4915,7 +4915,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp, * We can't fit all the extents into the fixed size buffer. * Allocate one that is large enough and use it instead. */ - host_ifconf = malloc(outbufsz); + host_ifconf = g_try_malloc(outbufsz); if (!host_ifconf) { return -TARGET_ENOMEM; } @@ -4963,7 +4963,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp, } if (free_buf) { - free(host_ifconf); + g_free(host_ifconf); } return ret; -- cgit 1.4.1