^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) #include <linux/compiler.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) #include <linux/context_tracking.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) #include <linux/errno.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include <linux/nospec.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include <linux/ptrace.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include <linux/syscalls.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <asm/daifflags.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <asm/debug-monitors.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <asm/fpsimd.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <asm/syscall.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <asm/thread_info.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <asm/unistd.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) long compat_arm_syscall(struct pt_regs *regs, int scno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) long sys_ni_syscall(void);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) static long do_ni_syscall(struct pt_regs *regs, int scno)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #ifdef CONFIG_COMPAT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) long ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) if (is_compat_task()) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) ret = compat_arm_syscall(regs, scno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) if (ret != -ENOSYS)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) return sys_ni_syscall();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) static long __invoke_syscall(struct pt_regs *regs, syscall_fn_t syscall_fn)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) return syscall_fn(regs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) static void invoke_syscall(struct pt_regs *regs, unsigned int scno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) unsigned int sc_nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) const syscall_fn_t syscall_table[])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) long ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) if (scno < sc_nr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) syscall_fn_t syscall_fn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) syscall_fn = syscall_table[array_index_nospec(scno, sc_nr)];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) ret = __invoke_syscall(regs, syscall_fn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) ret = do_ni_syscall(regs, scno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) syscall_set_return_value(current, regs, 0, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) static inline bool has_syscall_work(unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) return unlikely(flags & _TIF_SYSCALL_WORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) int syscall_trace_enter(struct pt_regs *regs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) void syscall_trace_exit(struct pt_regs *regs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) #ifdef CONFIG_ARM64_ERRATUM_1463225
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) DECLARE_PER_CPU(int, __in_cortex_a76_erratum_1463225_wa);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) static void cortex_a76_erratum_1463225_svc_handler(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) u32 reg, val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) if (!unlikely(test_thread_flag(TIF_SINGLESTEP)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) if (!unlikely(this_cpu_has_cap(ARM64_WORKAROUND_1463225)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) __this_cpu_write(__in_cortex_a76_erratum_1463225_wa, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) reg = read_sysreg(mdscr_el1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) val = reg | DBG_MDSCR_SS | DBG_MDSCR_KDE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) write_sysreg(val, mdscr_el1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) asm volatile("msr daifclr, #8");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) isb();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) /* We will have taken a single-step exception by this point */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) write_sysreg(reg, mdscr_el1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) __this_cpu_write(__in_cortex_a76_erratum_1463225_wa, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) static void cortex_a76_erratum_1463225_svc_handler(void) { }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) #endif /* CONFIG_ARM64_ERRATUM_1463225 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) const syscall_fn_t syscall_table[])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) unsigned long flags = current_thread_info()->flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) regs->orig_x0 = regs->regs[0];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) regs->syscallno = scno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) * BTI note:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) * The architecture does not guarantee that SPSR.BTYPE is zero
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) * on taking an SVC, so we could return to userspace with a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) * non-zero BTYPE after the syscall.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) * This shouldn't matter except when userspace is explicitly
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) * doing something stupid, such as setting PROT_BTI on a page
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) * that lacks conforming BTI/PACIxSP instructions, falling
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) * through from one executable page to another with differing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) * PROT_BTI, or messing with BTYPE via ptrace: in such cases,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) * userspace should not be surprised if a SIGILL occurs on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) * syscall return.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) * So, don't touch regs->pstate & PSR_BTYPE_MASK here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) * (Similarly for HVC and SMC elsewhere.)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) cortex_a76_erratum_1463225_svc_handler();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) local_daif_restore(DAIF_PROCCTX);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) if (flags & _TIF_MTE_ASYNC_FAULT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) * Process the asynchronous tag check fault before the actual
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) * syscall. do_notify_resume() will send a signal to userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) * before the syscall is restarted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) syscall_set_return_value(current, regs, -ERESTARTNOINTR, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) if (has_syscall_work(flags)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) * The de-facto standard way to skip a system call using ptrace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) * is to set the system call to -1 (NO_SYSCALL) and set x0 to a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) * suitable error code for consumption by userspace. However,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) * this cannot be distinguished from a user-issued syscall(-1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) * and so we must set x0 to -ENOSYS here in case the tracer doesn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) * issue the skip and we fall into trace_exit with x0 preserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) * This is slightly odd because it also means that if a tracer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) * sets the system call number to -1 but does not initialise x0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) * then x0 will be preserved for all system calls apart from a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) * user-issued syscall(-1). However, requesting a skip and not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) * setting the return value is unlikely to do anything sensible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) * anyway.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) if (scno == NO_SYSCALL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) syscall_set_return_value(current, regs, -ENOSYS, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) scno = syscall_trace_enter(regs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) if (scno == NO_SYSCALL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) goto trace_exit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) invoke_syscall(regs, scno, sc_nr, syscall_table);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) * The tracing status may have changed under our feet, so we have to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * check again. However, if we were tracing entry, then we always trace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * exit regardless, as the old entry assembly did.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) if (!has_syscall_work(flags) && !IS_ENABLED(CONFIG_DEBUG_RSEQ)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) local_daif_mask();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) flags = current_thread_info()->flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) if (!has_syscall_work(flags) && !(flags & _TIF_SINGLESTEP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) local_daif_restore(DAIF_PROCCTX);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) trace_exit:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) syscall_trace_exit(regs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) static inline void sve_user_discard(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) if (!system_supports_sve())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) clear_thread_flag(TIF_SVE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) * task_fpsimd_load() won't be called to update CPACR_EL1 in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) * ret_to_user unless TIF_FOREIGN_FPSTATE is still set, which only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) * happens if a context switch or kernel_neon_begin() or context
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) * modification (sigreturn, ptrace) intervenes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) * So, ensure that CPACR_EL1 is already correct for the fast-path case.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) sve_user_disable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) void do_el0_svc(struct pt_regs *regs)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) sve_user_discard();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) el0_svc_common(regs, regs->regs[8], __NR_syscalls, sys_call_table);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) #ifdef CONFIG_COMPAT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) void do_el0_svc_compat(struct pt_regs *regs)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) el0_svc_common(regs, regs->regs[7], __NR_compat_syscalls,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) compat_sys_call_table);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) #endif