diff options
Diffstat (limited to 'arch/x86/kernel/entry_64.S')
| -rw-r--r-- | arch/x86/kernel/entry_64.S | 193 |
1 files changed, 98 insertions, 95 deletions
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a6..42571baaca32 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -11,15 +11,15 @@ * * NOTE: This code handles signal-recognition, which happens every time * after an interrupt and after each system call. - * - * Normal syscalls and interrupts don't save a full stack frame, this is + * + * Normal syscalls and interrupts don't save a full stack frame, this is * only done for syscall tracing, signals or fork/exec et.al. - * - * A note on terminology: - * - top of stack: Architecture defined interrupt frame from SS to RIP - * at the top of the kernel process stack. + * + * A note on terminology: + * - top of stack: Architecture defined interrupt frame from SS to RIP + * at the top of the kernel process stack. * - partial stack frame: partially saved registers upto R11. - * - full stack frame: Like partial stack frame, but all register saved. + * - full stack frame: Like partial stack frame, but all register saved. * * Some macro usage: * - CFI macros are used to generate dwarf2 unwind information for better @@ -142,7 +142,7 @@ END(mcount) #ifndef CONFIG_PREEMPT #define retint_kernel retint_restore_args -#endif +#endif #ifdef CONFIG_PARAVIRT ENTRY(native_usergs_sysret64) @@ -161,14 +161,14 @@ ENTRY(native_usergs_sysret64) .endm /* - * C code is not supposed to know about undefined top of stack. Every time - * a C function with an pt_regs argument is called from the SYSCALL based + * C code is not supposed to know about undefined top of stack. Every time + * a C function with an pt_regs argument is called from the SYSCALL based * fast path FIXUP_TOP_OF_STACK is needed. * RESTORE_TOP_OF_STACK syncs the syscall state after any possible ptregs * manipulation. - */ - - /* %rsp:at FRAMEEND */ + */ + + /* %rsp:at FRAMEEND */ .macro FIXUP_TOP_OF_STACK tmp movq %gs:pda_oldrsp,\tmp movq \tmp,RSP(%rsp) @@ -244,8 +244,8 @@ ENTRY(native_usergs_sysret64) .endm /* * A newly forked process directly context switches into this. - */ -/* rdi: prev */ + */ +/* rdi: prev */ ENTRY(ret_from_fork) CFI_DEFAULT_STACK push kernel_eflags(%rip) @@ -255,8 +255,9 @@ ENTRY(ret_from_fork) call schedule_tail GET_THREAD_INFO(%rcx) testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),TI_flags(%rcx) + CFI_REMEMBER_STATE jnz rff_trace -rff_action: +rff_action: RESTORE_REST testl $3,CS-ARGOFFSET(%rsp) # from kernel_thread? je int_ret_from_sys_call @@ -264,10 +265,11 @@ rff_action: jnz int_ret_from_sys_call RESTORE_TOP_OF_STACK %rdi,ARGOFFSET jmp ret_from_sys_call + CFI_RESTORE_STATE rff_trace: movq %rsp,%rdi call syscall_trace_leave - GET_THREAD_INFO(%rcx) + GET_THREAD_INFO(%rcx) jmp rff_action CFI_ENDPROC END(ret_from_fork) @@ -278,20 +280,20 @@ END(ret_from_fork) * SYSCALL does not save anything on the stack and does not change the * stack pointer. */ - + /* - * Register setup: + * Register setup: * rax system call number * rdi arg0 - * rcx return address for syscall/sysret, C arg3 + * rcx return address for syscall/sysret, C arg3 * rsi arg1 - * rdx arg2 + * rdx arg2 * r10 arg3 (--> moved to rcx for C) * r8 arg4 * r9 arg5 * r11 eflags for syscall/sysret, temporary for C - * r12-r15,rbp,rbx saved by C code, not touched. - * + * r12-r15,rbp,rbx saved by C code, not touched. + * * Interrupts are off on entry. * Only called from user space. * @@ -301,7 +303,7 @@ END(ret_from_fork) * When user can change the frames always force IRET. That is because * it deals with uncanonical addresses better. SYSRET has trouble * with them due to bugs in both AMD and Intel CPUs. - */ + */ ENTRY(system_call) CFI_STARTPROC simple @@ -317,7 +319,7 @@ ENTRY(system_call) */ ENTRY(system_call_after_swapgs) - movq %rsp,%gs:pda_oldrsp + movq %rsp,%gs:pda_oldrsp movq %gs:pda_kernelstack,%rsp /* * No need to follow this irqs off/on section - it's straight @@ -325,7 +327,7 @@ ENTRY(system_call_after_swapgs) */ ENABLE_INTERRUPTS(CLBR_NONE) SAVE_ARGS 8,1 - movq %rax,ORIG_RAX-ARGOFFSET(%rsp) + movq %rax,ORIG_RAX-ARGOFFSET(%rsp) movq %rcx,RIP-ARGOFFSET(%rsp) CFI_REL_OFFSET rip,RIP-ARGOFFSET GET_THREAD_INFO(%rcx) @@ -339,19 +341,19 @@ system_call_fastpath: movq %rax,RAX-ARGOFFSET(%rsp) /* * Syscall return path ending with SYSRET (fast path) - * Has incomplete stack frame and undefined top of stack. - */ + * Has incomplete stack frame and undefined top of stack. + */ ret_from_sys_call: movl $_TIF_ALLWORK_MASK,%edi /* edi: flagmask */ -sysret_check: +sysret_check: LOCKDEP_SYS_EXIT GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF movl TI_flags(%rcx),%edx andl %edi,%edx - jnz sysret_careful + jnz sysret_careful CFI_REMEMBER_STATE /* * sysretq will re-enable interrupts: @@ -366,7 +368,7 @@ sysret_check: CFI_RESTORE_STATE /* Handle reschedules */ - /* edx: work, edi: workmask */ + /* edx: work, edi: workmask */ sysret_careful: bt $TIF_NEED_RESCHED,%edx jnc sysret_signal @@ -379,7 +381,7 @@ sysret_careful: CFI_ADJUST_CFA_OFFSET -8 jmp sysret_check - /* Handle a signal */ + /* Handle a signal */ sysret_signal: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) @@ -398,7 +400,7 @@ sysret_signal: DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp int_with_check - + badsys: movq $-ENOSYS,RAX-ARGOFFSET(%rsp) jmp ret_from_sys_call @@ -437,7 +439,7 @@ sysret_audit: #endif /* CONFIG_AUDITSYSCALL */ /* Do syscall tracing */ -tracesys: +tracesys: #ifdef CONFIG_AUDITSYSCALL testl $(_TIF_WORK_SYSCALL_ENTRY & ~_TIF_SYSCALL_AUDIT),TI_flags(%rcx) jz auditsys @@ -460,8 +462,8 @@ tracesys: call *sys_call_table(,%rax,8) movq %rax,RAX-ARGOFFSET(%rsp) /* Use IRET because user could have changed frame */ - -/* + +/* * Syscall return path ending with IRET. * Has correct top of stack, but partial stack frame. */ @@ -505,18 +507,18 @@ int_very_careful: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - /* Check for syscall exit trace */ + /* Check for syscall exit trace */ testl $_TIF_WORK_SYSCALL_EXIT,%edx jz int_signal pushq %rdi CFI_ADJUST_CFA_OFFSET 8 - leaq 8(%rsp),%rdi # &ptregs -> arg1 + leaq 8(%rsp),%rdi # &ptregs -> arg1 call syscall_trace_leave popq %rdi CFI_ADJUST_CFA_OFFSET -8 andl $~(_TIF_WORK_SYSCALL_EXIT|_TIF_SYSCALL_EMU),%edi jmp int_restore_rest - + int_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz 1f @@ -531,11 +533,11 @@ int_restore_rest: jmp int_with_check CFI_ENDPROC END(system_call) - -/* + +/* * Certain special system calls that need to save a complete full stack frame. - */ - + */ + .macro PTREGSCALL label,func,arg .globl \label \label: @@ -572,7 +574,7 @@ ENTRY(ptregscall_common) ret CFI_ENDPROC END(ptregscall_common) - + ENTRY(stub_execve) CFI_STARTPROC popq %r11 @@ -588,11 +590,11 @@ ENTRY(stub_execve) jmp int_ret_from_sys_call CFI_ENDPROC END(stub_execve) - + /* * sigreturn is special because it needs to restore all registers on return. * This cannot be done with SYSRET, so use the IRET return path instead. - */ + */ ENTRY(stub_rt_sigreturn) CFI_STARTPROC addq $8, %rsp @@ -685,12 +687,12 @@ exit_intr: GET_THREAD_INFO(%rcx) testl $3,CS-ARGOFFSET(%rsp) je retint_kernel - + /* Interrupt came from user space */ /* * Has a correct top of stack, but a partial stack frame * %rcx: thread info. Interrupts off. - */ + */ retint_with_reschedule: movl $_TIF_WORK_MASK,%edi retint_check: @@ -763,20 +765,20 @@ retint_careful: pushq %rdi CFI_ADJUST_CFA_OFFSET 8 call schedule - popq %rdi + popq %rdi CFI_ADJUST_CFA_OFFSET -8 GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp retint_check - + retint_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz retint_swapgs TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - movq $-1,ORIG_RAX(%rsp) + movq $-1,ORIG_RAX(%rsp) xorl %esi,%esi # oldset movq %rsp,%rdi # &pt_regs call do_notify_resume @@ -798,14 +800,14 @@ ENTRY(retint_kernel) jnc retint_restore_args call preempt_schedule_irq jmp exit_intr -#endif +#endif CFI_ENDPROC END(common_interrupt) - + /* * APIC interrupts. - */ + */ .macro apicinterrupt num,func INTR_FRAME pushq $~(\num) @@ -823,14 +825,14 @@ ENTRY(threshold_interrupt) apicinterrupt THRESHOLD_APIC_VECTOR,mce_threshold_interrupt END(threshold_interrupt) -#ifdef CONFIG_SMP +#ifdef CONFIG_SMP ENTRY(reschedule_interrupt) apicinterrupt RESCHEDULE_VECTOR,smp_reschedule_interrupt END(reschedule_interrupt) .macro INVALIDATE_ENTRY num ENTRY(invalidate_interrupt\num) - apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt + apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt END(invalidate_interrupt\num) .endm @@ -869,22 +871,22 @@ END(error_interrupt) ENTRY(spurious_interrupt) apicinterrupt SPURIOUS_APIC_VECTOR,smp_spurious_interrupt END(spurious_interrupt) - + /* * Exception entry points. - */ + */ .macro zeroentry sym INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 /* push error code/oldrax */ + pushq $0 /* push error code/oldrax */ CFI_ADJUST_CFA_OFFSET 8 - pushq %rax /* push real oldrax to the rdi slot */ + pushq %rax /* push real oldrax to the rdi slot */ CFI_ADJUST_CFA_OFFSET 8 CFI_REL_OFFSET rax,0 leaq \sym(%rip),%rax jmp error_entry CFI_ENDPROC - .endm + .endm .macro errorentry sym XCPT_FRAME @@ -998,13 +1000,13 @@ paranoid_schedule\trace: /* * Exception entry point. This expects an error code/orig_rax on the stack - * and the exception handler in %rax. - */ + * and the exception handler in %rax. + */ KPROBE_ENTRY(error_entry) _frame RDI CFI_REL_OFFSET rax,0 /* rdi slot contains rax, oldrax contains error code */ - cld + cld subq $14*8,%rsp CFI_ADJUST_CFA_OFFSET (14*8) movq %rsi,13*8(%rsp) @@ -1015,7 +1017,7 @@ KPROBE_ENTRY(error_entry) CFI_REL_OFFSET rdx,RDX movq %rcx,11*8(%rsp) CFI_REL_OFFSET rcx,RCX - movq %rsi,10*8(%rsp) /* store rax */ + movq %rsi,10*8(%rsp) /* store rax */ CFI_REL_OFFSET rax,RAX movq %r8, 9*8(%rsp) CFI_REL_OFFSET r8,R8 @@ -1025,29 +1027,29 @@ KPROBE_ENTRY(error_entry) CFI_REL_OFFSET r10,R10 movq %r11,6*8(%rsp) CFI_REL_OFFSET r11,R11 - movq %rbx,5*8(%rsp) + movq %rbx,5*8(%rsp) CFI_REL_OFFSET rbx,RBX - movq %rbp,4*8(%rsp) + movq %rbp,4*8(%rsp) CFI_REL_OFFSET rbp,RBP - movq %r12,3*8(%rsp) + movq %r12,3*8(%rsp) CFI_REL_OFFSET r12,R12 - movq %r13,2*8(%rsp) + movq %r13,2*8(%rsp) CFI_REL_OFFSET r13,R13 - movq %r14,1*8(%rsp) + movq %r14,1*8(%rsp) CFI_REL_OFFSET r14,R14 - movq %r15,(%rsp) + movq %r15,(%rsp) CFI_REL_OFFSET r15,R15 - xorl %ebx,%ebx + xorl %ebx,%ebx testl $3,CS(%rsp) je error_kernelspace -error_swapgs: +error_swapgs: SWAPGS error_sti: TRACE_IRQS_OFF - movq %rdi,RDI(%rsp) + movq %rdi,RDI(%rsp) CFI_REL_OFFSET rdi,RDI movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ movq $-1,ORIG_RAX(%rsp) call *%rax /* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ @@ -1056,7 +1058,7 @@ error_exit: RESTORE_REST DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF - GET_THREAD_INFO(%rcx) + GET_THREAD_INFO(%rcx) testl %eax,%eax jne retint_kernel LOCKDEP_SYS_EXIT_IRQ @@ -1072,7 +1074,7 @@ error_kernelspace: /* There are two places in the kernel that can potentially fault with usergs. Handle them here. The exception handlers after iret run with kernel gs again, so don't set the user space flag. - B stepping K8s sometimes report an truncated RIP for IRET + B stepping K8s sometimes report an truncated RIP for IRET exceptions returning to compat mode. Check for these here too. */ leaq irq_return(%rip),%rcx cmpq %rcx,RIP(%rsp) @@ -1084,17 +1086,17 @@ error_kernelspace: je error_swapgs jmp error_sti KPROBE_END(error_entry) - + /* Reload gs selector with exception handling */ - /* edi: new selector */ + /* edi: new selector */ ENTRY(native_load_gs_index) CFI_STARTPROC pushf CFI_ADJUST_CFA_OFFSET 8 DISABLE_INTERRUPTS(CLBR_ANY | ~(CLBR_RDI)) SWAPGS -gs_change: - movl %edi,%gs +gs_change: + movl %edi,%gs 2: mfence /* workaround */ SWAPGS popf @@ -1102,20 +1104,20 @@ gs_change: ret CFI_ENDPROC ENDPROC(native_load_gs_index) - + .section __ex_table,"a" .align 8 .quad gs_change,bad_gs .previous .section .fixup,"ax" /* running with kernelgs */ -bad_gs: +bad_gs: SWAPGS /* switch back to user gs */ xorl %eax,%eax movl %eax,%gs jmp 2b - .previous - + .previous + /* * Create a kernel thread. * @@ -1138,7 +1140,7 @@ ENTRY(kernel_thread) xorl %r8d,%r8d xorl %r9d,%r9d - + # clone now call do_fork movq %rax,RAX(%rsp) @@ -1149,14 +1151,14 @@ ENTRY(kernel_thread) * so internally to the x86_64 port you can rely on kernel_thread() * not to reschedule the child before returning, this avoids the need * of hacks for example to fork off the per-CPU idle tasks. - * [Hopefully no generic code relies on the reschedule -AK] + * [Hopefully no generic code relies on the reschedule -AK] */ RESTORE_ALL UNFAKE_STACK_FRAME ret CFI_ENDPROC ENDPROC(kernel_thread) - + child_rip: pushq $0 # fake return address CFI_STARTPROC @@ -1170,6 +1172,7 @@ child_rip: # exit mov %eax, %edi call do_exit + ud2 # padding for call trace CFI_ENDPROC ENDPROC(child_rip) @@ -1191,10 +1194,10 @@ ENDPROC(child_rip) ENTRY(kernel_execve) CFI_STARTPROC FAKE_STACK_FRAME $0 - SAVE_ALL + SAVE_ALL movq %rsp,%rcx call sys_execve - movq %rax, RAX(%rsp) + movq %rax, RAX(%rsp) RESTORE_REST testq %rax,%rax je int_ret_from_sys_call @@ -1213,7 +1216,7 @@ ENTRY(coprocessor_error) END(coprocessor_error) ENTRY(simd_coprocessor_error) - zeroentry do_simd_coprocessor_error + zeroentry do_simd_coprocessor_error END(simd_coprocessor_error) ENTRY(device_not_available) @@ -1225,12 +1228,12 @@ KPROBE_ENTRY(debug) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_debug, DEBUG_STACK paranoidexit KPROBE_END(debug) - /* runs on exception stack */ + /* runs on exception stack */ KPROBE_ENTRY(nmi) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME @@ -1264,7 +1267,7 @@ ENTRY(bounds) END(bounds) ENTRY(invalid_op) - zeroentry do_invalid_op + zeroentry do_invalid_op END(invalid_op) ENTRY(coprocessor_segment_overrun) @@ -1319,7 +1322,7 @@ ENTRY(machine_check) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_machine_check jmp paranoid_exit1 CFI_ENDPROC |
