x86: ret_from_fork - get rid of jump back (5b3eec0c) · Commits · 戴 / test

Impact: remove dead code If we take a closer look at the rff_trace/rff_action ret_from_fork code, we have to realize that it does all the wrong things: for example it checks the TIF flag - while later on jumping back to the ret-from-syscall path - duplicating the check needlessly. But checking for _TIF_SYSCALL_TRACE is completely unnecessary here because we clear that flag for every freshly forked task. So the whole "tracing" code here, for which there is a out of line jump optimization that makes it even harder to read, is in reality completely dead code ... Reported-by:

Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by:

Ingo Molnar <mingo@elte.hu> Tested-by:

Cyrill Gorcunov <gorcunov@gmail.com>

arch/x86/kernel/entry_64.S

+16 −15

Original line number	Diff line number	Diff line
		@@ -361,34 +361,35 @@ ENTRY(save_paranoid)
		END(save_paranoid)

		/*
		* A newly forked process directly context switches into this.
		* A newly forked process directly context switches into this address.
		*
		* rdi: prev task we switched from
		*/
		/* rdi: prev */
		ENTRY(ret_from_fork)
		DEFAULT_FRAME

		push kernel_eflags(%rip)
		CFI_ADJUST_CFA_OFFSET 8
		popf # reset kernel eflags
		CFI_ADJUST_CFA_OFFSET -8
		call schedule_tail

		call schedule_tail # rdi: 'prev' task parameter

		GET_THREAD_INFO(%rcx)
		testl $(_TIF_SYSCALL_TRACE\|_TIF_SYSCALL_AUDIT),TI_flags(%rcx)

		CFI_REMEMBER_STATE
		jnz rff_trace
		rff_action:
		RESTORE_REST

		testl $3, CS-ARGOFFSET(%rsp) # from kernel_thread?
		je int_ret_from_sys_call
		testl $_TIF_IA32,TI_flags(%rcx)

		testl $_TIF_IA32, TI_flags(%rcx) # 32-bit compat task needs IRET
		jnz int_ret_from_sys_call

		RESTORE_TOP_OF_STACK %rdi, -ARGOFFSET
		jmp ret_from_sys_call
		jmp ret_from_sys_call # go to the SYSRET fastpath

		CFI_RESTORE_STATE
		rff_trace:
		movq %rsp,%rdi
		call syscall_trace_leave
		GET_THREAD_INFO(%rcx)
		jmp rff_action
		CFI_ENDPROC
		END(ret_from_fork)

Admin message