Re: [PATCH for review] [127/145] i386: movekernel_thread_helper into entry.S

[Jan Beulich]

>> And this too - the value now in %eax has no relation with the
>> value known by the caller of this routine (which doesn't expect
>> any return from here anyway).
>
>Ok, but somehow it needs to be annotiated so that the unwinder stops
>and doesn't fall back. Can you please send a replacement patch that 
>does this correctly? 

Actually, with the previous attempt we still didn't achieve the
original
goal of terminating the frame chain at kernel_thread_helper. Thus
another try (the seemingly odd extra push can be avoided once we
start using CFI expressions, which the unwinder currently doesn't
support):

ENTRY(kernel_thread_helper)
	CFI_STARTPROC
	pushl $0			# fake return address
	movl %edx,%eax
	pushl %edx
	CFI_ADJUST_CFA_OFFSET 4
	call *%ebx
	pushl %eax
	CFI_ADJUST_CFA_OFFSET 4
	call do_exit
	CFI_ENDPROC
ENDPROC(kernel_thread_helper)

Jan
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/