Re: [RFC][PATCH 2/3] x86_64: Allow breakpoints to emulate call functions

[Josh Poimboeuf]

On Tue, May 07, 2019 at 01:42:29PM -0400, Steven Rostedt wrote:
> From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> 
> In order to allow breakpoints to emulate call functions, they need to push
> the return address onto the stack. But because the breakpoint exception
> frame is added to the stack when the breakpoint is hit, there's no room to
> add the address onto the stack and return to the address of the emulated
> called funtion.
> 
> To handle this, copy the exception frame on entry of the breakpoint handler
> and have leave a gap that can be used to add a return address to the stack
> frame and return from the breakpoint to the emulated called function,
> allowing for that called function to return back to the location after the
> breakpoint was placed.

This part is done by patch 1.

> 
> The helper functions were also added:

No longer "also" :-)

>   int3_emulate_push(): to push the address onto the gap in the stack
>   int3_emulate_jmp(): changes the location of the regs->ip to return there.
>   int3_emulate_call(): push the return address and change regs->ip
> 
> Cc: Andy Lutomirski <luto@xxxxxxxxxx>
> Cc: Nicolai Stange <nstange@xxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Borislav Petkov <bp@xxxxxxxxx>
> Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
> Cc: the arch/x86 maintainers <x86@xxxxxxxxxx>
> Cc: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> Cc: Jiri Kosina <jikos@xxxxxxxxxx>
> Cc: Miroslav Benes <mbenes@xxxxxxx>
> Cc: Petr Mladek <pmladek@xxxxxxxx>
> Cc: Joe Lawrence <joe.lawrence@xxxxxxxxxx>
> Cc: Shuah Khan <shuah@xxxxxxxxxx>
> Cc: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
> Cc: Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>
> Cc: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> Cc: Mimi Zohar <zohar@xxxxxxxxxxxxx>
> Cc: Juergen Gross <jgross@xxxxxxxx>
> Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> Cc: Nayna Jain <nayna@xxxxxxxxxxxxx>
> Cc: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
> Cc: Joerg Roedel <jroedel@xxxxxxx>
> Cc: "open list:KERNEL SELFTEST FRAMEWORK" <linux-kselftest@xxxxxxxxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> Fixes: b700e7f03df5 ("livepatch: kernel: add support for live patching")
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
> [ Modified to only work for x86_64 ]
> Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
> ---
>  arch/x86/include/asm/text-patching.h | 22 ++++++++++++++++++++++
>  1 file changed, 22 insertions(+)
> 
> diff --git a/arch/x86/include/asm/text-patching.h b/arch/x86/include/asm/text-patching.h
> index e85ff65c43c3..455bf9f88233 100644
> --- a/arch/x86/include/asm/text-patching.h
> +++ b/arch/x86/include/asm/text-patching.h
> @@ -39,4 +39,26 @@ extern int poke_int3_handler(struct pt_regs *regs);
>  extern void *text_poke_bp(void *addr, const void *opcode, size_t len, void *handler);
>  extern int after_bootmem;
>  
> +static inline void int3_emulate_jmp(struct pt_regs *regs, unsigned long ip)
> +{
> +	regs->ip = ip;
> +}
> +
> +#define INT3_INSN_SIZE 1
> +#define CALL_INSN_SIZE 5
> +
> +#ifdef CONFIG_X86_64
> +static inline void int3_emulate_push(struct pt_regs *regs, unsigned long val)
> +{
> +	regs->sp -= sizeof(unsigned long);
> +	*(unsigned long *)regs->sp = val;
> +}

How this works isn't really obvious.  A comment is probably warranted to
explain the fact that the int3 entry code reserved some space on the
stack.

-- 
Josh