Re: [PATCH] arm64: ftrace: don't dereference a probably invalid address

[Mark-PK Tsai]

> On Mon, 7 Jun 2021 22:15:22 +0800
> Mark-PK Tsai <mark-pk.tsai@xxxxxxxxxxxx> wrote:
> 
> > And the crash is becuase kernel trying to read *rec->ip in print_ip_ins() if
> > ftrace_bug() get error code -EINVAL.
> 
> Right, so the actual fix should be something like this:
> 
> [ not tested, nor even compiled ]
> 
> -- Steve
> 

I just test this patch, and turns out it work well.
All my modification is only rename copy_from_kernel_nofault
into probe_kernel_read because the older kernel version
I use.
Then the kernel will show the error mesage instead of panic as following.

[    0.000000] ftrace failed to modify
[    0.000000] [<0000000000000020>] 0x20
[    0.000000] actual:   [FAULT](0000000000000020)

Should I resend this patch as v2?
Or you will upstream this fix?

> 
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 2e8a3fde7104..72ef4dccbcc4 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -1967,12 +1967,18 @@ static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
>  
>  static void print_ip_ins(const char *fmt, const unsigned char *p)
>  {
> +	char ins[MCOUNT_INSN_SIZE];
>  	int i;
>  
> +	if (copy_from_kernel_nofault(ins, p, MCOUNT_INSN_SIZE)) {
> +		printk(KERN_CONT "%s[FAULT](%px)", fmt, p);
> +		return;
> +	}
> +
>  	printk(KERN_CONT "%s", fmt);
>  
>  	for (i = 0; i < MCOUNT_INSN_SIZE; i++)
> -		printk(KERN_CONT "%s%02x", i ? ":" : "", p[i]);
> +		printk(KERN_CONT "%s%02x", i ? ":" : "", ins[i]);
>  }
>  
>  enum ftrace_bug_type ftrace_bug_type;