Re: [PATCH] filter: Optimize instruction revalidation code.

[Hagen Paul Pfeifer <hagen@jauu.net>]

* Changli Gao | 2010-11-17 07:31:51 [+0800]:

>> Maybe I don't get it, but you increment the opcode by one, but you never
>> increment the opcode in sk_run_filter() - do I miss something? Did you test
>> the your patch (a trivial tcpdump rule should be sufficient)?
>>
>
>+               code = codes[code];
>+               if (!code--)
>+                       return -EINVAL;

Right, temporary in sk_chk_filter() but as I wrote earlier not in sk_run_filter().

>But how about this:
>
>enum {
>        BPF_S_RET_K = 1,

better.

Best regards, Hagen

BTW: you can verify your code by construct some artificial filter rules via

struct sock_filter {
	uint16_t  code;   /* Actual filter code */
	uint8_t   jt;     /* Jump true */
	uint8_t   jf;     /* Jump false */
	uint32_t  k;      /* Generic multiuse field
};

and attach them to a socket with setsockopt.