linux-kernel - Re: [PATCH v4 6/9] tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20230830162039.95c575460609cebdc34ab0c1@kernel.org>
Date:   Wed, 30 Aug 2023 16:20:39 +0900
From:   Masami Hiramatsu (Google) <mhiramat@...nel.org>
To:     "Masami Hiramatsu (Google)" <mhiramat@...nel.org>
Cc:     Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Florent Revest <revest@...omium.org>,
        linux-trace-kernel@...r.kernel.org,
        LKML <linux-kernel@...r.kernel.org>,
        Martin KaFai Lau <martin.lau@...ux.dev>,
        bpf <bpf@...r.kernel.org>, Sven Schnelle <svens@...ux.ibm.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Jiri Olsa <jolsa@...nel.org>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Alan Maguire <alan.maguire@...cle.com>,
        Mark Rutland <mark.rutland@....com>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH v4 6/9] tracing/fprobe: Enable fprobe events with
 CONFIG_DYNAMIC_FTRACE_WITH_ARGS

On Thu, 24 Aug 2023 00:16:37 +0900
"Masami Hiramatsu (Google)" <mhiramat@...nel.org> wrote:

> +#if !defined(CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS) || \
> +	defined(CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST)
> +
> +static __always_inline
> +struct pt_regs *perf_fprobe_partial_regs(struct ftrace_regs *fregs)
> +{
> +	/* See include/linux/ftrace.h, this returns &fregs->regs */
> +	return ftrace_partial_regs(fregs, NULL);
> +}
> +
> +#define perf_fprobe_return_regs(regs) do {} while (0)
> +
> +#else /* CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS && !CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST */
> +
> +/* Since fprobe handlers can be nested, pt_regs buffer need to be a stack */
> +#define PERF_FPROBE_REGS_MAX	4
> +
> +struct pt_regs_stack {
> +	struct pt_regs regs[PERF_FPROBE_REGS_MAX];
> +	int idx;
> +};
> +
> +static DEFINE_PER_CPU(struct pt_regs_stack, perf_fprobe_regs);
> +
> +static __always_inline
> +struct pt_regs *perf_fprobe_partial_regs(struct ftrace_regs *fregs)
> +{
> +	struct pt_regs_stack *stack = this_cpu_ptr(&perf_fprobe_regs);
> +	struct pt_regs *regs;
> +
> +	if (stack->idx < PERF_FPROBE_REGS_MAX) {
> +		regs = stack->regs[stack->idx++];
> +		return ftrace_partial_regs(fregs, regs);
> +	}
> +	return NULL;
> +}
> +
> +static __always_inline void perf_fprobe_return_regs(struct pt_regs *regs)
> +{
> +	struct pt_regs_stack *stack = this_cpu_ptr(&perf_fprobe_regs);
> +
> +	if (WARN_ON_ONCE(regs != stack->regs[stack->idx]))
> +		return;
> +
> +	--stack->idx;
> +}

Ah, I found that the perf_trace_buf_alloc() does the same thing. So

perf_trace_buf_alloc(size, &pt_regs, &rctx);

will give us the pt_regs at that point. Trace event does that so I think
it is OK to do that here.

Thank you,

> +
> +#endif /* !CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS || CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST */
> +
>  static int fentry_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> -			    struct pt_regs *regs)
> +			    struct ftrace_regs *fregs)
>  {
>  	struct trace_event_call *call = trace_probe_event_call(&tf->tp);
>  	struct fentry_trace_entry_head *entry;
>  	struct hlist_head *head;
>  	int size, __size, dsize;
> +	struct pt_regs *regs;
>  	int rctx;
>  
> +	regs = perf_fprobe_partial_regs(fregs);
> +	if (!regs)
> +		return -EINVAL;
> +
>  	head = this_cpu_ptr(call->perf_events);
>  	if (hlist_empty(head))
> -		return 0;
> +		goto out;
>  
> -	dsize = __get_data_size(&tf->tp, regs);
> +	dsize = __get_data_size(&tf->tp, fregs);
>  	__size = sizeof(*entry) + tf->tp.size + dsize;
>  	size = ALIGN(__size + sizeof(u32), sizeof(u64));
>  	size -= sizeof(u32);
>  
>  	entry = perf_trace_buf_alloc(size, NULL, &rctx);

Here, we can borrow the pt_regs.

>  	if (!entry)
> -		return 0;
> +		goto out;
>  
>  	entry->ip = entry_ip;
>  	memset(&entry[1], 0, dsize);
> -	store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> +	store_trace_args(&entry[1], &tf->tp, fregs, sizeof(*entry), dsize);
>  	perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
>  			      head, NULL);
> +out:
> +	perf_fprobe_return_regs(regs);
>  	return 0;
>  }
>  NOKPROBE_SYMBOL(fentry_perf_func);
>  
>  static void
>  fexit_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> -		unsigned long ret_ip, struct pt_regs *regs)
> +		unsigned long ret_ip, struct ftrace_regs *fregs)
>  {
>  	struct trace_event_call *call = trace_probe_event_call(&tf->tp);
>  	struct fexit_trace_entry_head *entry;
>  	struct hlist_head *head;
>  	int size, __size, dsize;
> +	struct pt_regs *regs;
>  	int rctx;
>  
> +	regs = perf_fprobe_partial_regs(fregs);
> +	if (!regs)
> +		return;
> +
>  	head = this_cpu_ptr(call->perf_events);
>  	if (hlist_empty(head))
> -		return;
> +		goto out;
>  
> -	dsize = __get_data_size(&tf->tp, regs);
> +	dsize = __get_data_size(&tf->tp, fregs);
>  	__size = sizeof(*entry) + tf->tp.size + dsize;
>  	size = ALIGN(__size + sizeof(u32), sizeof(u64));
>  	size -= sizeof(u32);
>  
>  	entry = perf_trace_buf_alloc(size, NULL, &rctx);

Ditto.

Thanks,
-- 
Masami Hiramatsu (Google) <mhiramat@...nel.org>