linux-kernel - Re: [PATCH v2 2/5] tracing: Add division and multiplication support for hist triggers

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: <20211019222756.1fde436b@gandalf.local.home>
Date:   Tue, 19 Oct 2021 22:27:56 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     Kalesh Singh <kaleshsingh@...gle.com>
Cc:     surenb@...gle.com, hridya@...gle.com, namhyung@...nel.org,
        kernel-team@...roid.com, Jonathan Corbet <corbet@....net>,
        Ingo Molnar <mingo@...hat.com>, Shuah Khan <shuah@...nel.org>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Tom Zanussi <zanussi@...nel.org>, linux-doc@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Subject: Re: [PATCH v2 2/5] tracing: Add division and multiplication support
 for hist triggers

On Tue, 19 Oct 2021 18:31:39 -0700
Kalesh Singh <kaleshsingh@...gle.com> wrote:

> +static u64 hist_field_div(struct hist_field *hist_field,
> +			   struct tracing_map_elt *elt,
> +			   struct trace_buffer *buffer,
> +			   struct ring_buffer_event *rbe,
> +			   void *event)
> +{
> +	struct hist_field *operand1 = hist_field->operands[0];
> +	struct hist_field *operand2 = hist_field->operands[1];
> +
> +	u64 val1 = operand1->fn(operand1, elt, buffer, rbe, event);
> +	u64 val2 = operand2->fn(operand2, elt, buffer, rbe, event);
> +
> +	/* Return -1 for the undefined case */
> +	if (!val2)
> +		return -1;
> +
> +	return div64_u64(val1, val2);
> +}
> +

I wonder if you should add a shift operator as well?

I mean, if for some reason you want to divide by a power of two, then why
us the division. Especially if this is on a 32 bit machine.

Of course, the parsing could detect that. If the divisor is a constant. Or
we could even optimize the above with:

	if (!val2)
		return -1;

	if (!(val2 & (val2 - 1))
		return val1 >> __ffs64(val2);

Which should be faster than a divide, and even if it isn't a power of two,
the subtract and & should be in the noise compared to the divide.

Note, the above can be added to this. I'm not suggesting changing this
patch.

-- Steve