[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4CBE206A.20702@zytor.com>
Date: Tue, 19 Oct 2010 15:49:14 -0700
From: "H. Peter Anvin" <hpa@...or.com>
To: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
CC: Steven Rostedt <rostedt@...dmis.org>,
Thomas Gleixner <tglx@...utronix.de>,
Koki Sanagi <sanagi.koki@...fujitsu.com>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...e.hu>,
Frederic Weisbecker <fweisbec@...il.com>,
nhorman@...driver.com, scott.a.mcmillan@...el.com,
laijs@...fujitsu.com, LKML <linux-kernel@...r.kernel.org>,
eric.dumazet@...il.com, kaneshige.kenji@...fujitsu.com,
David Miller <davem@...emloft.net>, izumi.taku@...fujitsu.com,
kosaki.motohiro@...fujitsu.com,
Heiko Carstens <heiko.carstens@...ibm.com>,
"Luck, Tony" <tony.luck@...el.com>, Jason Baron <jbaron@...hat.com>
Subject: Re: [PATCH] tracing: Cleanup the convoluted softirq tracepoints
On 10/19/2010 03:41 PM, Mathieu Desnoyers wrote:
>>
>> OK, first of all, there are some serious WTFs here:
>>
>> # define JUMP_LABEL_INITIAL_NOP ".byte 0xe9 \n\t .long 0\n\t"
>>
>> A jump instruction is one of the worst possible NOPs. Why are we doing
>> this?
>
> This code is dynamically patched at boot time (and module load time) with a
> better nop, just like the function tracer does.
>
That's just ridiculous... start out with something sane and you at least
have the chance of not having to patch it.
> Intel's manual "Intel 64 and IA-32 Architectures Optimization Reference Manual"
>
> http://www.intel.com/Assets/PDF/manual/248966.pdf
>
> Page C-33 (or 577 in the pdf)
>
> "7. Selection of conditional jump instructions should be based on the
> recommendation of section Section 3.4.1, “Branch Prediction Optimization,” to
> improve the predictability of branches. When branches are predicted
> successfully, the latency of jcc is effectively zero."
>
> So it mentions "jcc", but not jmp. Is there any reason for jmp to have a higher
> latency than jcc ?
>
> In this manual, the latency of predicted jcc is therefore 0 cycle, and its
> throughput is 0.5 cycle/insn.
>
> NOP (page C-29) is stated to have a latency of 0.5 to 1 cycle/insn (depending on
> the exact HW), and throughput of 0.5 cycle/insn.
>
> However, I have not found "jmp" explicitly in this listing.
>
> So if we were executing tracepoints in a maze of jumps, we could argue that
> instruction throughput is the most important there. However, if we expect the
> common case to be surrounded by some non-ALU instructions, latency tends to
> become the most important criterion.
>
> But I feel I might be missing something important that distinguish "jcc" from
> "jmp".
NOP has a latency of 0.5-1.0 cycle/insns, *but has no consumers*.
JMP/Jcc does have a consumer -- the IP -- and actually measuring shows
that it is much, much worse than NOP and other dummy instructions.
-hpa
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists