From: Roger Quadros <rogerq@ti.com>
To: Brian Norris <computersforpeace@gmail.com>
Cc: <tony@atomide.com>, <dwmw2@infradead.org>,
<ezequiel@vanguardiasur.com.ar>, <javier@dowhile0.org>,
<fcooper@ti.com>, <nsekhar@ti.com>,
<linux-mtd@lists.infradead.org>, <linux-omap@vger.kernel.org>,
<devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3 00/27] memory: omap-gpmc: mtd: nand: Support GPMC NAND on non-OMAP platforms
Date: Wed, 2 Dec 2015 10:42:12 +0530 [thread overview]
Message-ID: <565E7DAC.7060401@ti.com> (raw)
In-Reply-To: <20151202032643.GF64635@google.com>
Brian,
On 02/12/15 08:56, Brian Norris wrote:
> Hi Roger,
>
> On Tue, Dec 01, 2015 at 04:41:16PM +0200, Roger Quadros wrote:
>> On 30/11/15 21:54, Brian Norris wrote:
>>> On Tue, Oct 27, 2015 at 11:37:03AM +0200, Roger Quadros wrote:
>>>> On 26/10/15 23:23, Brian Norris wrote:
>>>>> On Fri, Sep 18, 2015 at 05:53:22PM +0300, Roger Quadros wrote:
>>>>>> - Remove NAND IRQ handling from omap-gpmc driver, share the GPMC IRQ
>>>>>> with the omap2-nand driver and handle NAND IRQ events in the NAND driver.
>>>>>> This causes performance increase when using prefetch-irq mode.
>>>>>> 30% increase in read, 17% increase in write in prefetch-irq mode.
>>>>>
>>>>> Have you pinpointed the exact causes for the performance increase, or
>>>>> can you give an educated guess? AIUI, you're reducing the number of
>>>>> interrupts needed for NAND prefetch mode, but you're also removing a bit
>>>>> of abstraction and implementing hooks that look awfully like the
>>>>> existing abstractions:
>>>>>
>>>>> + int (*nand_irq_enable)(enum gpmc_nand_irq irq);
>>>>> + int (*nand_irq_disable)(enum gpmc_nand_irq irq);
>>>>> + void (*nand_irq_clear)(enum gpmc_nand_irq irq);
>>>>> + u32 (*nand_irq_status)(void);
>>>>>
>>>>> That's not really a problem if there's a good reason for them (brcmnand
>>>>> implements similar hooks because of quirks in the implementation of
>>>>> interrupts across various BRCM SoCs, and it's not worth writing irqchip
>>>>> drivers for those cases). I'm mainly curious for an explanation.
>>>>
>>>> I have both implementations with me. My guess is that the 20% performance
>>>> gain is due to absence of irqchip/irqdomain translation code.
>>>> I haven't investigated further though.
>>>
>>> I don't have much context for whether this makes sense or not. According
>>> to your tests, you're getting ~800K interrupts over ~15 seconds. So
>>> should you start noticing performance hits due to abstraction at 53K
>>> interrupts per second?
>>
>> Yes, this was my understanding.
>
> Am I computing wrong, or is that a pretty insane rate of interrupts?
I don't have the test board with me right now and so can't tell you
for sure if the mtd tests took 15 seconds or more.
I can try it out on a different board that I have and let you know
for sure about how many interrupts we get per second.
>
>>> But anyway, I'm not sure that completely answered my question. My
>>> question was whether you were removing the irqchip code solely for
>>> performance reasons, or are there others?
>>
>> Yes. Only for performance reasons.
>
> Hmm, that's not my favorite answer. I'd prefer that more analysis was
> done here before scrapping irqchip...
I agree. We could retain the irqchip model till we have more satisfying
analysis.
>
> But maybe that's not too bad. It seems like your patch set overall is a
> net positive for disentangling some of arch/ and drivers/.
:)
>
> I'll take another pass over your patch set, but if things are looking
> better, how do you expect to merge this? There are significant portions
> that touch at least 2 or 3 different subsystem trees, AFAICT.
Tony could create an immutable branch with all the dts and memory changes.
You could base the mtd changes on top of that?
cheers,
-roger
next prev parent reply other threads:[~2015-12-02 5:13 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-18 14:53 [PATCH v3 00/27] memory: omap-gpmc: mtd: nand: Support GPMC NAND on non-OMAP platforms Roger Quadros
2015-09-18 14:53 ` [PATCH v3 01/27] ARM: OMAP2+: gpmc: Add platform data Roger Quadros
2015-09-18 14:53 ` [PATCH v3 02/27] ARM: OMAP2+: gpmc: Add gpmc timings and settings to " Roger Quadros
2015-09-18 14:53 ` [PATCH v3 03/27] memory: omap-gpmc: Introduce GPMC to NAND interface Roger Quadros
2015-09-18 14:53 ` [PATCH v3 04/27] mtd: nand: omap2: Use gpmc_omap_get_nand_ops() to get NAND registers Roger Quadros
2015-12-03 5:00 ` Brian Norris
2015-09-18 14:53 ` [PATCH v3 05/27] memory: omap-gpmc: Add GPMC-NAND ops to get writebufferempty status Roger Quadros
2015-09-18 14:53 ` [PATCH v3 06/27] mtd: nand: omap2: Switch to using GPMC-NAND ops for writebuffer empty check Roger Quadros
2015-09-18 14:53 ` [PATCH v3 07/27] memory: omap-gpmc: Remove NAND IRQ code Roger Quadros
2015-09-18 14:53 ` [PATCH v3 08/27] memory: omap-gpmc: Add IRQ ops for GPMC-NAND interface Roger Quadros
2015-09-18 14:53 ` [PATCH v3 09/27] mtd: nand: omap2: manage NAND interrupts Roger Quadros
2015-09-18 14:53 ` [PATCH v3 10/27] mtd: nand: omap: Copy platform data parameters to omap_nand_info data Roger Quadros
2015-09-18 14:53 ` [PATCH v3 11/27] mtd: nand: omap: Clean up device tree support Roger Quadros
2015-10-06 10:35 ` [PATCH v4 " Roger Quadros
2015-12-03 4:29 ` Brian Norris
2015-12-03 5:57 ` Roger Quadros
2015-12-03 6:09 ` Brian Norris
2015-09-18 14:53 ` [PATCH v3 12/27] mtd: nand: omap: Update DT binding documentation Roger Quadros
2015-09-18 14:53 ` [PATCH v3 13/27] memory: omap-gpmc: Prevent mapping into 1st 16MB Roger Quadros
2015-09-18 14:53 ` [PATCH v3 14/27] memory: omap-gpmc: Move device tree binding to correct location Roger Quadros
2015-09-18 14:53 ` [PATCH v3 15/27] memory: omap-gpmc: Support general purpose input for WAITPINs Roger Quadros
2015-09-18 14:53 ` [PATCH v3 16/27] memory: omap-gpmc: Reserve WAITPIN if needed for WAIT monitoring Roger Quadros
2015-09-18 14:53 ` [PATCH v3 17/27] memory: omap-gpmc: Add irqchip support to the gpiochip Roger Quadros
2015-09-18 14:53 ` [PATCH v3 18/27] mtd: nand: omap2: Implement NAND ready using gpiolib Roger Quadros
2015-10-26 20:49 ` Brian Norris
2015-10-27 8:03 ` Roger Quadros
2015-10-27 8:12 ` Boris Brezillon
2015-10-27 8:43 ` Roger Quadros
2015-10-27 8:28 ` Boris Brezillon
2015-12-03 4:45 ` Brian Norris
2015-12-03 8:41 ` Boris Brezillon
2015-09-18 14:53 ` [PATCH v3 19/27] memory: omap-gpmc: Prevent GPMC_STATUS from being accessed via gpmc_regs Roger Quadros
2015-09-18 14:53 ` [PATCH v3 20/27] ARM: dts: dra7: Fix NAND device nodes Roger Quadros
2015-10-14 13:34 ` Franklin S Cooper Jr.
2015-10-14 14:17 ` Roger Quadros
2015-10-14 14:37 ` Franklin S Cooper Jr.
2015-09-18 14:53 ` [PATCH v3 21/27] ARM: dts: dra7x-evm: Provide NAND ready pin Roger Quadros
2015-09-18 14:53 ` [PATCH v3 22/27] ARM: dts: am437x: Fix NAND device nodes Roger Quadros
2015-09-18 14:53 ` [PATCH v3 23/27] ARM: dts: am437x-gp-evm: Provide NAND ready pin Roger Quadros
2015-09-18 14:53 ` [PATCH v3 24/27] ARM: dts: am335x: Fix NAND device nodes Roger Quadros
2015-09-18 14:53 ` [PATCH v3 25/27] ARM: dts: am335x: Provide NAND ready pin Roger Quadros
2015-09-18 14:53 ` [PATCH v3 26/27] ARM: dts: dm816x: Fix gpmc and NAND node Roger Quadros
2015-09-18 14:53 ` [PATCH v3 27/27] ARM: dts: omap3: Fix gpmc and NAND nodes Roger Quadros
2015-10-13 0:43 ` Tony Lindgren
2015-10-13 6:29 ` Roger Quadros
2015-10-13 15:18 ` Tony Lindgren
2015-10-14 7:39 ` Roger Quadros
2015-10-14 8:55 ` [PATCH v4 " Roger Quadros
2015-09-30 7:39 ` [PATCH v3 00/27] memory: omap-gpmc: mtd: nand: Support GPMC NAND on non-OMAP platforms Roger Quadros
2015-09-30 11:00 ` Roger Quadros
2015-10-06 8:33 ` Tony Lindgren
2015-10-06 9:54 ` Roger Quadros
2015-10-06 10:00 ` Tony Lindgren
2015-10-06 10:05 ` Roger Quadros
2015-10-06 10:28 ` Roger Quadros
2015-10-06 11:01 ` Tony Lindgren
2015-10-06 11:09 ` Roger Quadros
2015-10-16 21:25 ` Tony Lindgren
2015-10-19 7:08 ` Roger Quadros
2015-10-21 8:31 ` Roger Quadros
2015-10-21 15:20 ` Tony Lindgren
2015-10-23 7:09 ` Roger Quadros
2015-11-30 17:26 ` Tony Lindgren
2015-10-26 21:23 ` Brian Norris
2015-10-27 9:37 ` Roger Quadros
2015-11-25 10:42 ` Roger Quadros
2015-11-30 19:54 ` Brian Norris
2015-12-01 14:41 ` Roger Quadros
2015-12-02 3:26 ` Brian Norris
2015-12-02 5:12 ` Roger Quadros [this message]
2015-12-02 15:03 ` Tony Lindgren
2015-12-02 18:13 ` Brian Norris
2015-12-02 20:05 ` Tony Lindgren
2015-12-02 18:43 ` Brian Norris
2015-12-03 5:09 ` Brian Norris
2015-12-03 6:08 ` Roger Quadros
2015-12-03 6:22 ` Brian Norris
2015-12-03 9:01 ` Roger Quadros
2015-12-03 15:17 ` Tony Lindgren
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=565E7DAC.7060401@ti.com \
--to=rogerq@ti.com \
--cc=computersforpeace@gmail.com \
--cc=devicetree@vger.kernel.org \
--cc=dwmw2@infradead.org \
--cc=ezequiel@vanguardiasur.com.ar \
--cc=fcooper@ti.com \
--cc=javier@dowhile0.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mtd@lists.infradead.org \
--cc=linux-omap@vger.kernel.org \
--cc=nsekhar@ti.com \
--cc=tony@atomide.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).