Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Leonard Crestez <leonard.crestez@nxp.com>
To:     Peng Fan <peng.fan@nxp.com>,
        "jassisinghbrar@gmail.com" <jassisinghbrar@gmail.com>,
        "o.rempel@pengutronix.de" <o.rempel@pengutronix.de>
CC:     "shawnguo@kernel.org" <shawnguo@kernel.org>,
        "s.hauer@pengutronix.de" <s.hauer@pengutronix.de>,
        "kernel@pengutronix.de" <kernel@pengutronix.de>,
        "festevam@gmail.com" <festevam@gmail.com>,
        dl-linux-imx <linux-imx@nxp.com>,
        Anson Huang <anson.huang@nxp.com>,
        Aisheng Dong <aisheng.dong@nxp.com>,
        "linux-arm-kernel@lists.infradead.org" 
        <linux-arm-kernel@lists.infradead.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH V6 0/4] mailbox/firmware: imx: support SCU channel type
Thread-Topic: [PATCH V6 0/4] mailbox/firmware: imx: support SCU channel type
Thread-Index: AQHV8emRabKKTJAuM0OK7xtU5vZU9A==
Date:   Wed, 18 Mar 2020 02:44:32 +0000
Message-ID: <VI1PR04MB694108DF36F465324BA45E05EEF70@VI1PR04MB6941.eurprd04.prod.outlook.com>
References: <1583300977-2327-1-git-send-email-peng.fan@nxp.com>
 <VI1PR04MB7023455D0FE9766FFBE1EE5EEEFD0@VI1PR04MB7023.eurprd04.prod.outlook.com>
 <AM0PR04MB448123609B2FE8F5ECAF1F4388FA0@AM0PR04MB4481.eurprd04.prod.outlook.com>
 <AM0PR04MB4481D74E3C38B047562F419988FA0@AM0PR04MB4481.eurprd04.prod.outlook.com>
Accept-Language: en-US
Content-Language: en-US
received-spf: None (protection.outlook.com: nxp.com does not designate
 permitted sender hosts)
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-Network-Message-Id: 10ae6edb-df68-4914-4b60-08d7cae6499b
X-MS-Exchange-CrossTenant-originalarrivaltime: 18 Mar 2020 02:44:32.1202
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 686ea1d3-bc2b-4c6f-a92c-d99c5c301635
X-MS-Exchange-CrossTenant-mailboxtype: HOSTED
X-MS-Exchange-CrossTenant-userprincipalname: E90VmxLc2pXfyYKWXdcVnTDYpz7eg2iTpg8WKYNsZ9rIL7s5BtNIOwgbYEzotsY2ou+jbiLoJSXP0GAc5ZpPXg==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: VI1PR04MB4447
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 2020-03-13 9:38 AM, Peng Fan wrote:=0A=
>> Subject: RE: [PATCH V6 0/4] mailbox/firmware: imx: support SCU channel=
=0A=
>> type=0A=
>>=0A=
>> Hi Leonard,=0A=
>>=0A=
>>> Subject: Re: [PATCH V6 0/4] mailbox/firmware: imx: support SCU channel=
=0A=
>>> type=0A=
>>>=0A=
>>> On 2020-03-04 7:55 AM, Peng Fan wrote:=0A=
>>>> From: Peng Fan <peng.fan@nxp.com>=0A=
>>>>=0A=
>>>> V6:=0A=
>>>>    Add Oleksij's R-b tag=0A=
>>>>    Patch 3/4, per=0A=
>>> https://www.kernel.org/doc/Documentation/printk-formats.txt=0A=
>>>>    should use %zu for printk sizeof=0A=
>>>>=0A=
>>>> V5:=0A=
>>>>    Move imx_mu_dcfg below imx_mu_priv=0A=
>>>>    Add init hooks to imx_mu_dcfg=0A=
>>>>    drop __packed __aligned=0A=
>>>>    Add more debug msg=0A=
>>>>    code style cleanup=0A=
>>>>=0A=
>>>> V4:=0A=
>>>>    Drop IMX_MU_TYPE_[GENERIC, SCU]=0A=
>>>>    Pack MU chans init to separate function=0A=
>>>>    Add separate function for SCU chans init and xlate=0A=
>>>>    Add santity check to msg hdr.size=0A=
>>>>    Limit SCU MU chans to 6, TX0/RX0/RXDB[0-3]=0A=
>>>>=0A=
>>>> V3:=0A=
>>>>    Rebase to Shawn's for-next=0A=
>>>>    Include fsl,imx8-mu-scu compatible=0A=
>>>>    Per Oleksij's comments, introduce generic tx/rx and added scu mu ty=
pe=0A=
>>>>    Check fsl,imx8-mu-scu in firmware driver for fast_ipc=0A=
>>>>=0A=
>>>> V2:=0A=
>>>>    Drop patch 1/3 which added fsl,scu property=0A=
>>>>    Force to use scu channel type when machine has node compatible=0A=
>>> "fsl,imx-scu"=0A=
>>>>    Force imx-scu to use fast_ipc=0A=
>>>>=0A=
>>>>    I not found a generic method to make SCFW message generic enough,=
=0A=
>>> SCFW=0A=
>>>>    message is not fixed length including TX and RX. And it use TR0/RR0=
=0A=
>>>>    interrupt.=0A=
>>>>=0A=
>>>> V1:=0A=
>>>> Sorry to bind the mailbox/firmware patch together. This is make it=0A=
>>>> to understand what changed to support using 1 TX and 1 RX channel=0A=
>>>> for SCFW message.=0A=
>>>>=0A=
>>>> Per i.MX8QXP Reference mannual, there are several message using=0A=
>>>> examples. One of them is:=0A=
>>>> Passing short messages: Transmit register(s) can be used to pass=0A=
>>>> short messages from one to four words in length. For example, when a=
=0A=
>>>> four-word message is desired, only one of the registers needs to=0A=
>>>> have its corresponding interrupt enable bit set at the receiver side.=
=0A=
>>>>=0A=
>>>> This patchset is to using this for SCFW message to replace four TX=0A=
>>>> and four RX method.=0A=
>>>=0A=
>>> Tested-by: Leonard Crestez <leonard.crestez@nxp.com>=0A=
>>>=0A=
>>=0A=
>> Thanks for the test.=0A=
>>=0A=
>>> My stress tests pass on imx8qxp with this patcheset, however=0A=
>>> performance is not greatly improved. My guess is that this happens=0A=
>>> because of too many interrupts.=0A=
>>=0A=
>> Might be. Could you share your testcase?=0A=
=0A=
https://github.com/cdleonard/imx-scu-test=0A=
=0A=
>>> Is there really a reason to enable TIE? Spinning on TE bits without=0A=
>>> any interrupts should be just plain faster.=0A=
>>=0A=
>> I could try to disable TIE and give a try. If performance improves lot, =
I could=0A=
>> change to non TX interrupt.=0A=
> =0A=
> After rethinking about this, we need TX interrupt, otherwise we have to=
=0A=
> use TX_POLL which is slower or let the client kick the TX state machine.=
=0A=
> =0A=
> Compared with original method, this already reduces to use 1 TX and 1 RX=
=0A=
> interrupt. This already good for system.=0A=
=0A=
Sorry, I missed that fact that your patches don't include the required =0A=
DTS changes. Indeed that is only one TX and one RX irq per call now.=0A=
=0A=
Running my test now results in RX timeout :(=0A=
=0A=
-----=0A=
=0A=
On an unrelated note: are you sure it is appropriate to change the =0A=
compat string here? Another way to implement direct SCU communication =0A=
would be as another channel type, IMX_MU_TYPE_SCUTX.=0A=
=0A=
It also strange that you're adding a bool fast_ipc in imx-scu, do we =0A=
really want to support the old path?=0A=
=0A=
If SCU protocol was implemented as a channel type then maybe we could =0A=
sidestep mbox_request_channel_by_name, parse mboxes manually and always =0A=
request MU_TYPE_SCUTX.=0A=
=0A=
--=0A=
Regards,=0A=
Leonard=0A=