[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID:
<MEYP282MB26975755ACB814BD556334BABB13A@MEYP282MB2697.AUSP282.PROD.OUTLOOK.COM>
Date: Thu, 10 Aug 2023 23:19:59 +0800
From: Jinjian Song <songjinjian@...mail.com>
To: jiri@...nulli.us
Cc: chandrashekar.devegowda@...el.com,
chiranjeevi.rapolu@...ux.intel.com,
danielwinkler@...gle.com,
davem@...emloft.net,
edumazet@...gle.com,
haijun.liu@...iatek.com,
ilpo.jarvinen@...ux.intel.com,
jesse.brandeburg@...el.com,
jinjian.song@...ocom.com,
johannes@...solutions.net,
kuba@...nel.org,
linuxwwan@...el.com,
linuxwwan_5g@...el.com,
loic.poulain@...aro.org,
m.chetan.kumar@...ux.intel.com,
netdev@...r.kernel.org,
pabeni@...hat.com,
ricardo.martinez@...ux.intel.com,
ryazanov.s.a@...il.com,
songjinjian@...mail.com,
soumya.prakash.mishra@...el.com,
vsankar@...ovo.com
Subject: Re: [net-next 2/6] net: wwan: t7xx: Driver registers with Devlink framework
>Mon, Aug 07, 2023 at 02:44:00PM CEST, songjinjian@...mail.com wrote:
>>Sat, Aug 05, 2023 at 02:12:13PM CEST, songjinjian@...mail.com wrote:
>>>>Thu, Aug 03, 2023 at 04:18:08AM CEST, songjinjian@...mail.com wrote:
>>>>
>>>>[...]
>>>>
>>>>>>+static const struct devlink_param t7xx_devlink_params[] = {
>>>>>>+ DEVLINK_PARAM_DRIVER(T7XX_DEVLINK_PARAM_ID_FASTBOOT,
>>>>>>+ "fastboot", DEVLINK_PARAM_TYPE_BOOL,
>>>>>>+ BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
>>>>>>+ NULL, NULL, NULL),
>>>>>
>>>>>driver init params is there so the user could configure driver instance
>>>>>and then hit devlink reload in order to reinitialize with the new param
>>>>>values. In your case, it is a device command. Does not make any sense to
>>>>>have it as param.
>>>>>
>>>>>NAK
>>>>
>>>>Thanks for your review, so drop this param and set this param insider driver when driver_reinit,
>>>>is that fine?
>>>
>>>I don't understand the question, sorry.
>>
>>Thanks for your review, I mean if I don't define fastboot param like devlink_param above, I will
>
>Could you not thank me in every reply, please?
>
>
>>define a global bool variable in driver, then when devlink ... driver_reinit, set this variable to true.
>
>That would be very wrong. Driver reinit should just reload the driver,
>recreate the entities created. should not be trigger to change
>behaviour.
Thanks for your review, if so, I hope it is possible to keep the devlink param as above.
>
>>
>>like:
>> t7xx_devlink {
>> ....
>> bool reset_to_fastboot;
>> }
>>
>>
>> t7xx_devlink_reload_down () {
>> ...
>> case DEVLINK_RELOAD_ACTION_DRIVER_REINIT:
>> t7xx_devlink.reset_to_fastboot = true;
>> ...
>> }
>>
>> other functions use this variable:
>>
>> if (t7xx_devlink.reset_to_fastboot) {
>> iowrite(reg, "reset to fastboot");
>> }
>>
>>Intel colleague has change to the way of devlink_param, so I hope to keep this.
>>
>>>
>
Powered by blists - more mailing lists