Received-SPF: pass (google.com: domain of linux-kernel+bounces-56080-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Message-ID: <978c8110-3083-4fc2-a348-d7f97fd7839c@ti.com>
Date: Wed, 7 Feb 2024 13:03:37 +0530
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v2] clk: keystone: sci-clk: Adding support for non
 contiguous clocks
To: CHANDRU DHAVAMANI <chandru@ti.com>, Kamlesh Gurudasani <kamlesh@ti.com>,
        Nishanth Menon <nm@ti.com>
CC: <kristo@kernel.org>, <ssantosh@kernel.org>, <rishabh@ti.com>,
        <vigneshr@ti.com>, <mturquette@baylibre.com>, <sboyd@kernel.org>,
        <linux-arm-kernel@lists.infradead.org>, <linux-kernel@vger.kernel.org>,
        <linux-clk@vger.kernel.org>
References: <20240206104357.3803517-1-u-kumar1@ti.com>
 <20240206131420.wtitflgav23jto2q@verbally>
 <871q9pzoiq.fsf@kamlesh.i-did-not-set--mail-host-address--so-tickle-me>
 <c5b6bd1d-dbb4-4bfb-8b3e-9b0733e2ba5d@ti.com>
 <c2b7f22d-f07d-4cac-8a01-af7b014e7ff4@ti.com>
 <a4fdbcfe-e0e0-4280-8638-e39b6b46778e@ti.com>
 <87ttmly4fa.fsf@kamlesh.i-did-not-set--mail-host-address--so-tickle-me>
 <e2ea03ce-0367-413b-aca9-7c1fabda173a@ti.com>
 <4398af20-3396-4a5c-8fcb-a7b3d0d6f15b@ti.com>
Content-Language: en-US
From: "Kumar, Udit" <u-kumar1@ti.com>
In-Reply-To: <4398af20-3396-4a5c-8fcb-a7b3d0d6f15b@ti.com>
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Transfer-Encoding: 8bit


On 2/7/2024 12:53 PM, CHANDRU DHAVAMANI wrote:
>
> On 07/02/24 11:03, Kumar, Udit wrote:
>>
>> On 2/6/2024 9:24 PM, Kamlesh Gurudasani wrote:
>>> "Kumar, Udit" <u-kumar1@ti.com> writes:
>>>
>>>>>>>> get_freq is a bit expensive as it has to walk the clock tree to 
>>>>>>>> find
>>>>>>>> the clock frequency (at least the first time?). just wondering if
>>>>>>>> there is lighter alternative here?
>>>>>>>>
>>>>>>> How about get_clock? Doesn't read the registers at least.
>>>>>> Said API needs, some flags to be passed,
>>>>>>
>>>>>> Can those flag be set to zero, Chandru ?
>>>>>
>>>>> get_clock doesn't require any flags to be passed.
>>>>
>>>> May be firmware does not need it but  I was referring to
>>>>
>>>> https://elixir.bootlin.com/linux/latest/source/drivers/clk/keystone/sci-clk.c#L78 
>>>>
>>> Just took a look,
>>>
>>> I now understand the reason for confusion,
>>>
>>> #define TI_SCI_MSG_SET_CLOCK_STATE    0x0100
>>> #define TI_SCI_MSG_GET_CLOCK_STATE    0x0101
>>>
>>> cops->get_clock = ti_sci_cmd_get_clock;  --> refers to
>>> TI_SCI_MSG_SET_CLOCK_STATE
>>> That's why we are passing the flag from linux for get_clock
>>>
>>> Linux is using terminology of get/put.
>>>
>>> As Chandru pointed, we don't have to pass flags, cause he is refering
>>> to TI_SCI_MSG_GET_CLOCK_STATE
>>>
>>> Below functions passes TI_SCI_MSG_GET_CLOCK_STATE to DM, which is what
>>> we actually want.
>>> cops->is_auto = ti_sci_cmd_clk_is_auto;
>>> cops->is_on = ti_sci_cmd_clk_is_on;
>>> cops->is_off = ti_sci_cmd_clk_is_off;
>>
>>
>> I think calling ti_sci_cmd_clk_is_auto should be good . other 
>> functions needs current state and requested state.
>>
>> Chandru ?
>>
>
> ti_sci_cmd_clk_is_auto is internal function to linux.
> For TI_SCI_MSG_GET_CLOCK_STATE, linux only needs to pass pointers to 
> the variables where result will be stored.


Yes this internal function calls TI_SCI_MSG_GET_CLOCK_STATE


>
>
>>>
>>> Which should be safe to call, Chandru can confirm.
>>>
>>> Regards,
>>> Kamlesh
>>>>
>>>>
>>>>>
>>>>>>
>>>>>>> Regards,
>>>>>>> Kamlesh