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;
Subject: Re: [PATCH RFC 0/2] mach-omap2: handle autoidle denial
To:     Tony Lindgren <tony@atomide.com>,
        Andreas Kemnade <andreas@kemnade.info>
CC:     <mturquette@baylibre.com>, <sboyd@kernel.org>,
        <linux-omap@vger.kernel.org>, <linux-clk@vger.kernel.org>,
        <linux-kernel@vger.kernel.org>, <paul@pwsan.com>,
        <letux-kernel@openphoenux.org>
References: <20181004055147.23048-1-andreas@kemnade.info>
 <20181004142558.GB5662@atomide.com>
From:   Tero Kristo <t-kristo@ti.com>
Message-ID: <013b01a1-2593-bdc0-dd9a-e5a114388067@ti.com>
Date:   Thu, 4 Oct 2018 17:42:25 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <20181004142558.GB5662@atomide.com>
Content-Type: text/plain; charset="utf-8"; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 04/10/18 17:25, Tony Lindgren wrote:
> * Andreas Kemnade <andreas@kemnade.info> [181004 05:56]:
>> On the gta04 with a dm3730 omap_hdq does not work properly when the
>> device enters lower power states. Idling uart1 and 2 is enough
>> to show up that problem, if there are no other things enabled.
>> Further research reveals that hdq iclk must not be turned off during
>> transfers, also according to the TRM. That fact is also correctly described
>> in the flags but the code to handle that is incomplete.
>>
>> Since the order is first disable all autoidles, then disable selected
>> and then enable all, we need to either change that order or add
>> a usecount. Since it is done only in init, we could think about changing
>> order.
> 
> These patches look OK to me, assuming Tero will review them more
> closely.

There is no locking whatsoever in the autoidle counting atm, that must 
be fixed otherwise you get races.

> It seems we should just provide a generic interface for
> clk_allow_autoidle() and clk_deny_autoidle()? Otherwise we'll
> be forever stuck with pdata callbacks it seems.

The TI clock driver is actually providing these APIs, so that should be 
fine. I don't think there is any use / need for pdata callbacks atm, it 
just happens hwmod core is calling these at the moment which might have 
confused you.

-Tero
--
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki