linux-kernel - Re: [PATCH 1/2] clk: mvebu: set flags in CP110 gate clock

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 30 Aug 2016 15:34:03 +0200
From:   Marcin Wojtas <mw@...ihalf.com>
To:     Thomas Petazzoni <thomas.petazzoni@...e-electrons.com>
Cc:     linux-kernel@...r.kernel.org,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>, linux-clk@...r.kernel.org,
        Stephen Boyd <sboyd@...eaurora.org>,
        Michael Turquette <mturquette@...libre.com>,
        Andrew Lunn <andrew@...n.ch>,
        Jason Cooper <jason@...edaemon.net>,
        Tomasz Nowicki <tn@...ihalf.com>, nadavh@...vell.com,
        Lior Amsalem <alior@...vell.com>,
        Grzegorz Jaszczyk <jaz@...ihalf.com>,
        Gregory Clément 
        <gregory.clement@...e-electrons.com>,
        Sebastian Hesselbarth <sebastian.hesselbarth@...il.com>
Subject: Re: [PATCH 1/2] clk: mvebu: set flags in CP110 gate clock

Hi Thomas,

2016-08-30 15:10 GMT+02:00 Thomas Petazzoni
<thomas.petazzoni@...e-electrons.com>:
> Hello,
>
> On Tue, 23 Aug 2016 08:26:48 +0200, Marcin Wojtas wrote:
>
>> diff --git a/drivers/clk/mvebu/cp110-system-controller.c b/drivers/clk/mvebu/cp110-system-controller.c
>> index 7fa42d6..0835e1d 100644
>> --- a/drivers/clk/mvebu/cp110-system-controller.c
>> +++ b/drivers/clk/mvebu/cp110-system-controller.c
>> @@ -144,6 +144,7 @@ static struct clk *cp110_register_gate(const char *name,
>>
>>       init.name = name;
>>       init.ops = &cp110_gate_ops;
>> +     init.flags = CLK_IS_BASIC;
>
> Is this really correct?
>
> The documentation for CLK_IS_BASIC is pretty slim, but it says:
>
>    #define CLK_IS_BASIC            BIT(5) /* Basic clk, can't do a to_clk_foo() */
>
> However, we *do* have a to_clk_*() macro in this driver:
>
> struct cp110_gate_clk {
>         struct clk_hw hw;
>         struct regmap *regmap;
>         u8 bit_idx;
> };
>
> #define to_cp110_gate_clk(clk) container_of(clk, struct cp110_gate_clk, hw)
>
> If you read the commit log of commit
> f7d8caadfd2813cbada82ce9041b13c38e8e5282, which introduced the flag, it
> says:
>
>     clk: Add CLK_IS_BASIC flag to identify basic clocks
>
>     Most platforms end up using a mix of basic clock types and
>     some which use clk_hw_foo struct for filling in custom platform
>     information when the clocks don't fit into basic types supported.
>
>     In platform code, its useful to know if a clock is using a basic
>     type or clk_hw_foo, which helps platforms know if they can
>     safely use to_clk_hw_foo to derive the clk_hw_foo pointer from
>     clk_hw.
>
>     Mark all basic clocks with a CLK_IS_BASIC flag.
>
>     Signed-off-by: Rajendra Nayak <rnayak@...com>
>     Signed-off-by: Mike Turquette <mturquette@...aro.org>
>
> We are in the case where we have our own clk_hw_foo structure, and a
> to_clk_hw_foo macro to derive the clk_hw_foo from clk_hw.
>
> According to this, the CP110 clocks are *not* basic clocks, and
> therefore we shouldn't have this flag. Perhaps just the memset() is
> missing.
>

I agree, from functional point of view and considering also not exact
fit to CLK_IS_BASIC definition, memset should be sufficient.

Best regards,
Marcin