[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f0aa7983-0300-ce21-8726-41d033f6afbe@tuxon.dev>
Date: Fri, 15 Sep 2023 08:51:41 +0300
From: claudiu beznea <claudiu.beznea@...on.dev>
To: Geert Uytterhoeven <geert@...ux-m68k.org>
Cc: mturquette@...libre.com, sboyd@...nel.org, robh+dt@...nel.org,
krzysztof.kozlowski+dt@...aro.org, conor+dt@...nel.org,
ulf.hansson@...aro.org, linus.walleij@...aro.org,
gregkh@...uxfoundation.org, jirislaby@...nel.org,
magnus.damm@...il.com, catalin.marinas@....com, will@...nel.org,
prabhakar.mahadev-lad.rj@...renesas.com,
biju.das.jz@...renesas.com, quic_bjorande@...cinc.com,
arnd@...db.de, konrad.dybcio@...aro.org, neil.armstrong@...aro.org,
nfraprado@...labora.com, rafal@...ecki.pl,
wsa+renesas@...g-engineering.com,
linux-renesas-soc@...r.kernel.org, linux-clk@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-mmc@...r.kernel.org, linux-gpio@...r.kernel.org,
linux-serial@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
Claudiu Beznea <claudiu.beznea.uj@...renesas.com>
Subject: Re: [PATCH 12/37] clk: renesas: rzg2l: reduce the critical area
On 14.09.2023 16:12, Geert Uytterhoeven wrote:
> Hi Claudiu,
>
> On Tue, Sep 12, 2023 at 6:52 AM Claudiu <claudiu.beznea@...on.dev> wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@...renesas.com>
>>
>> spinlock in rzg2l_mod_clock_endisable() is intended to protect the accesses
>> to hardware register. There is no need to protect the instructions that set
>> temporary variable which will be then written to register. Thus limit the
>> spinlock only to the hardware register access.
>>
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@...renesas.com>
>
> Thanks for your patch!
>
>> --- a/drivers/clk/renesas/rzg2l-cpg.c
>> +++ b/drivers/clk/renesas/rzg2l-cpg.c
>> @@ -912,13 +912,13 @@ static int rzg2l_mod_clock_endisable(struct clk_hw *hw, bool enable)
>>
>> dev_dbg(dev, "CLK_ON %u/%pC %s\n", CLK_ON_R(reg), hw->clk,
>> enable ? "ON" : "OFF");
>> - spin_lock_irqsave(&priv->rmw_lock, flags);
>>
>> value = bitmask << 16;
>> if (enable)
>> value |= bitmask;
>> - writel(value, priv->base + CLK_ON_R(reg));
>>
>> + spin_lock_irqsave(&priv->rmw_lock, flags);
>> + writel(value, priv->base + CLK_ON_R(reg));
>> spin_unlock_irqrestore(&priv->rmw_lock, flags);
>
> After this, it becomes obvious there is nothing to protect at all,
> so the locking can just be removed from this function?
I tend to be paranoid when writing to hardware resources thus I kept it.
Would you prefer to remove it at all?
>
> Gr{oetje,eeting}s,
>
> Geert
>
Powered by blists - more mailing lists