On 11/30, Yixun Lan wrote:
> Hi Stephen
>
> On 11/30/17 03:34, Stephen Boyd wrote:
> > On 11/28, Yixun Lan wrote:
> >> diff --git a/drivers/clk/meson/axg.c b/drivers/clk/meson/axg.c
> >> new file mode 100644
> >> index 000000000000..51c5b4062715
> >> --- /dev/null
> >> +++ b/drivers/clk/meson/axg.c
> >> @@ -0,0 +1,948 @@
> >> +/*
> >> + * AmLogic Meson-AXG Clock Controller Driver
> >> + *
> >> + * Copyright (c) 2016 Baylibre SAS.
> >> + * Author: Michael Turquette <[email protected]>
> >> + *
> >> + * Copyright (c) 2017 Amlogic, inc.
> >> + * Author: Qiufang Dai <[email protected]>
> >> + *
> >> + * SPDX-License-Identifier: GPL-2.0+
> >> + */
> >> +
> >> +#include <linux/clk.h>
> >> +#include <linux/clk-provider.h>
> >> +#include <linux/of_address.h>
> >> +#include <linux/of_device.h>
> >> +#include <linux/platform_device.h>
> >> +#include <linux/init.h>
> >> +
> >> +#include "clkc.h"
> >> +#include "axg.h"
> >> +
> >> +static DEFINE_SPINLOCK(clk_lock);
> >
> > meson_axg_clk_lock?
> >
> em... I'd leave it unchanged
>
> because the spinlock will be used at macro MESON_GATE() [1] which
> defined at drivers/clk/meson/clkc.h, and it assume using the generic
> name 'clk_lock', change name will break the code..
>
> and besides it's already defined as static, so I see no problem here
The problem is lockdep debugging and ctags/grep on the source
code. clk_lock is very generic when it should be more specific so
we can find this lock later on from a lockdep report with a
simple search of the code.
Maybe make another patch to rename it to meson_clk_lock so the
macro doesn't have to change too much, and we get it slightly
more unique. Looks like v4l2 also has a clk_lock.
>
> >> +};
> >> +
> >> +static int axg_clkc_probe(struct platform_device *pdev)
> >> +{
> >> + const struct clkc_data *clkc_data;
> >> + void __iomem *clk_base;
> >> + int ret, clkid, i;
> >> + struct device *dev = &pdev->dev;
> >> +
> >> + clkc_data = of_device_get_match_data(&pdev->dev);
> >> + if (!clkc_data)
> >> + return -EINVAL;
> >> +
> >> + /* Generic clocks and PLLs */
> >> + clk_base = of_iomap(dev->of_node, 0);
> >
> > Use platform device APIs for ioremapping?
> >
> I assume you are referring to 'platform_get_resource +
> devm_ioremap_resource' ?
>
> the idea sounds good to me.
Yes.
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project