[PATCH 3/3] mtd: gpmi: change the code for clocks
Huang Shijie
b32955 at freescale.com
Fri Jun 29 01:52:40 EDT 2012
于 2012年06月29日 10:06, Shawn Guo 写道:
>> gpmi_regs + HW_GPMI_TIMING1);
>> >
>> > /* Get the timing information we need. */
>> > - nfc->clock_frequency_in_hz = clk_get_rate(r->clock);
>> > + nfc->clock_frequency_in_hz = clk_get_rate(r->clock[0]);
>> > clock_period_in_ns = 1000000000 / nfc->clock_frequency_in_hz;
>> >
>> > gpmi_nfc_compute_hardware_timing(this,&hw);
>> > @@ -784,8 +812,7 @@ err_out:
>> >
>> > void gpmi_end(struct gpmi_nand_data *this)
>> > {
>> > - struct resources *r =&this->resources;
>> > - clk_disable_unprepare(r->clock);
>> > + gpmi_disable_clk(this);
>> > }
>> >
>> > /* Clears a BCH interrupt. */
>> > diff --git a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> > index 941cfb7..edda3b1 100644
>> > --- a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> > +++ b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> > @@ -464,9 +464,59 @@ acquire_err:
>> > return -EINVAL;
>> > }
>> >
>> > +static void gpmi_put_clks(struct gpmi_nand_data *this)
>> > +{
>> > + struct resources *r =&this->resources;
>> > + struct clk *clk;
>> > + int i;
>> > +
>> > + for (i = 0; i< GPMI_CLK_MAX; i++) {
>> > + clk = r->clock[i];
>> > + if (clk) {
>> > + clk_put(clk);
>> > + r->clock[i] = NULL;
>> > + }
>> > + }
>> > +}
>> > +
>> > +static char *extra_clks_for_mx6q[] = {
>> > + "gpmi_apb", "gpmi_bch", "gpmi_bch_apb", "per1_bch",
>> > +};
>> > +
>> > +static int __devinit gpmi_get_clks(struct gpmi_nand_data *this)
>> > +{
>> > + struct resources *r =&this->resources;
>> > + char **extra_clks = NULL;
>> > + struct clk *clk;
>> > + int i;
>> > +
>> > + r->clock[0] = clk_get(&this->pdev->dev, NULL);
>> > + if (IS_ERR(r->clock[0]))
>> > + goto err_clock;
>> > +
>> > + /* Get extra clocks */
>> > + if (GPMI_IS_MX6Q(this))
>> > + extra_clks = extra_clks_for_mx6q;
> We probably do not need this tweaking. We can have the driver always
> take all those 5 clocks, and I think the current imx28 clock driver
> can just work with it, because the gpmi-nand clkdev lookup has NULL
> con_id and all those 5 clocks can match the same one on imx28.
>
I think your method makes the code hard to understand.
My code is more clear in logic.
thanks
Huang Shijie
More information about the linux-mtd
mailing list