Re: [PATCH v6 2/2] mtd: nand: sunxi: add reset line support

[Boris Brezillon <boris.brezillon@free-electrons.com>]

On Fri, 24 Jun 2016 07:20:38 +0800
Icenowy Zheng <icenowy@aosc.xyz> wrote:

> In my opinion, return directly PTR_ERR(nfc->reset) is OK here.
> If devm_reset_control_get_optional() return -EPROBE_DEFER, the code here will also return it. However, if we get other error, why should it return -EPROBE_DEFER again?

Sorry, I just had a brainfart :-). Your implementation is correct.
BTW, can you avoid top-posting and reply inline?

> 
> 24.06.2016, 00:01, "Boris Brezillon" <boris.brezillon@free-electrons.com>:
> > On Mon, 20 Jun 2016 12:48:38 +0800
> > Icenowy Zheng <icenowy@aosc.xyz> wrote:
> >  
> >>  The NAND controller on some sun8i chips needs its reset line to be
> >>  deasserted before they can enter working state.
> >>
> >>  Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz>
> >>  ---
> >>    Changes in v2:
> >>      - Corrected the error checking code of reset line.
> >>
> >>    Changes in v3:
> >>      - Corrected a more serious error brought in the "fix" of v2.
> >>
> >>    Changes in v4:
> >>      - Removed unneeded code block after "else".
> >>
> >>    Changes in v5:
> >>      - Added reassertion code in case of initialization error and device
> >>        remove.
> >>
> >>    Changes in v6:
> >>      - Fixed a resource leak by not using goto to exit in case of error.
> >>
> >>   drivers/mtd/nand/sunxi_nand.c | 28 +++++++++++++++++++++++++---
> >>   1 file changed, 25 insertions(+), 3 deletions(-)
> >>
> >>  diff --git a/drivers/mtd/nand/sunxi_nand.c b/drivers/mtd/nand/sunxi_nand.c
> >>  index a83a690..08d5e88 100644
> >>  --- a/drivers/mtd/nand/sunxi_nand.c
> >>  +++ b/drivers/mtd/nand/sunxi_nand.c
> >>  @@ -39,6 +39,7 @@
> >>   #include <linux/gpio.h>
> >>   #include <linux/interrupt.h>
> >>   #include <linux/iopoll.h>
> >>  +#include <linux/reset.h>
> >>
> >>   #define NFC_REG_CTL 0x0000
> >>   #define NFC_REG_ST 0x0004
> >>  @@ -269,6 +270,7 @@ struct sunxi_nfc {
> >>           void __iomem *regs;
> >>           struct clk *ahb_clk;
> >>           struct clk *mod_clk;
> >>  + struct reset_control *reset;
> >>           unsigned long assigned_cs;
> >>           unsigned long clk_rate;
> >>           struct list_head chips;
> >>  @@ -1871,26 +1873,42 @@ static int sunxi_nfc_probe(struct platform_device *pdev)
> >>           if (ret)
> >>                   goto out_ahb_clk_unprepare;
> >>
> >>  + nfc->reset = devm_reset_control_get_optional(dev, "ahb");
> >>  +
> >>  + if (!IS_ERR(nfc->reset)) {
> >>  + ret = reset_control_deassert(nfc->reset);
> >>  + if (ret) {
> >>  + dev_err(dev, "reset err %d\n", ret);
> >>  + goto out_mod_clk_unprepare;
> >>  + }
> >>  + } else if (PTR_ERR(nfc->reset) != -ENOENT) {
> >>  + ret = PTR_ERR(nfc->reset);  
> >
> > You should return -EDEFER_PROBE here.
> >
> > And can you please rebase this series on top of nand/next [1]?
> >
> > [1]https://github.com/linux-nand/linux/tree/nand/next