On Sat, 3 Sep 2016 01:26:40 +0200
Arnd Bergmann <a...@arndb.de> wrote:
> Drivers should not use NO_IRQ, as we are trying to get rid of that.
> In this case, the call to irq_of_parse_and_map() is both wrong
> (as it returns '0' on failure, not NO_IRQ) and unnecessary
> (as platform_get_irq() does the same thing)
>
> This removes the call to irq_of_parse_and_map() and checks for
> the error code correctly.
>
> Signed-off-by: Arnd Bergmann <a...@arndb.de>
Acked-by: Boris Brezillon <boris.brezil...@free-electrons.com>
> ---
> drivers/crypto/mv_cesa.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> It would be good if someone could test this on a machine that boots
> from DT to ensure the conversion was correct.
>
> diff --git a/drivers/crypto/mv_cesa.c b/drivers/crypto/mv_cesa.c
> index e6b658faef63..104e9ce9400a 100644
> --- a/drivers/crypto/mv_cesa.c
> +++ b/drivers/crypto/mv_cesa.c
> @@ -1091,11 +1091,8 @@ static int mv_probe(struct platform_device *pdev)
>
> cp->max_req_size = cp->sram_size - SRAM_CFG_SPACE;
>
> - if (pdev->dev.of_node)
> - irq = irq_of_parse_and_map(pdev->dev.of_node, 0);
> - else
> - irq = platform_get_irq(pdev, 0);
> - if (irq < 0 || irq == NO_IRQ) {
> + irq = platform_get_irq(pdev, 0);
> + if (irq < 0) {
> ret = irq;
> goto err;
> }
