Re: [PATCH] Add support for IRQ_NOAUTOEN in __set_irq_handler

["Esben Haabendal" <esbenhaabendal@gmail.com>]

Currently, IRQ_NOAUTOEN is only supported with request_irq, and not fx.
set_irq_chained_handler(). As the IRQ_NOAUTOEN flag is silently ignored by
__set_irq_handler, you risk writing faulty code that start with irq enabled
although the developer expects irq disabled.

If the flag is not handled, wouldn't it make sense to return an error?

/Esben

On Thu, Jun 19, 2008 at 9:37 PM, Esben Haabendal <eha@doredevelopment.dk> wrote:
>
> Any reason for __set_irq_handler not supporting the IRQ_NOAUTOEN flag?
>
> /Esben
>
>  kernel/irq/chip.c |   11 ++++++++---
>  1 files changed, 8 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
> index 964964b..42bb163 100644
> --- a/kernel/irq/chip.c
> +++ b/kernel/irq/chip.c
> @@ -584,10 +584,15 @@ __set_irq_handler(unsigned int irq,
> irq_flow_handler_t handle, int is_chained,
>        desc->name = name;
>
>        if (handle != handle_bad_irq && is_chained) {
> -               desc->status &= ~IRQ_DISABLED;
>                desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE;
> -               desc->depth = 0;
> -               desc->chip->unmask(irq);
> +               if (!(desc->status & IRQ_NOAUTOEN)) {
> +                       desc->depth = 0;
> +                       desc->status &= ~IRQ_DISABLED;
> +                       desc->chip->unmask(irq);
> +               } else {
> +                       /* Undo nested disables: */
> +                       desc->depth = 1;
> +               }
>        }
>        spin_unlock_irqrestore(&desc->lock, flags);
>  }