On Mon, Oct 19, 2020 at 12:06:33PM +0200, Thomas Gleixner wrote:
> From: Ahmed S. Darwish <a.darwish@xxxxxxxxxxxxx>
>
> The usage of in_interrupt() in drivers is phased out and Linus clearly
> requested that code which changes behaviour depending on context should
> either be separated or the context be conveyed in an argument passed by the
> caller, which usually knows the context.
>
> The debug printk() in digi_write() prints in_interrupt() as context
> information. TTY writes happen always in preemptible task context and
> console writes can happen from almost any context, so in_interrupt() is not
> really helpful.
The above statement is not correct, TTY writes can and do happen from
other contexts, including soft IRQ (e.g. PPP).
> Aside of that issuing a printk() from a console->write() callback is not a
> really brilliant idea for obvious reasons.
True, but we don't need to sprinkle conditionals for the benefit of
people trying to debug USB serial drivers using a USB serial console.
They get what they deserve. ;)
> Remove the in_interrupt() printout and make the printk() depend on tty.
>
> Signed-off-by: Ahmed S. Darwish <a.darwish@xxxxxxxxxxxxx>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Johan Hovold <johan@xxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: linux-usb@xxxxxxxxxxxxxxx
>
> ---
> drivers/usb/serial/digi_acceleport.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> --- a/drivers/usb/serial/digi_acceleport.c
> +++ b/drivers/usb/serial/digi_acceleport.c
> @@ -911,9 +911,10 @@ static int digi_write(struct tty_struct
> unsigned char *data = port->write_urb->transfer_buffer;
> unsigned long flags = 0;
>
> - dev_dbg(&port->dev,
> - "digi_write: TOP: port=%d, count=%d, in_interrupt=%ld\n",
> - priv->dp_port_num, count, in_interrupt());
> + if (tty) {
> + dev_dbg(&port->dev, "digi_write: TOP: port=%d, count=%d\n",
> + priv->dp_port_num, count);
> + }
So just drop the in_interrupt() here.
Also note that we already have another unconditional dev_dbg() at the
end of this function.
>
> /* copy user data (which can sleep) before getting spin lock */
> count = min(count, port->bulk_out_size-2);
>
Johan
