On Wed, Jul 16, 2014 at 09:40:18PM +0100, Michalis Pappas wrote:
> On 07/09/2014 07:51 PM, Greg KH wrote:
> >> diff --git a/drivers/staging/gdm72xx/gdm_sdio.c
> >> b/drivers/staging/gdm72xx/gdm_sdio.c
> >> index 9d2de6f..914fd75 100644
> >> --- a/drivers/staging/gdm72xx/gdm_sdio.c
> >> +++ b/drivers/staging/gdm72xx/gdm_sdio.c
> >> @@ -280,9 +280,11 @@ static void send_sdu(struct sdio_func *func, struct
> >> tx_cxt *tx)
> >>
> >> spin_unlock_irqrestore(&tx->lock, flags);
> >>
> >> + #if defined(GDM72xx_DEBUG)
> >> print_hex_dump_debug("sdio_send: ", DUMP_PREFIX_NONE, 16, 1,
> >> tx->sdu_buf + TYPE_A_HEADER_SIZE,
> >> aggr_len - TYPE_A_HEADER_SIZE, false);
> >> + #endif
> >
> > This should be moved to use dev_dbg(), along with the other calls to
> > this function in this file.
> >
>
> But dev_dbg() gets eventually to be printk(), which cannot print the
> buffer, so using print_hex_dump_debug() seems to be correct for this
> case, no?
No, you don't want to print the message unless debugging is enabled, and
dev_dbg() uses the proper in-kernel dynamic debug infrastructure. There
should never be a separate config option for debugging a driver, that
ensures that no user will ever be able to help you out with it.
So delete the ifdef stuff, and the config option, and just use the
proper in-kernel infrastructure for this.
thanks,
greg k-h
_______________________________________________
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
