Re: [RFC][PATCH] Add suspend and resume support to uli526x

[Nigel Cunningham <nigel@nigel.suspend2.net>]

[-- Attachment #1: Type: text/plain, Size: 2962 bytes --]

Hi.

On Tue, 2007-06-05 at 00:53 +0200, Pavel Machek wrote:
> On Tue 2007-06-05 08:49:04, Nigel Cunningham wrote:
> > Hi.
> > 
> > On Tue, 2007-06-05 at 00:41 +0200, Pavel Machek wrote:
> > > Hi!
> > > 
> > > > > > > +#endif /* CONFIG_PM */
> > > > > > 
> > > > > > > @@ -1689,6 +1775,10 @@ static struct pci_driver uli526x_driver 
> > > > > > >  	.id_table	= uli526x_pci_tbl,
> > > > > > >  	.probe		= uli526x_init_one,
> > > > > > >  	.remove		= __devexit_p(uli526x_remove_one),
> > > > > > > +#ifdef CONFIG_PM
> > > > > > > +	.suspend	= uli526x_suspend,
> > > > > > > +	.resume		= uli526x_resume,
> > > > > > > +#endif
> > > > > > 
> > > > > > ...so that this ifdef is not needed?
> > > > > 
> > > > > OK, why not.
> > > > 
> > > > Because it's uglier and #ifdef is the established way of doing things?
> > > 
> > > Actually the way I suggested is nicer, IIRC akpm invented it. It keeps
> > > ifdefs localized around the block that _needs_ to be ifdefed.
> > 
> > The localised point is true. I'll also admit that 'nicer'/'uglier' is a
> > matter of aesthetics and therefore personal opinion.
> > 
> > I guess that leaves the question, "What's the precedent to follow?" or
> > "Is there a driver that's already got this new format?".
> 
> for example net/ne.c ...
> 
> pavel@amd:/data/l/linux/drivers$ grep "_suspend NULL" */*.c
> leds/leds-ams-delta.c:#define ams_delta_led_suspend NULL
> leds/leds-net48xx.c:#define net48xx_led_suspend NULL
> leds/leds-s3c24xx.c:#define s3c24xx_led_suspend NULL
> leds/leds-tosa.c:#define tosaled_suspend NULL
> leds/leds-wrap.c:#define wrap_led_suspend NULL
> misc/tifm_7xx1.c:#define tifm_7xx1_suspend NULL
> misc/tifm_core.c:#define tifm_device_suspend NULL
> net/3c509.c:#define el3_suspend NULL
> net/forcedeth.c:#define nv_suspend NULL
> net/ne.c:#define ne_drv_suspend NULL
> parport/parport_ax88796.c:#define parport_ax88796_suspend NULL
> rtc/rtc-at91rm9200.c:#define at91_rtc_suspend NULL
> rtc/rtc-omap.c:#define omap_rtc_suspend NULL
> rtc/rtc-s3c.c:#define s3c_rtc_suspend NULL
> serial/8250_pnp.c:#define serial_pnp_suspend NULL
> serial/atmel_serial.c:#define atmel_serial_suspend NULL
> serial/s3c2410.c:#define s3c24xx_serial_suspend NULL
> spi/pxa2xx_spi.c:#define pxa2xx_spi_suspend NULL
> spi/spi_bfin5xx.c:#define bfin5xx_spi_suspend NULL
> spi/spi_imx.c:#define spi_imx_suspend NULL
> spi/spi_s3c24xx.c:#define s3c24xx_spi_suspend NULL
> spi/spi_s3c24xx_gpio.c:#define s3c2410_spigpio_suspend NULL
> video/arkfb.c:#define ark_pci_suspend NULL
> video/au1100fb.c:#define au1100fb_drv_suspend NULL
> video/s3c2410fb.c:#define s3c2410fb_suspend NULL
> video/skeletonfb.c:#define xxxfb_suspend NULL
> video/skeletonfb.c:#define xxxfb_suspend NULL
> video/sm501fb.c:#define sm501fb_suspend NULL

Cool. That's a lot more than I would have expected. Ok. I'll drop my
objection then. (Still think it's ugly though :>).

Regards,

Nigel

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]