Re: [PATCH] Add pci_save_state() to ALSA

[Takashi Iwai <tiwai@suse.de>]

At Fri, 12 Nov 2004 14:45:40 +0100,
Arjan van de Ven wrote:
> 
> On Fri, 2004-11-12 at 14:26 +0100, Takashi Iwai wrote:
> > But pci_save_state() is called again after the driver's suspend
> > callback is called.  So, the final saved state must be anyway same.
> 
> no that changed recently in the upstream kernel.
> pci_save_state() is now only called if there is no suspend callback in the driver!

Ah, thanks, that explains why (I referred 2.6.10-rc1).

If so, the patch is almost correct, but pci_save_state() should be put
after the call of callback.


Takashi

--- linux/sound/core/init.c	8 Nov 2004 11:37:08 -0000	1.48
+++ linux/sound/core/init.c	12 Nov 2004 13:56:32 -0000
@@ -782,12 +782,15 @@
 int snd_card_pci_suspend(struct pci_dev *dev, u32 state)
 {
 	snd_card_t *card = pci_get_drvdata(dev);
+	int err;
 	if (! card || ! card->pm_suspend)
 		return 0;
 	if (card->power_state == SNDRV_CTL_POWER_D3hot)
 		return 0;
 	/* FIXME: correct state value? */
-	return card->pm_suspend(card, 0);
+	err = card->pm_suspend(card, 0);
+	pci_save_state(dev);
+	return err;
 }
 
 int snd_card_pci_resume(struct pci_dev *dev)