[alsa-devel] ice1712 suspend/resume for M-Audio Audiophile 2496

Tommi Uimonen tee.uimonen at gmail.com
Thu Aug 22 12:11:22 CEST 2013

Hello. I tried implementing suspend for ice1712, and at least it now
calls snd_ice1712_suspend during suspend, but I have no idea what data
I should send to the card to make it suspend properly.

static int snd_ice1712_suspend(struct pci_dev *pci, pm_message_t state)
        struct snd_card *card = pci_get_drvdata(pci);
        struct snd_ice1712 *ice = card->private_data;
        //      printk(KERN_INFO "snd_ice1712_suspend starts\n");

        if (!ice->pm_suspend_enabled)
            printk(KERN_INFO "snd_ice1712_suspend disabled\n");
                return 0;

        snd_power_change_state(card, SNDRV_CTL_POWER_D3hot);

        //      ice->pm_saved_is_spdif_master = ice->is_spdif_master(ice);
        //      ice->pm_saved_spdif_ctrl = inw(ICEMT(ice, SPDIF_CTRL));
        //      ice->pm_saved_spdif_cfg = inb(ICEREG(ice, SPDIF_CFG));
        ice->pm_saved_route = inl(ICEMT(ice, MONITOR_ROUTECTRL));

        if (ice->pm_suspend)

        pci_set_power_state(pci, pci_choose_state(pci, state));
        return 0;

ice->pm_suspend(ice) seems to hang the whole system. Commenting it out
makes system usable, but no sound after resume (rmmod & modprobe cycle

Does the card even support suspending properly? Am I wasting my time
here and should I just add 'fuser -k /dev/snd && rmmod' to the suspend
script and modprobe to the resume (ugly workaround that does not fix
the real problem)


2013/7/13 Tommi Uimonen <tee.uimonen at gmail.com>:
> Hello, the lack of suspend/resume support for ice1712 driver has been
> bugging me a lot, and I was wondering if it's possible to implement it
> by following how it's done in 1724.c?
> I don't really know much about the driver, but after some comparison
> between 1712.c and 1724.c it seems doable.
> At least the suspend part seems straightforward, just changing
> ICEMT1724 with ICEMT and ICEREG1724 with ICEREG. To resume, I should
> add some new functions like snd_ice1712_chip_reset (I guess it follows
> the same logic as in 1724.c, just use ICE1712_RESET instead) and maybe
> some spdif stuff but I think the 1724 has different spdif options. I
> don't have any 1724 cards so can't be sure, but 1712 does not have
> spdif rate. Maybe it will work without those, at least for me since I
> don't use spdif.
> And then add this to the pci_driver ice1712_driver struct:
> .driver = {
> .pm = SND_ICE1712_PM_OPS,
> }
> and define the SIMPLE_DEV_PM_OPS accordingly.
> Or is this already under work by someone?
> Tommi

More information about the Alsa-devel mailing list