Hi, Marcelo:
Here's a change for ymfpci, hopefuly the last one before 2.4.17.
It should fix artsd (KDE sound). Also I am slipping a Kai's
region request along with it.
I've got some good Kai's fixes lined up for 2.4.18.
(well, unless you delay .17 so much that it makes sense
to push them before that :)
Greetings,
-- Pete
--- linux-2.4.17-pre5/drivers/sound/ymfpci.c Thu Dec 6 11:23:44 2001
+++ linux-2.4.17-pre5-p3/drivers/sound/ymfpci.c Thu Dec 6 11:44:40 2001
@@ -1443,13 +1443,14 @@
{
struct ymf_state *state = (struct ymf_state *)file->private_data;
struct ymf_dmabuf *dmabuf;
+ int redzone;
unsigned long flags;
unsigned int mask = 0;
if (file->f_mode & FMODE_WRITE)
poll_wait(file, &state->wpcm.dmabuf.wait, wait);
- // if (file->f_mode & FMODE_READ)
- // poll_wait(file, &dmabuf->wait, wait);
+ if (file->f_mode & FMODE_READ)
+ poll_wait(file, &state->rpcm.dmabuf.wait, wait);
spin_lock_irqsave(&state->unit->reg_lock, flags);
if (file->f_mode & FMODE_READ) {
@@ -1458,12 +1459,21 @@
mask |= POLLIN | POLLRDNORM;
}
if (file->f_mode & FMODE_WRITE) {
+ redzone = ymf_calc_lend(state->format.rate);
+ redzone <<= state->format.shift;
+ redzone *= 3;
+
dmabuf = &state->wpcm.dmabuf;
if (dmabuf->mapped) {
if (dmabuf->count >= (signed)dmabuf->fragsize)
mask |= POLLOUT | POLLWRNORM;
} else {
- if ((signed)dmabuf->dmasize >= dmabuf->count + (signed)dmabuf->fragsize)
+ /*
+ * Don't select unless a full fragment is available.
+ * Otherwise artsd does GETOSPACE, sees 0, and loops.
+ */
+ if (dmabuf->count + redzone + dmabuf->fragsize
+ <= dmabuf->dmasize)
mask |= POLLOUT | POLLWRNORM;
}
}
@@ -1498,6 +1508,7 @@
return -EAGAIN;
dmabuf->mapped = 1;
+/* P3 */ printk(KERN_INFO "ymfpci: using memory mapped sound, untested!\n");
return 0;
}
@@ -2430,6 +2441,7 @@
static int __devinit ymf_probe_one(struct pci_dev *pcidev, const struct pci_device_id *ent)
{
u16 ctrl;
+ unsigned long base;
ymfpci_t *codec;
int err;
@@ -2438,6 +2450,7 @@
printk(KERN_ERR "ymfpci: pci_enable_device failed\n");
return err;
}
+ base = pci_resource_start(pcidev, 0);
if ((codec = kmalloc(sizeof(ymfpci_t), GFP_KERNEL)) == NULL) {
printk(KERN_ERR "ymfpci: no core\n");
@@ -2452,16 +2465,21 @@
codec->pci = pcidev;
pci_read_config_byte(pcidev, PCI_REVISION_ID, &codec->rev);
- codec->reg_area_virt = ioremap(pci_resource_start(pcidev, 0), 0x8000);
- if (codec->reg_area_virt == NULL) {
- printk(KERN_ERR "ymfpci: unable to map registers\n");
+
+ if (request_mem_region(base, 0x8000, "ymfpci") == NULL) {
+ printk(KERN_ERR "ymfpci: unable to request mem region\n");
goto out_free;
}
+ if ((codec->reg_area_virt = ioremap(base, 0x8000)) == NULL) {
+ printk(KERN_ERR "ymfpci: unable to map registers\n");
+ goto out_release_region;
+ }
+
pci_set_master(pcidev);
printk(KERN_INFO "ymfpci: %s at 0x%lx IRQ %d\n",
- (char *)ent->driver_data, pci_resource_start(pcidev, 0), pcidev->irq);
+ (char *)ent->driver_data, base, pcidev->irq);
ymfpci_aclink_reset(pcidev);
if (ymfpci_codec_ready(codec, 0, 1) < 0)
@@ -2491,8 +2509,7 @@
/* register /dev/dsp */
if ((codec->dev_audio = register_sound_dsp(&ymf_fops, -1)) < 0) {
- printk(KERN_ERR "ymfpci%d: unable to register dsp\n",
- codec->dev_audio);
+ printk(KERN_ERR "ymfpci: unable to register dsp\n");
goto out_free_irq;
}
@@ -2538,6 +2555,8 @@
ymfpci_writel(codec, YDSXGR_STATUS, ~0);
out_unmap:
iounmap(codec->reg_area_virt);
+ out_release_region:
+ release_mem_region(pci_resource_start(pcidev, 0), 0x8000);
out_free:
kfree(codec);
return -ENODEV;
@@ -2561,6 +2580,7 @@
ctrl = ymfpci_readw(codec, YDSXGR_GLOBALCTRL);
ymfpci_writew(codec, YDSXGR_GLOBALCTRL, ctrl & ~0x0007);
iounmap(codec->reg_area_virt);
+ release_mem_region(pci_resource_start(pcidev, 0), 0x8000);
#ifdef CONFIG_SOUND_YMFPCI_LEGACY
if (codec->iomidi) {
unload_uart401(&codec->mpu_data);
On Thu, 6 Dec 2001, Pete Zaitcev wrote:
> Hi, Marcelo:
>
> Here's a change for ymfpci, hopefuly the last one before 2.4.17.
> It should fix artsd (KDE sound). Also I am slipping a Kai's
> region request along with it.
>
> I've got some good Kai's fixes lined up for 2.4.18.
> (well, unless you delay .17 so much that it makes sense
> to push them before that :)
>
> Greetings,
> -- Pete
>
Should this ymfpci work with my ymf740C. I have never been able to get
ymfpci support to work directly from the kernel. I've always used alsa.
--
Chad Young
Registered Linux User #195191 @ http://counter.li.org
On Fri, 7 Dec 2001, skidley wrote:
> Should this ymfpci work with my ymf740C. I have never been able to get
> ymfpci support to work directly from the kernel. I've always used alsa.
If you're using ALSA's ymfpci.o, yes. The in-kernel ymfpci driver is
heavily based upon the ALSA driver and should therefore support the same
hardware (and it explicitly mentions the 740C). What problem do you have?
--Kai