[Openvortex-dev] Re: au8810 support patch

[Jeff Muizelaar]

On 6 May 2003, Pontus Fuchs wrote:

> > >What is the reason for this patch? Everything else looks good.
> > >
> > >diff -u -r1.11 -r1.12
> > >--- openvortex/alsa/pci/au88x0/au88x0.c    2003/05/05 01:21:38     1.11
> > >+++ openvortex/alsa/pci/au88x0/au88x0.c    2003/05/05 05:52:37     1.12
> > >@@ -18,6 +18,7 @@
> > > #include <linux/init.h>
> > > #include <linux/pci.h>
> > > #include <linux/slab.h>
> > >+#include <linux/interrupt.h>
> > >
> > This include ... dont know really. I think we could remove it. Maybe
> > useful for
> > linux-2.5 ? Pontus ... what can you tell about ?
>
> Without it the driver won't compile under kernel 2.5.

Fair enough.

> > > #define SNDRV_GET_ID
> > > #include <sound/initval.h>
> > >
> > >@@ -142,7 +143,6 @@
> > >   goto irq_out;
> > >     }
> > >     chip->irq = pci->irq;
> > >-    vortex_enable_int(chip);
> > >
> > >     pci_set_master(pci);
> > >     // End of PCI setup.
> > >@@ -272,6 +272,7 @@
> > >     // (7)
> > >     pci_set_drvdata(pci, chip);
> > >     dev++;
> > >+    vortex_enable_int(chip);
> > >     return 0;
> > > }
> > Moving the Interrupt enable ... really not needed. Maybe we should put
> > it back
> > since it doesn't matter if it is done later and it would be nicer to
> > have it near
> > the interrupt requesting code.
>
> What do you mean? The interrupt cannot be enabled until the mpu401 has
> been registered with alsa, otherwise the interrupt-handler can call the
> mpu401-code in alsa resulting in a crash.

Good point. I guess it will stay.

-Jeff