There is quite a bit of PCI stuff that gets compiled into parport_pc.c
even when CONFIG_PCI isn't enabled. This patch cuts down the size quite
a bit (more than 4k off the object, about 1k off the zImage) for the
older non-PCI machines which are typically resource starved anyway...
Patch is against 2.4.2
Paul.
--- drivers/parport/parport_pc.c~ Wed Feb 14 02:41:01 2001
+++ drivers/parport/parport_pc.c Thu Mar 1 00:54:19 2001
@@ -11,6 +11,7 @@
* Cleaned up include files - Russell King <[email protected]>
* DMA support - Bert De Jonghe <[email protected]>
* Many ECP bugs fixed. Fred Barnes & Jamie Lokier, 1999
+ * More PCI support now conditional on CONFIG_PCI, 03/2001, Paul G.
*/
/* This driver should work with any hardware that is broadly compatible
@@ -2182,6 +2183,7 @@
}
+#ifdef CONFIG_PCI
/* Via support maintained by Jeff Garzik <[email protected]> */
static int __devinit sio_via_686a_probe (struct pci_dev *pdev)
{
@@ -2547,7 +2549,6 @@
static int __init parport_pc_init_superio (void)
{
-#ifdef CONFIG_PCI
const struct pci_device_id *id;
struct pci_dev *pdev;
@@ -2558,10 +2559,13 @@
return parport_pc_superio_info[id->driver_data].probe (pdev);
}
-#endif /* CONFIG_PCI */
return 0; /* zero devices found */
}
+#else
+static struct pci_driver parport_pc_pci_driver;
+static int __init parport_pc_init_superio(void) {return 0;}
+#endif /* CONFIG_PCI */
/* This is called by parport_pc_find_nonpci_ports (in asm/parport.h) */
static int __init __attribute__((unused))
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
Forgot to mention that the patch also fixes the warning:
`parport_pc_superio_info' defined but not used
for non-PCI, which was the original reason why I was poking around in there.
Paul.
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com