Looks like your problem is caused by suspend. I see the exact same problem after S3 resume on my ACPI desktop system with a PCI 3c59x. I am attaching a patch which works on my PCI desktop, could you also test it on your cardbus/laptop setup?hda: start_power_step(step: 0) hda: start_power_step(step: 1) hda: complete_power_step(step: 1, stat: 50, err: 0) hda: completing PM request, suspend PCI: Guessed IRQ 10 for device 0000:00:0b.0 PCI: Sharing IRQ 10 with 0000:00:0a.0 hda: Wakeup request inited, waiting for !BSY... hda: start_power_step(step: 1000) blk: queue cb596e00, I/O limit 4095Mb (mask 0xffffffff) hda: completing PM request, resume spurious 8259A interrupt: IRQ7. eth0: command 0x5800 did not complete! Status=0xffff
The problem is that under the new PCI driver model, cards are required to always restore state and call pci_enable_device() on resume. So the patch changes the driver to do its restore state calls unconditionally (it used to only do them when it was configured for wake-on-lan) and adds a call to pci_enable_device().
Maybe we should also be disabling bus-master on suspend, but I didn't want to change any more than I had to.
(By CC to Jeff - is this ok for the netdev tree?)
Nathan
------------------------------------------------------------------------
===== drivers/net/3c59x.c 1.57 vs edited =====
--- 1.57/drivers/net/3c59x.c 2004-07-14 04:13:48 -04:00
+++ edited/drivers/net/3c59x.c 2004-08-11 22:32:58 -04:00
@@ -1476,7 +1476,7 @@
#ifdef CONFIG_NET_POLL_CONTROLLER
dev->poll_controller = poll_vortex; #endif
- if (pdev && vp->enable_wol) {
+ if (pdev) {
vp->pm_state_valid = 1;
pci_save_state(VORTEX_PCI(vp), vp->power_state);
acpi_set_WOL(dev);
@@ -1533,9 +1533,10 @@
unsigned int config;
int i;
- if (VORTEX_PCI(vp) && vp->enable_wol) {
+ if (VORTEX_PCI(vp)) {
pci_set_power_state(VORTEX_PCI(vp), 0); /* Go active */
pci_restore_state(VORTEX_PCI(vp), vp->power_state);
+ pci_enable_device(VORTEX_PCI(vp));
}
/* Before initializing select the active media port. */
@@ -2684,7 +2685,7 @@
if (vp->full_bus_master_tx)
outl(0, ioaddr + DownListPtr);
- if (final_down && VORTEX_PCI(vp) && vp->enable_wol) {
+ if (final_down && VORTEX_PCI(vp)) {
pci_save_state(VORTEX_PCI(vp), vp->power_state);
acpi_set_WOL(dev);
}
@@ -3052,15 +3053,17 @@
struct vortex_private *vp = netdev_priv(dev);
long ioaddr = dev->base_addr;
- /* Power up on: 1==Downloaded Filter, 2==Magic Packets, 4==Link Status. */
- EL3WINDOW(7);
- outw(2, ioaddr + 0x0c);
- /* The RxFilter must accept the WOL frames. */
- outw(SetRxFilter|RxStation|RxMulticast|RxBroadcast, ioaddr + EL3_CMD);
- outw(RxEnable, ioaddr + EL3_CMD);
+ if (vp->enable_wol) {
+ /* Power up on: 1==Downloaded Filter, 2==Magic Packets, 4==Link Status. */
+ EL3WINDOW(7);
+ outw(2, ioaddr + 0x0c);
+ /* The RxFilter must accept the WOL frames. */
+ outw(SetRxFilter|RxStation|RxMulticast|RxBroadcast, ioaddr + EL3_CMD);
+ outw(RxEnable, ioaddr + EL3_CMD);
+ pci_enable_wake(VORTEX_PCI(vp), 0, 1);
+ }
/* Change the power state to D3; RxEnable doesn't take effect. */
- pci_enable_wake(VORTEX_PCI(vp), 0, 1);
pci_set_power_state(VORTEX_PCI(vp), 3);
}
@@ -3083,7 +3086,7 @@
*/
unregister_netdev(dev);
- if (VORTEX_PCI(vp) && vp->enable_wol) {
+ if (VORTEX_PCI(vp)) {
pci_set_power_state(VORTEX_PCI(vp), 0); /* Go active */
if (vp->pm_state_valid)
pci_restore_state(VORTEX_PCI(vp), vp->power_state);
Andrew and Alan have been poking at this sort of thing off and on, so I would like to get their comments on it.
It looks OK to me.
Jeff
- : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
