On Thu, Jan 29, 2009 at 02:06:35PM -0800, Parag Warudkar wrote:

On Thu, 29 Jan 2009, Matt Carlson wrote:

quoted

Can you apply the following test patch and see if it helps?  The patch
does two things.  First, it enables a bit which should restore firmware
communication.  If that fixes the problem, then let me know and I'll
spin a proper patch.

In the event that it doesn't work, the patch goes on to test the memory
mapping by simply printing the register value at offset 0x0.  The value
should be the device's vendor ID and device ID.  Please post the
results so that I can verify it.

diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
index 8b3f846..39fce42 100644
--- a/drivers/net/tg3.c
+++ b/drivers/net/tg3.c

@@ -7227,6 +7227,11 @@ static int tg3_init_hw(struct tg3 *tp, int reset_phy)
 {
 	tg3_switch_clocks(tp);
 
+	printk( KERN_NOTICE "%s: Reg value at offset 0x0 is 0x%x\n",
+		tp->dev->name, tr32(0x0) );
+
+	tw32(MEMARB_MODE, tr32(MEMARB_MODE) | MEMARB_MODE_ENABLE);
+
 	tw32(TG3PCI_MEM_WIN_BASE_ADDR, 0);
 
 	return tg3_reset_hw(tp, reset_phy);

Hi Matt,

Thanks for the patch. It didn't help with resume - but below is the 
output after patching, let me know if you need more details.

( Looks like 0xffffffff is invalid/corrupted device id /vendor id? )

[  163.856001] tg3 0000:0e:00.0: restoring config space at offset 0xc (was 0x0, writing 0x20040000)                                                                                                            
[  163.856001] tg3 0000:0e:00.0: restoring config space at offset 0x3 (was 0x0, writing 0x10)                                                                                                                  
[  163.856001] tg3 0000:0e:00.0: restoring config space at offset 0x1 (was 0x100000, writing 0x100006)                                                                                                         

[snip]

[  164.450277] pcieport-driver 0000:1e:00.0: setting latency timer to 64
[  164.450415] pcieport-driver 0000:1e:01.0: setting latency timer to 64
[  164.450493] tg3 0000:0e:00.0: restoring config space at offset 0xc (was 0x0, writing 0x20040000)
[  164.451110] serial 00:08: activated

[snip]

[  168.913863] Restarting tasks ... done.
[  170.332953] tg3 0000:0e:00.0: wake-up capability disabled by ACPI
[  170.332960] tg3 0000:0e:00.0: PME# disabled
[  170.333047] tg3 0000:0e:00.0: irq 54 for MSI/MSI-X
[  170.333250] eth0: Reg value at offset 0x0 is 0xffffffff

                                           ^^^^^^^^^^^^^^^^^
So here is our problem.  For some reason the memory mapped IO is
failing.  I'll have to think about how and why that might happen.

FWIW, I can suspend and resume using the latest linux-2.6 kernel
on a machine with a similar chip here.  The problem doesn't seem to
affect all Broadcom devices.

[  170.394281] [drm] Loading R500 Microcode
[  170.394330] [drm] Num pipes: 1
[  171.726650] tg3: eth0: No firmware running.
[  183.119745] ADDRCONF(NETDEV_UP): eth0: link is not ready


Parag