Re: e1000_intr in request_irq faults in 2.6.20-git

From: Auke Kok
Date: Fri Feb 16 2007 - 00:21:30 EST

[Adding Dimitri Mishin to the CC - he proposed the same patch earlier]

Len Brown wrote:
On Thursday 15 February 2007 21:10, Brandeburg, Jesse wrote:
Eric W. Biederman wrote:
Len Brown <lenb@xxxxxxxxxx> writes:

e1000 faults in 2.6.20-git, while 2.6.20 worked fine.

System is a D875PBZ with LOM.

I'm guessing this is an old bug found by the following bit of
debug coded added into since v2.6.20

+ if (irqflags & IRQF_SHARED) {
+ /*
+ * It's a shared IRQ -- the driver ought to be
prepared for it + * to happen immediately, so let's
make sure.... + * We do this before actually
registering it, to make sure that + * a 'real' IRQ
doesn't run in parallel with our fake + */
+ if (irqflags & IRQF_DISABLED) {
+ unsigned long flags;
+ local_irq_save(flags);
+ handler(irq, dev_id);
+ local_irq_restore(flags);
+ } else
+ handler(irq, dev_id);
+ }

I don't have a clue why the e1000 wasn't ready though.

our code is clearly calling request_irq before we have assigned the
function pointer adapter->clean_rx as well as adapter->alloc_rx_buf

That would be a bug, a possible patch would be (inline and attached):
compile tested, *but* I couldn't test this patch to make sure it worked
because I couldn't boot 2.6.20-git due to it not finding my RAID0 + lvm

[PATCH] e1000: fix shared interrupt warning message

From: Jesse Brandeburg <jesse.brandeburg@xxxxxxxxx>

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@xxxxxxxxx>

drivers/net/e1000/e1000_main.c | 13 +++++++------
1 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/net/e1000/e1000_main.c
index 619c892..b8c4d5c 100644
--- a/drivers/net/e1000/e1000_main.c
+++ b/drivers/net/e1000/e1000_main.c
@@ -1417,10 +1417,6 @@ e1000_open(struct net_device *netdev)
if ((err = e1000_setup_all_rx_resources(adapter)))
goto err_setup_rx;

- err = e1000_request_irq(adapter);
- if (err)
- goto err_req_irq;

if ((err = e1000_up(adapter)))
@@ -1431,6 +1427,10 @@ e1000_open(struct net_device *netdev)

+ err = e1000_request_irq(adapter);
+ if (err)
+ goto err_req_irq;
/* If AMT is enabled, let the firmware know that the network
* interface is now open */
if (adapter->hw.mac_type == e1000_82573 &&
@@ -1439,10 +1439,11 @@ e1000_open(struct net_device *netdev)

return E1000_SUCCESS;

+ e1000_down(adapter);
+ e1000_free_irq(adapter);
- e1000_free_irq(adapter);

Works for me(tm) on latest 2.6.20-git and D875PBZ.

If there are no objections I'll push this patch to Jeff Garzik together with two other changes I have for him.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at