some random coverage testing found that when changing mtu
under heavy traffic load, NAPI would use the rx_buffer_len variable
after it had been changed by change_mtu.

Similar to e1000 bugs found long ago.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

a user pointed out that setting variables out of order with respect
to the checks we make for tx timeout handling could result in a race
where ->dma was set but ->time_stamp was set to the old value.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

in order to prevent the case where poll_disable is waiting
on our device to permanently, check the flag to make sure we're not
down or closing down before re-enabling interrupts.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

ixgb hardware (not ixgbe) has a problem where it might dma past the
end of a buffer in certain cases.  leave 8 bytes extra room.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

there was one more bug hidden in the prefetch routines in ixgb hardware
that force us to remove it completely.  Writebacks were being done on
descriptors with stale data due to internal hardware fifo corruption.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

this patch has been made to many other drivers in kernel to fix
the storage of 64 bit resources in 32 bit variables.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

back when maybe stop tx was added to the ixgb driver some mistakes
were made and the driver
a) didn't remove the tx lock, which is now un-necessary
b) didn't change the restart code to be compliant with maybe_stop
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Compile-tested only.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Compile-tested only.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Compile-tested only.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

- Updated version number
Signed-off-by: NSantosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: NRamkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

- Enable msi-x link interrupts because the timer based scheduler was getting
cancelled causing the link state to be lost with repetitive card up/downs
when changing the mtu.
- Unmask mac_rmac_link interrupts only for Xframe I and prevent a spurious
link interrupt in Xframe II.
- Stop the tx queue and indicate link down when card is down
Signed-off-by: NSantosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: NRamkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

- Fix IOMMU overflow checking. As reported by Andi Kleen <ak@linux.intel.com>
removed check for zero dma address.
Signed-off-by: NSantosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: NRamkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

It removes a dependancy from velocity_init_rd_ring to dev->mtu.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Updates of the RBRDU have two different meanings depending on their
context:
1. the receiving process has not started - the value which is written
   into the RBRDU register is supposed to be the free rx descriptors
   count (rounded to a multiple of 4)
2. the receiving process is running - the value increments the count
   above (sic)

The update is currently issued deep inside the rx replenish chain (see
velocity_give_many_rx_descs).

Let's propagate enough information to the caller so that the rx
replenish functions do not depend on hardware any more.

It is needed to perform the Rx/Tx buffers housekeeping when MTU changes.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

- PCI consistent areas need no memset
- use dev_err instead of plain printk
- avoid a few casts
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Executive summary: the bounce buffers are in my way

- they use something like a 64 * 1500 bytes area of PCI
  consistent area
- they are not resized when the MTU changes
- they are used
  - to hand-pad undersized packets. skb_pad anyone ?
  - to linearize fragmented skbs whose fragment count
    goes beyond the 7 fragments hardware limit in order
    to claim scatter-gather support

Actually the SG code is commented out and I wonder if it
could not be implemented (ab-)using the large send feature
of the chipset since the latter should support some
multi-descriptor packet transmitting.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Fixed-by: NSéguier Régis <rseguier@e-teleport.net>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Multiple TX queue support is a core networking feature.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Accesses are mostly structured such that when there are multiple TX
queues the code transformations will be a little bit simpler.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This allows us to use this calling convention all the way down into
qdisc_restart().
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Only plain netif_schedule() remains taking a net_device, mostly as a
compatability item while we transition the rest of these interfaces.

Everything else calls netif_schedule_queue() or __netif_schedule(),
both of which take a netdev_queue pointer.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Accomplish this by using local variables.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This indicates if the NOOP scheduler is what is active for TX on a
given device.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

First, we add a qdisc_tx_changing() helper which returns true if the
qdisc attachment is in transition.

Second, we remove an assertion warning which is of limited value and
is hard to express precisely in a multiqueue environment.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is a helper function, currently used by IRDA.

This is being added so that we can contain and isolate as many
explicit ->tx_queue references in the tree as possible.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Isolate callers that want to simply reset all the TX qdiscs from the
details of TX queues.

Use this in the ISDN code.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We schedule queues, not the device, for output queue processing in BH.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It just wants the root qdisc given an arbitrary qdisc,
and that is simply qdisc->dev_queue->qdisc
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Acked-by: NStephen Hemminger <shemminger@vyatta.com>

It is always equal to qdisc->dev_queue->lock
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now that our qdisc management is bi-directional, per-queue, and fully
orthogonal, there is no reason to have a special ingress qdisc pointer
in struct net_device.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now qdisc, qdisc_sleeping, and qdisc_list also live there.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Every qdisc is assosciated with a queue, and in the case of ingress
qdiscs that will now be netdev->rx_queue so using that queue's lock is
the thing to do.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The lock is now an attribute of the device queue.

One thing to notice is that "suspicious" places
emerge which will need specific training about
multiple queue handling.  They are so marked with
explicit "netdev->rx_queue" and "netdev->tx_queue"
references.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It can be obtained via the netdev_queue.  So create a helper routine,
qdisc_dev(), to make the transformations nicer looking.

Now, qdisc_alloc() now no longer needs a net_device pointer argument.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A netdev_queue is an entity managed by a qdisc.

Currently there is one RX and one TX queue, and a netdev_queue merely
contains a backpointer to the net_device.

The Qdisc struct is augmented with a netdev_queue pointer as well.

Eventually the 'dev' Qdisc member will go away and we will have the
resulting hierarchy:

	net_device --> netdev_queue --> Qdisc

Also, qdisc_alloc() and qdisc_create_dflt() now take a netdev_queue
pointer argument.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We haven't had netdev->tbusy in many years :)
Signed-off-by: NDavid S. Miller <davem@davemloft.net>