The cxgb3 driver is incorrectly configuring the HW CQ context for CQ's
that use overflow-avoidance.  Namely the RDMA control CQ.  This results
in a bad DMA from the device to bus address 0.  The solution is to set
the CQ_ERR bit in the context for these types of CQs.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Qualify toggling of xgmac tx enable with not getting pause frames,
we might not make forward progress because the peer is sending
lots of pause frames.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Clear pciE PEX errors late at module load time.
Log details when PEX errors occur.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Update firmware version.
Allow the driver to be up and running with older FW image
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Since E100 timer is 2HZ, use rounding to make timer occur on the
correct boundary.
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NAuke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

This patch adds support for the Intel 82598 PCI-Express 10GbE
chipset. Devices will be available on the market soon.

This version of the driver is largely the same as the last release:

  * Driver uses a single RX and single TX queue, each using 1 MSI-X
  irq vector.
  * Driver runs in NAPI mode only
  * Driver is largely multiqueue-ready (TM)

Changes since 20070803:
  * removed wrappers for hardware functions
  * incorporated e1000e-style HW api reorganization code
  * sparse/checkpatch cleanups, namespace cleanups
  * driver prints out extra debugging information at load time
    identifying adapter board number, mac, phy types
  * removed ixgbe_api.c, ixgbe_api.h, ixgbe_osdep.h
  * driver update to 1.1.18
  * removed ixgbe.txt which contained no useful info anymore

[ Integrated napi_struct changes from Auke as well... -DaveM ]
Signed-off-by: NAuke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: NAyyappan Veeraiyan <ayyappan.veeraiyan@intel.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Removed the workaround that was needed for PS3 firmware versions
prior to the first release.
Signed-off-by: NMasakazu Mokuno <mokuno@sm.sony.co.jp>
CC: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

The PS3 virtual network device requires a vlan tag in the sending packet
to select the destination device, ethernet port or wireless.
As the vlan tag field is in the middle of the passed data,
we should insert it into the packet data.
To avoid copying much of the packet data, the driver used two tx descriptors
for one tx skb; one descriptor was for sending a small static
buffer which contained vlan tag and copied header (two mac addresses),
one was for the residual data after the vlan field.

This patch changes the way to insert the vlan tag.  By changing
netdev->hard_header_len, we can make the headroom for moving mac address
fields in the skb buffer. Then we can send one tx skb with
one tx descriptor.  This also gives us a tx throughut gain of approx.
20% according to netperf results.
Signed-off-by: NMasakazu Mokuno <mokuno@sm.sony.co.jp>
CC: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Unfortunately there's no timeout for how long a packet can sit on
the TX ring after completion before an interrupt is generated, and
we want to have a threshold that's larger than one packet per interrupt.

So we have to have a timer that occasionally cleans the TX ring even
though there hasn't been an interrupt. Instead of setting up a dedicated
timer for this, just clean it in the NAPI poll routine instead.

[ Resolved conflicts with napi_struct changes... -DaveM ]
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Enable LLTX on pasemi_mac: we're already doing sufficient locking
in the driver to enable it.

[ Resolved merge conflicts with napi_struct changes... -DaveM ]
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

RX side flag to use is CHECKSUM_UNNECESSARY, not CHECKSUM_COMPLETE.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

The old logic didn't detect full (tx) ring cases properly, causing
overruns and general badness. Clean it up a bit and abstract out the
ring size checks, always making sure to leave 1 slot open.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Postpone pci unmap and skb free of the transmitted buffers to outside of
the tx ring lock, batching them up 32 at a time.

Also increase the count threshold to 128.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Various RX performance tweaks, do some explicit prefetching of packet
data, etc.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Fix up memcpy for short receives.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Enable settings to target l2 for the first few cachelines of the packet, since
we'll access them to get to the various headers.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Move away from using the pci config access functions for simple register
access.  Our device has all of the registers in the config space (hey,
from the hardware point of view it looks reasonable :-), so we need to
somehow get to it. Newer firmwares have it in the device tree such that
we can just get it and ioremap it there (in case it ever moves in future
products). For now, provide a hardcoded fallback for older firmwares.

[ Resolved napi_struct conflicts... -DaveM ]
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Abstract out the PCI config read/write accesses into reg read/write ones,
still calling the pci accessors on the back end.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

NetworkManager will not start dhcpd on an interface unless it reports
link-up state via ethtool.
Signed-off-by: NOlaf Hering <olaf@aepfle.de>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Use the PCI layer config access functions. The driver was using the
memory mapped window in device, to workaround issues accessing the
advanced error reporting registers.
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use the kernel interfaces for advanced error reporting.
This should be cleaner and clear up errors on boot.

For those systems with busted BIOS's that don't correctly
support mmconfig, advanced error reporting will be disabled.
The PCI registers for advanced error reporting start at 0x100 which
is too large to be accessed by legacy functions.
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Take out the code that protects driver from accessing the
PCI config space.
We are old enough to run with scissors now.
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add documentation of GPHY_CTRL register bits even if driver
is not using them (yet).
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Use debugfs rename to handle device neame changes.
Signed-off-by: NStephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Set PM1 internal memory to round robin mode
It balances access to this internal memory for multiport adapters.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Log HW serial number when cxgb3 module is loaded.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Stop the MAC when a fatal error is detected.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Enforce validity checks on connection ids
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

A HW issue requires limiting the receive window size
to 23 pages of internal memory.
These pages can be configured to different sizes,
thus the RDMA driver needs to know the
page size to enforce the upper limit.

Also assign explicit enum values.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Send small TX_DATA work requests as immediate data even when
there are fragments. this avoids doing multiple DMAs for
small fragmented packets.
The driver already implements this optimization for small
contiguous packets.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Log doorbell Fifo overflow
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Reduce Rx coalescing length to 12288
Large bursts from the adapter to the host create back pressure
on the chip. Reducing the burst size avoids the issue.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Check return of pci_enable_device in vortex_up().

Also modify vortex_up to return error to callers. Handle failure of
vortex_up in vortex_open and vortex_resume.
Signed-off-by: NMark Hindley <mark@hindley.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Removes the use of bitfields from the ibmveth driver. This results
in slightly smaller object code.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Removes dead frag processing code from ibmveth. Since NETIF_F_SG was
not set, this code was never executed. Also, since the ibmveth
interface can only handle 6 fragments, core networking code would need
to be modified in order to efficiently enable this support.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add ethtool hooks to ibmveth to retrieve driver statistics.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add handlers for get_tso and get_ufo to prevent errors being printed
by ethtool.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

This patch adds the appropriate ethtool hooks to allow for enabling/disabling
of hypervisor assisted checksum offload for TCP.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

This patchset enables TCP checksum offload support for IPV4
on ibmveth. This completely eliminates the generation and checking of
the checksum for packets that are completely virtual and never
touch a physical network. A simple TCP_STREAM netperf run on
a virtual network with maximum mtu set yielded a ~30% increase
in throughput. This feature is enabled by default on systems that
support it, but can be disabled with a module option.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Actually, D-Link modified the VendorID/ProductID of the TC902x.
The TC902x is the original chipset.
Signed-off-by: NKomuro <komurojun-mbn@nifty.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>