Users report problems w/ auto-negotiation disabled and the link set
to 100/Half or 10/Half.  Problems range from poor performance to no
link at all.

The current sky2 code does not set things properly on link up if
autonegotiation is disabled.  Plus it does not contemplate a 10Mbit
setting at all.  This patch corrects that.
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Increase version, and get rid of out-dated comment.
Speed setting has worked for quite a while.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This hardware supports Message Signaled interrupts.
When setting up, use software interrupt to check for bad hardware.

Signed-off-by: Stephen Hemminger <shemminger @osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

The sky2 interrupt can be used to add entropy.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Move the interrupt clear to before processing, this avoids a
possible races with status delaying.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Using the sky2 driver with bonding can result in oopses related to
reinitializing the PHY when the MAC address is changed (which bonding
is wont to do).  This patch changes sky2_set_mac_address to take less
drastic measures.

This is analagous to the skge patch here:

	http://lkml.org/lkml/2005/9/29/399

which fixed the issue here:

	http://bugzilla.kernel.org/show_bug.cgi?id=5271Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This fixes setting rx_coalesce_usecs_irq via ethtool in sky2.
The write was directed to the wrong register.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NCarl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

There were bugs in mmconfig access to PCI space, up to and
include 2.6.16-rc1. These prevented the sky2 driver from being
able to clear PCI express errors.

This patch makes the driver check (during probe), for errors
in PCI config access and fail.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Fix suspend/resume for sky2. The status ring was getting reallocated
and a bunch of other mistakes. Also, check return from power_state
on resume.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Be more careful about transmit locking, this solves a possible race
between tx_complete and transmit, that would cause a tx timeout.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't need to inline quite so many routines, let the compiler
decide
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Can use kzalloc here.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Make sure and rate limit all the error messages that might occur. If a problem
occurs then a few messages are enough.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Small optimization, if dma addresses are 32 bits, then high
bits are always zero.
Signed-off-by: NStephen Hemminger <shemminger@osdl.or>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't need to zero out the status ring entries after processing.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Be more careful about memory barriers. The only place we really
need them is before and after updating the chip's ring interface.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Fix problems with Yukon FE rev 2 chipset. Don't cut and paste bugs in from
sk98lin driver. Change how the ram buffer is divided up, and make the math
clearer. Also, set the thresholds where rx takes precedence. The threshold
values are just guesses at this point, it might be worth tuning them later.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Version update.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Need to call pci_set_consistent_dma_mask in the case of 64 bit
DMA.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Need to make sure that sky2 receive buffers are 64 bit
aligned. Also, don't need to start off with GFP_ATOMIC
on initial setup.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Version number change.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

It is possible for hardware to get confused when an oversized frame
is received. In that case, just drop the packet and increment a counter.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't lose multicast addresses when link goes down or ring
parameters change.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't die if we run out of memory on mtu or ring parameter change.
For other admin operations, don't rebuild Rx ring, just restart the PHY.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Change version number.  Still don't have enough history on this
to call it 1.0 yet.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Under load, my SysKonnect SK-9S22 sees duplicate checksums from
earlier packets.  Doesn't happen on other platforms so just disable
receive checksum there.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Turn symmetric flow control on by default. This was an oversight
in the initial versions.
Signed-off-by: NStephen HEmminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Add a couple more prefetches to where we walk the rings.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't print ring full message if we lose race.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Put more logic in to try and handle transmit timeout recovery.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Adjust interrupt deferral values for better performance and avoid
starvation issues under stress.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Optimize poll routine by not clearing interrupt till after
processing, and checking for more work before re-enable.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Hand optimize transmit completion routine.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Change transmit complete handling to use the status values
in the poll list because they are in-cache, rather than reading
non-cached memory for chips status.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Some transmit side small fixes:
 * When computing number of list elements per transmit, do full
   comparision to check for checksuming.
 * Get rid of racy check for tx_complete
 * Change stop test to match wake condition.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Split Tx and Rx ring into two different data structures.
Tx needs the next value (to handle partial status), and
Rx always needs the mapaddr (to handle resubmitting same buffer).
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't need to keep track of mapping length in ring structure
because we can get the same info from other info.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Faster way to compute ring distance.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't need to disable interrupts during the transmit routine.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>