Commit graph

46 commits

Author SHA1 Message Date
Stephen Hemminger
d70cd51ac0 [PATCH] sky2: prefetch tuning
Add a couple more prefetches to where we walk the rings.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
8c463ef792 [PATCH] sky2: quiet ring full message in case of race
Don't print ring full message if we lose race.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
8cc048e338 [PATCH] sky2: handle tx timeout
Put more logic in to try and handle transmit timeout recovery.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
69634ee736 [PATCH] sky2: interrupt coalescing tuning
Adjust interrupt deferral values for better performance and avoid
starvation issues under stress.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
3e4b32e116 [PATCH] sky2: interrupt/poll optimization
Optimize poll routine by not clearing interrupt till after
processing, and checking for more work before re-enable.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
af2a58acb1 [PATCH] sky2: transmit complete routine optimization
Hand optimize transmit completion routine.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
13b97b74b9 [PATCH] sky2: transmit complete index optimization
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: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
0e3ff6aab1 [PATCH] sky2: transmit logic fixes
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: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:21 -05:00
Stephen Hemminger
6cdbbdf305 [PATCH] sky2: tx/rx ring data structure split
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: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
734d186846 [PATCH] sky2: map length optimization
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: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
129372d052 [PATCH] sky2: ring distance optimization
Faster way to compute ring distance.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
f2e46561cc [PATCH] sky2: no irq disable needed during tx
Don't need to disable interrupts during the transmit routine.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
91c86df5a8 [PATCH] sky2: phy processing in workqueue rather than tasklet
Do phy processing in a work queue rather than a tasklet.
This means we can let bottom halves run.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
fb17358fe3 [PATCH] sky2: ethtool get/set interrupt coalescing
Add support for get/set the interrupt coalescing settings.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
bdb5c58ebe [PATCH] sky2: copy threshold as module parameter
Make the copy threshold a module parameter for those who
may want to turn it off.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:20 -05:00
Stephen Hemminger
92f965e8c5 [PATCH] sky2: don't die if we see chip rev 0xb5
There maybe new versions of Yukon2 in the future, so make the
driver more robust to handle this.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-12 15:27:19 -05:00
Stephen Hemminger
f89c2b4645 [PATCH] sky2: interrupt not cleared.
One of the rearrangements dropped this and it means
sky2 spins in napi/interrupt.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-03 20:29:03 -05:00
shemminger@osdl.org
5f4f9dc113 [PATCH] sky2: update version number
Update version number and print version in boot message.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:21 -05:00
shemminger@osdl.org
50241c4c59 [PATCH] sky2: use pci_register_driver
Switch to using pci_register_driver as per current convention.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:21 -05:00
shemminger@osdl.org
af4ed7e6ba [PATCH] sky2: remove pci-express hacks
Eliminate special case tuning for PCI-Express. This code
causes receive hangs and doesn't help performance much anyway.

Signed-off-by: Stephen Hemminger <shemmnger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:21 -05:00
shemminger@osdl.org
2224795d7e [PATCH] sky2: dual port tx completion
Sometimes on dual port cards, one tx complete may cover both
ports. To handle that rearrange poll routine to lookup at
end.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
018d1c667e [PATCH] sky2: race with MTU change
Avoid possible race conditions when doing MTU and change and shutdown.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
0a12257697 [PATCH] sky2: change netif_rx_schedule_test to __netif_schedule_prep
I didn't like the name netif_rx_schedule_test(), in earlier patches
and changed to __netif_rx_schedule_prep to be more consistent.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
a018e3305f [PATCH] sky2: handle DMA boundary crossing
Handle the case of DMA spanning 32 bit boundary.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
5a5b1ea026 [PATCH] sky2: add Yukon-EC ultra support
Add support for Yukon-EC Ultra chip as implemented in SysKonnect's
driver version 8.26.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
13210ce5c0 [PATCH] sky2: improve receive performance
Changes to receive side processing:
	* bigger receive ring
	* clean up polling loop

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:20 -05:00
shemminger@osdl.org
42eeea0145 [PATCH] sky2: fix receive flush/pause issues
Fix issues with pause and flush. This code works on
all chip versions tested.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:19 -05:00
shemminger@osdl.org
ef743d3359 [PATCH] sky2: add MII support
Add MII ioctl interface to sky2.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:19 -05:00
shemminger@osdl.org
2143764348 [PATCH] sky2: eliminate special case for EC-A1
Eliminate special case for EC-A1. The overhead isn't so
great that having config option is worth it.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-12-01 02:20:19 -05:00
Andrew Morton
d0bbccfa32 [PATCH] sky2 needs dma_mapping.h
On alpha:

drivers/net/sky2.c: In function `sky2_probe':
drivers/net/sky2.c:2819: error: `DMA_64BIT_MASK' undeclared (first use in this function)
drivers/net/sky2.c:2819: error: (Each undeclared identifier is reported only once
drivers/net/sky2.c:2819: error: for each function it appears in.)
drivers/net/sky2.c:2825: error: `DMA_32BIT_MASK' undeclared (first use in this function)

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-11-11 05:52:37 -05:00
shemminger@osdl.org
f1e691a249 [PATCH] sky2: version 0.7
Change version number

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-28 12:18:53 -04:00
shemminger@osdl.org
bea8610331 [PATCH] sky2: fix NAPI and receive handling
Speed up the receive and interrupt processing and eliminate a
couple of race conditions from NAPI code.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-28 12:18:53 -04:00
shemminger@osdl.org
d571b694df [PATCH] sky2: spelling fixes
Cosmetic cleanup's: mostly spelling fixes etc.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-28 12:18:53 -04:00
shemminger@osdl.org
b2f5ad4fec [PATCH] sky2: use kzalloc
Can use kzalloc to save a little code.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-28 12:18:52 -04:00
Stephen Hemminger
2995bfb785 [PATCH] sky2: add permanent address support.
Add permanent address support

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-04 07:59:43 -04:00
Stephen Hemminger
9a7ae0a978 [PATCH] sky2: nway reset (BONUS FEATURE)
Here is support for ethtool controlled renegotiation.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:52:58 -04:00
shemminger@osdl.org
724bca3ca4 [PATCH] sky2: version 0.6
Verion number change, comment update and one simple optimization

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:52:57 -04:00
shemminger@osdl.org
5afa0a9cfb [PATCH] sky2: explicit set power state
Add better power management, and power down the chip on device removal

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:52:57 -04:00
shemminger@osdl.org
d1f1370863 [PATCH] sky2: add hardware VLAN acceleration support
Use the hardware to do VLAN.
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:49:32 -04:00
shemminger@osdl.org
d11c13e752 [PATCH] sky2: cleanup interrupt processing
The receive processing can be cleaned up by not using local variables
to store checksum status, instead just put it directly onto the expected
skb.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:49:32 -04:00
shemminger@osdl.org
6b1a3aefd8 [PATCH] sky2: changing mtu doesn't have to reset link
Changing the MTU of the network device doesn't mean the whole link
has to be brought down and back up again.  Just stopping the receive
engine is good enough.

Signed-off-by: Stephen Hemminger <shmminger @osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-28 11:49:32 -04:00
Stephen Hemminger
ecfd7f32aa [PATCH] sky2: version 0.5
Bump version (we are half way to a stable driver??)

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-21 22:30:52 -04:00
Stephen Hemminger
6e4cbb34d4 [PATCH] sky2: allow ethtool debug access to all of PCI space
For debugging, it is useful to allow ethtool to dump all of
the PCI mapped space. There is one small set of registers for RAM
diagnostics that needs to be avoided to prevent hanging, but all
the rest is accessible.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-21 22:30:52 -04:00
Stephen Hemminger
79e57d32fe [PATCH] sky2: fix FIFO DMA alignment problems
The Yukon2 chip FIFO has a problem recovering from a pause frame
if the receive buffer is not aligned. The workaround is to just
leave receive buffer at original alignment, which will make IP
header unaligned. Rework receive logic for clarity.

Need to ignore receive overrun interrupts (or mac hangs).

These fixes make the sky2 driver generally usable.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-21 22:30:52 -04:00
Stephen Hemminger
793b883ed1 [PATCH] sky2: driver update.
Here is revised patch against netdev sky2 branch.
It includes whitespace fixes, all the changes from the previous
review as well as some optimizations and timing fixes to
solve some of the hangs.

The stall problem is better but not perfect. It appears that
under stress the chip can't keep up with the bus
and sends a pause frame, then hangs. This version is for
testing, and hopefully other eyes might see the root
cause of the problem.

I don't want to reinvent the ugly watchdog code in the syskonnect
version of sk98lin.  If you read it you will see, the original
driver writer and the hardware developer obviously didn't
understand each other.

Dual port support is included, but not tested yet. It did
require small change to NAPI since both ports share same
IRQ.
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-16 02:48:03 -04:00
Stephen Hemminger
cd28ab6a4e [PATCH] sky2: new experimental Marvell Yukon2 driver
New driver for the Marvell Yukon2 Gigabit Ethernet chipset.
This driver is based on the skge driver, but using the logic
from the SysKonnect version of the sk98lin driver.
It should support all the Yukon2 chipsets that are available
in many current Intel and AMD motherboards.

The driver does support ethtool, tx and rx checksum, and
tcp segmentation offload.  But it has only been tested for
a short while and is known to stop receiving under heavy load.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-08-17 00:53:12 -04:00