Francois Romieu [Sun, 29 Jun 2008 11:10:54 +0000 (13:10 +0200)]
r8169: change default behavior for mildly identified 8168c chipsets
The addition of a new device has so far implied a specialization of
these masks. While they identify 8168c devices, they can be expected
to be further refined as they have been by Realtek so far.
The change should bring the driver closer to the version 8.006.00 of
Realtek's 8168 driver.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Cc: Edward Hsu <edward_hsu@realtek.com.tw>
Francois Romieu [Sat, 28 Jun 2008 09:58:39 +0000 (11:58 +0200)]
r8169: make room for more specific 8168 hardware start procedure
Broadly speaking the 8168c* share some common code which will
be factored in __rtl_hw_start_8168cp. The 8168b* share some
code too but it will be a bit different.
Any change of behavior should be confined to the currently
unidentified 8168 chipsets. They will not be applied the Tx
performance tweak and will emit a warning instead.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Cc: Edward Hsu <edward_hsu@realtek.com.tw>
Francois Romieu [Sun, 1 Jun 2008 10:31:57 +0000 (12:31 +0200)]
r8169: shuffle some registers handling around (8168 operation only)
I can not argue strongly for (or against) a specific ordering
on a purely technical ground but the patch avoids to swallow
Realtek's changes in one big, hard-to-read gulp.
Let aside the way the RxConfig register is written (see
rtl_set_rx_tx_config_registers / RxConfig / rtl_set_rx_mode),
this change brings the registers write ordering closer with
Realtek's driver one (version 8.006.00) for the 8168 chipsets.
More 8168 specific code which touches the Configx registers will
be added in the section covered by Cfg9346_UnLock / Cfg9346_Lock.
This code should not be the cause of regression for 810x and
8110 users.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Cc: Edward Hsu <edward_hsu@realtek.com.tw>
Vlad Yasevich [Thu, 9 Oct 2008 21:33:26 +0000 (14:33 -0700)]
sctp: update SNMP statiscts when T5 timer expired.
The T5 timer is the timer for the over-all shutdown procedure. If
this timer expires, then shutdown procedure has not completed and we
ABORT the association. We should update SCTP_MIB_ABORTED and
SCTP_MIB_CURRESTAB when aborting.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vlad Yasevich [Thu, 9 Oct 2008 21:33:01 +0000 (14:33 -0700)]
sctp: Fix SNMP number of SCTP_MIB_ABORTED during violation handling.
If ABORT chunks require authentication and a protocol violation
is triggered, we do not tear down the association. Subsequently,
we should not increment SCTP_MIB_ABORTED.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Yongjun [Thu, 9 Oct 2008 21:32:24 +0000 (14:32 -0700)]
sctp: Fix the SNMP number of SCTP_MIB_CURRESTAB
RFC3873 defined SCTP_MIB_CURRESTAB:
sctpCurrEstab OBJECT-TYPE
SYNTAX Gauge32
MAX-ACCESS read-only
STATUS current
DESCRIPTION
"The number of associations for which the current state is
either ESTABLISHED, SHUTDOWN-RECEIVED or SHUTDOWN-PENDING."
REFERENCE
"Section 4 in RFC2960 covers the SCTP Association state
diagram."
If the T4 RTO timer expires many times(timeout), the association will enter
CLOSED state, so we should dec the number of SCTP_MIB_CURRESTAB, not inc the
number of SCTP_MIB_CURRESTAB.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dave Graham [Thu, 9 Oct 2008 21:29:26 +0000 (14:29 -0700)]
e1000: don't generate bad checksums for tcp packets with 0 csum
When offloading transmit checksums only, the driver was not
correctly configuring the hardware to handle the case of a zero
checksum. For UDP the correct behavior is to leave it alone, but
for tcp the checksum must be changed from 0x0000 to 0xFFFF. The
hardware takes care of this case but only if it is told the
packet is tcp.
same patch as e1000e
Signed-off-by: Dave Graham <david.graham@intel.com> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dave Graham [Thu, 9 Oct 2008 21:28:58 +0000 (14:28 -0700)]
e1000e: don't generate bad checksums for tcp packets with 0 csum
When offloading transmit checksums only, the driver was not
correctly configuring the hardware to handle the case of a zero
checksum. For UDP the correct behavior is to leave it alone, but
for tcp the checksum must be changed from 0x0000 to 0xFFFF. The
hardware takes care of this case but only if it is told the
packet is tcp.
Signed-off-by: Dave Graham <david.graham@intel.com> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Benjamin Li [Thu, 9 Oct 2008 19:26:41 +0000 (12:26 -0700)]
bnx2: Handle DMA mapping errors.
Before, the driver would not care about the return codes from pci_map_*
functions. This could be potentially dangerous if a mapping failed.
Now, we will check all pci_map_* calls. On the transmit side, we switch
to use the new function skb_dma_map(). On the receive side, we add
pci_dma_mapping_error().
Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Thu, 9 Oct 2008 19:21:46 +0000 (12:21 -0700)]
bnx2: Check netif_running() in all ethtool operations.
We need to check netif_running() state in most ethtool operations
and properly handle the !netif_running() state where the chip is
in an uninitailzed state or low power state that may not accept
any MMIO.
Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Thu, 9 Oct 2008 19:21:08 +0000 (12:21 -0700)]
bnx2: Add bnx2_shutdown_chip().
This logic is used in bnx2_close() and bnx2_suspend() and
so should be separated out into a separate function.
Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Oct 2008 19:03:17 +0000 (12:03 -0700)]
inet: Make tunnel RX/TX byte counters more consistent
This patch makes the RX/TX byte counters for IPIP, GRE and SIT more
consistent. Previously we included the external IP headers on the
way out but not when the packet is inbound.
The new scheme is to count payload only in both directions. For
IPIP and SIT this simply means the exclusion of the external IP
header. For GRE this means that we exclude the GRE header as
well.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Oct 2008 19:00:17 +0000 (12:00 -0700)]
gre: Add Transparent Ethernet Bridging
This patch adds support for Ethernet over GRE encapsulation.
This is exposed to user-space with a new link type of "gretap"
instead of "gre". It will create an ARPHRD_ETHER device in
lieu of the usual ARPHRD_IPGRE.
Note that to preserver backwards compatibility all Transparent
Ethernet Bridging packets are passed to an ARPHRD_IPGRE tunnel
if its key matches and there is no ARPHRD_ETHER device whose
key matches more closely.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Oct 2008 18:59:55 +0000 (11:59 -0700)]
gre: Add netlink interface
This patch adds a netlink interface that will eventually displace
the existing ioctl interface. It utilises the elegant rtnl_link_ops
mechanism.
This also means that user-space no longer needs to rely on the
tunnel interface being of type GRE to identify GRE tunnels. The
identification can now occur using rtnl_link_ops.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Oct 2008 18:59:32 +0000 (11:59 -0700)]
gre: Move MTU setting out of ipgre_tunnel_bind_dev
This patch moves the dev->mtu setting out of ipgre_tunnel_bind_dev.
This is in prepartion of using rtnl_link where we'll need to make
the MTU setting conditional on whether the user has supplied an
MTU. This also requires the move of the ipgre_tunnel_bind_dev
call out of the dev->init function so that we can access the user
parameters later.
This patch also adds a check to prevent setting the MTU below
the minimum of 68.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Oct 2008 18:58:54 +0000 (11:58 -0700)]
gre: Use needed_headroom
Now that we have dev->needed_headroom, we can use it instead of
having a bogus dev->hard_header_len. This also allows us to
include dev->hard_header_len in the MTU computation so that when
we do have a meaningful hard_harder_len in future it is included
automatically in figuring out the MTU.
Incidentally, this fixes a bug where we ignored the needed_headroom
field of the underlying device in calculating our own hard_header_len.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
phylib: two dynamic mii_bus allocation fallout fixes
1. arch/powerpc/platforms/pasemi/gpio_mdio.c also needs to be
converted over to mdiobus_{alloc,free}().
2. drivers/net/phy/fixed.c used to embed a struct mii_bus into its
struct fixed_mdio_bus and then use container_of() to go from the
former to the latter. Since mii bus structures are no longer
embedded, we need to do something like use the mii bus private
pointer to go from mii_bus to fixed_mdio_bus instead.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
akeemting [Thu, 9 Oct 2008 02:50:03 +0000 (19:50 -0700)]
jme: Faulty IRQ handle bug fix
Fix IRQ handle bug when interrupt mode.
The driver was incorrectly handled and returned IRQ_HANDLED
while the device is not generating the interrupt.
It happened due to faulty determination of interrupt status register.
Found by: "Ethan" <ethanhsiao@jmicron.com>
Fixed by: "akeemting" <akeem@jmicron.com> Signed-off-by: Guo-Fu Tseng <cooldavid@cooldavid.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Guo-Fu Tseng [Thu, 9 Oct 2008 02:48:58 +0000 (19:48 -0700)]
jme: Added half-duplex mode and IPv6 RSS fix
1. Set bit 5 of GPREG1 to 1 to enable hardware workaround for half-duplex
mode. Which the MAC processor generates CRS/COL by itself instead of
receive it from PHY processor.
2. Set bit 6 of GPREG1 to 1 to enable hardware workaround that masks the
MAC processor working right while calculating IPv6 RSS in 10/100
mode.
3. Group the workaround codes all together.
Signed-off-by: Guo-Fu Tseng <cooldavid@cooldavid.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 9 Oct 2008 00:39:31 +0000 (17:39 -0700)]
cxgb3: commnonize LASI phy code
Add generic code to manage interrupt driven PHYs.
Do not reset the phy after link parameters update,
the new values might get lost.
Return early from link change notification
when the link parameters remain unchanged.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 9 Oct 2008 00:39:00 +0000 (17:39 -0700)]
cxgb3: More flexible support for PHY interrupts.
Do not require PHY interrupts to be connected to GPIs in ascending order.
Base interrupt availability both on PHYs supporting them and on GPIs being
hooked up. Allows boards to specify interrupt GPIs though the PHYs don't
use them.
Remove spurious PHY interrupts due to clearing T3DBG interrupts before
setting their polarity.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 9 Oct 2008 00:38:29 +0000 (17:38 -0700)]
cxgb3: simplify port type struct and usage
Second step in overall phy layer reorganization.
Clean up the port_type_info structure.
Support coextistence of clause 22 and clause 45 MDIO devices.
Select the type of MDIO transaction on a per transaction basis.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 9 Oct 2008 00:37:33 +0000 (17:37 -0700)]
cxgb3: Allocate multiqueues at init time
Allocate a queue set per core, up to the maximum of available qsets.
Share the queue sets on multi port adapters.
Rename MSI-X interrupt vectors ethX-N, N being the queue set number.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
dsa: add support for the Marvell 88E6060 switch chip
Add support for the Marvell 88E6060 switch chip. This chip only
supports the Header and Trailer tagging formats, and we use it in
Trailer mode since that mode is slightly easier to handle than
Header mode.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Tested-by: Byron Bradley <byron.bbradley@gmail.com> Tested-by: Tim Ellis <tim.ellis@mac.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This adds support for the Trailer switch tagging format. This is
another tagging that doesn't explicitly mark tagged packets with a
distinct ethertype, so that we need to add a similar hack in the
receive path as for the Original DSA tagging format.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Tested-by: Byron Bradley <byron.bbradley@gmail.com> Tested-by: Tim Ellis <tim.ellis@mac.com> Signed-off-by: David S. Miller <davem@davemloft.net>
dsa: add support for the Marvell 88E6131 switch chip
Add support for the Marvell 88E6131 switch chip. This chip only
supports the original (ethertype-less) DSA tagging format.
On the 88E6131, there is a PHY Polling Unit (PPU) which has exclusive
access to each of the PHYs's MII management registers. If we want to
talk to the PHYs from software, we have to disable the PPU and wait
for it to complete its current transaction before we can do so, and we
need to re-enable the PPU afterwards to make sure that the switch will
notice changes in link state and speed on the individual ports as they
occur.
Since disabling the PPU is rather slow, and since MII management
accesses are typically done in bursts, this patch keeps the PPU disabled
for 10ms after a software access completes. This makes handling the
PPU slightly more complex, but speeds up something like running ethtool
on one of the switch slave interfaces from ~300ms to ~30ms on typical
hardware.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Tested-by: Nicolas Pitre <nico@marvell.com> Tested-by: Peter van Valderen <linux@ddcrew.com> Tested-by: Dirk Teurlings <dirk@upexia.nl> Signed-off-by: David S. Miller <davem@davemloft.net>
Most of the DSA switches currently in the field do not support the
Ethertype DSA tagging format that one of the previous patches added
support for, but only the original DSA tagging format.
The original DSA tagging format carries the same information as the
Ethertype DSA tagging format, but with the difference that it does not
have an ethertype field. In other words, when receiving a packet that
is tagged with an original DSA tag, there is no way of telling in
eth_type_trans() that this packet is in fact a DSA-tagged packet.
This patch adds a hook into eth_type_trans() which is only compiled in
if support for a switch chip that doesn't support Ethertype DSA is
selected, and which checks whether there is a DSA switch driver
instance attached to this network device which uses the old tag format.
If so, it sets the protocol field to ETH_P_DSA without looking at the
packet, so that the packet ends up in the right place.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Tested-by: Nicolas Pitre <nico@marvell.com> Tested-by: Peter van Valderen <linux@ddcrew.com> Tested-by: Dirk Teurlings <dirk@upexia.nl> Signed-off-by: David S. Miller <davem@davemloft.net>
net: Distributed Switch Architecture protocol support
Distributed Switch Architecture is a protocol for managing hardware
switch chips. It consists of a set of MII management registers and
commands to configure the switch, and an ethernet header format to
signal which of the ports of the switch a packet was received from
or is intended to be sent to.
The switches that this driver supports are typically embedded in
access points and routers, and a typical setup with a DSA switch
looks something like this:
The switch driver presents each port on the switch as a separate
network interface to Linux, polls the switch to maintain software
link state of those ports, forwards MII management interface
accesses to those network interfaces (e.g. as done by ethtool) to
the switch, and exposes the switch's hardware statistics counters
via the appropriate Linux kernel interfaces.
This initial patch supports the MII management interface register
layout of the Marvell 88E6123, 88E6161 and 88E6165 switch chips, and
supports the "Ethertype DSA" packet tagging format.
(There is no officially registered ethertype for the Ethertype DSA
packet format, so we just grab a random one. The ethertype to use
is programmed into the switch, and the switch driver uses the value
of ETH_P_EDSA for this, so this define can be changed at any time in
the future if the one we chose is allocated to another protocol or
if Ethertype DSA gets its own officially registered ethertype, and
everything will continue to work.)
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Tested-by: Nicolas Pitre <nico@marvell.com> Tested-by: Byron Bradley <byron.bbradley@gmail.com> Tested-by: Tim Ellis <tim.ellis@mac.com> Tested-by: Peter van Valderen <linux@ddcrew.com> Tested-by: Dirk Teurlings <dirk@upexia.nl> Signed-off-by: David S. Miller <davem@davemloft.net>
The write barrier should be used before starting a DMA transfer. This fixes
a problem, where almost all packets received on another machine had garbled
content. Tested with an RTL8100C on a MIPS machine.
Signed-off-by: Andreas Oberritter <obi@linuxtv.org> Signed-off-by: David S. Miller <davem@davemloft.net>
there's several drivers that have use "tx_timeout" for the .. tx
timeout function. All fine with that, they're static, however for
doing stats on how often which driver hits the timeout it's a tad
unfortunate. The patch below gives the ones I found in the
kerneloops.org database unique names.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Bruno Prémont [Thu, 9 Oct 2008 00:06:25 +0000 (17:06 -0700)]
r8169: WoL fixes, part 2.
Since recent kernel (2.6.26 or 2.6.27) the PCI wakeup functions are
influenced by generic device ability and configuration when enabling
PCI-device triggered wake-up.
This patch causes WoL setting to enable/disable device's wish to
be permitted to wake-up the host when changing WoL options and
also during device probing.
Without this patch one has write 'enabled' to
/sys/bus/pci/devices/0000:02:08.0/power/wakeup
Signed-off-by: Bruno Prémont <bonbons@linux-vserver.org> Acked-by: Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
Trent Piepho [Thu, 2 Oct 2008 11:12:24 +0000 (11:12 +0000)]
gianfar: Create net device with carrier down
The device's carrier status is controlled via the functions
netif_carrier_on() and netif_carrier_off(). These set or clear a bit
indicating the carrier (aka lower level link) is down, and if the state
changed, they fire off a routing netlink event.
Add a call to netif_carrier_off() before register_netdev() so that the
newly created device will be set to carrier down. Then when the carrier
comes up for the first time, a netlink event will be generated, as the
carrier changed from down to up. Otherwise the initial carrier up will
appear to be changing the status from up to up, and so no event is
generated since that's not a change.
Signed-off-by: Trent Piepho <tpiepho@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
mv643xx_eth uses ip_hdr() (defined in linux/ip.h), but relied on
another header file to include the needed header file indirectly.
In latest net-next this indirect include chain is gone, so the
driver fails to build. Include linux/ip.h explicitly to fix this.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
is caused by unconditional local_irq_disable()/local_irq_enable() in
disable_irq_lockdep()/enable_irq_lockdep() used by __ei_poll(). Since
netconsole/netpoll always calls dev->poll_controller() with local irqs
disabled, disable_irq()/enable_irq() instead is safe and enough (like
e.g. in 3c509 or 8139xx drivers).
Reported-and-tested-by: Bernard Pidoux F6BVP <f6bvp@free.fr> Signed-off-by: Jarek Poplawski <jarkao2@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
It was pointed out by Breno Leitao <leitao@linux.vnet.ibm.com> that
ixgb would crash on PPC when an IOMMU was in use, if change_mtu was
called.
It appears to be a pretty simple issue in the driver that wasn't discovered
because most systems don't run with an IOMMU. The driver needs to only unmap
buffers that are mapped (duh).
CC: Breno Leitao <leitao@linux.vnet.ibm.com> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Arthur Jones [Mon, 29 Sep 2008 11:51:09 +0000 (11:51 +0000)]
e1000e: release hw semaphore after successfully writing EEPROM
Since e1000e has been existance in linux-2.6, we've
never released the hardware semaphore after a successful
write to the SPI EEPROM. I guess we don't write to
SPI EEPROM much -- but those few of us that do appreciate
it when we can later read from the EEPROM without having
to reboot.
Found-by: Nick Van Fossen <Nick.VanFossen@riverbed.com> Signed-off-by: Arthur Jones <ajones@riverbed.com> Reviewed-by: Auke Kok <auke-jan.h.kok@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Add mdiobus_{read,write} routines to allow direct reading/writing
of registers on an mii bus without having to go through the PHY
abstraction, and make phy_{read,write} use these primitives.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce the mdio_bus class, and give each 'struct mii_bus' its own
'struct device', so that mii_bus objects are represented in the device
tree and can be found by querying the device tree.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Andy Fleming <afleming@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
phylib: move to dynamic allocation of struct mii_bus
This patch introduces mdiobus_alloc() and mdiobus_free(), and
makes all mdio bus drivers use these functions to allocate their
struct mii_bus'es dynamically.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Andy Fleming <afleming@freescale.com>
In preparation of giving mii_bus objects a device tree presence of
their own, rename struct mii_bus's ->dev argument to ->parent, since
having a 'struct device *dev' that points to our parent device
conflicts with introducing a 'struct device dev' representing our own
device.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Andy Fleming <afleming@freescale.com>
Jay Cliburn [Sat, 27 Sep 2008 04:17:22 +0000 (04:17 +0000)]
atl1: remove EXPERIMENTAL label
Remove the EXPERIMENTAL label from the atl1 driver and change the vendor
name to include Attansic's successor, Atheros. We'll leave Attansic in
the name since Attansic's PCI ID (1969) is encoded in the PCI config and
is what users encounter on their systems.
Signed-off-by: Jay Cliburn <jacliburn@bellsouth.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[10/21] driver/net/skge.c: restart the interface when it's options or
pauseparam is set
On Wednesday 24 September 2008 07:47, Stephen Hemminger wrote:
> On Mon, 22 Sep 2008 14:52:17 -0700
>
> akpm@linux-foundation.org wrote:
> > From: "Xiaoming.Zhang" <Xiaoming.Zhang@resilience.com>
> >
> > We have an issue of the skge driver: The card won't work when it's
> > options are changed. Here's the hardware info:
> >
> > # lspci -v
> > 05:04.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001
> > Gigabit Ethernet Controller (rev 13) Subsystem: Marvell Technology Group
> > Ltd. Marvell RDK-8001 Flags: bus master, 66MHz, medium devsel, latency
> > 32, IRQ 16 Memory at d042c000 (32-bit, non-prefetchable) [size=16K] I/O
> > ports at d000 [size=256]
> > [virtual] Expansion ROM at 20400000 [disabled] [size=128K]
> > Capabilities: [48] Power Management version 2
> > Capabilities: [50] Vital Product Data
> >
> > The happens in both Linux-2.6.26(skge version 1.23) and RHEL5.2(skge
> > version 1.6).
> >
> > For example, at first it is set to "speed 1000 duplex full auto-neg on"
> > and it works, then run
> >
> > ethtool -s <ethx> autoneg off
> > or ethtool -s <ethx> speed 100 duplex full autoneg off
> >
> > Then it will stop working. After that if we restart the interface:
> >
> > ifconifg <ethx> down
> > ifconfig <ethx> up
> >
> > It will work again. And `ethtool -A' has the same issue.
> >
> > So we think after setting the options, the interface should be restarted.
> >
> > Signed-off-by: Zhang Xiaoming <xiaoming.zhang@resilience.com>
> > Cc: Stephen Hemminger <shemminger@vyatta.com>
> > Cc: Jeff Garzik <jeff@garzik.org>
> > Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> > ---
> >
> > drivers/net/skge.c | 12 ++++++++----
> > 1 file changed, 8 insertions(+), 4 deletions(-)
> >
> > diff -puN
> > drivers/net/skge.c~driver-net-skgec-restart-the-interface-when-its-option
> >s-or-pauseparam-is-set drivers/net/skge.c ---
> > a/drivers/net/skge.c~driver-net-skgec-restart-the-interface-when-its-opti
> >ons-or-pauseparam-is-set +++ a/drivers/net/skge.c
> > @@ -353,8 +353,10 @@ static int skge_set_settings(struct net_
> > skge->autoneg = ecmd->autoneg;
> > skge->advertising = ecmd->advertising;
> >
> > - if (netif_running(dev))
> > - skge_phy_reset(skge);
> > + if (netif_running(dev)) {
> > + skge_down(dev);
> > + skge_up(dev);
> > + }
> >
> > return (0);
> > }
> > @@ -595,8 +597,10 @@ static int skge_set_pauseparam(struct ne
> > skge->flow_control = FLOW_MODE_NONE;
> > }
> >
> > - if (netif_running(dev))
> > - skge_phy_reset(skge);
> > + if (netif_running(dev)) {
> > + skge_down(dev);
> > + skge_up(dev);
> > + }
> >
> > return 0;
> > }
>
> Since skge_up can fail because of out of memory, this code needs to
> check the return value. And then if it fails the "limbo state" needs
> to be handled in skge_down.
How about like this? It is tested.
Thank you.
Signed-off-by: Zhang Xiaoming <xiaoming.zhang@resilience.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kevin Hao [Thu, 25 Sep 2008 16:20:11 +0000 (16:20 +0000)]
net: remove LLTX in atl2 driver
When NETIF_F_LLTX is set, the atlx driver will use a private lock.
But in recent kernels this implementation seems redundant and
can cause problems where AF_PACKET sees things twice. Since
NETIF_F_LLTX is marked as deprecated and shouldn't be used in
new driver, this patch removes NETIF_F_LLTX and adds a mmiowb
before sending packet. I have tested this driver on a Eee PC.
It works well.
Signed-off-by: Kevin Hao <kexin.hao@windriver.com> Acked-by: Jay Cliburn <jacliburn@bellsouth.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Trent Piepho [Wed, 8 Oct 2008 22:46:57 +0000 (15:46 -0700)]
phylib: Wake PHY state machine on interrupt
This way the phy layer will respond to a change in phy state immediately,
instead of up to one second later when the state machine timer runs.
Signed-off-by: Trent Piepho <tpiepho@freescale.com> Acked-by: Andy Fleming <afleming@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The PHY's aneg is configured and restarted whenever the link is brought up,
e.g. when DHCP is started after the kernel has booted. This can take the
link down for several seconds while auto-negotiation is redone.
If the advertised features haven't changed, then it shouldn't be necessary
to bring down the link and start auto-negotiation over again.
genphy_config_advert() is enhanced to return 0 when the advertised features
haven't been changed and >0 when they have been.
genphy_config_aneg() then uses this information to not call
genphy_restart_aneg() if there has been no change.
Signed-off-by: Trent Piepho <tpiepho@freescale.com> Acked-by: Andy Fleming <afleming@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kevin Hao [Fri, 19 Sep 2008 21:56:44 +0000 (21:56 +0000)]
net: add net poll support for atl2 driver
Add netconsole support for Atheros L2 10/100 network device.
Signed-off-by: Kevin Hao <kexin.hao@windriver.com> Acked-by: Jay Cliburn <jacliburn@bellsouth.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Make the SIOCGMIIPHY case fall through properly (it is supposed
to not only return the ID of the default PHY but also to read from
that PHY), and make phy_mii_ioctl() return the same error code as
generic_mii_ioctl() in case of an unsupported operation.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Andy Fleming <afleming@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vlad Yasevich [Wed, 8 Oct 2008 21:19:01 +0000 (14:19 -0700)]
sctp: shrink sctp_tsnmap some more by removing gabs array
The gabs array in the sctp_tsnmap structure is only used
in one place, sctp_make_sack(). As such, carrying the
array around in the sctp_tsnmap and thus directly in
the sctp_association is rather pointless since most
of the time it's just taking up space. Now, let
sctp_make_sack create and populate it and then throw
it away when it's done.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vlad Yasevich [Wed, 8 Oct 2008 21:18:39 +0000 (14:18 -0700)]
sctp: Rework the tsn map to use generic bitmap.
The tsn map currently use is 4K large and is stuck inside
the sctp_association structure making memory references REALLY
expensive. What we really need is at most 4K worth of bits
so the biggest map we would have is 512 bytes. Also, the
map is only really usefull when we have gaps to store and
report. As such, starting with minimal map of say 32 TSNs (bits)
should be enough for normal low-loss operations. We can grow
the map by some multiple of 32 along with some extra room any
time we receive the TSN which would put us outside of the map
boundry. As we close gaps, we can shift the map to rebase
it on the latest TSN we've seen. This saves 4088 bytes per
association just in the map alone along savings from the now
unnecessary structure members.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 8 Oct 2008 21:18:04 +0000 (14:18 -0700)]
inet: cleanup of local_port_range
I noticed sysctl_local_port_range[] and its associated seqlock
sysctl_local_port_range_lock were on separate cache lines.
Moreover, sysctl_local_port_range[] was close to unrelated
variables, highly modified, leading to cache misses.
Moving these two variables in a structure can help data
locality and moving this structure to read_mostly section
helps sharing of this data among cpus.
Cleanup of extern declarations (moved in include file where
they belong), and use of inet_get_local_port_range()
accessor instead of direct access to ports values.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 8 Oct 2008 18:44:17 +0000 (11:44 -0700)]
udp: Improve port randomization
Current UDP port allocation is suboptimal.
We select the shortest chain to chose a port (out of 512)
that will hash in this shortest chain.
First, it can lead to give not so ramdom ports and ease
give attackers more opportunities to break the system.
Second, it can consume a lot of CPU to scan all table
in order to find the shortest chain.
Third, in some pathological cases we can fail to find
a free port even if they are plenty of them.
This patch zap the search for a short chain and only
use one random seed. Problem of getting long chains
should be addressed in another way, since we can
obtain long chains with non random ports.
Based on a report and patch from Vitaly Mayatskikh
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
tcp: Fix tcp_hybla zero congestion window growth with small rho and large cwnd.
net: Fix netdev_run_todo dead-lock
tcp: Fix possible double-ack w/ user dma
net: only invoke dev->change_rx_flags when device is UP
netrom: Fix sock_orphan() use in nr_release
ax25: Quick fix for making sure unaccepted sockets get destroyed.
Revert "ax25: Fix std timer socket destroy handling."
[Bluetooth] Add reset quirk for A-Link BlueUSB21 dongle
[Bluetooth] Add reset quirk for new Targus and Belkin dongles
[Bluetooth] Fix double frees on error paths of btusb and bpa10x drivers
Jarek Poplawski [Wed, 8 Oct 2008 18:36:22 +0000 (11:36 -0700)]
pkt_sched: Update qdisc requeue stats in dev_requeue_skb()
After the last change of requeuing there is no info about such
incidents in tc stats. This patch updates the counter, but we should
consider this should differ from previous stats because of additional
checks preventing to repeat this. On the other hand, previous stats
didn't include requeuing of gso_segmented skbs.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ilpo Järvinen [Wed, 8 Oct 2008 18:34:06 +0000 (11:34 -0700)]
tcp: fix length used for checksum in a reset
While looking for some common code I came across difference
in checksum calculation between tcp_v6_send_(reset|ack) I
couldn't explain. I checked both v4 and v6 and found out that
both seem to have the same "feature". I couldn't find anything
in rfc nor anywhere else which would state that md5 option
should be ignored like it was in case of reset so I came to
a conclusion that this is probably a genuine bug. I suspect
that addition of md5 just was fooled by the excessive
copy-paste code in those functions and the reset part was
never tested well enough to find out the problem.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-2.6-omap-h63xx.git / log