22 Mar, 2009
23 commits
-
Fix compiler warning about non-const format string.
Signed-off-by: Stephen Hemminger
Acked-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
Protocols should be able to use constant value for the descriptor.
Minor whitespace cleanup as wellSigned-off-by: Stephen Hemminger
Acked-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
The igb driver was using map_single to map the skbs and then unmap_page to
unmap them. This update changes that so instead uses skb_dma_map and
skb_dma_unmap.In addition the next_to_watch member of the buffer_info struct was being
set uneccesarily. I removed the spots where it was set without being needed.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
Currently if there are no multicast addresses programmed into the PF then
the VFs cannot have their multicast filters reset. This change makes it so
the code path that updates vf multicast is always called along with the pf
updates.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
igb was previously setting up all of the timer members itself. It is
easier to just call setup_timer and reduce the calls to one line.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
This patch removes IGB_DESC_UNUSED and replaces it with a function call
instead in order to cleanup some of the ugliness introduced by the macro.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
When the 82575 is fed 802.1q packets, it chokes with
an error of the form:igb 0000:08:00.1 partial checksum but proto=81!
As the logic there was not smart enough to look into
the vlan header to pick out the encapsulated protocol.There are times when we'd like to send these packets
out without having to configure a vlan on the interface.
Here we check for the vlan tag and allow the packet to
go out with the correct hardware checksum.Thanks to Kand Ly for discovering the
issue and the coming up with a solution. This patch is
based upon his work.Signed-off-by: Arthur Jones
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
Signed-off-by: Yang Hongyang
Signed-off-by: David S. Miller -
Hi,
IBM EMAC driver performs device reset (drivers/net/ibm_newemac/core.c:
emac_probe() -> emac_init_phy() -> emac_reset()) before registering
appropriate net_device (emac_probe() -> register_netdev()), so
net_device name contains raw format string during EMAC reset ("eth%d").If the case of reset timeout, emac_report_timeout_error() function is
called to report an error. The problem is this function uses net_device
name to report device related, which is not correct, as a result in the
kernel log buffer we see:eth%d: reset timeout
The solution is to print device_node full_name instead. After applying
the patch proposed, error string is like the following:/plb/opb/ethernet@ef600e00: reset timeout
Signed-off-by: Mikhail Zolotaryov
Acked-by: Benjamin Herrenschmidt
Signed-off-by: David S. Miller -
There is no gain using prefetch() in dev_hard_start_xmit(), since
we already had to read ops->ndo_select_queue pointer in dev_pick_tx(),
and both pointers are probably located in the same cache line.This prefetch call slows down fast path because of a stall in address
computation.Signed-off-by: Eric Dumazet
Signed-off-by: David S. Miller -
Remove 2 TEST_FRAME hacks that are no longer needed. These allowed
sctp regression tests to compile before, but are no longer needed.Signed-off-by: Vlad Yasevich
Signed-off-by: David S. Miller -
Convert ixgbe to use net_device_ops properly.
Rather than changing the select_queue function pointer
just check the flag.Signed-off-by: Stephen Hemminger
Acked-by: Peter P Waskiewicz Jr
Signed-off-by: David S. Miller -
Some minor changes to queue hashing:
1. Use const on accessor functions
2. Export skb_tx_hash for use in drivers (see ixgbe)Signed-off-by: Stephen Hemminger
Signed-off-by: David S. Miller -
Rather than calling device pointer directly (which is incorrect with
net_device_ops), use the standard dev_change_mtu. Compile tested only.Signed-off-by: Stephen Hemminger
Signed-off-by: David S. Miller -
tcp_sack_swap seems unnecessary so I pushed swap to the caller.
Also removed comment that seemed then pointless, and added include
when not already there. Compile tested.Signed-off-by: Ilpo Järvinen
Signed-off-by: David S. Miller -
A zero length frame filter was recently introduced in ROSE protocole.
Previous commit makes the same at AX25 protocole level.
This patch has the same purpose for NetRom protocole.
The reason is that empty frames have no meaning in NetRom protocole.Signed-off-by: Bernard Pidoux
Signed-off-by: David S. Miller -
In previous commit 244f46ae6e9e18f6fc0be7d1f49febde4762c34b
was introduced a zero length frame filter for ROSE protocole.
This patch has the same purpose at AX25 frame level for the same
reason. Empty frames have no meaning in AX25 protocole.Signed-off-by: Bernard Pidoux
Signed-off-by: David S. Miller -
This patch condenses two debug messages in one.
Signed-off-by: Bernard Pidoux
Signed-off-by: David S. Miller -
commit 1577ecef766650a57fceb171acee2b13cbfaf1d3 ("netdev: Merge UCC
and gianfar MDIO bus drivers") broke the TSEC TBI PHY support: the
driver now refuses to probe TBI MDIO buses as it doesn't know about
"fsl,gianfar-tbi" compatible entry, and thus _probe() fails with
-ENODEV status.Fix this by adding "fsl,gianfar-tbi" to the list of known Gianfar
MDIO buses.Signed-off-by: Anton Vorontsov
Acked-by: Andy Fleming
Signed-off-by: David S. Miller -
reorder struct inet6_ifaddr to remove padding on 64 bit builds
remove 8 bytes of padding so inet6_ifaddr becomes 192 bytes & fits into
a smaller slab.Signed-off-by: Richard Kennedy
Signed-off-by: David S. Miller -
commit 4826857f1bf07f9c0f1495e9b05d125552c88a85 ("gianfar: pass the
proper dev to DMA ops") introduced this build breakage:CC drivers/net/gianfar.o
drivers/net/gianfar.c: In function 'gfar_suspend':
drivers/net/gianfar.c:552: error: 'struct gfar_private' has no member named 'dev'
drivers/net/gianfar.c: In function 'gfar_resume':
drivers/net/gianfar.c:601: error: 'struct gfar_private' has no member named 'dev'
make[2]: *** [drivers/net/gianfar.o] Error 1Fix this by converting suspend and resume routines to use
gfar_private->ndev.Signed-off-by: Anton Vorontsov
Acked-by: Kumar Gala
Signed-off-by: David S. Miller -
This patch enables support for the new Intel 82552 adapter (new PHY paired
with the existing MAC in the ICH7 chipset). No new features are added to
the driver, however there are minor changes due to updated registers and a
few workarounds for hardware errata.Signed-off-by: Bruce Allan
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller
21 Mar, 2009
1 commit
-
When merging into Jeff's tree:
commit 5f66f208064f083aab5e55935d0575892e033b59
Author: Arthur Jones
Date: Thu Mar 19 01:13:08 2009 +0000e1000e: allow tx of pre-formatted vlan tagged packets
We lost one line, this fixes that missing
piece...Signed-off-by: Arthur Jones
Signed-off-by: David S. Miller
20 Mar, 2009
16 commits
-
Conflicts:
drivers/net/virtio_net.c -
dev_queue_xmit() needs to dirty fields "state", "q", "bstats" and "qstats"
On x86_64 arch, they currently span three cache lines, involving more
cache line ping pongs than necessary, making longer holding of queue spinlock.We can reduce this to one cache line, by grouping all read-mostly fields
at the beginning of structure. (Or should I say, all highly modified fields
at the end :) )Before patch :
offsetof(struct Qdisc, state)=0x38
offsetof(struct Qdisc, q)=0x48
offsetof(struct Qdisc, bstats)=0x80
offsetof(struct Qdisc, qstats)=0x90
sizeof(struct Qdisc)=0xc8After patch :
offsetof(struct Qdisc, state)=0x80
offsetof(struct Qdisc, q)=0x88
offsetof(struct Qdisc, bstats)=0xa0
offsetof(struct Qdisc, qstats)=0xac
sizeof(struct Qdisc)=0xc0Signed-off-by: Eric Dumazet
Signed-off-by: David S. Miller -
Removed duplicated #include in drivers/net/dnet.c.
Signed-off-by: Huang Weiyi
Signed-off-by: David S. Miller -
The watchdog will schedule an interrupt to help make sure queues are
cleaned in the case when an interrupt is missed, most likely due to very
high load. On 82599, there are extra interrupt registers to account for
the larger number of MSI-X vectors (64 total for 82599 vs. 18 total for
82598). These must be taken into account when performing this operation in
the watchdog.Signed-off-by: Jesse Brandeburg
Signed-off-by: Peter P Waskiewicz Jr
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
ethtool isn't reporting the support level of WoL for 82599 KX4 devices.
While the device does support WoL, ethtool was never updated to properly
report the level of support, nor will it allow ethtool to modify the type
of packets to listen for.Signed-off-by: Peter P Waskiewicz Jr
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
82598's PCI bus reporting on driver load was broken after 82599 merged.
This results in incorrect reporting, and an erroneous warning message
that the 82598 is in a PCIe slot that isn't fast enough to run 10GbE.Signed-off-by: Peter P Waskiewicz Jr
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
This code updates ixgb so that it can use the skb_dma_map/unmap functions
to map the buffers. In addition it also updates the tx hang logic to use
time_stamp instead of dma to determine if it has detected a tx hang.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
As with igb, when the e1000e driver is fed 802.1q
packets with hardware checksum on, it chokes with an
error of the form:checksum_partial proto=81!
As the logic there was not smart enough to look into
the vlan header to pick out the encapsulated protocol.There are times when we'd like to send these packets
out without having to configure a vlan on the interface.
Here we check for the vlan tag and allow the packet to
go out wiht the correct hardware checksum.Thanks to Kand Ly for discovering the
issue and the coming up with a solution. This patch is
based upon his work.Fixups from Stephen Hemminger and
Alexander DuyckSigned-off-by: Arthur Jones
Signed-off-by: Jeff Kirsher
CC: Stephen Hemminger
CC: Alexander Duyck
Signed-off-by: David S. Miller -
There were a few issues I noticed in e1000e. These include a double free
of the skb if mapping fails, and the fact that context descriptors appear
to be left in the descriptor ring after the failure.Signed-off-by: Alexander Duyck
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
Add device ID and related support for 82583 mac.
Signed-off-by: Alexander Duyck
Acked-by: Radheka Godse
Acked-by: Bruce Allan
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
The cpu_to_le16 here looks suspicious to me, I don't think we need it
because put_unaligned_le16 also does this.I don't currently have any big endian hardware with a PCI bus available
to test on, so I haven't been able to verify this.Signed-off-by: Steve Glendinning
Signed-off-by: David S. Miller -
Signed-off-by: David S. Miller
-
This patch replaces the generic eth_mac_addr function with one that
also updates the hardware mac address registers. It also renames the
existing smsc911x_set_mac_address function to
smsc911x_hw_set_mac_address for clarity.Newer LAN911x and all LAN921x devices also support changing the mac
address while the device is running, which is useful for some bonding
modes.Signed-off-by: Steve Glendinning
Signed-off-by: David S. Miller -
Signed-off-by: Steve Glendinning
Signed-off-by: David S. Miller -
The vast majority of bit constants in this driver are defined in the
header file, but TX and RX status word bits are not. This patch (which
should make no functional change) defines these, to make the driver
slightly more readable.Signed-off-by: Steve Glendinning
Signed-off-by: David S. Miller -
To improve manageability, it would be good to be able to disambiguate routes
added by administrator from those added by DHCP client. The only necessary
kernel change is to add value to rtnetlink include file so iproute2 utility
can use it.Signed-off-by: Stephen Hemminger
Signed-off-by: David S. Miller
