Conflicts:
drivers/nfc/microread/mei.c
net/netfilter/nfnetlink_queue_core.c
Pull in 'net' to get Eric Biederman's AF_UNIX fix, upon which
some cleanups are going to go on-top.
Signed-off-by: David S. Miller <davem@davemloft.net>
The suggestion as following:
- initial settings or default settings
- rtl_hw_start_xxx. rtl_hw_start_xxx may change some default settings.
- enable tx/rx. This has to be after the above two steps.
- rtl_set_rx_mode. AcceptXXXs have to be enabled after enabling tx/rx.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- replace rtl8168g-1.fw with rtl8168g-2.fw which support new method.
- fix PHY power down is useless.
- disable rx early which causes the rx abnormal.
- enable auto fifo.
- set 10M IFG to default value.
- fix the conflict between jumbo frame and flow control.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove useless action PHY_READ_EFUSE, PHY_READ_MAC_BYTE, PHY_WRITE_MAC_BYTE,
PHY_WRITE_ERI_WORD. And define the new action PHY_MDIO_CHG.
PHY_MDIO_CHG is used to modify the mdio operation. By the way, the
firmware could support setting mac ocp.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the new settings and correct the wrong settings.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace the current settings with rtl_writephy and rtl_readphy.
For the hardware, the settings are same with previous ones. This
make the setting method like the previous chips.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some codes are belong to binary codes and should be removed.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It would cause no link after suspending or shutdowning when the
nic changes the speed to 10M and connects to a link partner which
forces the speed to 100M.
Check the link partner ability to determine which speed to set.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Emitting netdev_alloc_skb and netdev_alloc_skb_ip_align OOM
messages is unnecessary as there is already a dump_stack
after allocation failures.
Other trivial changes around these removals:
Convert a few comparisons of pointer to 0 to !pointer.
Change flow to remove unnecessary label.
Remove now unused variable.
Hoist assignment from if.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some hardware start settings implicitely assume an usual 1500 bytes mtu
that can't be guaranteed because changes of mtu may be requested both
before and after the hardware is started.
Reported-by: Tomi Orava <tomimo@ncircle.nullnet.fi>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Here is the big driver core merge for 3.9-rc1
There are two major series here, both of which touch lots of drivers all
over the kernel, and will cause you some merge conflicts:
- add a new function called devm_ioremap_resource() to properly be
able to check return values.
- remove CONFIG_EXPERIMENTAL
If you need me to provide a merged tree to handle these resolutions,
please let me know.
Other than those patches, there's not much here, some minor fixes and
updates.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iEYEABECAAYFAlEmV0cACgkQMUfUDdst+yncCQCfbmnQZju7kzWXk6PjdFuKspT9
weAAoMCzcAtEzzc4LXuUxxG/sXBVBCjW
=yWAQ
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core patches from Greg Kroah-Hartman:
"Here is the big driver core merge for 3.9-rc1
There are two major series here, both of which touch lots of drivers
all over the kernel, and will cause you some merge conflicts:
- add a new function called devm_ioremap_resource() to properly be
able to check return values.
- remove CONFIG_EXPERIMENTAL
Other than those patches, there's not much here, some minor fixes and
updates"
Fix up trivial conflicts
* tag 'driver-core-3.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (221 commits)
base: memory: fix soft/hard_offline_page permissions
drivercore: Fix ordering between deferred_probe and exiting initcalls
backlight: fix class_find_device() arguments
TTY: mark tty_get_device call with the proper const values
driver-core: constify data for class_find_device()
firmware: Ignore abort check when no user-helper is used
firmware: Reduce ifdef CONFIG_FW_LOADER_USER_HELPER
firmware: Make user-mode helper optional
firmware: Refactoring for splitting user-mode helper code
Driver core: treat unregistered bus_types as having no devices
watchdog: Convert to devm_ioremap_resource()
thermal: Convert to devm_ioremap_resource()
spi: Convert to devm_ioremap_resource()
power: Convert to devm_ioremap_resource()
mtd: Convert to devm_ioremap_resource()
mmc: Convert to devm_ioremap_resource()
mfd: Convert to devm_ioremap_resource()
media: Convert to devm_ioremap_resource()
iommu: Convert to devm_ioremap_resource()
drm: Convert to devm_ioremap_resource()
...
Conflicts:
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
The bnx2x gso_type setting bug fix in 'net' conflicted with
changes in 'net-next' that broke the gso_* setting logic
out into a seperate function, which also fixes the bug in
question. Thus, use the 'net-next' version.
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit d64ec84151.
Jörg Otte reported his 8168evl to increase boot time link detection
from 1.6 to 10 s.
Hayes suggests reverting it for the time being.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Hayes Wang <hayeswang@realtek.com>
Cc: Jörg Otte <jrg.otte@gmail.com>
This reverts commit e0c0755779.
Jörg Otte reported his 8168evl to fail boot time link detection.
Hayes suggests reverting it for the time being.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Hayes Wang <hayeswang@realtek.com>
Cc: Jörg Otte <jrg.otte@gmail.com>
Bring in the 'net' tree so that we can get some ipv4/ipv6 bug
fixes that some net-next work will build upon.
Signed-off-by: David S. Miller <davem@davemloft.net>
Control of receive descriptor must not be returned to ethernet chipset
before vlan tag processing is done.
VLAN tag receive word is now reset both in normal and error path.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Spotted-by: Timo Teras <timo.teras@iki.fi>
Cc: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This was introduced in commit 6dccd16 "r8169: merge with version
6.001.00 of Realtek's r8169 driver". I did not find the version
6.001.00 online, but in 6.002.00 or any later r8169 from Realtek
this hunk is no longer present.
Also commit 05af214 "r8169: fix Ethernet Hangup for RTL8110SC
rev d" claims to have fixed this issue otherwise.
The magic compare mask of 0xfffe000 is dubious as it masks
parts of the Reserved part, and parts of the VLAN tag. But this
does not make much sense as the VLAN tag parts are perfectly
valid there. In matter of fact this seems to be triggered with
any VLAN tagged packet as RxVlanTag bit is matched. I would
suspect 0xfffe0000 was intended to test reserved part only.
Finally, this hunk is evil as it can cause more packets to be
handled than what was NAPI quota causing net/core/dev.c:
net_rx_action(): WARN_ON_ONCE(work > weight) to trigger, and
mess up the NAPI state causing device to hang.
As result, any system using VLANs and having high receive
traffic (so that NAPI poll budget limits rtl_rx) would result
in device hang.
Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After commit 6f0333b ("r8169: use 50% less ram for RX ring") the rx
ring buffers are always copied making dirty_rx useless.
Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The CONFIG_EXPERIMENTAL config item has not carried much meaning for a
while now and is almost always enabled by default. As agreed during the
Linux kernel summit, remove it from any "depends on" lines in Kconfigs.
CC: "David S. Miller" <davem@davemloft.net>
CC: Paul Gortmaker <paul.gortmaker@windriver.com>
CC: Cesar Eduardo Barros <cesarb@cesarb.net>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: David S. Miller <davem@davemloft.net>
perm_addr is initialized correctly in register_netdevice() so to init it in
drivers is no longer needed.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit: cb64edb6b8 upstream
Above commit may introduce a race between cp_interrupt and dev_close
/ change MTU / dev_open up state. Changes cp_interrupt to tolerate
this. Change spin_locking in cp_interrupt to avoid possible
but unobserved race.
Reported-by: "Francois Romieu" <romieu@fr.zoreil.com>
Tested on virtual hardware, Tx MTU size up to 4096, max tx payload
was ping -s 4068 for MTU of 4096. No real hardware, need test
assist.
Signed-off-by: "John Greene" <jogreene@redhat.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: "David Woodhouse" <David.Woodhouse@intel.com>
Tested-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking changes from David Miller:
1) Allow to dump, monitor, and change the bridge multicast database
using netlink. From Cong Wang.
2) RFC 5961 TCP blind data injection attack mitigation, from Eric
Dumazet.
3) Networking user namespace support from Eric W. Biederman.
4) tuntap/virtio-net multiqueue support by Jason Wang.
5) Support for checksum offload of encapsulated packets (basically,
tunneled traffic can still be checksummed by HW). From Joseph
Gasparakis.
6) Allow BPF filter access to VLAN tags, from Eric Dumazet and
Daniel Borkmann.
7) Bridge port parameters over netlink and BPDU blocking support
from Stephen Hemminger.
8) Improve data access patterns during inet socket demux by rearranging
socket layout, from Eric Dumazet.
9) TIPC protocol updates and cleanups from Ying Xue, Paul Gortmaker, and
Jon Maloy.
10) Update TCP socket hash sizing to be more in line with current day
realities. The existing heurstics were choosen a decade ago.
From Eric Dumazet.
11) Fix races, queue bloat, and excessive wakeups in ATM and
associated drivers, from Krzysztof Mazur and David Woodhouse.
12) Support DOVE (Distributed Overlay Virtual Ethernet) extensions
in VXLAN driver, from David Stevens.
13) Add "oops_only" mode to netconsole, from Amerigo Wang.
14) Support set and query of VEB/VEPA bridge mode via PF_BRIDGE, also
allow DCB netlink to work on namespaces other than the initial
namespace. From John Fastabend.
15) Support PTP in the Tigon3 driver, from Matt Carlson.
16) tun/vhost zero copy fixes and improvements, plus turn it on
by default, from Michael S. Tsirkin.
17) Support per-association statistics in SCTP, from Michele
Baldessari.
And many, many, driver updates, cleanups, and improvements. Too
numerous to mention individually.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1722 commits)
net/mlx4_en: Add support for destination MAC in steering rules
net/mlx4_en: Use generic etherdevice.h functions.
net: ethtool: Add destination MAC address to flow steering API
bridge: add support of adding and deleting mdb entries
bridge: notify mdb changes via netlink
ndisc: Unexport ndisc_{build,send}_skb().
uapi: add missing netconf.h to export list
pkt_sched: avoid requeues if possible
solos-pci: fix double-free of TX skb in DMA mode
bnx2: Fix accidental reversions.
bna: Driver Version Updated to 3.1.2.1
bna: Firmware update
bna: Add RX State
bna: Rx Page Based Allocation
bna: TX Intr Coalescing Fix
bna: Tx and Rx Optimizations
bna: Code Cleanup and Enhancements
ath9k: check pdata variable before dereferencing it
ath5k: RX timestamp is reported at end of frame
ath9k_htc: RX timestamp is reported at end of frame
...
GigaMAC registers have been reported left unitialized in several
situations:
- after cold boot from power-off state
- after S3 resume
Tweaking rtl_hw_phy_config takes care of both.
This patch removes an excess entry (",") at the end of the exgmac_reg
array as well.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Wang YanQing <udknight@gmail.com>
Cc: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is a series of patches that remove the dev* attributes for all
networking drivers, with the exception of wireless drivers, those are in
a different branch.
Use of __devinit, __devexit_p, __devinitdata, __devinitconst, and
__devexit are no longer needed since CONFIG_HOTPLUG is being removed as
an option.
Note, there are some devinit compiler section mismatch warnings due to
this series, but they are fixed up when merged with my driver-next
branch, which fixes the PCI device id warnings, and removes the modpost
detection, as it's no longer needed.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iEYEABECAAYFAlC9COEACgkQMUfUDdst+ykOEgCeKZsrtKrbjMFM5jEROqnk33FI
SwEAoMSZWFJ7M1/27FGRkylmmypWXi/p
=HH3A
-----END PGP SIGNATURE-----
Merge tag 'dev_removal' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/net-next
Networking: Remove __dev* markings from the networking drivers
This is a series of patches that remove the dev* attributes for all
networking drivers, with the exception of wireless drivers, those are in
a different branch.
Use of __devinit, __devexit_p, __devinitdata, __devinitconst, and
__devexit are no longer needed since CONFIG_HOTPLUG is being removed as
an option.
Note, there are some devinit compiler section mismatch warnings due to
this series, but they are fixed up when merged with my driver-next
branch, which fixes the PCI device id warnings, and removes the modpost
detection, as it's no longer needed.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
CONFIG_HOTPLUG is going away as an option. As result the __dev*
markings will be going away.
Remove use of __devinit, __devexit_p, __devinitdata, __devinitconst,
and __devexit.
Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CONFIG_HOTPLUG is going away as an option. As result the __dev*
markings will be going away.
Remove use of __devinit, __devexit_p, __devinitdata, __devinitconst,
and __devexit.
Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Cc: Realtek linux nic maintainers <nic_swsd@realtek.com>
Cc: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The 8139cp driver has a change_mtu function that has not been
enabled since the dawn of the git repository. However, the
generic eth_change_mtu is not used in its place, so that
invalid MTU values can be set on the interface.
Original patch salvages the broken code for the single case of
setting the MTU while the interface is down, which is safe
and also includes the range check. Now enhanced to support up
or down interface.
v2: fix case where rxbufsz isn't changed in the up state case
Original patch from
http://lkml.indiana.edu/hypermail/linux/kernel/1202.2/00770.html
Testing: has been test on virtual 8139cp setup without issue,
have no access real hardware 8139cp, need testing help.
Signed-off-by: "John Greene" <jogreene@redhat.com>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
cp_open
[...]
rc = cp_alloc_rings(cp);
if (rc)
return rc;
cp_alloc_rings
[...]
mem = dma_alloc_coherent(&cp->pdev->dev, CP_RING_BYTES,
&cp->ring_dma, GFP_KERNEL);
- cp_alloc_rings never frees the coherent mapping it allocates
- neither do cp_open when cp_alloc_rings fails
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recovery doesn't work too well if we leave interrupts disabled...
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This adds support for byte queue limits on RTL8139C+
Tested on real hardware.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Acked-By: Dave Täht <dave.taht@bufferbloat.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This fixes (for me) a regression introduced by commit b01af457 ("8139cp:
set ring address before enabling receiver"). That commit configured the
descriptor ring addresses earlier in the initialisation sequence, in
order to avoid the possibility of triggering stray DMA before the
correct address had been set up.
Unfortunately, it seems that the hardware will scribble garbage into the
TxRingAddr registers when we enable "plus mode" Tx in the CpCmd
register. Observed on a Traverse Geos router board.
To deal with this, while not reintroducing the problem which led to the
original commit, we augment cp_start_hw() to write to the CpCmd register
*first*, then set the descriptor ring addresses, and then finally to
enable Rx and Tx in the original 8139 Cmd register. The datasheet
actually indicates that we should enable Tx/Rx in the Cmd register
*before* configuring the descriptor addresses, but that would appear to
re-introduce the problem that the offending commit b01af457 was trying
to solve. And this variant appears to work fine on real hardware.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: stable@kernel.org [3.5+]
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit b26623dab7.
This reverts the revert, in net-next we'll try another scheme
to fix this bug using patches from David Woodhouse.
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/wireless/iwlwifi/pcie/tx.c
Minor iwlwifi conflict in TX queue disabling between 'net', which
removed a bogus warning, and 'net-next' which added some status
register poking code.
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch reverts b01af4579e.
The original patch was tested with emulated hardware. Real
hardware chokes.
Fixes https://bugzilla.kernel.org/show_bug.cgi?id=47041
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Acked-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since eab6d18d (vlan: Don't check for vlan group before
vlan_tx_tag_present.) we don't check tp->vlgrp and thus
tp is not needed in this function.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Conflicts:
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
Minor conflict between the BCM_CNIC define removal in net-next
and a bug fix added to net. Based upon a conflict resolution
patch posted by Stephen Rothwell.
Signed-off-by: David S. Miller <davem@davemloft.net>
The following chips need to enable internal settings to let ASPM
and clock request work.
RTL8111E-VL, RTL8111F, RTL8411, RTL8111G
RTL8105, RTL8402, RTL8106
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Building atp.o triggers this GCC warning:
drivers/net/ethernet/realtek/atp.c: In function ‘set_rx_mode’:
drivers/net/ethernet/realtek/atp.c:871:26: warning: ‘mc_filter[0]’ may be used uninitialized in this function [-Wuninitialized]
GCC is correct. In promiscuous mode 'mc_filter' will be used
uninitialized in set_rx_mode_8012(), which is apparently inlined into
set_rx_mode().
But it turns out set_rx_mode_8012() will never be called, since
net_local.chip_type will always be RTL8002. So we can just remove
set_rx_mode_8012() and do some related cleanups.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
RTL_GIGA_MAC_VER_35 includes no multicast hardware filter.
Signed-off-by: Nathan Walp <faceprint@faceprint.com>
Suggested-by: Hayes Wang <hayeswang@realtek.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This regression was spotted between Debian squeeze and Debian wheezy
kernels (respectively based on 2.6.32 and 3.2). More info about
Wake-on-LAN issues with Realtek's 816x chipsets can be found in the
following thread: http://marc.info/?t=132079219400004
Probable regression from d4ed95d796e5126bba51466dc07e287cebc8bd19;
more chipsets are likely affected.
Tested on top of a 3.2.23 kernel.
Reported-by: Florent Fourcot <florent.fourcot@enst-bretagne.fr>
Tested-by: Florent Fourcot <florent.fourcot@enst-bretagne.fr>
Hinted-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Cyril Brulebois <kibi@debian.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
After d58d46b5 (r8169: jumbo fixes.) max frame len is stored in
rtl_chip_infos[].jumbo_max for each chip and SafeMtu should be gone.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
use the module_pci_driver macro to make the code simpler
by eliminating the module_init and module_exit calls
Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Enable ALDPS function to save power when link down. Note that the
feature should be set after the other PHY settings. And the firmware
is necessary. Don't enable it without loading the firmware.
None of the firmware-free chipsets support ALDPS. Neither do the
RTL8168d/8111d.
For 8136 series, make sure the ALDPS is disabled before loading the
firmware. For 8168 series, the ALDPS would be disabled automatically
when loading firmware. You must not disable it directly.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current GRO can hold packets in gro_list for almost unlimited
time, in case napi->poll() handler consumes its budget over and over.
In this case, napi_complete()/napi_gro_flush() are not called.
Another problem is that gro_list is flushed in non friendly way :
We scan the list and complete packets in the reverse order.
(youngest packets first, oldest packets last)
This defeats priorities that sender could have cooked.
Since GRO currently only store TCP packets, we dont really notice the
bug because of retransmits, but this behavior can add unexpected
latencies, particularly on mice flows clamped by elephant flows.
This patch makes sure no packet can stay more than 1 ms in queue, and
only in stress situations.
It also complete packets in the right order to minimize latencies.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Jesse Gross <jesse@nicira.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking changes from David Miller:
1) GRE now works over ipv6, from Dmitry Kozlov.
2) Make SCTP more network namespace aware, from Eric Biederman.
3) TEAM driver now works with non-ethernet devices, from Jiri Pirko.
4) Make openvswitch network namespace aware, from Pravin B Shelar.
5) IPV6 NAT implementation, from Patrick McHardy.
6) Server side support for TCP Fast Open, from Jerry Chu and others.
7) Packet BPF filter supports MOD and XOR, from Eric Dumazet and Daniel
Borkmann.
8) Increate the loopback default MTU to 64K, from Eric Dumazet.
9) Use a per-task rather than per-socket page fragment allocator for
outgoing networking traffic. This benefits processes that have very
many mostly idle sockets, which is quite common.
From Eric Dumazet.
10) Use up to 32K for page fragment allocations, with fallbacks to
smaller sizes when higher order page allocations fail. Benefits are
a) less segments for driver to process b) less calls to page
allocator c) less waste of space.
From Eric Dumazet.
11) Allow GRO to be used on GRE tunnels, from Eric Dumazet.
12) VXLAN device driver, one way to handle VLAN issues such as the
limitation of 4096 VLAN IDs yet still have some level of isolation.
From Stephen Hemminger.
13) As usual there is a large boatload of driver changes, with the scale
perhaps tilted towards the wireless side this time around.
Fix up various fairly trivial conflicts, mostly caused by the user
namespace changes.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1012 commits)
hyperv: Add buffer for extended info after the RNDIS response message.
hyperv: Report actual status in receive completion packet
hyperv: Remove extra allocated space for recv_pkt_list elements
hyperv: Fix page buffer handling in rndis_filter_send_request()
hyperv: Fix the missing return value in rndis_filter_set_packet_filter()
hyperv: Fix the max_xfer_size in RNDIS initialization
vxlan: put UDP socket in correct namespace
vxlan: Depend on CONFIG_INET
sfc: Fix the reported priorities of different filter types
sfc: Remove EFX_FILTER_FLAG_RX_OVERRIDE_IP
sfc: Fix loopback self-test with separate_tx_channels=1
sfc: Fix MCDI structure field lookup
sfc: Add parentheses around use of bitfield macro arguments
sfc: Fix null function pointer in efx_sriov_channel_type
vxlan: virtual extensible lan
igmp: export symbol ip_mc_leave_group
netlink: add attributes to fdb interface
tg3: unconditionally select HWMON support when tg3 is enabled.
Revert "net: ti cpsw ethernet: allow reading phy interface mode from DT"
gre: fix sparse warning
...
The r8169 driver currently limits the DMA burst for TX to 1024 bytes. I have
a box where this prevents the interface from using the gigabit line to its full
potential. This patch solves the problem by setting TX_DMA_BURST to unlimited.
The box has an ASRock B75M motherboard with on-board RTL8168evl/8111evl
(XID 0c900880). TSO is enabled.
I used netperf (TCP_STREAM test) to measure the dependency of TX throughput
on MTU. I did it for three different values of TX_DMA_BURST ('5'=512, '6'=1024,
'7'=unlimited). This chart shows the results:
http://michich.fedorapeople.org/r8169/r8169-effects-of-TX_DMA_BURST.png
Interesting points:
- With the current DMA burst limit (1024):
- at the default MTU=1500 I get only 842 Mbit/s.
- when going from small MTU, the performance rises monotonically with
increasing MTU only up to a peak at MTU=1076 (908 MBit/s). Then there's
a sudden drop to 762 MBit/s from which the throughput rises monotonically
again with further MTU increases.
- With a smaller DMA burst limit (512):
- there's a similar peak at MTU=1076 and another one at MTU=564.
- With unlimited DMA burst:
- at the default MTU=1500 I get nice 940 Mbit/s.
- the throughput rises monotonically with increasing MTU with no strange
peaks.
Notice that the peaks occur at MTU sizes that are multiples of the DMA burst
limit plus 52. Why 52? Because:
20 (IP header) + 20 (TCP header) + 12 (TCP options) = 52
The Realtek-provided r8168 driver (v8.032.00) uses unlimited TX DMA burst too,
except for CFG_METHOD_1 where the TX DMA burst is set to 512 bytes.
CFG_METHOD_1 appears to be the oldest MAC version of "RTL8168B/8111B",
i.e. RTL_GIGA_MAC_VER_11 in r8169. Not sure if this MAC version really needs
the smaller burst limit, or if any other versions have similar requirements.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
