Andrew Lunn [Sat, 2 Jan 2010 10:30:50 +0000 (11:30 +0100)]
Staging: batman-adv: Use printk(%pM) for MAC addresses
printk() since kernel version 2.6.29 has supported printing MAC
addresses directly, as an extension to the %p processing. This patch
makes use of this for printk() and bat_dbg(). This will remove the
overhead of using addr_to_string() which is normally never actually
output.
Fixed a typo found by Gus Wirth.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Lunn [Sat, 2 Jan 2010 10:30:49 +0000 (11:30 +0100)]
Staging: batman-adv: Allow the MAC address to be set
Some embedded devices have very limited sources of entropy for the
random number generator. It has been observed that the random MAC
address on the interface bat0 is not always random. When testing with
a collection of identical hardware, sometimes the bat0 device the same
MAC address on multiple devices, causing mayhem. This patch allows the
MAC address to be set by the user.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Staging: batman-adv: receive packets directly using skbs
This patch removes the (ugly and racy) packet receiving thread and the
kernel socket usage. Instead, packets are received directly by registering
the ethernet type and handling skbs instead of self-allocated buffers.
Some consequences and comments:
* we don't copy the payload data when forwarding/sending/receiving data
anymore. This should boost performance.
* packets from/to different interfaces can be (theoretically) processed
simultaneously. Only the big originator hash lock might be in the way.
* no more polling or sleeping/wakeup/scheduling issues when receiving
packets
* this might introduce new race conditions.
* aggregation and vis code still use packet buffers and are not (yet)
converted.
* all spinlocks were converted to irqsave/restore versions to solve
some lifelock issues when preempted. This might be overkill, some
of these locks might be reverted later.
* skb copies are only done if neccesary to avoid overhead
performance differences:
* we made some "benchmarks" with intel laptops.
* bandwidth on Gigabit Ethernet increased from ~500 MBit/s to ~920 MBit/s
* ping latency decresed from ~2ms to ~0.2 ms
I did some tests on my 9 node qemu environment and could confirm that
usual sending/receiving, forwarding, vis, batctl ping etc works.
Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Acked-by: Sven Eckelmann <sven.eckelmann@gmx.de> Acked-by: Marek Lindner <lindner_marek@yahoo.de> Acked-by: Linus Lüssing <linus.luessing@web.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
there are some kmallocs left which are not checked whether they succeeds or
not, which might lead to corrupted data structures if the system memory is
full. This patch should clean up the remaining unchecked kmalloc()s.
Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Marek Lindner [Sat, 2 Jan 2010 10:30:46 +0000 (11:30 +0100)]
Staging: batman-adv: splitting /proc vis file into vis_server and vis_data
The /proc vis file was used to enable/disable the vis server and to output
the vis data at the same time. This behaviour was confusing and lacked a
proper method to display the current vis server status.
This patch seperates the 2 functionalities:
* use vis_server to enable/disable the vis server and to retrieve its status
* use vis_data to retrieve the vis raw data (if the server is enabled)
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
batman-adv used to export the vis data using different formats that were
switchable via /proc/net/batman-adv/vis_format. The various formats moved
to user space and rendered this configuration switch useless.
Signed-off-by: Linus Luessing <linus.luessing@web.de> Signed-off-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Lunn [Sat, 2 Jan 2010 10:30:44 +0000 (11:30 +0100)]
Staging: batman-adv: moving vis output formats out of the kernel
The batman-adv kernel module is able to output visualization data using the
dot draw or JSON format. This patch transforms the output into a generic
format (called vis raw). User space tool may convert the raw data to support
a variety of formats without the need of modifying the kernel module.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Linus Lüssing [Sat, 2 Jan 2010 10:30:42 +0000 (11:30 +0100)]
Staging: batman-adv: Fixes rounding issues in vis.c
This patches fixes two rounding issues in vis.c for sending and
purging vis packets. Before, the timers and timeouts always got
rounded down to seconds, though we want a precision in
milliseconds.
This also fixes a kernel panic that occures when lowering the
timer for sending vis packets (vis_interval) to less than 1000ms
in main.c manually.
Signed-off-by: Linus Lüssing <linus.luessing@web.de> Acked-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Sven Eckelmann [Sat, 2 Jan 2010 10:30:41 +0000 (11:30 +0100)]
Staging: batman-adv: Use forw_bcast_list_lock always with disabled interrupts
forw_bcast_list_lock is spin_locked in both process and softirq context.
SoftIRQ calls the spinlock with disabled IRQ and normal process context
with enabled IRQs.
When process context is inside an spin_locked area protected by
forw_bcast_list_lock and gets interrupted by an IRQ, it could happen
that something tries to lock forw_bcast_list_lock again in SoftIRQ
context. It cannot proceed further since the lock is already taken
somewhere else, but no reschedule will happen inside the SoftIRQ
context. This leads to an complete kernel hang without any chance of
resurrection.
All functions called in process context must disable IRQs when they try
to get get that lock to to prevent any reschedule due to IRQs.
Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Acked-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Jarod Wilson [Mon, 4 Jan 2010 23:02:27 +0000 (18:02 -0500)]
staging: add Broadcom Crystal HD driver
This patch supersedes the earlier ones sent by Manu Abraham to add
the Broadcom Crystal HD driver to the staging tree, per discussion
with him about it. I've been working with Broadcom's Naren Sankar
on this driver for a number of months, and had already talked Naren
about submitting this on Broadcom's behalf, didn't expect anyone
else to jump on submitting it as quickly as Manu did. ;)
This version is a one-shot deal, incorporating the original driver,
Manu's coding style clean-ups, udev device creation support from
Edgar 'gimli' Hucek, and a number of other small tweaks from myself
and Scott Davilla, the other individual who has been working closely
on this code with Naren and I.
I've tested this iteration of the code lightly on a mini pci-e board
in a ThinkPad T61p running x86_64 Fedora 12, with the expected results,
and will test further on other systems with other variants of the card
(I have three varieties of this device currently in hand). Scott has
also tested on assorted primarily i686 varieties of Ubuntu, and Naren
has tested with both Fedora and Ubuntu, iirc.
Note: only the 70012 is currently supported by this driver, 70015
support will follow later. Also note that Blu-Ray support isn't
enabled (at the firmware level), due to misc fun related to the
BD encryption scheme, DRM, etc. :\
I *do* have a git tree containing the driver, lib, gst plugin and
firmware that I'm working from at the moment[*], as there are inter-
dependencies between the driver and lib, and the driver can be used
with kernels going a ways back (I've only tested back to 2.6.18 as
it exists in Red Hat Enterprise Linux 5). I'm exporting from there,
into a linux-next tree, then generating patches from there. The goal
is to feed everything upstream as quickly as possible, but there are
users who want this code for earlier kernels too...
The firmware will be submitted for inclusion in dwmw2's linux-firmware
tree once there is a suitable redistribution-no-modification type of
license on it (I believe Naren is working with Broadcom legal to get
that in place).
include: lib doesn't build w/o the removed stdint include
So add it back...
Signed-off-by: Jarod Wilson <jarod@redhat.com>
commit c181070a330530b792d2b80e3ec6ab12a5a57394
Author: Scott Davilla <davilla@4pi.com>
Date: Mon Jan 4 13:38:37 2010 -0500
include: don't define VOID if its already defined
Signed-off-by: Scott Davilla <davilla@4pi.com> Signed-off-by: Jarod Wilson <jarod@redhat.com>
commit 33d8a2b691e81212e398f53770578d79650bf0bc
Author: Jarod Wilson <jarod@redhat.com>
Date: Mon Jan 4 13:12:10 2010 -0500
driver: create crystalhd device using udev
Based on:
http://sourceforge.net/apps/trac/archvdr/browser/trunk/archvdr/crystalhd/use_udev.patch
Signed-off-by: Edgar ( gimli ) Hucek <ebsi4711 at gmail dot com>
Formatting tweaks, error-handling path fixups and any bugs added by Jarod.
Signed-off-by: Jarod Wilson <jarod@redhat.com>
commit c44c64dea5537814796fcbe2d9db0209383c78b9
Author: Manu Abraham <abraham.manu@gmail.com>
Date: Mon Jan 4 10:32:47 2010 -0500
crystalhd: coding style cleanups
Signed-off-by: Manu Abraham <abraham.manu@gmail.com> Signed-off-by: Jarod Wilson <jarod@redhat.com>
commit cffa6da7467ff697a656d1dfff54bb0513a053dc
Author: Jarod Wilson <jarod@redhat.com>
Date: Mon Jan 4 10:17:27 2010 -0500
crystalhd: run dos2unix over everything, this is linux source...
Signed-off-by: Jarod Wilson <jarod@redhat.com>
commit 7fa38a282db7af5a5746055f7c6cef8a9b8ee138
Author: Jarod Wilson <jarod@redhat.com>
Date: Mon Jan 4 10:02:33 2010 -0500
crystalhd: initial import of released Broadcom code
Unfortunately, we're unable to publicly publish all the history that got
us from the initial internal code to what was released here, but such is
life, we can just be happy we've got this open-sourced now. :)
Signed-off-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Naren Sankar <nsankar@broadcom.com> Signed-off-by: Scott Davilla <davilla@4pi.com> Signed-off-by: Manu Abraham <abraham.manu@gmail.com> Signed-off-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Simon Horman [Mon, 21 Dec 2009 04:09:53 +0000 (15:09 +1100)]
Staging: wlan-ng: initialise mibitem
$ gcc --version
gcc (Debian 4.4.2-5) 4.4.2
Copyright (C) 2009 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ make
...
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_dorequest’:
drivers/staging/wlan-ng/p80211wext.c:139: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:139: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwretry’:
drivers/staging/wlan-ng/p80211wext.c:967: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:967: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:967: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwretry’:
drivers/staging/wlan-ng/p80211wext.c:1057: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:1057: warning: ‘mibitem.status’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwtxpow’:
drivers/staging/wlan-ng/p80211wext.c:1149: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:1149: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:1149: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwtxpow’:
drivers/staging/wlan-ng/p80211wext.c:1123: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:1123: warning: ‘mibitem.status’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwfrag’:
drivers/staging/wlan-ng/p80211wext.c:891: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:891: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:891: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwfrag’:
drivers/staging/wlan-ng/p80211wext.c:933: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:933: warning: ‘mibitem.status’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwrts’:
drivers/staging/wlan-ng/p80211wext.c:826: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:826: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:826: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwrts’:
drivers/staging/wlan-ng/p80211wext.c:866: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:866: warning: ‘mibitem.status’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwrate’:
drivers/staging/wlan-ng/p80211wext.c:775: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:775: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:775: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_giwfreq’:
drivers/staging/wlan-ng/p80211wext.c:273: warning: ‘mibitem.len’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:273: warning: ‘mibitem.status’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:273: warning: ‘mibitem.data’ is used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwfreq’:
drivers/staging/wlan-ng/p80211wext.c:320: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c: In function ‘p80211wext_siwmode’:
drivers/staging/wlan-ng/p80211wext.c:401: warning: ‘mibitem.len’ may be used uninitialized in this function
drivers/staging/wlan-ng/p80211wext.c:401: warning: ‘mibitem.status’ may be used uninitialized in this function
...
Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Simon Horman [Mon, 21 Dec 2009 04:07:50 +0000 (15:07 +1100)]
Staging: rtl8192su: Use z format qualifier for output of sizeof()
$ gcc --version
gcc (Debian 4.4.2-5) 4.4.2
Copyright (C) 2009 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ make
...
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c: In function ‘ieee80211_rx_ADDBAReq’:
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c:342: warning: format ‘%ld’ expects type ‘long int’, but argument 3 has type ‘unsigned int’
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c: In function ‘ieee80211_rx_ADDBARsp’:
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c:443: warning: format ‘%ld’ expects type ‘long int’, but argument 3 has type ‘unsigned int’
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c: In function ‘ieee80211_rx_DELBA’:
drivers/staging/rtl8192su/ieee80211/rtl819x_BAProc.c:573: warning: format ‘%ld’ expects type ‘long int’, but argument 3 has type ‘unsigned int’
...
Also some style fixes for these lines:
+ Fix excessively long lines
+ Remove leading space before struct
+ Remove unnecessary parentheses
Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Justin Madru [Fri, 18 Dec 2009 23:52:53 +0000 (15:52 -0800)]
Staging: s5k3e2fx.c: simplify complexity by factoring
the code was looping, setting s_move[i] to the following calculations
if (actual_step>= 0)
s_move[i] = ((((i + 1) * gain + 0x200) - (i * gain + 0x200)) / 0x400);
else
s_move[i] = ((((i + 1) * gain - 0x200) - (i * gain - 0x200)) / 0x400);
but, this code reduces to the expression
s_move[i] = gain>> 10;
The reason for the complexity was to generate a step function with
integer division and rounding to land on specific values. But these calculations
can be simplified to the following code:
gain = ((actual_step<< 10) / 5)>> 10;
for (i = 0; i<= 4; i++)
s_move[i] = gain;
Signed-off-by: Justin Madru<jdm64@gawab.com> Reviewed-by: Ray Lee<ray-lk@madrabbit.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Bernie Thompson [Tue, 24 Nov 2009 23:52:21 +0000 (15:52 -0800)]
Staging: udlfb: add dynamic modeset support
Add dynamic modeset support
udlfb uses EDID to find the monitor’s preferred mode
udlfb no longer has fixed mode tables – it’s able to set any mode
dynamically, from the standard VESA timing characteristics of the monitor.
Draws from probe and setmode code of both displaylink-mod 0.3 branch of
Roberto De Ioris, and Jaya Kumar's displaylinkfb.
Lays foundation for defio support and making backbuffer optional.
With additional changes to minimize diffs and clean for checkpatch.pl style.
Does not yet include new ioctls or refcount/mutex code from displaylink-mod.
Tested to work with existing xf-video-displaylink X server unmodified.
Radu Voicilas [Sat, 5 Dec 2009 04:42:53 +0000 (20:42 -0800)]
Staging: rtl8192e: fix some codingstyle issues
dot11d.h needed a good refactoring - I've dropped some of the //
comments or transformed them to match the kernel documentation.
r8180_93cx6.h - fixed a little bit the copyright section.
Signed-off-by: Radu Voicilas <rvoicilas@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Lunn [Sat, 12 Dec 2009 22:39:43 +0000 (23:39 +0100)]
Staging: batman-adv: stop persistent warnings if a device is deactivated.
Without this change we spam the kernel log on every packet received on
any other interface when an interface has been added, but is not yet
active, ie UP.
Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Lunn [Sat, 12 Dec 2009 22:39:42 +0000 (23:39 +0100)]
Staging: batman-adv: Refactor routing.c
Break up a lot of the big functions up into many smaller ones. This
helps with readability and there is now a lot less code squashed
against the right hand margin.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Acked-by: Marek Lindner <lindner_marek@yahoo.de> Acked-by: Simon Wunderlich <simon.wunderlich@s2003.tu-chemnitz.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
batman-adv used its own logging infrastructure. Replace this with
standard kernel logging, printk(), with compile time and runtime
options to enable/disable different debug levels.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Peter Huewe [Tue, 15 Dec 2009 04:04:22 +0000 (05:04 +0100)]
Staging: asus_oled: Add defines for asus vendor_id and device_id to be consistent with hid-ids.h
This almost trivial patch replaces the hardcoded values for the vendor and
device ids with defines, as they are used in drivers/hid/hid-ids.h
For me this seems to be more consistent, however as drivers/hid/hid-ids.h
is not within the default include directory I had to redefine the defines here
(maybe move the hid-ids.h to include/linux ?)
Signed-off-by: Peter Huewe <peterhuewe@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Martyn Welch [Tue, 15 Dec 2009 08:42:56 +0000 (08:42 +0000)]
Staging: vme: Allow override of geographical address on ca91c142
Geographical addressing only works if the VME backplane supports it. There
are a large number of old backplanes which do not support geographical
addressing. These boards will generally report a slot ID of zero - which is
an invalid ID in the slot numbering scheme.
Allow the geographical address to be over-ridden on the ca91c142 so that a
slot ID can be provided manually in these circumstances.
Martyn Welch [Tue, 15 Dec 2009 08:42:49 +0000 (08:42 +0000)]
Staging: vme: Allow override of geographical address on tsi148
Geographical addressing only works if the VME backplane supports it. There
are a large number of old backplanes which do not support geographical
addressing. These boards will generally report a slot ID of zero - which is
an invalid ID in the slot numbering scheme.
Allow the geographical address to be over-ridden on the tsi148 so that a
slot ID can be provided manually in these circumstances.
Carlos O'Donell [Mon, 22 Feb 2010 23:25:59 +0000 (23:25 +0000)]
parisc: Set PCI CLS early in boot.
Set the PCI CLS early in the boot process to prevent
device failures. In pcibios_set_master use the new
pci_cache_line_size instead of a hard-coded value.
Signed-off-by: Carlos O'Donell <carlos@codesourcery.com> Reviewed-by: Grant Grundler <grundler@google.com> Signed-off-by: Kyle McMartin <kyle@redhat.com>
Michal Simek [Mon, 15 Feb 2010 09:50:42 +0000 (10:50 +0100)]
microblaze: Fix cache loop function for cache range
I create wrong asm code but none test shows that this part of code is wrong.
I am not convinces that were good idea to create asm optimized macros
for caches. The reason is that there is not optimization with previous code
that's why make sense to add old code and do some benchmarking which
functions are faster.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
net: bug fix for vlan + gro issue
tc35815: Remove a wrong netif_wake_queue() call which triggers BUG_ON
cdc_ether: new PID for Ericsson C3607w to the whitelist (resubmit)
IPv6: better document max_addresses parameter
MAINTAINERS: update mv643xx_eth maintenance status
e1000: Fix DMA mapping error handling on RX
iwlwifi: sanity check before counting number of tfds can be free
iwlwifi: error checking for number of tfds in queue
iwlwifi: set HT flags after channel in rxon
Ajit Khaparde [Tue, 16 Feb 2010 20:25:43 +0000 (20:25 +0000)]
net: bug fix for vlan + gro issue
Traffic (tcp) doesnot start on a vlan interface when gro is enabled.
Even the tcp handshake was not taking place.
This is because, the eth_type_trans call before the netif_receive_skb
in napi_gro_finish() resets the skb->dev to napi->dev from the previously
set vlan netdev interface. This causes the ip_route_input to drop the
incoming packet considering it as a packet coming from a martian source.
I could repro this on 2.6.32.7 (stable) and 2.6.33-rc7.
With this fix, the traffic starts and the test runs fine on both vlan
and non-vlan interfaces.
CC: Herbert Xu <herbert@gondor.apana.org.au> CC: Patrick McHardy <kaber@trash.net> Signed-off-by: Ajit Khaparde <ajitk@serverengines.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>