Anna Neal [Tue, 9 Dec 2008 21:23:45 +0000 (13:23 -0800)]
libertas: Create sysfs entry for changing the mesh probe response limit
This patch adds the ability to change the number of probe response retries sent
by the mesh interface.
In dense networks it is recommended to change this value to zero to reduce
traffic congestion.
Signed-off-by: Anna Neal <anna@cozybit.com> Signed-off-by: Andrey Yurovsky <andrey@cozybit.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Wed, 10 Dec 2008 12:51:47 +0000 (14:51 +0200)]
mac80211: Disable requests for new scans in AP mode
AP mode operations are seriously affected if mac80211 runs through a
multi-second scan while the AP is trying to send Beacon frames on the
operation channel. While this could be implemented in a way that does
not cause too many problems, it is not very simple and will require
synchronization with Beacon frame scheduling in the drivers (scan one
channel at a time between Beacon frames). Furthermore, such scanning
takes quite a bit longer time and existing userspace applications
would be likely to timeout while waiting for the results.
For now, just refuse requests for new scans (SIOCSIWSCAN) when in AP
mode. In practice, this moves the rejection from iwl* drivers into
mac80211 to make it apply to every mac80211-based driver.
This issue shows up in associated stations getting disconnected when
something (e.g., Network Manager) requests a scan while the interface
is in AP mode. When doing this continuously (e.g., NM does it every 120
seconds), the network gets close to useless.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com> Acked-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Wed, 10 Dec 2008 12:17:26 +0000 (13:17 +0100)]
ipw2200: fix scanning while associated
This patch fixes sporadic firmware restarts when scanning while associated.
The firmware will quietly cancel a scan (while associated) if the dwell time
for a channel to be scanned is larger than the time it may stay away from the
operating channel (because of DTIM catching). Unfortunately the driver is not
notified about the canceled scan and therefore the scan watchdog timeout will
be hit and the driver causes a firmware restart which results in
disassociation. This mainly happens on passive channels which use a dwell time
of 120 whereas a typical beacon interval is around 100.
The patch changes the dwell time for passive channels to be slightly smaller
than the actual beacon interval to work around the firmware issue. Furthermore
the number of allowed beacon misses is increased from one to three as otherwise
most scans (while associated) won't complete successfully.
However scanning while associated will still fail in corner cases such as a
beacon intervals below 30.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Wed, 10 Dec 2008 05:34:27 +0000 (23:34 -0600)]
rtl8187: Use usb anchor facilities to manage urbs
When SLUB debugging is enabled in the kernel, and the boot command includes
the option "slub_debug=P", rtl8187 encounters a GPF due to a read-after-free
of a urb.
Following the example of changes in p54usb to fix the same problem, the code
has been modified to use the usb_anchor_urb() method. With this change, the
USB core handles the freeing of urb's.
This patch fixes the problem reported in Kernel Bugzilla #12185
(http://bugzilla.kernel.org/show_bug.cgi?id=12185).
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Tested-by: Hin-Tak Leung <htl10@users.sourceforge.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Bob Copeland [Wed, 10 Dec 2008 04:05:38 +0000 (23:05 -0500)]
ath5k: fix endianness of bitwise ops when installing mic
Fix these bugs found by sparse:
ath5k/pcu.c:1102:21: warning: restricted __le32 degrades to integer
ath5k/pcu.c:1102:13: warning: incorrect type in assignment (different base types)
ath5k/pcu.c:1102:13: expected restricted __le32 <noident>
ath5k/pcu.c:1102:13: got unsigned int
ath5k/pcu.c:1104:20: warning: restricted __le32 degrades to integer
ath5k/pcu.c:1104:13: warning: incorrect type in assignment (different base types)
ath5k/pcu.c:1104:13: expected restricted __le32 <noident>
ath5k/pcu.c:1104:13: got unsigned int
Changes-licensed-under: ISC
Reported-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: Bob Copeland <me@bobcopeland.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
David Kilroy [Tue, 9 Dec 2008 21:46:29 +0000 (21:46 +0000)]
orinoco: Fix inappropriate use of IRQ_BAP
This hardware buffer should only be used from an interrupt. The
wireless event generation functions are called from a workqueue, so use
USER_BAP instead.
Signed-off-by: David Kilroy <kilroyd@googlemail.com> Signed-off-by: Pavel Roskin <proski@gnu.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
David Kilroy [Tue, 9 Dec 2008 21:46:28 +0000 (21:46 +0000)]
orinoco: Fix function names used in debug strings
Signed-off-by: David Kilroy <kilroyd@googlemail.com> Signed-off-by: Pavel Roskin <proski@gnu.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Winkler, Tomas [Tue, 9 Dec 2008 19:29:01 +0000 (11:29 -0800)]
iwlwifi: remove qos module parameter
The ability of disabling qos from module params is not required anymore.
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Winkler, Tomas [Tue, 9 Dec 2008 19:29:00 +0000 (11:29 -0800)]
iwlwifi: 3945 remove qos module parameter
The ability of disabling qos from module params is not required
anymore.
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Winkler, Tomas [Tue, 9 Dec 2008 19:28:58 +0000 (11:28 -0800)]
iwlwifi: change email contact information
This patch replaces personal emails with hopefully
always valid Intel Linux Wireless, which will be routed
to a current maintainer
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211: integrate sta_notify_ps cmds into sta_notify
This patch replaces the newly introduced sta_notify_ps function,
which can be used to notify the driver about every power state
transition for all associated stations, by integrating its functionality
back into the original sta_notify callback.
Signed-off-by: Christian Lamparter <chunkeey@web.de> Acked-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
p54usb: rewriting rx/tx routines to make use of usb_anchor's facilities
Alan Stern found several flaws in p54usb's implementation and annotated:
"usb_kill_urb() and similar routines do not expect an URB's completion
routine to deallocate it. This is almost obvious -- if the URB is deallocated
before the completion routine returns then there's no way for usb_kill_urb
to detect when the URB actually is complete."
This patch addresses all known limitations in the old implementation and fixes
khub's "use-after-freed" hang, when SLUB debug's poisoning option is enabled.
Signed-off-by: Christian Lamparter <chunkeey@web.de> Cc: stable@kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k: BH shouldn't be enabled when hardirqs are disabled.
ath_tx_complete_buf uses a BH version of spinlock and so releasing
the lock enables BH which is incorrect when called from sta_notify
callback as MAC80211 disables hardirqs before the driver callback
is called.
As ath_tx_complete_buf is shared between user and softirq context
using normal spinlock may not be appropriate. Though the proper
fix would be to cleanup the context properly in the driver code,
this would be an interim fix to avoid kernel warning.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
There's no driver that actually does fragmentation on the
device, and the callback is buggy (when it returns an error,
mac80211's fragmentation status is changed so reading the
frag threshold from userspace reads the new value despite
the error). Let's just remove it, if we really find some
hardware supporting it we can add it back later.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rami Rosen [Tue, 9 Dec 2008 06:37:28 +0000 (08:37 +0200)]
iwlwifi: fix build error (iwl-rx.c).
This patch adds #ifdef CONFIG_IWLWIFI_DEBUG in iwl_rx_reply_rx()
method in iwlwifi/iwl-rx.c) to avoid build error caused by "iwlwifi:
implement iwl_debug_report_frame for iwlagn".
Signed-off-by: Rami Rosen <ramirose@gmail.com> Acked-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Mon, 8 Dec 2008 23:10:30 +0000 (00:10 +0100)]
mac80211: only create default STA interface if supported
Drivers will support this, obviously, but this forces them to
set it up properly.
(This includes the fix posted as "mac80211: fix ifmodes check" and
tested in wireless-testing by Hin-Tak and others. -- JWL)
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Reported-by: Larry Finger <Larry.Finger@lwfinger.net> Tested-by: Hin-Tak Leung <htl10@users.sourceforge.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k: Incorrect band enum causes invalid antenna configuration.
It's a regression from the patch titled "ath9k: Use mac80211's band
macros and remove enum hal_freq_band". MAC80211 band macros can't be
mapped directly with our hal band enum.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Mon, 8 Dec 2008 11:39:04 +0000 (12:39 +0100)]
cfg80211: fix nl80211 frequency handling
Fix two small bugs with HT frequency setting:
* HT is accepted even when the driver is incapable
* HT40 is accepted when the driver cannot do 40 MHz
(both on the selected band)
Also simplify the code a little.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith [Sun, 7 Dec 2008 16:12:10 +0000 (21:42 +0530)]
ath9k: Maintain rate table choice after association
A scan run after association would change sc_curmode which is
used to get the current rate table. This patch fixes it
by removing sc_curmode and setting the rate table in usage in cur_rate_table
on association.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith [Sun, 7 Dec 2008 16:11:37 +0000 (21:41 +0530)]
ath9k: Fix bug in rate table management
The proper rate table wouldn't be used if a disassoc
happens and a new attempt is made to associate using
wpa_supplicant. This patch fixes it by storing the rate
table to be used on association.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Fri, 5 Dec 2008 18:43:27 +0000 (20:43 +0200)]
ath9k: Disable staggered Beacon frame scheduling
It looks like there are some issues in the current ath9k version as
far as staggered Beacon frame scheduling is concerned. This results in
Beacon frame timestamp being off by 25 milliseconds or so which can
cause issues with multicast power save buffering in AP mode. Some
client stations fail to receive multicast frames when there is this
large an offset between TBTT and actual Beacon transmission time.
Since ath9k does not yet support multiple BSSes anyway, the staggered
Beacon frame scheduling is just increasing the number of interrupts at
this point. The easiest fix for the timestamp offset is to disable
staggered Beacon frames for now. We need to redesign beacon.c anyway
when adding multi-BSS support with mac80211 since the existing code
depends on different mechanism for Beacon frame transmission.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Fri, 5 Dec 2008 18:42:45 +0000 (20:42 +0200)]
ath9k: Free Beacon skbs in AP mode
Fix a memory leak where AP mode did not free transmitted Beacon frame
skbs.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com> Acked-by: Pat Erley <pat-lkml@erley.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Zhu, Yi [Fri, 5 Dec 2008 15:58:41 +0000 (07:58 -0800)]
iwlwifi: use rmb/wmb to protect indirect mmio operation
This patch protects iwlwifi indirect mmio operations with rmb() and
wmb(). It makes sure CPU reordering won't affect our indirect mmio
access.
Signed-off-by: Zhu Yi <yi.zhu@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Zhu, Yi [Fri, 5 Dec 2008 15:58:40 +0000 (07:58 -0800)]
iwlwifi: iwl_poll_{direct_}bit cleanup
The patch merges implementation of iwl_poll_bit() and
iwl_poll_direct_bit() by letting the latter be a special case of
the former.
Signed-off-by: Zhu Yi <yi.zhu@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Zhu, Yi [Fri, 5 Dec 2008 15:58:39 +0000 (07:58 -0800)]
iwlwifi: fix "MAC in deep sleep" error
This patch fixes the misue of CSR_GP_CNTRL with CSR_RESET address
in polling the CSR_GP_CNTRL_REG_FLAG_MAC_CLOCK_READY bit in
iwl4965_apm_reset(). This causes "MAC in deep sleep" error sometimes.
The patch also fixes the timeout value and the iwl_poll_bit() return
value check.
Signed-off-by: Zhu Yi <yi.zhu@intel.com> Acked-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Zhu, Yi [Fri, 5 Dec 2008 15:58:38 +0000 (07:58 -0800)]
iwlwifi: fix time interval misuse in iwl_poll_{direct_}bit
The patch fixes the misuse of microsecond with millisecond in the
polling mechanism of the iwlwifi driver. The impact of this problem
is the unacceptable latency for the whole system (especially during
bringing down the wlan interface).
Signed-off-by: Zhu Yi <yi.zhu@intel.com> Acked-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Abbas, Mohamed [Fri, 5 Dec 2008 15:58:37 +0000 (07:58 -0800)]
iwl3945: Fix rate scale for B-mode connection
This patch make sure we use only CCK rate for B only network.
For 3945 driver, it sets REPLY_RATE_SCALE command every time
we connect to a new network. In this command we set for every
rate the number of try and next rate. The problem mac80211 reports
same mode for both B and G mode as IEEE80211_BAND_2GHZ which will cause
using invalid rate other than CCK in B only network. THis patch
on association will examine sta valid rate on association, if no
OFDM rate in valid available rates it considers it as B only mode so
we can set REPLY_RATE_SCALE command with valid B only network and only
choose CCK rate in rate scaling.
Signed-off-by: Mohamed Abbas <mohamed.abbas@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
iwlwifi: implement iwl_debug_report_frame for iwlagn
The old version of this code was copied from 3945 and never updated. This patch
cleans up structs that do not apply and accounts for other changes from
3945->iwlagn (e.g., PHY info can come in a separate cmd response from the
ucode.)
Signed-off-by: Daniel Halperin <daniel.c.halperin@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reinette Chatre [Thu, 4 Dec 2008 22:49:08 +0000 (14:49 -0800)]
mac80211: remove WARN_ON() from ieee80211_hw_config
ieee80211_hw_config can return an error when the hardware
has rfkill enabled. A WARN_ON() is too harsh for this
failure as it is a valid scenario. Only comment this warning
as we would like to have it back when rfkill is integrated into
mac80211.
Also reintroduce propagation of error if ieee80211_hw_config fails
in ieee80211_config_beacon.
This patch fixes a regression introduced by
"wireless: avoid some net/ieee80211.h vs. linux/ieee80211.h conflicts"
LEAP authentication algorithm identifier should be 128.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Thu, 4 Dec 2008 04:21:20 +0000 (22:21 -0600)]
rtl8187: Improve wireless statistics for RTL8187
The current wireless statistics for the RTL8187 poorly indicate the signal
strength and quality. With testing, I found that the AGC value is inversely
correlated with the strength as in the RTL8187B. By implementing a similar
calculation, much more code becomes common to the two devices.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Tested by: Martín Ernesto Barreyro <barreyromartin@gmail.com> Acked-by: Hin-Tak Leung <htl10@users.sourceforge.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Implements the KS8695 ethernet device (ks8695net).
This driver is only of use on the KS8695 which is an ARM9 based SoC. The
documentation on this SoC is sparse and poor, with barely a register
description and a rough outline of how the ethernet works, this driver was
therefore written with strong reference to the Micrel supplied Linux 2.6.9
port, and to Andrew Victor's ks8695eth driver.
Signed-off-by: Daniel Silverstone <dsilvers@simtec.co.uk> Signed-off-by: Vincent Sanders <vince@simtec.co.uk> Acked-by: Ben Dooks <ben@simtec.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Atsushi Nemoto [Fri, 12 Dec 2008 04:58:04 +0000 (20:58 -0800)]
tc35815: Enable StripCRC feature
The chip can strip CRC automatically on receiving. Enable it.
Also fix potential RX_BUF_SIZE calculation bug which was obscured by
alignment. And use proper symbols (NET_IP_ALIGN, ETH_FCS_LEN, etc.)
instead of magic numbers.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp> Signed-off-by: David S. Miller <davem@davemloft.net>
Jeff Kirsher [Fri, 12 Dec 2008 04:34:19 +0000 (20:34 -0800)]
igb: fixup AER with proper error handling
Based on Peter Waskiewicz patch for ixgbe. Add error handling based on
Hemminger's recommendation.
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Acked-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Leblond [Thu, 11 Dec 2008 01:24:33 +0000 (17:24 -0800)]
netfilter: xt_NFLOG is dependant of nfnetlink_log
The patch "don't call nf_log_packet in NFLOG module" make xt_NFLOG
dependant of nfnetlink_log. This patch forces the dependencies to fix
compilation in case only xt_NFLOG compilation was asked and modifies the
help message accordingly to the change.
Signed-off-by: Eric Leblond <eric@inl.fr> Signed-off-by: David S. Miller <davem@davemloft.net>
Benjamin Thery [Thu, 11 Dec 2008 00:30:15 +0000 (16:30 -0800)]
netns: ip6mr: enable namespace support in ipv6 multicast forwarding code
This last patch makes the appropriate changes to use and propagate the
network namespace where needed in IPv6 multicast forwarding code.
This consists mainly in replacing all the remaining init_net occurences
with current netns pointer retrieved from sockets, net devices or
mfc6_caches depending on the routines' contexts.
Some routines receive a new 'struct net' parameter to propagate the current
netns:
* ip6mr_get_route
* ip6mr_cache_report
* ip6mr_cache_find
* ip6mr_cache_unresolved
* mif6_add/mif6_delete
* ip6mr_mfc_add/ip6mr_mfc_delete
* ip6mr_reg_vif
All the IPv6 multicast forwarding variables moved to struct netns_ipv6 by
the previous patches are now referenced in the correct namespace.
Changelog:
==========
* Take into account the net associated to mfc6_cache when matching entries in
mfc_unres_queue list.
* Call mroute_clean_tables() in ip6mr_net_exit() to free memory allocated
per-namespace.
* Call dev_net_set() in ip6mr_reg_vif() to initialize dev->nd_net
correctly.
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Preliminary work to make IPv6 multicast forwarding netns-aware.
Declare variable cache_resolve_queue_len per-namespace: moves it into
struct netns_ipv6.
This variable counts the number of unresolved cache entries queued in the
list mfc_unres_queue. This list is kept global to all netns as the number
of entries per namespace is limited to 10 (hardcoded in routine
ip6mr_cache_unresolved).
Entries belonging to different namespaces in mfc_unres_queue will be
identified by matching the mfc_net member introduced previously in
struct mfc6_cache.
Keeping this list global to all netns, also allows us to keep a single
timer (ipmr_expire_timer) to handle their expiration.
In some places cache_resolve_queue_len value was tested for arming
or deleting the timer. These tests were equivalent to testing
mfc_unres_queue value instead and are replaced in this patch.
At the moment, cache_resolve_queue_len is only referenced in init_net.
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Convert the driver to select 16-bit or 32-bit bus access at runtime,
at a small performance cost.
Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements the PCIe Advanced Error Reporting callbacks in
ixgbe. The 82598 hardware supports AER, so we enable it.
Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Alexander Duyck [Wed, 10 Dec 2008 09:09:53 +0000 (01:09 -0800)]
igb: Correctly determine pci-e function number in virtual environment
When running in a virtual environment the ports of an 82575/6 can appear to
be single function devices which is not correct. To resolve this we rely
on the function number stored in the status register.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Alexander Duyck [Wed, 10 Dec 2008 09:08:59 +0000 (01:08 -0800)]
igb: update handling of RCTL for smaller buffer sizes
As we begin to move into virtualization the use of a global value for
buffer sizes becomes undesirable. This change makes it so only 256 and 512
byte buffers require use of the RCTL register.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: make tbl argument to bond_parse_parm() const
bond_parse_parm() parses a parameter table for a particular value and
is therefore not modifying the table at all. Therefore make the 2nd
argument const, thus allowing to make the tables const later.
Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org> Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/bonding/bond_ipv6.c: In function 'bond_na_send':
drivers/net/bonding/bond_ipv6.c:75: error: 'slave' undeclared (first use in this function)
drivers/net/bonding/bond_ipv6.c:75: error: (Each undeclared identifier is reported only once
drivers/net/bonding/bond_ipv6.c:75: error: for each function it appears in.)
This patch fixes that.
Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org> Signed-off-by: David S. Miller <davem@davemloft.net>
htb_id_find_next_upper() is usually called to find a class with next
id after some previously removed class, so let's move a check for
equality to the end: it's the least likely here.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:19:06 +0000 (01:19 -0800)]
dccp ccid-2: Phase out the use of boolean Ack Vector sysctl
This removes the use of the sysctl and the minisock variable for the Send Ack
Vector feature, as it now is handled fully dynamically via feature negotiation
(i.e. when CCID-2 is enabled, Ack Vectors are automatically enabled as per
RFC 4341, 4.).
Using a sysctl in parallel to this implementation would open the door to
crashes, since much of the code relies on tests of the boolean minisock /
sysctl variable. Thus, this patch replaces all tests of type
if (dccp_msk(sk)->dccpms_send_ack_vector)
/* ... */
with
if (dp->dccps_hc_rx_ackvec != NULL)
/* ... */
The dccps_hc_rx_ackvec is allocated by the dccp_hdlr_ackvec() when feature
negotiation concluded that Ack Vectors are to be used on the half-connection.
Otherwise, it is NULL (due to dccp_init_sock/dccp_create_openreq_child),
so that the test is a valid one.
The activation handler for Ack Vectors is called as soon as the feature
negotiation has concluded at the
* server when the Ack marking the transition RESPOND => OPEN arrives;
* client after it has sent its ACK, marking the transition REQUEST => PARTOPEN.
Adding the sequence number of the Response packet to the Ack Vector has been
removed, since
(a) connection establishment implies that the Response has been received;
(b) the CCIDs only look at packets received in the (PART)OPEN state, i.e.
this entry will always be ignored;
(c) it can not be used for anything useful - to detect loss for instance, only
packets received after the loss can serve as pseudo-dupacks.
There was a FIXME to change the error code when dccp_ackvec_add() fails.
I removed this after finding out that:
* the check whether ackno < ISN is already made earlier,
* this Response is likely the 1st packet with an Ackno that the client gets,
* so when dccp_ackvec_add() fails, the reason is likely not a packet error.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:18:37 +0000 (01:18 -0800)]
dccp: Remove manual influence on NDP Count feature
Updating the NDP count feature is handled automatically now:
* for CCID-2 it is disabled, since the code does not use NDP counts;
* for CCID-3 it is enabled, as NDP counts are used to determine loss lengths.
Allowing the user to change NDP values leads to unpredictable and failing
behaviour, since it is then possible to disable NDP counts even when they
are needed (e.g. in CCID-3).
This means that only those user settings are sensible that agree with the
values for Send NDP Count implied by the choice of CCID. But those settings
are already activated by the feature negotiation (CCID dependency tracking),
hence this form of support is redundant.
At startup the initialisation of the NDP count feature uses the default
value of 0, which is done implicitly by the zeroing-out of the socket when
it is allocated. If the choice of CCID or feature negotiation enables NDP
count, this will then be updated via the NDP activation handler.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:18:05 +0000 (01:18 -0800)]
dccp: Remove obsolete parts of the old CCID interface
The TX/RX CCIDs of the minisock are now redundant: similar to the Ack Vector
case, their value equals initially that of the sysctl, but at the end of
feature negotiation may be something different.
The old interface removed by this patch thus has been replaced by the newer
interface to dynamically query the currently loaded CCIDs.
Also removed are the constructors for the TX CCID and the RX CCID, since the
switch "rx <-> non-rx" is done by the handler in minisocks.c (and the handler
is the only place in the code where CCIDs are loaded).
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:17:32 +0000 (01:17 -0800)]
dccp: Clean up old feature-negotiation infrastructure
The code removed by this patch is no longer referenced or used, the added
lines update documentation and copyrights.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:16:27 +0000 (01:16 -0800)]
dccp: Integration of dynamic feature activation - part 3 (client side)
This integrates feature-activation in the client:
1. When dccp_parse_options() fails, the reset code is already set; request_sent\
_state_process() currently overrides this with `Packet Error', which is not
intended - changed to use the reset code supplied by dccp_parse_options().
2. When feature negotiation fails, the socket should be marked as not usable,
so that the application is notified that an error occurred. This is achieved
by a new label 'unable_to_proceed': generating an error code of `Aborted',
setting the socket state to CLOSED, returning with ECOMM in sk_err.
3. Avoids parsing the Ack twice in Respond state by not doing option processing
again in dccp_rcv_respond_partopen_state_process (as option processing has
already been done on the request_sock in dccp_check_req).
Since this addresses congestion-control initialisation, a corresponding
FIXME has been removed.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:15:55 +0000 (01:15 -0800)]
dccp: Integration of dynamic feature activation - part 2 (server side)
This patch integrates the activation of features at the end of negotiation
into the server-side code.
Note regarding the removal of 'const':
--------------------------------------
The 'const' attribute has been removed from 'dreq' since dccp_activate_values()
needs to operate on dreq's feature list. Part of the activation is to remove
those options from the list that have already been confirmed, hence it is not
purely read-only.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Gerrit Renker [Mon, 8 Dec 2008 09:15:26 +0000 (01:15 -0800)]
dccp: Integration of dynamic feature activation - part 1 (socket setup)
This first patch out of three replaces the hardcoded default settings with
initialisation code for the dynamic feature negotiation.
The patch also ensures that the client feature-negotiation queue is flushed
only when entering the OPEN state.
Since confirmed Change options are removed as soon as they are confirmed
(in the DCCP-Response), this ensures that Confirm options are retransmitted.
Note on retransmitting Confirm options:
---------------------------------------
Implementation experience showed that it is necessary to retransmit Confirm
options. Thanks to Leandro Melo de Sales who reported a bug in an earlier
revision of the patch set, resulting from not retransmitting these options.
As long as the client is in PARTOPEN, it needs to retransmit the Confirm
options for the Change options received on the DCCP-Response from the server.
Otherwise, if the packet containing the Confirm options gets dropped in the
network, the connection aborts due to undefined feature negotiation state.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Acked-by: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: David S. Miller <davem@davemloft.net>
Wang Chen [Mon, 8 Dec 2008 09:14:16 +0000 (01:14 -0800)]
netdevice: Kill netdev->priv
This is the last shoot of this series.
After I removing all directly reference of netdev->priv, I am killing
"priv" of "struct net_device" and fixing relative comments/docs.
Anyone will not be allowed to reference netdev->priv directly.
If you want to reference the memory of private data, use netdev_priv()
instead.
If the private data is not allocted when alloc_netdev(), use
netdev->ml_priv to point that memory after you creating that private
data.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wang Chen [Mon, 8 Dec 2008 09:13:25 +0000 (01:13 -0800)]
staging-winbond: Kill directly reference of netdev->priv
This driver is not yet finished.
At this time, we don't know how netdev be created and how
private data be allocated.
So, simply use netdev_priv() now and leave some temp comment.
Compile test only.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wang Chen [Sun, 7 Dec 2008 07:58:37 +0000 (23:58 -0800)]
staging-p80211: Kill directly reference of netdev->priv
In this driver, netdev's private data is wlandevice_t. And the
wlandev(type of wlandevice_t) is exist before netdev be allocated. So
use netdev->ml_priv to point to the private data.
I am not sure whether I should consider the kernel version older than
2.3.38. Because in those kernels, netdevice_t is "structure dev"
instead of "structure net_device" and of course "dev->ml_priv" will
cause compile error. But before my patch, in function wlan_setup(),
there is a ether_setup(net_device) which already broke kernels which
older than 2.3.38.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com> Acked-by: Greg Kroah-Hartman <gregkh@suse.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Wang Chen [Sun, 7 Dec 2008 07:57:49 +0000 (23:57 -0800)]
s390_net: Kill directly reference of netdev->priv
The private data comes from ccwgroup_device.
So just don't allocate private data memory when do alloc_netdev()
and use netdev->ml_priv to reference private data.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ilpo Järvinen [Sat, 6 Dec 2008 06:49:37 +0000 (22:49 -0800)]
tcp: fix tso_should_defer in 64bit
Since jiffies is unsigned long, the types get expanded into
that and after long enough time the difference will therefore
always be > 1 (and that probably happens near boot as well as
iirc the first jiffies wrap is scheduler close after boot to
find out problems related to that early).
This was originally noted by Bill Fink in Dec'07 but nobody
never ended fixing it.
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