Commit Graph

119963 Commits

Author SHA1 Message Date
Senthil Balasubramanian
9f80420217 ath9k: INI update for atheros chipets.
init values update for various atheros chipsets.

Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:28 -05:00
Luis R. Rodriguez
ea4a82dcee iwlwifi: enable custom fw regulatory solution
This enables the custom firmware regulatory solution option
on iwlwifi drivers. These devices are uncapable of mapping their
EEPROM regulatory domain to a specific ISO / IEC alpha2.
Although the new 11n devices (>= iwl 5000) have only
3 regultaory SKUs -- MOW, ABG (no N) and BG -- the older
devices (3945 and 4965) have a more complex SKU arrangement
and therefore its not practical to move this to the driver.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:27 -05:00
Luis R. Rodriguez
14b9815af3 cfg80211: add support for custom firmware regulatory solutions
This adds API to cfg80211 to allow wireless drivers to inform
us if their firmware can handle regulatory considerations *and*
they cannot map these regulatory domains to an ISO / IEC 3166
alpha2. In these cases we skip the first regulatory hint instead
of expecting the driver to build their own regulatory structure,
providing us with an alpha2, or using the reg_notifier().

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:27 -05:00
Luis R. Rodriguez
3f2355cb91 cfg80211/mac80211: Add 802.11d support
This adds country IE parsing to mac80211 and enables its usage
within the new regulatory infrastructure in cfg80211. We parse
the country IEs only on management beacons for the BSSID you are
associated to and disregard the IEs when the country and environment
(indoor, outdoor, any) matches the already processed country IE.

To avoid following misinformed or outdated APs we build and use
a regulatory domain out of the intersection between what the AP
provides us on the country IE and what CRDA is aware is allowed
on the same country.

A secondary device is allowed to follow only the same country IE
as it make no sense for two devices on a system to be in two
different countries.

In the case the AP is using country IEs for an incorrect country
the user may help compliance further by setting the regulatory
domain before or after the IE is parsed and in that case another
intersection will be performed.

CONFIG_WIRELESS_OLD_REGULATORY is supported but requires CRDA
present.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:26 -05:00
Luis R. Rodriguez
88dc1c3f7f cfg80211: mark regdomains with > NL80211_MAX_SUPP_REG_RULES invalid
Lets remain consistent and mark rds with > NL80211_MAX_SUPP_REG_RULES
number of reg rules as invalid in is_valid_rd().

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:25 -05:00
Luis R. Rodriguez
02ba0b3263 cfg80211: call_crda() won't tell us if CRDA was present
kobject_uevent_env() can return an error but it just tells us
if the uvent was built/sent or not, it doesn't tell us anything
about what happened in userspace, whether the udev rule was present
nor does it tell us if CRDA was present or not. So remove
the informative complaint about it assuming it will tell us
such things.

Note that you can determine if CRDA is present after loading cfg80211
by using:

is_old_static_regdom(cfg80211_regdomain)

but this doesn't account for possible user install after initial
boot, and also for when the user uses the static EU regulatory
domain.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:25 -05:00
Luis R. Rodriguez
a01ddafd43 cfg80211: expect different rd in cfg80211 when intersecting
When intersecting it is possible that set_regdom() was called
with a regulatory domain which we'll only use as an aid to
build a final regulatory domain.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:24 -05:00
Luis R. Rodriguez
b8295acdc3 cfg80211: separate intersection section in __set_regdom()
So far the __set_regdom() code is pretty generic as the
intersection case is fairly straight forward; this will however
change when 802.11d support is added so lets separate intersection
code for now in preparation for 802.11d support.

This patch only has slight functional changes.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:23 -05:00
Luis R. Rodriguez
8375af3ba2 cfg80211: remove switch from __set_regdom()
We have control over the REGDOM_SET_BY_* macros passed
so remove the switch.

This patch has no functional changes.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:22 -05:00
Luis R. Rodriguez
5203cdb6ad cfg80211: remove switch from __regulatory_hint()
We have complete control over REGDOM_SET_BY_* enum passed
down to __regulatory_hint() as such there is no need to
account for unexpected REGDOM_SET_BY_*'s, lets just remove
the switch statement as this code does not change and
won't change even when we add 802.11d support.

This patch has no functional changes.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:22 -05:00
Luis R. Rodriguez
91e9900418 cfg80211: mark negative frequencies as invalid
Regulatory rules with negative frequencies are now
marked as invalid in is_valid_reg_rule().

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:21 -05:00
Tomas Winkler
e0737a77d6 iwlwifi: iwl-fh.h cleanup
This patch fix value of upper FH register bound plus
it reorders and groups registers in more readable way

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Acked-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>
2008-11-25 16:41:20 -05:00
Zhu, Yi
34faf780cf iwlwifi: some fh document fix and cleanup
This patch cleans up some flow handler related document. It also
removes some blank lines.

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>
2008-11-25 16:41:20 -05:00
Zhu, Yi
352bc8de19 iwlwifi: configure_filter rewrite
The patch rewrites the mac80211 configure_filter handler to better mapping
mac80211 filter flags to iwlwifi hardware filter flags. We now can support
5 mac80211 filter flags: FIF_OTHER_BSS, FIF_ALLMULTI, FIF_PROMISC_IN_BSS,
FIF_BCN_PRBRESP_PROMISC and FIF_CONTROL. This patch also avoids reconnecting
if the filter flags are changed when the STA is associated. Because rx_assoc
is used when full rxon is not necessary.

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>
2008-11-25 16:41:19 -05:00
Abhijeet Kolekar
c305606540 iwlwifi : fix checkpatch.pl errors
Patch fixes checkpatch.pl errors for iwlwifi.

Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:18 -05:00
Chatre, Reinette
dbce56a456 iwlwifi: replace magic constants with define
use IWL_CCK_RATES_MASK and IWL_OFDM_RATES_MASK instead of
their values directly.

Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
cc: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:41:18 -05:00
Tomas Winkler
417f114bf2 iwlwifi: rs: remove fc variable and other cleanups
This patch
1. Removes use once use only fc variables, they are useless after refactoring
ieee80211 frame control handlers
2. Other trivial cleanups

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>
2008-11-25 16:41:17 -05:00
Tomas Winkler
9f58671e8d iwlwifi: consolidate station management code
This patch moves code around and group most of the station
management code into iwl-sta.c

No functional changes (yet)

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>
2008-11-25 16:41:06 -05:00
Kolekar, Abhijeet
cee53ddb46 iwl3945 : Simplify iwl3945_pci_probe
Patch aligns iwl3945_pci_probe with iwlwifi's iwl_pci_probe.
Added few comments and code simplified to make readable.

Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:35:20 -05:00
Ivo van Doorn
0e3de99846 rt2x00: Fix TX failure path
The callback function write_tx_data() can only fail
when our ENTRY_OWNER_DEVICE_DATA flag on a queue entry
failed to determine the entry was not available and
it is in fact still owned by the hardware.
This means that if that function fails the queue
must be stopped in mac80211.

When rt2x00queue_get_queue() returns NULL in the TX
path, it means mac80211 has passed us an invalid queue,
although this should be impossible, it shouldn't hurt
if we send mac80211 a signal to stop the queue either.

Both issues can simply be resolved by removing their
manual failure handler and making them use the failure path
provided in rt2x00mac_tx().

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:32:54 -05:00
Ivo van Doorn
0f829b1d6f rt2x00: Move rt73usb register access wrappers into rt2x00usb
rt2500usb and rt73usb have different register word sizes,
for that reason the register access wrappers were never
moved into rt2x00usb.
With rt2800usb on its way, we should favor the 32bit
register access and move those wrappers into rt2x00usb.
That saves duplicate code, since only rt2500usb will
need the special 16bit wrappers.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:32:53 -05:00
Ivo van Doorn
c9c3b1a5de rt2x00: Cleanup indirect register access
All code which accessed indirect registers was similar
in respect to the for-loop, the given timeout, etc.
Move it into a seperate function, which for PCI drivers
can be moved into rt2x00pci.

This allows us to cleanup the cleanup the code further
by removing the goto statementsand making the codepath
look a bit nicer.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:32:53 -05:00
Johannes Berg
9764f3f9c3 ath5k: name pci driver "ath5k" too
Call the ath5k pci driver struct "ath5k" too to be less
confusing in sysfs.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2008-11-25 16:32:53 -05:00
David S. Miller
2f9889a20c Revert "hso: Fix crashes on close."
This reverts commit 4a3e818181.

On request from Alan Cox.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 03:53:09 -08:00
David S. Miller
ab153d84d9 Revert "hso: Fix free of mutexes still in use."
This reverts commit 52429eb216.

On request from Alan Cox.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 03:52:46 -08:00
David S. Miller
cd90ee1799 Revert "hso: Add TIOCM ioctl handling."
This reverts commit 7ea3a9ad9b.

On request from Alan Cox.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 03:52:17 -08:00
Alexey Dobriyan
fb7e06748c xfrm: remove useless forward declarations
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 01:05:54 -08:00
Alexey Dobriyan
6daad37230 ah4/ah6: remove useless NULL assignments
struct will be kfreed in a moment, so...

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 01:05:09 -08:00
Alexander Duyck
69d728baf6 igb: loopback bits not correctly cleared from RCTL register
This change forces the bits to 0 by using an &= operation with an inverted
mask of all options instead of using an |= with a value of 0.

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>
2008-11-25 01:04:03 -08:00
Alexander Duyck
9b07f3d315 igb: remove unneeded bit refrence when enabling jumbo frames
There is a reference to a Buffer Size extention bit that is unneded by
82575/82576 hardware.  Since it is not needed it should be removed from the
code.

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>
2008-11-25 01:03:26 -08:00
Jeff Kirsher
7a6b6f515f DCB: fix kconfig option
Since the netlink option for DCB is necessary to actually be useful,
simplified the Kconfig option.  In addition, added useful help text for the
Kconfig option.

Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 01:02:08 -08:00
Harvey Harrison
411c41eea5 aoe: remove private mac address format function
Add %pm to omit the colons when printing a mac address.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:40:37 -08:00
Denis Joseph Barrow
9c8f92aed1 hso: Hook up ->reset_resume
Made usb_drivers reset_resume function point to hso_resume this 
fixes problems a usb reset is done when the network interface
is left idle for a few minutes. Possibly reset_resume should
initialise hardware more but this works in the common case.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:36:10 -08:00
Denis Joseph Barrow
7ea3a9ad9b hso: Add TIOCM ioctl handling.
Makes TIOCM ioctls for Data Carrier Detect & related functions
work like /drivers/serial/serial-core.c potentially needed 
for pppd & similar user programs.   

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:35:26 -08:00
Denis Joseph Barrow
52429eb216 hso: Fix free of mutexes still in use.
A new structure hso_mutex_table had to be declared statically
& used as as hso_device mutex_lock(&serial->parent->mutex) etc
is freed in hso_serial_open & hso_serial_close by kref_put while
the mutex is still in use.

This is a substantial change but should make the driver much stabler.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:33:13 -08:00
Denis Joseph Barrow
89930b7b5e hso: Fix URB submission -EINVAL.
Added check for IFF_UP in hso_resume, this should eliminate -EINVAL (-22)
errors caused from urb's being submitted twice, once by hso_resume
& once in hso_net_open, if suspend/resume USB power saving  mode is enabled

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:30:48 -08:00
Denis Joseph Barrow
4a3e818181 hso: Fix crashes on close.
Moved serial_open_count in hso_serial_open to
prevent crashes owing to the serial structure being made NULL
when hso_serial_close is called even though hso_serial_open
returned -ENODEV, Alan Cox pointed out this happens,
also put in sanity check in hso_serial_close
to check for a valid serial structure which should prevent
the most reproducable crash in the driver when the hso device
is disconnected while in use.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:27:50 -08:00
Denis Joseph Barrow
bab04c3adb hso: Add new usb device id's.
Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:26:12 -08:00
Stephen Hemminger
47fd5b8373 netdev: add HAVE_NET_DEVICE_OPS
As a concession to vendors who have to deal with one source for different
kernel versions, add a HAVE_NET_DEVICE_OPS so they don't end up hard
coding ifdef against kernel version.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 00:20:43 -08:00
Ilpo Järvinen
0ace285605 tcp: handle shift/merge of cloned skbs too
This caused me to get repeatably:

  tcpdump: pcap_loop: recvfrom: Bad address

Happens occassionally when I tcpdump my for-looped test xfers:
  while [ : ]; do echo -n "$(date '+%s.%N') "; ./sendfile; sleep 20; done

Rest of the relevant commands:
  ethtool -K eth0 tso off
  tc qdisc add dev eth0 root netem drop 4%
  tcpdump -n -s0 -i eth0 -w sacklog.all

Running net-next under kvm, connection goes to the same host
(basically just out of kvm). The connection itself works ok
and data gets sent without corruption even with a large
number of tests while tcpdump fails usually within less than
5 tests.

Whether it only happens because of this change or not, I
don't know for sure but it's the only thing with which
I've seen that error. The non-cloned variant works w/o it
for much longer time. I'm yet to debug where the error
actually comes from.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:30:21 -08:00
Ilpo Järvinen
111cc8b913 tcp: add some mibs to track collapsing
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:27:22 -08:00
Ilpo Järvinen
92ee76b6d9 tcp: Make shifting not clear the hints
The earlier version was just very basic one which is "playing
safe" by always clearing the hints. However, clearing of a hint
is extremely costly operation with large windows, so it must be
avoided at all cost whenever possible, there is a way with
shifting too achieve not-clearing.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:26:56 -08:00
Ilpo Järvinen
832d11c5cd tcp: Try to restore large SKBs while SACK processing
During SACK processing, most of the benefits of TSO are eaten by
the SACK blocks that one-by-one fragment SKBs to MSS sized chunks.
Then we're in problems when cleanup work for them has to be done
when a large cumulative ACK comes. Try to return back to pre-split
state already while more and more SACK info gets discovered by
combining newly discovered SACK areas with the previous skb if
that's SACKed as well.

This approach has a number of benefits:

1) The processing overhead is spread more equally over the RTT
2) Write queue has less skbs to process (affect everything
   which has to walk in the queue past the sacked areas)
3) Write queue is consistent whole the time, so no other parts
   of TCP has to be aware of this (this was not the case with
   some other approach that was, well, quite intrusive all
   around).
4) Clean_rtx_queue can release most of the pages using single
   put_page instead of previous PAGE_SIZE/mss+1 calls

In case a hole is fully filled by the new SACK block, we attempt
to combine the next skb too which allows construction of skbs
that are even larger than what tso split them to and it handles
hole per on every nth patterns that often occur during slow start
overshoot pretty nicely. Though this to be really useful also
a retransmission would have to get lost since cumulative ACKs
advance one hole at a time in the most typical case.

TODO: handle upwards only merging. That should be rather easy
when segment is fully sacked but I'm leaving that as future
work item (it won't make very large difference anyway since
this current approach already covers quite a lot of normal
cases).

I was earlier thinking of some sophisticated way of tracking
timestamps of the first and the last segment but later on
realized that it won't be that necessary at all to store the
timestamp of the last segment. The cases that can occur are
basically either:
  1) ambiguous => no sensible measurement can be taken anyway
  2) non-ambiguous is due to reordering => having the timestamp
     of the last segment there is just skewing things more off
     than does some good since the ack got triggered by one of
     the holes (besides some substle issues that would make
     determining right hole/skb even harder problem). Anyway,
     it has nothing to do with this change then.

I choose to route some abnormal looking cases with goto noop,
some could be handled differently (eg., by stopping the
walking at that skb but again). In general, they either
shouldn't happen at all or are rare enough to make no difference
in practice.

In theory this change (as whole) could cause some macroscale
regression (global) because of cache misses that are taken over
the round-trip time but it gets very likely better because of much
less (local) cache misses per other write queue walkers and the
big recovery clearing cumulative ack.

Worth to note that these benefits would be very easy to get also
without TSO/GSO being on as long as the data is in pages so that
we can merge them. Currently I won't let that happen because
DSACK splitting at fragment that would mess up pcounts due to
sk_can_gso in tcp_set_skb_tso_segs. Once DSACKs fragments gets
avoided, we have some conditions that can be made less strict.

TODO: I will probably have to convert the excessive pointer
passing to struct sacktag_state... :-)

My testing revealed that considerable amount of skbs couldn't
be shifted because they were cloned (most likely still awaiting
tx reclaim)...

[The rest is considering future work instead since I got
repeatably EFAULT to tcpdump's recvfrom when I added
pskb_expand_head to deal with clones, so I separated that
into another, later patch]

...To counter that, I gave up on the fifth advantage:

5) When growing previous SACK block, less allocs for new skbs
   are done, basically a new alloc is needed only when new hole
   is detected and when the previous skb runs out of frags space

...which now only happens of if reclaim is fast enough to dispose
the clone before the SACK block comes in (the window is RTT long),
otherwise we'll have to alloc some.

With clones being handled I got these numbers (will be somewhat
worse without that), taken with fine-grained mibs:

                  TCPSackShifted 398
                   TCPSackMerged 877
            TCPSackShiftFallback 320
      TCPSACKCOLLAPSEFALLBACKGSO 0
  TCPSACKCOLLAPSEFALLBACKSKBBITS 0
  TCPSACKCOLLAPSEFALLBACKSKBDATA 0
    TCPSACKCOLLAPSEFALLBACKBELOW 0
    TCPSACKCOLLAPSEFALLBACKFIRST 1
 TCPSACKCOLLAPSEFALLBACKPREVBITS 318
      TCPSACKCOLLAPSEFALLBACKMSS 1
   TCPSACKCOLLAPSEFALLBACKNOHEAD 0
    TCPSACKCOLLAPSEFALLBACKSHIFT 0
          TCPSACKCOLLAPSENOOPSEQ 0
  TCPSACKCOLLAPSENOOPSMALLPCOUNT 0
     TCPSACKCOLLAPSENOOPSMALLLEN 0
             TCPSACKCOLLAPSEHOLE 12

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:20:15 -08:00
Ilpo Järvinen
f58b22fd3c tcp: make tcp_sacktag_one able to handle partial skb too
This is preparatory work for SACK combiner patch which may
have to count TCP state changes for only a part of the skb
because it will intentionally avoids splitting skb to SACKed
and not sacked parts.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:14:43 -08:00
Ilpo Järvinen
adb92db857 tcp: Make SACK code to split only at mss boundaries
Sadly enough, this adds possible divide though we try to avoid
it by checking one mss as common case.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:13:50 -08:00
Ilpo Järvinen
e8bae275d9 tcp: more aggressive skipping
I knew already when rewriting the sacktag that this condition
was too conservative, change it now since it prevent lot of
useless work (especially in the sack shifter decision code
that is being added by a later patch). This shouldn't change
anything really, just save some processing regardless of the
shifter.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:12:28 -08:00
Ilpo Järvinen
e1aa680fa4 tcp: move tcp_simple_retransmit to tcp_input
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:11:55 -08:00
Ilpo Järvinen
4a17fc3add tcp: collapse more than two on retransmission
I always had thought that collapsing up to two at a time was
intentional decision to avoid excessive processing if 1 byte
sized skbs are to be combined for a full mtu, and consecutive
retransmissions would make the size of the retransmittee
double each round anyway, but some recent discussion made me
to understand that was not the case. Thus make collapse work
more and wait less.

It would be possible to take advantage of the shifting
machinery (added in the later patch) in the case of paged
data but that can be implemented on top of this change.

tcp_skb_is_last check is now provided by the loop.

I tested a bit (ss-after-idle-off, fill 4096x4096B xfer,
10s sleep + 4096 x 1byte writes while dropping them for
some a while with netem):

. 16774097:16775545(1448) ack 1 win 46
. 16775545:16776993(1448) ack 1 win 46
. ack 16759617 win 2399
P 16776993:16777217(224) ack 1 win 46
. ack 16762513 win 2399
. ack 16765409 win 2399
. ack 16768305 win 2399
. ack 16771201 win 2399
. ack 16774097 win 2399
. ack 16776993 win 2399
. ack 16777217 win 2399
P 16777217:16777257(40) ack 1 win 46
. ack 16777257 win 2399
P 16777257:16778705(1448) ack 1 win 46
P 16778705:16780153(1448) ack 1 win 46
FP 16780153:16781313(1160) ack 1 win 46
. ack 16778705 win 2399
. ack 16780153 win 2399
F 1:1(0) ack 16781314 win 2399

While without drop-all period I get this:

. 16773585:16775033(1448) ack 1 win 46
. ack 16764897 win 9367
. ack 16767793 win 9367
. ack 16770689 win 9367
. ack 16773585 win 9367
. 16775033:16776481(1448) ack 1 win 46
P 16776481:16777217(736) ack 1 win 46
. ack 16776481 win 9367
. ack 16777217 win 9367
P 16777217:16777218(1) ack 1 win 46
P 16777218:16777219(1) ack 1 win 46
P 16777219:16777220(1) ack 1 win 46
  ...
P 16777247:16777248(1) ack 1 win 46
. ack 16777218 win 9367
. ack 16777219 win 9367
  ...
. ack 16777233 win 9367
. ack 16777248 win 9367
P 16777248:16778696(1448) ack 1 win 46
P 16778696:16780144(1448) ack 1 win 46
FP 16780144:16781313(1169) ack 1 win 46
. ack 16780144 win 9367
F 1:1(0) ack 16781314 win 9367

The window seems to be 30-40 segments, which were successfully
combined into: P 16777217:16777257(40) ack 1 win 46

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 21:03:43 -08:00
Eric Dumazet
a21bba9454 net: avoid a pair of dst_hold()/dst_release() in ip_push_pending_frames()
We can reduce pressure on dst entry refcount that slowdown UDP transmit
path on SMP machines. This pressure is visible on RTP servers when
delivering content to mediagateways, especially big ones, handling
thousand of streams. Several cpus send UDP frames to the same
destination, hence use the same dst entry.

This patch makes ip_push_pending_frames() steal the refcount its
callers had to take when filling inet->cork.dst.

This doesnt avoid all refcounting, but still gives speedups on SMP,
on UDP/RAW transmit path.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 16:07:50 -08:00
Eric Dumazet
2e77d89b2f net: avoid a pair of dst_hold()/dst_release() in ip_append_data()
We can reduce pressure on dst entry refcount that slowdown UDP transmit
path on SMP machines. This pressure is visible on RTP servers when
delivering content to mediagateways, especially big ones, handling
thousand of streams. Several cpus send UDP frames to the same
destination, hence use the same dst entry.

This patch makes ip_append_data() eventually steal the refcount its
callers had to take on the dst entry.

This doesnt avoid all refcounting, but still gives speedups on SMP,
on UDP/RAW transmit path

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 15:52:46 -08:00