Dan Carpenter [Thu, 16 Aug 2012 03:14:04 +0000 (03:14 +0000)]
gre: information leak in ip6_tnl_ioctl()
There is a one byte hole between p->hop_limit and p->flowinfo where
stack memory is leaked to the user. This was introduced in c12b395a46
"gre: Support GRE over IPv6".
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Dan Carpenter [Thu, 16 Aug 2012 03:16:19 +0000 (03:16 +0000)]
sctp: fix bogus if statement in sctp_auth_recv_cid()
There is an extra semi-colon here, so we always return 0 instead of
calling __sctp_auth_cid().
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ulrich Weber [Thu, 16 Aug 2012 01:24:49 +0000 (01:24 +0000)]
sctp: fix compile issue with disabled CONFIG_NET_NS
struct seq_net_private has no struct net
if CONFIG_NET_NS is not enabled
Signed-off-by: Ulrich Weber <ulrich.weber@sophos.com> Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 15 Aug 2012 22:26:05 +0000 (15:26 -0700)]
Merge branch 'for-davem' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next
John W. Linville says:
====================
This is a batch of updates intended for 3.7. The ath9k, mwifiex,
and b43 drivers get the bulk of the commits this time, with a handful
of other driver bits thrown-in. It is mostly just minor fixes and
cleanups, etc.
Also included is a Bluetooth pull, with a lot of refactoring.
Gustavo says:
"These are the changes I queued for 3.7. There are a many
small fixes/improvements by Andre Guedes. A l2cap channel
refcounting refactor by Jaganath. Bluetooth sockets now
appears in /proc/net, by Masatake Yamato and Sachin Kamat
changes ours drivers to use devm_kzalloc()."
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Fan Du [Wed, 15 Aug 2012 02:13:47 +0000 (10:13 +0800)]
xfrm: remove redundant parameter "int dir" in struct xfrm_mgr.acquire
Sematically speaking, xfrm_mgr.acquire is called when kernel intends to ask
user space IKE daemon to negotiate SAs with peers. IOW the direction will
*always* be XFRM_POLICY_OUT, so remove int dir for clarity.
Signed-off-by: Fan Du <fan.du@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Cong Wang [Wed, 15 Aug 2012 10:18:11 +0000 (18:18 +0800)]
sctp: fix a compile error in sctp.h
I got the following compile error:
In file included from include/net/sctp/checksum.h:46:0,
from net/ipv4/netfilter/nf_nat_proto_sctp.c:14:
include/net/sctp/sctp.h: In function ‘sctp_dbg_objcnt_init’:
include/net/sctp/sctp.h:370:88: error: parameter name omitted
include/net/sctp/sctp.h: In function ‘sctp_dbg_objcnt_exit’:
include/net/sctp/sctp.h:371:88: error: parameter name omitted
Cc: David S. Miller <davem@davemloft.net> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Cong Wang <amwang@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Push struct net down into sctp_verify_ext_param
Add struct net as a parameter to sctp_verify_param so it can be passed
to sctp_verify_ext_param where struct net will be needed when the sctp
tunables become per net tunables.
Add struct net as a parameter to sctp_verify_init so struct net can be
passed to sctp_verify_param.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Push struct net down into all of the state machine functions
There are a handle of state machine functions primarily those dealing
with processing INIT packets where there is neither a valid endpoint nor
a valid assoication from which to derive a struct net. Therefore add
struct net * to the parameter list of sctp_state_fn_t and update all of
the state machine functions.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Push struct net down to sctp_chunk_event_lookup
This trickles up through sctp_sm_lookup_event up to sctp_do_sm
and up further into sctp_primitiv_NAME before the code reaches
places where struct net can be reliably found.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Start with an empty sctp_net_table that will be populated as the various
tunable sysctls are made per net.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
- Fix the sctp_af operations to work in all namespaces
- Enable sctp socket creation in all network namespaces.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
- Convert all of the files under /proc/net/sctp to be per
network namespace.
- Don't print anything for /proc/net/sctp/snmp except in
the initial network namespaces as the snmp counters still
have to be converted to be per network namespace.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Move the percpu sockets counter out of sctp_proc_init
The percpu sctp socket counter has nothing at all to do with the sctp
proc files, and having it in the wrong initialization is confusing,
and makes network namespace support a pain.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
- Kill sctp_get_ctl_sock, it is useless now.
- Pass struct net where needed so net->sctp.ctl_sock is accessible.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Make the address lists per network namespace
- Move the address lists into struct net
- Add per network namespace initialization and cleanup
- Pass around struct net so it is everywhere I need it.
- Rename all of the global variable references into references
to the variables moved into struct net
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Make the association hashtable handle multiple network namespaces
- Use struct net in the hash calculation
- Use sock_net(association.base.sk) in the association lookups.
- On receive calculate the network namespace from skb->dev.
- Pass struct net from receive down to the functions that actually
do the association lookup.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Make the endpoint hashtable handle multiple network namespaces
- Use struct net in the hash calculation
- Use sock_net(endpoint.base.sk) in the endpoint lookups.
- On receive calculate the network namespace from skb->dev.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: Make the port hash table use struct net in it's key.
- Add struct net into the port hash table hash calculation
- Add struct net inot the struct sctp_bind_bucket so there
is a memory of which network namespace a port is allocated in.
No need for a ref count because sctp_bind_bucket only exists
when there are sockets in the hash table and sockets can not
change their network namspace, and sockets already ref count
their network namespace.
- Add struct net into the key comparison when we are testing
to see if we have found the port hash table entry we are
looking for.
With these changes lookups in the port hash table becomes
safe to use in multiple network namespaces.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Mon, 13 Aug 2012 05:57:44 +0000 (05:57 +0000)]
packet: Report socket mclist info via diag module
The info is reported as an array of packet_diag_mclist structures. Each
includes not only the directly configured values (index, type, etc), but
also the "count".
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Mon, 13 Aug 2012 05:53:28 +0000 (05:53 +0000)]
packet: Diag core and basic socket info dumping
The diag module can be built independently from the af_packet.ko one,
just like it's done in unix sockets.
The core dumping message carries the info available at socket creation
time, i.e. family, type and protocol (in the same byte order as shown in
the proc file).
The socket inode number and cookie is reserved for future per-socket info
retrieving. The per-protocol filtering is also reserved for future by
requiring the sdiag_protocol to be zero.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Mon, 13 Aug 2012 05:49:39 +0000 (05:49 +0000)]
packet: Introduce net/packet/internal.h header
The diag module will need to access some private packet_sock data, so
move it to a header in advance. This file will be shared between the
af_packet.c and the diag.c
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Priyanka Jain [Sun, 12 Aug 2012 21:22:29 +0000 (21:22 +0000)]
Replace rwlock on xfrm_policy_afinfo with rcu
xfrm_policy_afinfo is read mosly data structure.
Write on xfrm_policy_afinfo is done only at the
time of configuration.
So rwlocks can be safely replaced with RCU.
RCUs usage optimizes the performance.
Signed-off-by: Priyanka Jain <Priyanka.Jain@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Amerigo Wang [Thu, 9 Aug 2012 22:14:56 +0000 (22:14 +0000)]
net: move and rename netif_notify_peers()
I believe net/core/dev.c is a better place for netif_notify_peers(),
because other net event notify functions also stay in this file.
And rename it to netdev_notify_peers().
Cc: David S. Miller <davem@davemloft.net> Cc: Ian Campbell <Ian.Campbell@citrix.com> Signed-off-by: Cong Wang <amwang@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
My commit "p54: parse output power table" introduced
the following compiler warnings for powerpc-allmodconfig
eeprom.c: In function 'p54_get_maxpower':
eeprom.c:291 warning: comparison of distinct pointer types lacks a cast
eeporm.c:292 warning: comparison of distinct pointer types lacks a cast
eeprom.c:293 warning: comparison of distinct pointer types lacks a cast
eeprom.c:294 warning: comparison of distinct pointer types lacks a cast
This patch fixes those by using max_t(u16
which forces a type cast.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Tested-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Christian Lamparter <chunkeey@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Marco Porsch [Wed, 8 Aug 2012 05:58:43 +0000 (07:58 +0200)]
mac80211: fix unnecessary beacon update after peering status change
ieee80211_bss_info_change_notify is called everytime a peer link is established
or closed, because the accepting_plinks flag in the meshconf IE *might* have changed.
With this patch the corresponding functions return the BSS_CHANGED_BEACON flag when a beacon update is necessary.
Also it makes mesh_accept_plinks_update the common place to update the accepting_plinks flag.
mesh_accept_plinks_update is called upon plink change and also periodically from ieee80211_mesh_housekeeping.
Thus, it also picks up changes of local->num_sta.
Signed-off-by: Marco Porsch <marco.porsch@etit.tu-chemnitz.de> Acked-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jeff Mahoney [Mon, 6 Aug 2012 19:17:26 +0000 (15:17 -0400)]
brcmsmac: document firmware dependencies
The brcmsmac driver requests firmware but doesn't document the
dependency. This means that software that analyzes the modules to
determine if firmware is needed won't detect it.
Specifically, (at least) openSUSE won't install the kernel-firmware
package if no hardware requires it.
This patch adds the MODULE_FIRMWARE directives.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Fri, 10 Aug 2012 11:17:23 +0000 (16:47 +0530)]
ath9k: Trim rate table
Remove ctrl_rate, cw40index, sgi_index, ht_index and calculate
the rate index for TX status from the valid_rate_index that
is populated at initialization time.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Fri, 10 Aug 2012 11:16:57 +0000 (16:46 +0530)]
ath9k: Fix RTS/CTS rate selection
The current method of assigning the RTS/CTS rate is completely
broken for HT mode and breaks P2P operation. Fix this by using
the basic_rates provided to the driver by mac80211. For now,
choose the lowest supported basic rate for HT frames.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Fri, 10 Aug 2012 11:16:31 +0000 (16:46 +0530)]
ath9k: Change rateset calculation
Commit "ath9k: Change rate control to use legacy rate as last MRR"
resulted in the mixing of HT/legacy rates in a single rateset,
which is undesirable. Revert this behavior.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Dan Carpenter [Thu, 9 Aug 2012 06:57:57 +0000 (09:57 +0300)]
mwifiex: use GFP_ATOMIC under spin lock
We're holding the sta_list_spinlock here so we can't sleep.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Bing Zhao [Thu, 9 Aug 2012 02:01:52 +0000 (19:01 -0700)]
mwifiex: fix 'smatch' warning in preparing key_material cmd
The key length can be 32 bytes for TKIP and 16 bytes for AES_CMAC.
'smatch' warns on memcpy using key_len variable to copy data to
a 16 bytes buffer. Use fixed length to avoid the warning.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Wed, 8 Aug 2012 17:37:04 +0000 (19:37 +0200)]
ssb: check for flash presentence
We can not assume parallel flash is always present, there are boards
with *serial* flash and probably some without flash at all.
Define some bits by the way.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com> Reviewed-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
These warnings can be detected by using powerpc64-linux toolchain
(gcc-4.6.3-nolibc).
CC [M] drivers/net/wireless/mwifiex/sta_event.o
drivers/net/wireless/mwifiex/sta_event.c: In function 'mwifiex_process_sta_event':
drivers/net/wireless/mwifiex/sta_event.c:388:4: warning: comparison of distinct pointer types lacks a cast [enabled by default]
CC [M] drivers/net/wireless/mwifiex/uap_event.o
drivers/net/wireless/mwifiex/uap_event.c: In function 'mwifiex_process_uap_event':
drivers/net/wireless/mwifiex/uap_event.c:258:11: warning: comparison of distinct pointer types lacks a cast [enabled by default]
Use min_t() instead of min() to fix the warnings.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Tue, 7 Aug 2012 19:46:44 +0000 (21:46 +0200)]
iwlegacy: clean up suspend/resume
There's no need to export the il_pci_suspend
and il_pci_resume functions since they're only
referenced from il_pm_ops. The latter can also
be defined using SIMPLE_DEV_PM_OPS instead of
open-coding it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Nick Kossifidis [Sun, 5 Aug 2012 19:35:34 +0000 (22:35 +0300)]
ath5k: Fix range scaling when setting rate power table
rates[i] is unsigned but txp_offset can be negative for newer parts
with PDADC table. We cover the case when rates[i] + txp_offset > 63
but we must also cover the case when its < 0 or else rates[i] will overflow.
Signed-off-by: Nick Kossifidis <mickflemm@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Nick Kossifidis [Sun, 5 Aug 2012 19:35:33 +0000 (22:35 +0300)]
ath5k: Use correct value for min_pwr and cur_pwr
Make sure we don't store the table offsets for min and cur power levels,
store the 0.25dB values instead. This way we don't clamp the tx power level
to max (because now cur_pwr holds the 0.25dB value, not the table offset) after
re-using cur_pwr on reset.
Signed-off-by: Nick Kossifidis <mickflemm@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Hauke Mehrtens [Sun, 5 Aug 2012 14:54:41 +0000 (16:54 +0200)]
bcma: make some functions static
The functions and structs are not used in an other file and the
prototypes are in no header file, just make them static so the compiler
is able to optimize them better.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
For the upcoming tpc changes, the driver needs
to provide sensible max output values for each
supported channel.
And while the eeprom always had a output_limit
table, which defines the upper limit for each
frequency and modulation, it was never really
useful for anything... until now.
Note: For anyone wondering about what your card
is calibrated for: check "iw list".
* 2412 MHz [1] (18.0 dBm)
* 2437 MHz [6] (19.0 dBm)
[...]
* 5180 MHz [36] (18.0 dBm)
* 5260 MHz [52] (17.0 dBm) (radar detection)
* 5680 MHz [136] (19.0 dBm) (radar detection)
(for a Dell Wireless 1450 USB Adapter)
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Pavel Emelyanov [Wed, 8 Aug 2012 21:53:36 +0000 (21:53 +0000)]
net: Loopback ifindex is constant now
As pointed out, there are places, that access net->loopback_dev->ifindex
and after ifindex generation is made per-net this value becomes constant
equals 1. So go ahead and introduce the LOOPBACK_IFINDEX constant and use
it where appropriate.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 8 Aug 2012 21:53:19 +0000 (21:53 +0000)]
net: Make ifindex generation per-net namespace
Strictly speaking this is only _really_ required for checkpoint-restore to
make loopback device always have the same index.
This change appears to be safe wrt "ifindex should be unique per-system"
concept, as all the ifindex usage is either already made per net namespace
of is explicitly limited with init_net only.
There are two cool side effects of this. The first one -- ifindices of
devices in container are always small, regardless of how many containers
we've started (and re-started) so far. The second one is -- we can speed
up the loopback ifidex access as shown in the next patch.
v2: Place ifindex right after dev_base_seq : avoid two holes and use the
same cache line, dirtied in list_netdevice()/unlist_netdevice()
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 8 Aug 2012 21:53:03 +0000 (21:53 +0000)]
veth: Allow to create peer link with given ifindex
The ifinfomsg is in there (thanks kaber@ for foreseeing this long time ago),
so take the given ifidex and register netdev with it.
Ben noticed, that this code path previously ignored ifmp->ifi_index and
userland could be passing in garbage. Thus it may now fail occasionally
because the value clashes with an existing interface.
To address this it's assumed that if the caller specifies the ifindex for
the veth master device, then it's aware of this possibility and should
explicitly specify (or set to 0 for auto-assignment) the peer's ifindex as
well. With this the compatibility with old tools not setting ifindex is
preserved.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 8 Aug 2012 21:52:46 +0000 (21:52 +0000)]
net: Allow to create links with given ifindex
Currently the RTM_NEWLINK results in -EOPNOTSUPP if the ifinfomsg->ifi_index
is not zero. I propose to allow requesting ifindices on link creation. This
is required by the checkpoint-restore to correctly restore a net namespace
(i.e. -- a container).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 8 Aug 2012 21:52:28 +0000 (21:52 +0000)]
net: Dont use ifindices in hash fns
Eric noticed, that when there will be devices with equal indices, some
hash functions that use them will become less effective as they could.
Fix this in advance by mixing the net_device address into the hash value
instead of the device index.
This is true for arp and ndisc hash fns. The netlabel, can and llc ones
are also ifindex-based, but that three are init_net-only, thus will not
be affected.
Many thanks to David and Eric for the hash32_ptr implementation!
Signed-off-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
commit cbbc719fccdb8cb (time: Change jiffies_to_clock_t() argument type
to unsigned long) only got around the problem.
As we cant output negative values in /proc/net/tcp without breaking
various tools, I suggest adding a jiffies_delta_to_clock_t() wrapper
that caps the negative delta to a 0 value.
Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Maciej Żenczykowski <maze@google.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Paul Gortmaker <paul.gortmaker@windriver.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: hank <pyu@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 7 Aug 2012 02:19:56 +0000 (02:19 +0000)]
net: output path optimizations
1) Avoid dirtying neighbour's confirmed field.
TCP workloads hits this cache line for each incoming ACK.
Lets write n->confirmed only if there is a jiffie change.
2) Optimize neigh_hh_output() for the common Ethernet case, were
hh_len is less than 16 bytes. Replace the memcpy() call
by two inlined 64bit load/stores on x86_64.
Bench results using udpflood test, with -C option (MSG_CONFIRM flag
added to sendto(), to reproduce the n->confirmed dirtying on UDP)
Eric Dumazet [Mon, 6 Aug 2012 11:04:43 +0000 (11:04 +0000)]
tcp: ecn: dont delay ACKS after CE
While playing with CoDel and ECN marking, I discovered a
non optimal behavior of receiver of CE (Congestion Encountered)
segments.
In pathological cases, sender has reduced its cwnd to low values,
and receiver delays its ACK (by 40 ms).
While RFC 3168 6.1.3 (The TCP Receiver) doesn't explicitly recommend
to send immediate ACKS, we believe its better to not delay ACKS, because
a CE segment should give same signal than a dropped segment, and its
quite important to reduce RTT to give ECE/CWR signals as fast as
possible.
Note we already call tcp_enter_quickack_mode() from TCP_ECN_check_ce()
if we receive a retransmit, for the same reason.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Neal Cardwell <ncardwell@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Sun, 5 Aug 2012 22:34:50 +0000 (22:34 +0000)]
net: tcp: GRO should be ECN friendly
While doing TCP ECN tests, I discovered GRO was reordering packets if it
receives one packet with CE set, while previous packets in same NAPI run
have ECT(0) for the same flow :
09:25:25.857620 IP (tos 0x2,ECT(0), ttl 64, id 27893, offset 0, flags
[DF], proto TCP (6), length 4396)
172.30.42.19.54550 > 172.30.42.13.44139: Flags [.], seq
233801:238145, ack 1, win 115, options [nop,nop,TS val 3397779 ecr 1990627], length 4344
09:25:25.857626 IP (tos 0x3,CE, ttl 64, id 27892, offset 0, flags [DF],
proto TCP (6), length 1500)
172.30.42.19.54550 > 172.30.42.13.44139: Flags [.], seq
232353:233801, ack 1, win 115, options [nop,nop,TS val 3397779 ecr 1990627], length 1448
If NIC gave packet1, then packet2, which happen to be from "different
flows" GRO feeds stack with packet2, then packet1. I have yet to
understand how to solve this problem.
2) GRO is not ECN friendly
Delivering packets out of order makes TCP stack not as fast as it could
be.
In this patch I suggest we make the tos test not part of the 'same_flow'
determination, but part of the 'should flush' logic
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Ying Luo [Sat, 4 Aug 2012 01:06:14 +0000 (18:06 -0700)]
mwifiex: add AES_CMAC support in key_material cmd
The sequence counter will be sent to firmware via key_material
command.
Signed-off-by: Ying Luo <luoy@marvell.com> Signed-off-by: Stone Piao <piaoyun@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ying Luo [Sat, 4 Aug 2012 01:06:13 +0000 (18:06 -0700)]
mwifiex: pass key_params pointer in mwifiex_set_encode
'cipher' and 'seq' coming from cfg80211 add_key handler will be
parsed in mwifiex_set_encode() to handle AES_CMAC cipher suite.
Signed-off-by: Ying Luo <luoy@marvell.com> Signed-off-by: Stone Piao <piaoyun@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ying Luo [Sat, 4 Aug 2012 01:06:12 +0000 (18:06 -0700)]
mwifiex: rename wapi_rxpn to pn
This array was used for wapi_rxpn only. Now it will be used for
AES_CMAC as well. So make a generic name for it.
Signed-off-by: Ying Luo <luoy@marvell.com> Signed-off-by: Stone Piao <piaoyun@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Avinash Patil [Sat, 4 Aug 2012 01:06:11 +0000 (18:06 -0700)]
mwifiex: cleanup TX/RX BA tables for uAP
Cleanup TX/RX BA tables when AP receives deauthentication from
associated station. During BSS_IDLE event, all wmm queues, BA
streams created for AP interface are deleted.
Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Kiran Divekar <dkiran@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Avinash Patil [Sat, 4 Aug 2012 01:06:10 +0000 (18:06 -0700)]
mwifiex: add 11n Block Ack support for uAP
This patch adds support for handling BA request and BA setup
events for AP interface.
RA list is marked as either 11n enabled or disabled from station's
capabilities in association request. BA setup is initiated only
after some specific number of packets for particular RA list are
transmitted.
Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Kiran Divekar <dkiran@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Avinash Patil [Sat, 4 Aug 2012 01:06:09 +0000 (18:06 -0700)]
mwifiex: support RX AMSDU aggregation for uAP
This patch adds support for reception and decoding of AMSDU
aggregation frames for AP interface.
Patch also adds support for handling AMSDU aggregation event.
Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Kiran Divekar <dkiran@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Avinash Patil [Sat, 4 Aug 2012 01:06:08 +0000 (18:06 -0700)]
mwifiex: improve uAP RX handling
1. Separate file for uAP RX handling.
2. If received packet is broadcast/multicast, send it to kernel
as well as requeue it back to uAP TX queue.
3. If received packet is for associated STA (intra-BSS), requeue
it back to uAP TX queue.
4. In all other cases (packets for AP or inter-BSS packets),
pass packet to kernel to handle it accordingly.
Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Kiran Divekar <dkiran@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Avinash Patil [Sat, 4 Aug 2012 01:06:07 +0000 (18:06 -0700)]
mwifiex: create list for associated stations in AP mode
After station is associated an entry would be added to station
list. This entry would have station specific information such as
11n support, AMSDU size. The entry would be deleted after
deauthentication. All station entries would be deleted during
BSS_IDLE event.
Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Kiran Divekar <dkiran@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
mwifiex: improve scan delay logic during Tx traffic
Earlier scan command was postponed by 20msec multiple times to give
preference to Tx traffic until we find empty wmm queue.
There is a corner case in which wmm queue becomes empty immediately
after processing the packet(before 20msec) and there may be next
packet coming after some time. In this case we should not resume scan
operation.
We will use new flag to check Tx traffic and resume scan operation
only if there is no traffic for 200msec.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
mwifiex: do le_to_cpu conversion for Rx packet header elements
We do similar conversion for Tx packet header but it was missing
for Rx packet header. Without this fix, Rx packet header won't
be correctly parsed on big endian platform.
Padding arithmetic will always work for MSDUs provided first MSDU
ends on 4-byte boundary. Fixing it by making sure that all MSDU ends
on 4-byte boundary.
Since commit: fb3c19bc9687d972b83faf366b38ac807eca8f25, adding
extra padding to AMSDU is redundant since same is being performed at
mwifiex_11n_aggregate_pkt after forming the AMSDU packet. Fixing it
by removing it.
All MSDU subframes, except the first one in AMSDU, were being written
wrongly at the location of first MSDU frame. Fixing that by copying
the MSDU at skb->tail of AMSDU.