Samuel Ortiz [Tue, 1 Sep 2009 13:14:05 +0000 (15:14 +0200)]
iwmc3200wifi: Handle UMAC stalls and UMAC assert properly
When UMAC stalls or asserts, we want to reset the device. But when we're
associated, the current reset worker will end up calling
cfg80211_connect_result() with the cfg80211 sme layer knowing that we're
reassociating. That ends up with some ugly warnings.
With this patch we're telling the upper layer that we've roamed if
reassociation succeeds, and that we're disconnected if it fails.
Samuel Ortiz [Tue, 1 Sep 2009 13:14:04 +0000 (15:14 +0200)]
iwmc3200wifi: New initial LMAC calibration
The LMAC calibration API got broken mostly by having a configuration bitmap
being different than the result one.
This patch tries to address that issue by correctly running calibrations with
the newest firmwares, and keeping a backward compatibility fallback path for
older firmwares, where the configuration and result bitmaps were identical.
Zhu Yi [Tue, 1 Sep 2009 13:14:02 +0000 (15:14 +0200)]
iwmc3200wifi: add disconnect work
When the driver receives "connection terminated" event from device,
it could be caused by 2 reasons: the firmware is roaming or the
connection is lost (AP disappears). For the former, an association
complete event is supposed to come within 3 seconds. For the latter,
the driver won't receive any event except the connection terminated.
So we kick a delayed work (5*HZ) when we receive the connection
terminated event. It will be canceled if it turns out to be a roaming
event later. Otherwise we notify SME and userspace the disconnection.
Zhu Yi [Tue, 1 Sep 2009 13:14:01 +0000 (15:14 +0200)]
iwmc3200wifi: use cfg80211_roamed to send roam event
The device sends connection terminated and [re]association success
(or failure) events when roaming occours. The patch uses
cfg80211_roamed instead of cfg80211_connect_result to notify SME
for roaming.
Samuel Ortiz [Tue, 1 Sep 2009 13:13:59 +0000 (15:13 +0200)]
iwmc3200wifi: Set WEP key from connect
When connect is called with the LEGACY_PSK authentication type set, and a
proper sme->key, we need to set the WEP key straight after setting the
profile otherwise the authentication will never start.
Jussi Kivilinna [Tue, 1 Sep 2009 12:33:00 +0000 (15:33 +0300)]
rndis_wlan: cleanup
- remove double newlines between functions
- remove commented out function (rndis_set_config_parameter_u32())
- coding style fix in rndis_set_config_parameter_str()
- add comment banners between function sections
Ivo van Doorn [Sat, 29 Aug 2009 18:30:45 +0000 (20:30 +0200)]
rt2x00: Reorganize padding & L2 padding
The old function rt2x00queue_payload_align() handled
both adding and removing L2 padding and some basic
frame alignment. The entire function was being abused
because it had multiple functions and the header length
argument was somtimes used to align the header instead
of the payload.
Additionally there was a bug when inserting L2 padding
that only the payload was aligned but not the header. This
happens when the header wasn't aligned properly by mac80211,
but rt2x00lib only moves the payload.
A secondary problem was that when removing L2 padding during
TXdone or RX the skb wasn't resized to the proper size.
Split the function into seperate functions each handling
its task as it should.
Ivo van Doorn [Sat, 29 Aug 2009 17:10:14 +0000 (19:10 +0200)]
rt2x00: Fix TX status reporting
Not all values of the TX status enumeration were
covered during updating of the TX statistics. This
could lead to wrong bitrate tuning but also wrong
behavior in tools like hostapd.
Reinette Chatre [Fri, 28 Aug 2009 16:58:50 +0000 (09:58 -0700)]
iwl3945: reduce debug noise when default debug flags used
Significant literature suggests users use debug flags 0x43fff - this causes
the debug flags to be set that causes information to be printed for every
received frame - including beacons. In the best case it fills up the logs,
at worst it slows driver down and causes failures due to timeouts.
In the RX handler, print debugging only if user requested RX debugging.
Reinette Chatre [Fri, 28 Aug 2009 16:44:49 +0000 (09:44 -0700)]
iwlwifi: fix situation in which debug message is printed
3945 does not have update_chain_flags defined and because if this we always
see the debug message that does not apply to it. Add a check to be specific
about what is actually happening.
Johannes Berg [Fri, 28 Aug 2009 16:44:48 +0000 (09:44 -0700)]
iwlwifi: use sleep interval succession
Some concerns were raised about the automatic adjustment
of sleep intervals to all the same, potentially high,
value, and I can imagine the hardware behaving better
when we don't ask too much of it.
So let's convert to use a succession of sleep levels
when requesting to go to deeper sleeps (which can only
happen with large DTIM intervals), using the succession
values from power level three, which have the benefit of
also having been tested extensively already.
As a result, the automatic sleep level adjustment will
now be mostly equivalent to power level three, except
for the RX/TX timeouts and possibly using smaller sleep
vectors to account for networking latency.
iwlwifi: use iwl_hwrate_get_mac80211_idx where appropriate
For HT packets, mac80211 expects the rate_idx to be an MCS number, which is the
lower byte of rate_n_flags. However, iwl_hwrate_to_plcp_idx takes the MCS
number and reduces it down to the range 0-8 (6 to 60 Mbps), removing the bits
that signify multiply streams, HT40 Duplicate mode, or unequal modulation.
This version is used for various internal purposes through the driver.
Add the function iwl_hwrate_get_mac80211_idx, an alternate version which takes
the rate and the band and returns the mac80211 index (MCS, for HT packets, and
PLCP rate, for legacy packets).
iwlwifi: remove incorrect uses of ieee80211_get_tx_rate to prevent TX stall
Refactor and correct rate selection for outgoing transmitted
packets.
First, note that HT rates in the mac80211 rate table do not provide valid
indices when ieee80211_get_tx_rate is called; the check to see if we could to
abort a transmission early in iwl_tx_skb() would thus occasionally read invalid
memory and occasionally stall transmission (if the erroneous byte was 0xff).
We remove that code; the check wasn't valid anyway.
Second, iwl_tx_cmd_build_rate() also called ieee80211_get_tx_rate to be used
for sending management packets, which do not use the uCode station table. This
patch refactors that function and adds comments to enhance legibility, replaces
the call to ieee80211_get_tx_rate() with a direct lookup, and adds error
handling in case the table entry is invalid.
ieee80211_supported_band is supposed to only contain legacy rates in the
bitrates table (HT rates go in the ieee80211_sta_ht_cap substruct). Make
iwlwifi driver obey this restriction by removing the 60 Mbps rate. Also, clean
up a few pieces of other code that formerly relied on 60 Mbps being in
sband->bitrates.
Joe Perches [Mon, 24 Aug 2009 17:29:47 +0000 (17:29 +0000)]
s2io.c: Standardize statistics accessors
Regularize the declaration and uses of
struct config_param *config = &sp->config;
struct mac_info *mac_control = &sp->mac_control;
and use
struct stat_block *stats = mac_control->stats_info;
struct swStat *swstats = &stats->sw_stat;
struct xpakStat *xstats = &stats->xpak_stat;
and convert the longish uses like
nic->mac_control.stats_info->sw_stat.<foo>
to
swstats-><foo>
etc.
This also makes the statistics code marginally smaller
and presumably faster.
Old:
$ size s2io.o
text data bss dec hex filename
114289 516 33360 148165 242c5 s2io.o
New:
$ size s2io.o
text data bss dec hex filename
114097 516 33360 147973 24205 s2io.o
Krishna Kumar [Mon, 31 Aug 2009 05:20:28 +0000 (22:20 -0700)]
pkt_sched: Fix resource limiting in pfifo_fast
pfifo_fast_enqueue has this check:
if (skb_queue_len(list) < qdisc_dev(qdisc)->tx_queue_len) {
which allows each band to enqueue upto tx_queue_len skbs for a
total of 3*tx_queue_len skbs. I am not sure if this was the
intention of limiting in qdisc.
Anton Vorontsov [Thu, 27 Aug 2009 07:35:57 +0000 (07:35 +0000)]
ucc_geth: Implement suspend/resume and Wake-On-LAN support
This patch implements suspend/resume and WOL support for UCC Ethernet
driver.
We support two wake up events: wake on PHY/link changes and wake
on magic packet.
In some CPUs (like MPC8569) QE shuts down during sleep, so magic packet
detection is unusable, and also on resume we should fully reinitialize
UCC structures.
Anton Vorontsov [Thu, 27 Aug 2009 07:35:50 +0000 (07:35 +0000)]
powerpc/qe: Implement qe_alive_during_sleep() helper function
In some CPUs (i.e. MPC8569) QE shuts down completely during sleep,
drivers may want to know that to reinitialize registers and buffer
descriptors.
This patch implements qe_alive_during_sleep() helper function, so far
it just checks if MPC8569-compatible power management controller is
present, which is a sign that QE turns off during sleep.
Anton Vorontsov [Thu, 27 Aug 2009 07:35:47 +0000 (07:35 +0000)]
ucc_geth: Fix NULL pointer dereference in uec_get_ethtool_stats()
In commit 3e73fc9a12679a546284d597c1f19165792d0b83 ("ucc_geth: Fix IO
memory (un)mapping code") I fixed ug_regs IO memory leak by properly
freeing the allocated memory. But ethtool_stats() callback doesn't
check for ug_regs being NULL, and that causes following oops if
'ethtool -S' is executed on a closed eth device:
Unable to handle kernel paging request for data at address 0x00000180
Faulting instruction address: 0xc0208228
Oops: Kernel access of bad area, sig: 11 [#1]
...
NIP [c0208228] uec_get_ethtool_stats+0x38/0x140
LR [c02559a0] ethtool_get_stats+0xf8/0x23c
Call Trace:
[ef87bcd0] [c025597c] ethtool_get_stats+0xd4/0x23c (unreliable)
[ef87bd00] [c025706c] dev_ethtool+0xfe8/0x11bc
[ef87be00] [c0252b5c] dev_ioctl+0x454/0x6a8
...
---[ end trace 77fff1162a9586b0 ]---
Segmentation fault
Matt Carlson [Fri, 28 Aug 2009 14:03:21 +0000 (14:03 +0000)]
tg3: Move per-int tx members to a per-int struct
This patch moves the tx_prod, tx_cons, tx_pending, tx_ring, and
tx_buffers transmit ring device members to a per-interrupt structure.
It also adds a new transmit producer mailbox member (prodmbox) and
converts the code to use it rather than a preprocessor constant.
Matt Carlson [Fri, 28 Aug 2009 14:03:01 +0000 (14:03 +0000)]
tg3: Move per-int rx members to per-int struct
This patch moves the rx_rcb, rx_rcb_mapping, and rx_rcb_ptr return ring
device members to a per-interrupt structure. It also adds a new return
ring consumer mailbox register member (consmbox) and converts the code
to use it rather than a preprocessor constant.
Matt Carlson [Fri, 28 Aug 2009 14:02:40 +0000 (14:02 +0000)]
tg3: Move general int members to a per-int struct
This patch moves the last_tag, last_tag_irq, and hw_status device
members to a per-interrupt structure. It also adds a new interrupt
mailbox member (int_mbox) and converts the code to use it rather than a
direct preprocessor constant.
Matt Carlson [Fri, 28 Aug 2009 14:01:57 +0000 (14:01 +0000)]
tg3: Convert ISR parameter to tnapi
This patch migrates the ISR parameter from struct net_device to struct
tg3_napi. Checkpatch complains about the existence of the preexisting
IRQF_SAMPLE_RANDOM flag. I've opted to keep this patch conservative and
let it continue to exist until the flag gets officially purged from the
kernel.
Matt Carlson [Fri, 28 Aug 2009 14:01:15 +0000 (14:01 +0000)]
tg3: Cleanup interrupt setup / teardown
Later patches will be adding MSIX support, which will complicate
interrupt initialization. This patch prepares for the integration by
breaking out the interrupt setup and teardown code into separate
functions and cleaning up the error return paths.
Matt Carlson [Fri, 28 Aug 2009 14:00:55 +0000 (14:00 +0000)]
tg3: Use ext rx bds
The 5717 only uses extended buffer descriptors for the jumbo producer
ring. Extended buffer descriptors are available on all devices that
support a separate jumbo producer ring so make the change universal.
Matt Carlson [Fri, 28 Aug 2009 13:59:57 +0000 (13:59 +0000)]
tg3: Create rx producer ring setup routines
Later patches are going to complicate the ring initialization routines.
This patch breaks out the setup and teardown of the rx producer rings
into separate functions to make the code more readable.
Matt Carlson [Fri, 28 Aug 2009 13:57:12 +0000 (13:57 +0000)]
tg3: Break out mini producer ring handling
This patch separates the code that sets up the mini producer ring from
the code that sets up the jumbo producer rings. The 5717 asic rev
devices do not have a mini ring, but do have a jumbo frame
implementation similar to the 5704 and previous devices.
Matt Carlson [Fri, 28 Aug 2009 12:27:50 +0000 (12:27 +0000)]
tg3: Delay mdio bus init until fw finishes
The device firmware uses the MDIO bus during early setup. If the driver
modifies the MDIO bus configuration while it is in use by the firmware,
any number of bad things can happen. This patch delays MDIO setup until
after the firmware posts its magic signature, signifying initialization
is complete.
net: sh_eth: add value of ether_link pin in platform_data
The method of ETHER_LINK pin is board dependence.
This patch adding paramters are:
- no_ether_link : If set to 1, do not use ETHER_LINK
- ether_link_active_low : If set to 1, ETHER_LINK is active low.
TI DaVinci EMAC: delay DaVinci EMAC initialization
On TI's DA850/OMAP-L138 EVM, MAC address is stored in SPI
flash which is accessed using MTD interface.
This patch delays the initialization of DaVinci EMAC driver
by changing module_init to late_initcall. This helps SPI and
MTD drivers to get initialized before EMAC thereby enabling
EMAC driver to read the MAC address while booting and use it.
Tested with NFS on DM644x, DM6467, DA830/OMAP-L137 and
DA850/OMAP-L138 EVMs.
John Dykstra [Wed, 19 Aug 2009 09:47:41 +0000 (09:47 +0000)]
tcp: Remove redundant copy of MD5 authentication key
Remove the copy of the MD5 authentication key from tcp_check_req().
This key has already been copied by tcp_v4_syn_recv_sock() or
tcp_v6_syn_recv_sock().
Krishna Kumar [Tue, 18 Aug 2009 21:55:59 +0000 (21:55 +0000)]
Speed-up pfifo_fast lookup using a private bitmap
Maintain a per-qdisc bitmap for pfifo_fast giving availability
of skbs for each band. This allows faster lookup for a skb when
there are no high priority skbs. Also, it helps in (rare) cases
when there are no skbs on the list, where an immediate lookup is
faster than iterating through the three bands.
David Ward [Sat, 29 Aug 2009 07:04:09 +0000 (00:04 -0700)]
ipv6: Update Neighbor Cache when IPv6 RA is received on a router
When processing a received IPv6 Router Advertisement, the kernel
creates or updates an IPv6 Neighbor Cache entry for the sender --
but presently this does not occur if IPv6 forwarding is enabled
(net.ipv6.conf.*.forwarding = 1), or if IPv6 Router Advertisements
are not accepted (net.ipv6.conf.*.accept_ra = 0), because in these
cases processing of the Router Advertisement has already halted.
This patch allows the Neighbor Cache to be updated in these cases,
while still avoiding any modification to routes or link parameters.
This continues to satisfy RFC 4861, since any entry created in the
Neighbor Cache as the result of a received Router Advertisement is
still placed in the STALE state.
Octavian Purdila [Sat, 29 Aug 2009 07:00:35 +0000 (00:00 -0700)]
tcp: fix premature termination of FIN_WAIT2 time-wait sockets
There is a race condition in the time-wait sockets code that can lead
to premature termination of FIN_WAIT2 and, subsequently, to RST
generation when the FIN,ACK from the peer finally arrives:
Say twdr->slot = 1 and we are running inet_twdr_hangman and in this
instance inet_twdr_do_twkill_work returns 1. At that point we will
mark slot 1 and schedule inet_twdr_twkill_work. We will also make
twdr->slot = 2.
Next, a connection is closed and tcp_time_wait(TCP_FIN_WAIT2, timeo)
is called which will create a new FIN_WAIT2 time-wait socket and will
place it in the last to be reached slot, i.e. twdr->slot = 1.
At this point say inet_twdr_twkill_work will run which will start
destroying the time-wait sockets in slot 1, including the just added
TCP_FIN_WAIT2 one.
To avoid this issue we increment the slot only if all entries in the
slot have been purged.
This change may delay the slots cleanup by a time-wait death row
period but only if the worker thread didn't had the time to run/purge
the current slot in the next period (6 seconds with default sysctl
settings). However, on such a busy system even without this change we
would probably see delays...
Jens Låås [Sat, 29 Aug 2009 06:57:15 +0000 (23:57 -0700)]
fib_trie: resize rework
Here is rework and cleanup of the resize function.
Some bugs we had. We were using ->parent when we should use
node_parent(). Also we used ->parent which is not assigned by
inflate in inflate loop.
Also a fix to set thresholds to power 2 to fit halve
and double strategy.
max_resize is renamed to max_work which better indicates
it's function.
Reaching max_work is not an error, so warning is removed.
max_work only limits amount of work done per resize.
(limits CPU-usage, outstanding memory etc).
The clean-up makes it relatively easy to add fixed sized
root-nodes if we would like to decrease the memory pressure
on routers with large routing tables and dynamic routing.
If we'll need that...
Sascha Hlusiak [Sat, 29 Aug 2009 06:53:53 +0000 (23:53 -0700)]
sit: allow ip fragmentation when using nopmtudisc to fix package loss
if tunnel parameters have frag_off set to IP_DF, pmtudisc on the ipv4 link
will be performed by deriving the mtu from the ipv4 link and setting the
DF-Flag of the encapsulating IPv4 Header. If fragmentation is needed on the
way, the IPv4 pmtu gets adjusted, the ipv6 package will be resent eventually,
using the new and lower mtu and everyone is happy.
If the frag_off parameter is unset, the mtu for the tunnel will be derived
from the tunnel device or the ipv6 pmtu, which might be higher than the ipv4
pmtu. In that case we must allow the fragmentation of the IPv4 packet because
the IPv6 mtu wouldn't 'learn' from the adjusted IPv4 pmtu, resulting in
frequent icmp_frag_needed and package loss on the IPv6 layer.
This patch allows fragmentation when tunnel was created with parameter
nopmtudisc, like in ipip/gre tunnels.
If not using delay then no need to update next_tx after
each packet sent. This allows pktgen to send faster especially
on systems with slower clock sources.
Do some reorganization of transmit logic path:
* move transmit queue full idle to separate routine
* add a cpu_relax()
* eliminate some of the uneeded goto's
* if queue is still stopped, go back to main thread loop.
* don't give up transmitting if quantum is exhausted (be greedy)
projects / linux.git / log