This simplifies both the transport layer and the upper layer.
Kill the union in the device command, which avoids the funny syntax
we had: cmd->cmd.payload.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Before this patch, the upper layer could register a callback for each
host command. This mechanism allowed the upper layer to have
different callbacks for the same command ID. In fact, it wasn't used
and the rx_handlers is enough: same callback for all the command with
a specific command ID.
The iwl_send_add_station needs the access the command that was sent
while handling the response (regardless if the command was sent in
SYNC or ASYNC mode). So now, all the handlers receive the host
command that was sent. This implies a change in the handler signature.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
and do a few clean up fixes on the way
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Trivial fixes to allow compilation without warnings when debug
compilation flag isn't set.
Also fix the compilation when debugfs flag isn't set.
Fix a warning: unused priv pointer on the way.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This finalizes the move of the data path to the transport layer.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
priv->mac80211_registered and priv->hw needed to move to shared.
stop_queue API was added in order to allow the upper layer to stop
the SW queues for regulatory purposes.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since all the check_empty logic is now in the transport layer,
the upper layer doesn't need to know anything about tx queues.
The disable aggregation flow was the last to know what a tx queue
is, so move it too.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This logic is responsible to tell mac80211 when the HW queues are
empty and the BA session can be started / torn down.
Fix a bug on the way:
When the the Tx BA session is stopped and the HW queues aren't empty,
we stop the SW queue to drain the HW queue and then switch to the
legacy HW queue. This is the IWL_EMPTYING_HW_QUEUE_DELBA state.
While in this state, we never wake the SW queue, even when the HW
queue is almost empty, since we need to drain it completely. Look
at iwl_trans_pcie_reclaim regarding this.
Once the HW queue is really empty, we must wake the SW queue in order
to get traffic to the legacy queue.
This step was missing leading to an odd situation were the traffic
would just stall after we tore down a Tx BA session while the HW
queue was not empty.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The queues and all the related logic suits to the transport layer.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This mapping is transport related.
This allows us to remove the notion of tx queue from the tx path in
the upper layer.
iwl_wake_any_queue moved to transport layer since it needs to access
these mappings.
The TX API is nicer now:
int (*tx)(struct iwl_trans *trans, struct sk_buff *skb,
struct iwl_device_cmd *dev_cmd, u8 ctx, u8 sta_id);
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In a near future, the upper layer won't be aware of the tx queues.
This allows to remove one place where the upper layer needed to
provide the tx queue index to the transport layer.
This also saves around 1.5MB.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
the mapping tx_queue -> fifo is really transport related. The upper
layer should be involved in such things.
Note that upon agg_disable, the queue is always mapped to fifo 0, but
this doesn't matter since when the queue will be setup again for a
new BA session, it will be configured to the good fifo anyway.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This removes the need for iwl_tx_info.
Each tx queue holds an array of skbs, the transport layer doesn't
need to know anything about the context in which a specific skb is
sent.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
There are still quite a few, but much less.
A few fields have been moved /copied to hw_params which sits in the
shared area:
* priv->cfg->base_params->num_of_ampdu_queues
* priv->cfg->base_params->shadow_reg_enable
* priv->cfg->sku
* priv->ucode_owner
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The rate scaling and the transport need to access the data in
iwl_tid_data, hence the move.
Note that the only component in the upper layer that needs this data
is the rate scaling. Refactoring the rate scaling may help to move
iwl_tid_data from the shared area to the transport area.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Logic move after all priv->status moved to struct iwl_shared
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Now that the reclaim flow has been moved to the transport layer, a lot of
functions can be made static or don't need to be exported outside the transport
layer.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The reclaim flow is really transport related. Define a simple API to allow the
upper layer to request from the transport layer to reclaim packets until an
index written in the Tx response / BA notification.
The transport layer prepares a list of the packets that are being freed and
passes this list to the upper layer.
Between the two layers, the CB of the skb is used to pass a pointer to the
context (BSS / PAN) in which the skb was sent.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Remove iwl_transport_register which was a W/A. The bus layer knows what
transport to use. So now, the bus layer gives the upper layer a pointer to the
iwl_trans_ops struct that it wants to use. The upper layer then, allocates the
desired transport layer using iwl_trans_ops->alloc function.
As a result of this, priv->trans, no longer exists, priv holds a pointer to
iwl_shared, which holds a pointer to iwl_trans. This required to change all the
calls to the transport layer from upper layer. While we were at it, trans_X
inlines have been renamed to iwl_trans_X to avoid confusions, which of course
required to rename the functions inside the transport layer because of
conflicts in names. So the static API functions inside the transport layer
implementation have been renamed to iwl_trans_pcie_X.
Until now, the IRQ / Tasklet were initialized in iwl_transport_layer. This is
confusing since the registration doesn't mean to request IRQ, so I added a
handler for that.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since it is used by all the layers, it needs to move to iwl_shared.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since it is used by all the layers, it needs to move to iwl_shared.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since it is used by all the layers, it needs to move to iwl_shared.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The previous P2P implementation turned out to
not work well and new uCode capabilities were
added to support P2P. Modify the driver to
take advantage of those, and also discover P2P
support automatically based on a uCode flag
instead of having a Kconfig symbol for P2P.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Move tm_fixed_rate inside CONFIG_IWLWIFI_DEVICE_SVTOOL and only
available when the option is enable.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
It still holds a pointer to iwl_priv. But hopefully this will disappear at some point.
Also add the multiple inclusion protection to iwl-trans.h that was forgotten.
Move iwl-trans structures to iwl-trans.h
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
All the configurations of the HW for AMPDU are now in the transport layer.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
tx start will start the tx queues: basically configure the SCD
Remove the IWLAGN prefix to SCD defines on the way.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.guy@intel.com>
The first assignment of TX_CMD_FLG_SEQ_CTL_MSK for
ack-expected mgmt frames is overwritten later in
the function, so it's useless. Also, probe response
frames, BACK request and others there are mutually
exclusive so can be moved into an else branch.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Split the Tx datapath in two parts:
* the first deals with the Tx cmd composition
* the second attaches the skb + Tx cmd to the queues
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Not needed since the driver split. Move single use routines to
calling location and keep static where possible.
Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
For testing purpose, we need better control of msc from user application.
Separate the fixed_rate between debugfs and testmode and enforce it.
Signed-off-by: Kenny Hsu <kenny.hsu@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Code duplication was needed during the move, not needed any more.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tx stop moves to transport layer.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tx free functions move to the transport layer. Unify the functions that deal with tx queues and cmd queue.
Since the CMD queue is not fully allocated, but uses the q->n_bd / q->window trick, the release flow of TX queue and CMD queue was different.
iwlagn_txq_free_tfd receives now the index of the TFD to be freed, which allows to unify the release flow for all the queues.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Our current TKIP code races against itself on TX
since we can process multiple packets at the same
time on different ACs, but they all share the TX
context for TKIP. This can lead to bad IVs etc.
Also, the crypto offload helper code just obtains
the P1K/P2K from the cache, and can update it as
well, but there's no guarantee that packets are
really processed in order.
To fix these issues, first introduce a spinlock
that will protect the IV16/IV32 values in the TX
context. This first step makes sure that we don't
assign the same IV multiple times or get confused
in other ways.
Secondly, change the way the P1K cache works. I
add a field "p1k_iv32" that stores the value of
the IV32 when the P1K was last recomputed, and
if different from the last time, then a new P1K
is recomputed. This can cause the P1K computation
to flip back and forth if packets are processed
out of order. All this also happens under the new
spinlock.
Finally, because there are argument differences,
split up the ieee80211_get_tkip_key() API into
ieee80211_get_tkip_p1k() and ieee80211_get_tkip_p2k()
and give them the correct arguments.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
These functions allocate all the Tx context. Only the simple tx_init is exported as API.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Get a pointer to the struct device during probe and get the rid of all the PCI
specific DMA wrappers.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Bus specific layer must know how to return the struct device* of the device.
Implement that as a callback of iwl_bus_ops and use that callback instead of
using the priv->pdev pointer which is meant to disappear soon.
Since the struct device * is needed in hot path, iwl_bus holds a pointer to it
instead of calling get_dev all the time.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Multiple A-MPDU actions will received from mac80211 while setting up the
aggregation queue, change the message log to better represent the states.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The exact same error message is used in three different functions in
iwlagn. Add the function name to the error string to disambiguate where
the error is coming from.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The pad argument to iwlagn_txq_free_tfd
isn't used, remove it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
All of these functions no longer need to be
accessed indirectly since they're shared in
all AGN devices.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
DMA mappings can fail, but the current code
doesn't check for that. Add checking, which
requires some restructuring for proper error
paths.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
All AGN devices need the bytecount table, so
remove the indirection and make the functions
static again.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
The device doesn't use the bytecount table for the
command queue, only for aggregation queues to make
aggregation decisions. So don't update it for the
command queue (and we even updated it with wrong
values).
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Need to use broadcast sta_id for management frames, otherwise we broke
BA session in the firmware and get messages like that:
"Received BA when not expected"
or (on older kernels):
"BA scd_flow 0 does not match txq_id 10"
This fix regression introduced in 2.6.35 during station management
code rewrite by:
commit 2a87c26bbe
Author: Johannes Berg <johannes.berg@intel.com>
Date: Fri Apr 30 11:30:45 2010 -0700
iwlwifi: use iwl_find_station less
Patch partially resolve:
https://bugzilla.kernel.org/show_bug.cgi?id=16691
However, there are still 11n performance problems on 4965 and 5xxx
devices that need to be investigated.
Cc: stable@kernel.org # 2.6.35+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Resolved logic conflicts causing a build failure due to
drivers/net/r8169.c changes using a patch from Stephen Rothwell.
Signed-off-by: David S. Miller <davem@davemloft.net>
Since
commit a120e912eb
Author: Stanislaw Gruszka <sgruszka@redhat.com>
Date: Fri Feb 19 15:47:33 2010 -0800
iwlwifi: sanity check before counting number of tfds can be free
we use skb->data after calling ieee80211_tx_status_irqsafe(), which
could free skb instantly.
On current kernels I do not observe practical problems related with
bug, but on 2.6.35.y it cause random system hangs when stressing
wireless link.
Cc: stable@kernel.org # 2.6.32+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since the driver split, there's no need for no_agg_framecnt_info since
all devices have this set to false. Secondly, the compressed block ack
handling code was broken. Fix this.
(1) A shift less than zero simply implies that the buffer wrapped, this
is expected. Remove the incorrect comment.
(2) The (agg->frame_count > (64-sh)) condition can happen if the last
frame is dropped. E.g., if I send 7 frames and the 6th is received but
the 7th is lost, the other side may only shift the window 6, not 7
frames since the last bit is a 0. This is perfectly fine behavior and
doesn't invalidate the feedback.
(3) Store the feedback from a Compressed BA in the first newly received
frame, rather than the start of the window. This way it will get
processed by the rate selection code. Feedback stored in a non-received
frame is likely to get overwritten by the retransmission.
This is based on the approach taken by minstrel_ht.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
iwlagn's hardware scheduler needs to be set up
with the right aggregation frame limit and
buffer sizes. To achieve this, we need to move
the hardware queue setup to when the session
becomes operational.
Tested-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since the driver split, there's no need for
function pointers any more for aggregation
queue setup and teardown as all devices now
share the same code. Simplify this.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since commit commit 470058e0ad
"iwlwifi: avoid Tx queue memory allocation in interface down" we do
not unmap dma and free skbs when down device and there is pending
transfer. What in consequence may cause that system hung (waiting
for free skb's) when performing shutdown at iptables module unload.
DMA leak manifest itself following warning:
WARNING: at lib/dma-debug.c:689 dma_debug_device_change+0x15a/0x1b0()
Hardware name: HP xw8600 Workstation
pci 0000:80:00.0: DMA-API: device driver has pending DMA allocations while released from device [count=240]
Modules linked in: iwlagn(-) aes_x86_64 aes_generic fuse cpufreq_ondemand acpi_cpufreq freq_table mperf xt_physdev ipt_REJECT nf_conntrack_ipv4 nf_defrag_ipv4 iptable_filter ip_tables ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables ipv6 ext3 jbd dm_mirror dm_region_hash dm_log dm_mod uinput hp_wmi sparse_keymap sg wmi microcode serio_raw tg3 arc4 ecb shpchp mac80211 cfg80211 rfkill ext4 mbcache jbd2 sr_mod cdrom sd_mod crc_t10dif firewire_ohci firewire_core crc_itu_t mptsas mptscsih mptbase scsi_transport_sas pata_acpi ata_generic ata_piix ahci libahci floppy nouveau ttm drm_kms_helper drm i2c_algo_bit i2c_core video [last unloaded: iwlagn]
Pid: 9131, comm: rmmod Tainted: G W 2.6.38-rc6-wl+ #33
Call Trace:
[<ffffffff810649ef>] ? warn_slowpath_common+0x7f/0xc0
[<ffffffff81064ae6>] ? warn_slowpath_fmt+0x46/0x50
[<ffffffff812320ab>] ? dma_debug_device_change+0xdb/0x1b0
[<ffffffff8123212a>] ? dma_debug_device_change+0x15a/0x1b0
[<ffffffff8149dc18>] ? notifier_call_chain+0x58/0xb0
[<ffffffff8108e370>] ? __blocking_notifier_call_chain+0x60/0x90
[<ffffffff8108e3b6>] ? blocking_notifier_call_chain+0x16/0x20
[<ffffffff812f570c>] ? __device_release_driver+0xbc/0xe0
[<ffffffff812f5808>] ? driver_detach+0xd8/0xe0
[<ffffffff812f45d1>] ? bus_remove_driver+0x91/0x100
[<ffffffff812f6022>] ? driver_unregister+0x62/0xa0
[<ffffffff8123d5d4>] ? pci_unregister_driver+0x44/0xa0
[<ffffffffa05632d1>] ? iwl_exit+0x15/0x1c [iwlagn]
[<ffffffff810ab492>] ? sys_delete_module+0x1a2/0x270
[<ffffffff81498da9>] ? trace_hardirqs_on_thunk+0x3a/0x3f
[<ffffffff8100bf42>] ? system_call_fastpath+0x16/0x1b
I still can observe above warning after apply patch, but it is very
hard to reproduce it, and have count=1. Whereas that one is easy to
reproduce using debugfs force_reset while transmitting data, and have
very big counts eg. 240, like quoted here. So count=1 WARNING seems
to be different issue that need to be resolved separately.
v1 -> v2: fix infinity loop bug I made during "for" to "while" loop transition.
v2 -> v3: remove unneeded EXPORT_SYMBOL
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
For device supporting PAN/P2P, use the PAN
context to implement the remain-on-channel
operation using device offloads so that the
filters in the device will be programmed
correctly -- otherwise we cannot receive
any probe request frames during off-channel
periods.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The compiler correctly warns:
iwl-agn-tx.c: In function ‘iwlagn_tx_status_reply_compressed_ba’:
iwl-agn-tx.c:1240: warning: ‘bitmap’ may be used uninitialized in this function
Move the debug print to the branch that reads the
bitmap, and move the variables too so it's more
obvious where they are needed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Instead of hardcoding the numbers that must
match mac80211, use the constants. Not that
this means we could change the constants,
but at least this way it's clearer.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since aggregation queues are station-specific, the
device will not reject packets in them but rather
will stop the appropriate aggregation queues when
a station goes to sleep. I forgot to account for
this in the driver, so if a station went to sleep
that had aggregation enabled, traffic would stop
indefinitely.
Fix this by only accounting frames queued on the
normal AC queues for associated station.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Through races, a packet may be enqueued for
transmission to a station while that station
is going to sleep, in which case the warning
here triggers. Instead of warning, check the
condition -- if this packet is not a PS-poll
response then we still enqueue it but it will
be rejected by the device since the station
is marked as asleep.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
We can simplify length calculation in iwlagn_tx_skb, that function
is enough complex, without fuzz it more than necessary.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Previously, we used the swq_id's mechanism
to have AC and HW queue different only for
aggregation queues. To be able to fix a bug
with iPAN simply always build the swq_id as
ac | (hwq << 2) and remove the flag bit.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Instead of passing the txq->swq_id, pass the
txq struct directly to make sure that in the
future nobody will pass an invalid number.
Only three places actually change from using
the txq_id or the skb's queue_mapping to now
using txq->swq_id as well.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
For newer devices, uCode provide both "number of frames sent"
and "number of frames acked" information inside the compressed_ba
packet. So instead of figure the success/failure information through
the bitmap, use those information which is much betrer approach.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Timing issues in microcode for some devices can cause a compressed BA to
be sent to the driver prior to returning any a-MPDU notification.
Traces show RTS-CTS is exchanged and then the timer fires which causes an
empty BA to be sent which acknowledges nothing. This results in a noisy
printk. Only print the message if the bitmap is non-zero.
Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The code to print out TX failure reasons is
AGN specific, so it can be in the AGN module.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
move paramater definitions to a device paramater structure only
leaving the device name, which antennas are used and what firmware
file to use in the iwl_cfg structure. this will not completely
remove the redundancies but greatly reduce them for devices that
only vary by name or antennas. the parameters that are more
likely to change within a given device family are left in iwl_cfg.
also separate bt param structure added to help reduce more.
Signed-off-by: Jay Sternberg <jay.e.sternberg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Frames for different contexts need to be put
on different queues, and multicast after DTIM
frames have a special queue yet which also
depends on the context, so put all this into
the context.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
iwlwifi occasionally needs to find the virtual
interface pointer to give it to mac80211, but right
now it only keeps one. Move it into the context so
that we can keep one pointer each.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
In status processing we'll need to find the context
for a given frame, so add a context pointer to the
TX info for each frame.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
The broadcast station ID is per context, so
add a variable for the ID in the context and
use it everywhere we previously hardcoded it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
PAN ucode will require a different queue assignment,
in particular queue 9 instead of 4 should be used for
commands.
This is required because the ucode will stop/start
queues 4 and 8 depending on the PAN state, since
queue 8 will be used for PAN multicast (after DTIM).
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
When aggregation is stopped again for some reason
before the queue we selected has drained, we will
currently leak the TX queue and keep it enabled
for aggregation. Normally this doesn't happen, so
the problem is rarely seen.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Adding the bluetooth full concurrency support for WiFi/BT combo devices.
Driver should configure uCode to operate in "full concurrency" mode (via
LUT) if both conditions are met:
- Antenna Coupling is more than 35dB
- WiFi Channel Inhibition Request is hornored by BT Core
Currently, there is no antenna coupling information provided by uCode;
use module parameter to specified the antenna coupling in dB.
When in "full concurrency" mode, driver need to download different LUT
to uCode while sending bt configuration command; also, driver need to
configure the device operate in 1x1 while in full concurrency mode.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The BT ignore bit should be set when transmitting
auth, assoc response and eap frames.
Also, scanning should set the BT ignore bit for the
probe request transmission; Note that we only use
the non-shared antenna.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently, mac80211 translates the cfg80211
cipher suite selectors into ALG_* values.
That isn't all too useful, and some drivers
benefit from the distinction between WEP40
and WEP104 as well. Therefore, convert it
all to use the cipher suite selectors.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently the driver will try to protect all frames,
which leads to a lot of odd things like sending an
RTS with a zeroed RA before multicast frames, which
is clearly bogus.
In order to fix all of this, we need to take a step
back and see what we need to achieve:
* we need RTS/CTS protection if requested by
the AP for the BSS, mac80211 tells us this
* in that case, CTS-to-self should only be
enabled when mac80211 tells us
* additionally, as a hardware workaround, on
some devices we have to protect aggregated
frames with RTS
To achieve the first two items, set up the RXON
accordingly and set the protection required flag
in the transmit command when mac80211 requests
protection for the frame.
To achieve the last item, set the rate-control
RTS-requested flag for all stations that we have
aggregation sessions with, and set the protection
required flag when sending aggregated frames (on
those devices where this is required).
Since otherwise bugs can occur, do not allow the
user to override the RTS-for-aggregation setting
from sysfs any more.
Finally, also clean up the way all these flags get
set in the driver and move everything into the
device-specific functions.
Cc: stable@kernel.org [2.6.35]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
spin_is_locked() can return zero on some (UP?)
configurations because locks don't exist, and
that causes an endless amount of warnings. Use
lockdep_assert_held() instead, which has two
advantages:
1) it verifies the current task is holding
the lock or mutex
2) it compiles away completely when lockdep
is not enabled
Cc: stable@kernel.org [2.6.34+, maybe only parts of patch]
Reported-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
It is a uCode bug which cause the tx queue id not match scd_flow
in compressed block ack frame, and it need to be addressed in uCode.
Currently, driver will log the information when it happen.
Since it is possible happen very often and we do not want to fill the syslog,
so don't enable the logging by default.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When we stop the Tx DMA channels, we poll bits 16:31 in
FH_TSSR_RX_STATUS_REG. From 4965 and up, only the bits 16:26 are legal.
Bits 27:31 are not used and are always unset.
Polling them will lead to fail on timeout but since the timeout is quite
small, the stall was not felt.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Now that the ampdu_action callback can sleep,
we can use the mutex to properly protect the
aggregation data, and return useful errors if
they should happen.
Also, add some sleep and mutex debugging so
we won't call any of the functions that now
require being able to sleep and/or the mutex
to be held in an invalid context.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
The flow id (scd_flow) in a compressed BA packet should match the txq_id
of the queue from which the aggregated packets were sent. However, in
some hardware like the 1000 series, sometimes the flow id is 0 for the
txq_id (10 to 19). This can cause the annoying message:
[ 2213.306191] iwlagn 0000:01:00.0: Received BA when not expected
[ 2213.310178] iwlagn 0000:01:00.0: Read index for DMA queue txq id (0),
index 5, is out of range [0-256] 7 7.
And even worse, if agg->wait_for_ba is true when the bad BA is arriving,
this can cause system hang due to NULL pointer dereference because the
code is operating in a wrong tx queue!
Signed-off-by: Shanyu Zhao <shanyu.zhao@intel.com>
Signed-off-by: Pradeep Kulkarni <pradeepx.kulkarni@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Using ieee80211_find_sta() needs to be under
RCU read lock, which iwlwifi currently misses,
so fix it.
Cc: stable@kernel.org
Reported-by: Miles Lane <miles.lane@gmail.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Acked-by: Reinette Chatre <reinette.chatre@intel.com>
Tested-by: Miles Lane <miles.lane@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Compressed BlockAck frames store the ACKs/NACKs in a 64-bit bitmap that starts
at the sequence number of the first frame sent in the aggregated batch. Note
that this is a selective ACKnowledgement following selective retransmission;
e.g., if frames 1,4-5 in a batch are ACKed then the next transmission will
include frames 2-3,6-10 (7 frames). In this latter case, the Compressed
BlockAck will not have all meaningful information in the low order bits -- the
semantically meaningful bits of the BA will be 0x1f3 (where the low-order frame
is seq 2).
The driver code originally just looked at the lower (in this case, 7) bits of
the BlockAck. In this case, the lower 7 bits of 0x1f3 => only 5 packets,
maximum, could ever be ACKed. In reality it should be looking at all of the
bits, filtered by those corresponding to packets that were actually sent. This
flaw meant that the number of correctly ACked packets could be significantly
underreported and might result in asynchronous state between TX and RX sides as
well as driver and uCode.
Fix this and also add a shortcut that doesn't require the code to loop through
all 64 bits of the bitmap but rather stops when no higher packets are ACKed.
In my experiments this fix greatly reduces throughput swing, making throughput
stable and high. It is also likely related to some of the stalls observed in
aggregation mode and maybe some of the buffer underruns observed, e.g.,
http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=1968http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=2098http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=2018
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
In "iwlwifi: make scan antenna forcing more generic"
I introduced generic scan RX antenna forcing, which
here I rename to make it more evident. Also add scan
TX antenna forcing, since I will need that as well.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Currently, the driver allocates up to 19 skb pointers
for each TFD, of which we have 256 per queue. This
means that for each TX queue, we allocate 19k/38k
(an order 4 or 5 allocation on 32/64 bit respectively)
just for each queue's "txb" array, which contains only
the SKB pointers.
However, due to the way we use these pointers only the
first one can ever be assigned. When the driver was
initially written, the idea was that it could be
passed multiple SKBs for each TFD and attach all
those to implement gather DMA. However, due to
constraints in the userspace API and lack of TCP/IP
level checksumming in the device, this is in fact not
possible. And even if it were, the SKBs would be
chained, and we wouldn't need to keep pointers to
each anyway.
Change this to only keep track of one SKB per TFD,
and thereby reduce memory consumption to just one
pointer per TFD, which is an order 0 allocation per
transmit queue.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
This can be cleanly applied to wireless-2.6 and iwlwifi git trees.
=
From: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Subject: [PATCH] iwlwifi: use the DMA state API instead of the pci equivalents
This replace the PCI DMA state API (include/linux/pci-dma.h) with the
DMA equivalents since the PCI DMA state API will be obsolete.
No functional change.
For further information about the background:
http://marc.info/?l=linux-netdev&m=127037540020276&w=2
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
There's a single use of this struct member, but
as it is write-only it clearly not necessary.
Thus we can free up some space here, even if we
don't need it right now it seems pointless to
carry around the variable.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
There are a few places where sta_lock is used, but the
station information protected by it is accessed outside
of the lock. Address this in two ways, if the access
won't sleep then just move the access into the lock, if
the access can sleep then copy the needed station
information to the stack to be accessed without risk of
it changing while access in progress.
Additionally, a number of other places access station
station information without holding the sta_lock, fix
those as well.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
There are now five places where we need to
look up the station ID, but the sta pointer
may be NULL due to mac80211 passing that to
indicate a certain special state.
Replace all these by a new inline function,
called iwl_sta_id_or_broadcast(), and add
documentation about when to use it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
With the station ID being stored in the
station struct, which mac80211 gives us
for aggregation callbacks, we can also
remove the use of iwl_find_station() in
those code paths.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Since we now store the station ID in each station
struct, many places need not look at the station
table any more since they can just pull the station
ID out of the struct. Remove iwl_get_sta_id() and
use iwl_sta_id() instead as appropriate.
This reduces the amount of code needed to find the
right station significantly, and works since
mac80211 passes the station only after it has been
fully initialised, ie. even if TX races with
station addition it will only be passed to TX once
the addition is complete.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Most of the TX aggregation handling can be passed
the virtual interface directly instead of having
to rely on priv->vif.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
The TX status code is currently abusing the ampdu_ack_map field (a bitmap) to
count the number of successfully received frames. The comments in mac80211.h
show there are actually three different, relevant variables, of which we are
currently using two, both incorrectly. Fix this by making
- ampdu_ack_len -> the number of ACKed frames (i.e. successes)
- ampdu_ack_map -> the bitmap
- ampdu_len -> the total number of frames sent (i.e., attempts)
to match the header file (and verified with ath9k's usage) and updating Intel's
RS code to match.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
When starting an aggregation session, the swq_id is generated in function
iwl_virtual_agg_queue_num() where the first parameter is supposed to be
the Access Class, but it used the tx fifo ID instead. This means the AC
value stored in swq_id is incorrect. To test this, look at the tx_queue
file in debugfs while transmitting Best Effort flow (ac=2), it shows:
hwq 10: read=0 write=0 stop=0 swq_id=0xa9 (ac 1/hwq 10)
After this fix, it will show:
hwq 10: read=0 write=0 stop=0 swq_id=0xaa (ac 2/hwq 10)
Signed-off-by: Shanyu Zhao <shanyu.zhao@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Emmanuel Grumbach