| Age | Commit message (Collapse) | Author |
|---|
|
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (37 commits)
smc91c92_cs: fix the problem of "Unable to find hardware address"
r8169: clean up my printk uglyness
net: Hook up cxgb4 to Kconfig and Makefile
cxgb4: Add main driver file and driver Makefile
cxgb4: Add remaining driver headers and L2T management
cxgb4: Add packet queues and packet DMA code
cxgb4: Add HW and FW support code
cxgb4: Add register, message, and FW definitions
netlabel: Fix several rcu_dereference() calls used without RCU read locks
bonding: fix potential deadlock in bond_uninit()
net: check the length of the socket address passed to connect(2)
stmmac: add documentation for the driver.
stmmac: fix kconfig for crc32 build error
be2net: fix bug in vlan rx path for big endian architecture
be2net: fix flashing on big endian architectures
be2net: fix a bug in flashing the redboot section
bonding: bond_xmit_roundrobin() fix
drivers/net: Add missing unlock
net: gianfar - align BD ring size console messages
net: gianfar - initialize per-queue statistics
...
|
|
Signed-off-by: Benjamin Larsson <banan@ludd.ltu.se>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
|
|
This converts mwl8k to use the new station
add/remove callbacks instead of using the
old sta_notify callback.
The new callbacks can sleep, so a lot of
code can be removed now.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Use spin_[un]lock_bh in mwl8k_sta_notify(). The sta_notify handler is
required to be atomic, yet it can be called in process context, so make
sure one call won't preempt another.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
get_tx_stats() will be removed from mac80211.
mwl8k used struct ieee80211_tx_queue_stats internally to track the queue
lenght. Replace struct ieee80211_tx_queue_stats with a simple len field
in struct mwl8k_tx_queue. Limit and count fields seemed to be unused.
Compile-tested only.
Cc: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Because DTIM information is required for powersave
but is only conveyed in beacons, wait for a beacon
before enabling powersave, and change the way the
information is conveyed to the driver accordingly.
mwl8k doesn't currently seem to implement PS but
requires the DTIM period in a different way; after
talking to Lennert we agreed to just have mwl8k do
the parsing itself in the finalize_join work.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
As follows:
- GET_HW_SPEC is now responsible for setting
priv->{ap,sta}_macids_supported, which are bitmasks of supported
macids for AP and STA mode. (Typically, STA firmware images will
support only one macid, #0, in STA mode, and AP firmware images
will support macids #0-7, in AP mode.)
- Our wiphy ->interfaces_modes is now set based on the non-zero-ness
of these two bitmasks.
- We main priv->macids_used, a bitmask of which macids are currently
in use. ->add_interface() will assign the lowest free macid for
this interface type as it is created, or bail out if there are no
more free macids to assign. ->delete_interface() will mark the
macid as being free again.
This enables the multi-BSS code added in the previous commits.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
SET_BEACON, SET_MAC_ADDR, BSS_START and SET_NEW_STN are the currently
supported firmware commands that are actually per-vif commands. Use
mwl8k_post_pervif_cmd() for these commands, so that the macid of the
vif they operate on gets passed down into the firmware.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
One of the bytes in the mwl8k firmware command header is the 'macid'
byte, which for per-vif commands indicates which of the BSSes this
command is intended for. (For commands that are not per-vif commands,
this byte can just be 0.)
This patch adds mwl8k_post_pervif_cmd(), which will take the macid
assigned to this interface (to be done in ->add_interface()), copy it
into the command packet macid field, and post the command as usual.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
To prepare for adding multi-BSS support.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Make the decision about whether to register the 2.4 and 5 GHz bands
with mac80211 by looking at the capability field in GET_HW_SPEC (STA
firmware only for now). This enables 5 GHz STA operation.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Whenever mac80211 gives us a legacy rate bitmap in the context of the
5 GHz band, we need to remember to shift the bitmap left by 5 positions
before giving it to the firmware, as the firmware follows the bitmap
bit assignment of the 2.4 GHz rate table even if we're on the 5 GHz
band, and the 2.4 GHz rate table includes five non-OFDM rates at the
start that are not valid in the 5 GHz band.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
The mwl8k firmware uses indices into the 2.4 GHz band rate table for
the receive descriptor channel field even if the packet was received
on a 5 GHz channel, while mac80211 expects an index into the 5 GHz
band rate table when packets are received on the 5 GHz band, which
presents a mismatch as the 5 GHz band rate table lacks the five
non-OFDM rates that the 2.4 GHz rate table starts with.
To handle this properly, we need to substract 5 from the rate index
field if the packet was received on a 5 GHz channel (and was not
received at an MCS rate).
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
So that we can make 2.4 GHz and 5 GHz band registration conditional
on the capability bitmask returned by the firmware.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
To prepare for adding 5 GHz band/channels/rates.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
0x2a43 is a single-band (2.4GHz only) 88w8366 mini-PCIe card.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
By storing the sequence counter in << 4 format, like other drivers do.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Add MODULE_FIRMWARE tags for the mwl8k firmware images that don't
have them yet, and move them to where the firmware image names are
declared.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Like how TX reclaim is done in a tasklet, move receive processing
to tasklet context as well. This can have nice benefits for CPU
utilisation and throughput, especially at 3-stream rates.
(Use the same CLEAR_SEL trick as the TX reclaim tasklet does, to
avoid having to touch the interrupt mask registers.)
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
By making use of the CLEAR_SEL feature of the mwl8k host interface
interrupt controller, we can keep the TX_DONE interrupt source masked
while the transmit reclaim tasklet is running (NAPI style) without
having to touch the interrupt controller's interrupt mask register
when entering or exiting polling mode, and without having to do any
more register reads/writes than we do now.
When CLEAR_SEL is enabled on the TX_DONE interrupt source, reading
the interrupt status register will clear the TX_DONE status bit if
it was set, allowing it to be set again if a new TX_DONE event arrives
while we are running the TX reclaim tasklet, but such a new event will
then not trigger another PCI interrupt until a zero is written to the
TX_DONE interrupt status register bit.
I.e., if we write a zero to the TX_DONE interrupt source bit in the
interrupt status register when the TX reclaim tasklet thinks it's
done, a PCI interrupt will be triggered if a new TX_DONE event arrived
from the hardware between us deciding that there is no more work to do
and re-enabling the TX_DONE interrupt source, thereby avoiding the
classic NAPI poll mode exit race that would otherwise occur.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Add a limit argument to mwl8k_txq_reclaim(), to allow limiting the
number of packets that it will reclaim, and make it return the number
of packets that it reclaimed.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Add support for creating AP interfaces, and enabling beaconing.
This allows running a basic AP (11b/g mode only for now).
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
STA firmware uses UPDATE_STADB to manipulate the hardware station
database, whereas AP firmware uses SET_NEW_STN -- this implements the
latter, and hooks it into mwl8k_sta_notify(), to be used if we're
running on AP firmware.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
As with the STA version, unicast will use auto rate adaptation, but
the AP version allows setting the rates to be used for management and
multicast transmissions, which can be set based on the BSS basic rate
set.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
As we always use the auto rate adaptation feature and never pass in
a rate table, USE_FIXED_RATE can be simplified somewhat. While we're
at it, rename it to *_sta, as this is the STA version of the command.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
While it is reasonable to expect that at least one transmit ring entry
will be processed per second while we are waiting for the transmit
rings to drain, the firmware can end up doing batching of transmit ring
status writeback, which means that the transmit rings can appear stuck
for more than a second at a time.
Bump the TX drain wait timeout up from 1 to 5 seconds to account for
this.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
The comment and code in mwl8k_cmd_set_edca_params() suggest that the
mapping between SET_EDCA_PARAMS queue numbers and transmit rings isn't
actually 1:1, while tests show that the mapping is in fact 1:1. So,
get rid of the transmit queue 0/1 swapping.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Conflicts:
net/mac80211/iface.c
|
|
This enables HT association and AMPDU in the receive direction for
STA firmware images on hardware that supports it.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
AMPDU receive doesn't need any special handling, so let's enable
this before tackling the transmit side.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Pass the AP's MCS rate mask to SET_RATE when associating, and make
UPDATE_STADB pass in the peer's HT caps and rates when adding a new
hardware station database entry.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Previously, mwl8k_bss_info_changed() would refuse to do anything if
the 'changed' argument indicated that the association status hadn't
changed. Fix this up so that it will allow changing things like the
preamble type, the slot time and the CTS-to-self protection method
without having to reassociate.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
When calling SET_RATE, SET_AID, or when creating a station database
entry for our AP, pass in the AP's rate set instead of just blindly
enabling all legacy rates, so as to end up doing the right thing when
talking to 11b-only APs.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
For STA firmware, move the per-peer hardware station ID to the
driver-private part of struct ieee80211_sta, where it belongs.
(Since issuing a hardware station database maintenance command sleeps,
we can't hold a reference to the ieee80211_sta * across the command,
and since we won't know the station ID until after the command
completes, we need to re-lookup the sta when the command is done to
write the returned station ID back to its driver-private part.)
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Inserting and removing a hardware station database entry for the AP
when we are in managed mode is currently done in ->bss_info_changed().
To prepare for adding AP mode support, implement the ->sta_notify()
driver method, and let that handle inserting and removing the hardware
station database entry for our AP instead.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Old:
> phy0: timeout waiting for tx rings to drain (9 -> 5 pkts), retrying
> phy0: timeout waiting for tx rings to drain (5 -> 2 pkts), retrying
> phy0: tx rings drained
New:
> phy0: waiting for tx rings to drain (9 -> 5 pkts)
> phy0: waiting for tx rings to drain (5 -> 2 pkts)
> phy0: tx rings drained
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
|
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
|
