Emilio G. Cota [Fri, 3 Dec 2010 09:05:08 +0000 (09:05 +0000)]
staging/vme_user: fix usage of the slave resources after they've been freed
buf_unalloc() frees the memory buffers allocated with vme_alloc_consistent.
The associated VME resource is needed in both vme_alloc_consistent and
vme_free_consistent; however the slave VME resources are being freed before
the calls to vme_free_consistent are made, which means the buffers
are never returned.
Fix this by freeing the VME resources only after the consistent buffers have
been returned.
Peter Huewe [Wed, 1 Dec 2010 23:43:34 +0000 (00:43 +0100)]
staging/sep: Fix printk format warning
This patch fixes gcc's complaints about the wrong format string for
size_t arguments:
"format '%x' expects type 'unsigned int', but argument has type
'size_t'"
Signed-off-by: Peter Huewe <peterhuewe@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Peter Huewe [Wed, 1 Dec 2010 23:43:32 +0000 (00:43 +0100)]
staging/sep: Fix sparse warning 'do-while statement is not a compound statement'
This patch fixes the warning generated by sparse: 'do-while statement is
not a compound statement' by adding the necessary brackets around the do
block
Signed-off-by: Peter Huewe <peterhuewe@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Brett Rudley [Wed, 1 Dec 2010 23:36:39 +0000 (15:36 -0800)]
staging: brcm80211: absorb linux_osl.h into osl.h
Move code from linux_osl.h into osl.h and delete linux_osl.h, which
created a bunch of checkpatch errors...
I fixed most but volatile warnings will have to be dealt with later.
Merge the functions into vmbus_drv.c as there's no need to have them
separate anymore. It will also make unwinding some of the function and
pointer mess easier, as well as making functions static in the future.
It's only ever set to one function, so just call that function instead.
Actually, that wrapper function only ever called vmbus_request_offers()
so just call that function instead, no need for a do-nothing
intermediate step here.
Roland Vossen [Wed, 1 Dec 2010 20:37:34 +0000 (21:37 +0100)]
staging: brcm80211: replaced typedef struct wl_info_t by struct wl_info.
Part of the code cleanup effort. CodingStyle doc advises to restrict the usage
of typedefs. Also moved several function declarations from .c to .h file
because this typedef replace operation induced checkpatch.pl warnings.
Signed-off-by: Roland Vossen <rvossen@broadcom.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Randy Dunlap [Wed, 1 Dec 2010 19:16:35 +0000 (11:16 -0800)]
staging: brcm80211 needs lots of delay.h
Fix lots of errors like:
drivers/staging/brcm80211/phy/wlc_phy_n.c:17613: error: implicit declaration of function 'mdelay'
drivers/staging/brcm80211/util/nicpci.c:246: error: implicit declaration of function 'udelay'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Cc: Brett Rudley <brudley@broadcom.com> Cc: Henry Ptasinski <henryp@broadcom.com> Cc: Dowan Kim <dowan@broadcom.com> Cc: Roland Vossen <rvossen@broadcom.com> Cc: Arend van Spriel <arend@broadcom.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Chosrova [Wed, 1 Dec 2010 12:42:16 +0000 (13:42 +0100)]
Staging: rtl8192u: check return code kmalloc.
This patch checks the return code of kmalloc when trying to allocate
memory for priv->rx_urb in rtl8192_usb_initendpoints(), return -ENOMEM
when failed.
Signed-off-by: David Chosrova <david.chosrova@libertysurf.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Chosrova [Mon, 29 Nov 2010 20:50:48 +0000 (21:50 +0100)]
Staging: rtl8192u: remove dead code to disable dot11d
This patch removes all the ENABLE_DOT11D ifdefs.
It is always defined for driver. DOT11D has to do with regulatory domains.
What prompted this patch was a warning message in Sparse.
drivers/staging/rtl8192u/r8192U_core.c:247:1: warning: "eqMacAddr"
redefined in file included from drivers/staging/rtl8192u/:81:81:
drivers/staging/rtl8192u/dot11d.h:35:1: warning: this is the location
of the previous definition
Now there are no ifdefs around dot11d.h it made no sense to have this
second definition, so I removed that macro as well. ( Thanks Dan ;-) ).
Acked-by. Dan Carpenter <error27@gmail.com> Signed-off-by: David Chosrova <dada2372@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Pekka Enberg [Sun, 28 Nov 2010 21:00:00 +0000 (23:00 +0200)]
Staging: w35und: Use pr_debug() for debugging
Use pr_debug() for debugging printk's and kill the FULL_DEBUG macro. It would
be even better to use dev_dbg() but unfortunately looking up struct device in
the current code structure makes things very ugly.
Please note that I dropped the DataDmp() calls from
RFSynthesizer_SwitchingChannel() because that function doesn't exist.
Cc: Pavel Machek <pavel@ucw.cz> Signed-off-by: Pekka Enberg <penberg@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Pekka Enberg [Sun, 28 Nov 2010 20:59:57 +0000 (22:59 +0200)]
Staging: w35und: Kill _USE_FALLBACK_RATE_ macro
The _USE_FALLBACK_RATE_ macro parametrizes DEFAULT_RATE_RETRY_LIMIT. It's only
used in Mxx_initial() in reg.c where _USE_FALLBACK_RATE_ is always defined
because the reg.c file includes sysdef.h at the top. It's therefore safe to
remove the _USE_FALLBACK_RATE_ macro.
Cc: Pavel Machek <pavel@ucw.cz> Signed-off-by: Pekka Enberg <penberg@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Marek Belisko [Wed, 24 Nov 2010 09:42:39 +0000 (10:42 +0100)]
staging: ft1000: Fix goto error logic.
Fix goto error logic which could lead to kernel panics
because kthread_stop() is called in not correct error
conditions. Seen it sometimes when dsp_reload() fails
then I got kernel panic.
Signed-off-by: Marek Belisko <marek.belisko@open-nandra.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Randy Dunlap [Wed, 24 Nov 2010 19:46:09 +0000 (11:46 -0800)]
staging: fix intel_sst/intelmid sound build
SND_INTELMID selects 3 SND_* kconfig symbols, but that is useless
if the SOUND & SND symbols in their hierarchy are not enabled,
so make this symbol depend on SOUND & SND.
Arend van Spriel [Tue, 23 Nov 2010 13:06:24 +0000 (14:06 +0100)]
staging: brcm80211: removed packet macros for accessing sk_buff fields
With the packet storage type changed from void pointer to struct sk_buff
pointer there is no need for macros for accessing these fields through
casting. These can now be accessed directly.
Arend van Spriel [Tue, 23 Nov 2010 13:06:23 +0000 (14:06 +0100)]
staging: brcm80211: change packet buffer type to native struct sk_buff
The packet queues now store struct sk_buff pointer and subsequently
all driver code handling packets now use struct sk_buff as package
storage type. Next step will be getting rid of packet macros.
Alan Cox [Wed, 24 Nov 2010 19:39:20 +0000 (19:39 +0000)]
Staging: sep: Fix crash if a device is not found
The existing code works mostly by luck. The PCI probe is done by the
register and completes before the register returns thus allowing the other
init code to run in time. Without a SEP or if unlucky this doesn't occur
and you get an OOPS which for some reason causes grumpiness.
As the season of good b^Hcheer is supposed to be approaching we should
probably fix it.
Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Cox [Wed, 24 Nov 2010 19:38:02 +0000 (19:38 +0000)]
Staging: sep: clean up some of the obvious sillies
Various assignments are done but not used
dev_dbg formatting is a bit weird and wraps when not needed
Take out some of the blank lines and reformat a bit to reduce view size
Remove some comments that are obvious
Fix else formatting
Remove some user triggerable dev_warns
Fix copy_from_user error return cases (-EFAULT not bytes copied)
Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Vinod Koul [Mon, 22 Nov 2010 10:33:51 +0000 (10:33 +0000)]
Staging: sst: add ioctls for post processing algorithm interface
This patch adds two new ioctls to intel_sst_ctrl device.
This i/f can be used by application to send algorithm parameters
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
[This will need further discussion in the context of the final ALSA interface
but is fine for staging, ie anyone who relies on it should expect changes
Also fixed a missing kmalloc fail check] Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Vinod Koul [Mon, 22 Nov 2010 10:33:40 +0000 (10:33 +0000)]
Staging: sst: Construct fw string name runtime
The firmware name for each platform is appended by PCI id of device.
This patch makes use of pci id to construct the string rather than
hardcode the string.
Sven Eckelmann [Sun, 21 Nov 2010 23:56:07 +0000 (00:56 +0100)]
Staging: batman-adv: Use kernel functions to identify broadcasts
linux/etherdevice.h already provides functions to classify different
ethernet addresses. These inlineable functions should be used instead of
custom functions.
The check for multicast together with multicast can also be replaced
with a single test for multicast because for every ethernet address x
following is always true:
Marek Lindner [Sun, 21 Nov 2010 23:56:05 +0000 (00:56 +0100)]
Staging: batman-adv: add gateway IPv6 support by filtering DHCPv6 messages
Some additional checks will be needed in case of extension headers
like the fragmentation or hop-by-hop (for jumbo frames for example)
headers or ipsec stuff. But this patch should do for most people
for now, the rest can be added with a later one.
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de> Acked-by: Linus Lüssing <linus.luessing@web.de> Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Marek Lindner [Sun, 21 Nov 2010 23:56:04 +0000 (00:56 +0100)]
Staging: batman-adv: best gw DHCP filter 802.1Q support
The DHCP filter inspects packets to determine whether or not to send
them via ethernet unicast. This patch adds 802.1Q (vlan) support for
this check.
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
[sven.eckelmann@gmx.de: Rework on top of current version] Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Marek Lindner [Sun, 21 Nov 2010 23:56:03 +0000 (00:56 +0100)]
Staging: batman-adv: send DHCP requests directly to the chosen gw
If the gateway client mode is active batman-adv will send the
broadcasted DHCP requests via unicast to the currently selected best
gateway. Therefore attached clients can profit from batman's knowledge
about the network topology.
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
[sven.eckelmann@gmx.de: Rework on top of current version] Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Marek Lindner [Sun, 21 Nov 2010 23:56:02 +0000 (00:56 +0100)]
Staging: batman-adv: adding gateway functionality
Via the /sys filesystem you can change the gateway mode of a node using
gw_mode. Adjustments to it can be done using gw_bandwidth for server
mode and gw_sel_class for client mode.
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
[sven.eckelmann@gmx.de: Rework on top of current version] Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Sven Eckelmann [Sun, 21 Nov 2010 23:56:01 +0000 (00:56 +0100)]
Staging: batman-adv: Limit spin_locks to spin_lock_bh
spin_lock_irqsave disables the IRQs and stores them inside the flags
provided by the caller. This is needed to protect a bottom half handler
or a user context critical section from being interrupted by an
interrupt handler which also tries to acquire the spinlock and locks
forever.
The linux device drivers will receive the packets inside an interrupt
handler and the network infrastructure will process them inside bottom
half. Thus batman-adv will only run in user context and bottom half
handlers. We can conclude that batman-adv doesn't share its own
spinlocks with real interrupt handlers.
This makes it possible to exchange the quite complex spin_lock_irqsave
with spin_lock_bh which only stops bottom halves from running on the
current cpu, but allows interrupt handlers to take over to keep the
interrupt latency low.
Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Sven Eckelmann [Sun, 21 Nov 2010 23:56:00 +0000 (00:56 +0100)]
Staging: batman-adv: Rewrite hash using hlist_*
The hash implementation is a complete implementation of a hash using
buckets as hash entries and overflow buckets attached to them.
The kernel already provides datastructures hlist_head and hlist_node
which can be used to implement an hash using lists as hash buckets. So
it is better to implement heavily used functionality on top of those
instead of providing a full hash implementation.
The rewrite changes the behavior of some functions slightly:
* hash_add add elements to the front instead of the tail
* hash_iterate doesn't provide pointer to access bucket->data directly,
but it can be accessed using hlist_entry
Reported-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sven Eckelmann <sven.eckelmann@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>