staging: android/lowmemorykiller: Don't unregister notifier from atomic context
The lowmemorykiller registers an atomic notifier for notfication of when
the task is freed. From this atomic notifier callback, it removes the
atomic notifier via task_free_unregister(). This is incorrect because
atomic_notifier_chain_unregister() calls syncronize_rcu(), which can
sleep, which shouldn't be done from an atomic notifier.
Fix this by registering the notifier during init, and only unregister it
if the lowmemorykiller is unloaded.
Rebased to -next by Paul E. McKenney.
Rebased to -next again by Anton Vorontsov.
Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Christian Bejram <christian.bejram@stericsson.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Reported-by: John Stultz <john.stultz@linaro.org> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: stable <stable@vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Zhengwang Ruan [Wed, 7 Mar 2012 02:36:57 +0000 (10:36 +0800)]
Staging:android: Change type for binder_debug_no_lock switch to bool
GCC warns that module_param_named() indirectly returns a bool type value
which is different from 'int' type binder_debug_no_lock declared. Change
it to bool because it is a internal switch for debugging.
The kernel already has a debug allocator, no need to have one unique to
a single driver. So delete it, replace with kfree, kmalloc, and, in a
few places that need it, kzalloc().
Cc: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kevin McKinney [Fri, 2 Mar 2012 05:17:04 +0000 (00:17 -0500)]
Staging: bcm: fix possible memory leak of 'pstAddIndication' in CmHost.c and removes a whitespace
Memory is being allocated by kmalloc and stored in
variable pstAddIndication. However, this memory is
not being freed in all cases. Therefore, this patch
frees it on several exit paths. This patch also
removes a whitespace.
Signed-off-by: Kevin McKinney <klmckinney1@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Martin Krause [Fri, 2 Mar 2012 12:01:40 +0000 (13:01 +0100)]
Staging: rtl8712/rtl8192u: move USB device ID within staging drivers.
It seems the USB device ID 0bda:8192 is wrongly assigned to the
RTL8192SU chip and not to the RTL8191SU chip in the USB database.
So this patch moves this device ID from the rtl8192u staging
driver to the rtl8712 staging driver.
This patch was tested with a Radicom WIFIHU embedded wireless
module with a RTL8191SU chip and the USB device ID 0bda:8192.
Without the patch the rtl8192u driver claims this device, but
it does not work. With the patch the rtl8712 driver services
this device and it works.
Signed-off-by: Martin Krause <martin.krause@tqs.de> Acked-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Fri, 2 Mar 2012 06:59:55 +0000 (09:59 +0300)]
Staging: ozwpan: prevent bogus dereference
app_id comes from the network and can't be trusted. If it's zero then
it will lead to a kernel crash.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Larry Finger [Mon, 27 Feb 2012 04:08:36 +0000 (22:08 -0600)]
staging: r8712u: Fix regression in signal level after commit c6dc001
In commit c6dc001 "staging: r8712u: Merging Realtek's latest (v2.6.6).
Various fixes", the returned qual.qual member of the iw_statistics
struct was changed. For strong signals, this change made no difference;
however for medium and weak signals it results in a low signal that
shows considerable fluctuation, When using wicd for a medium-strength
AP, the value reported in the status line is reduced from 100% to 60% by
this bug.
This problem is reported in https://bugzilla.kernel.org/show_bug.cgi?id=42826.
Reported-and-tested-by: Robert Crawford <wrc1944@gmail.com> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Cc: Stable <stable@vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Larry Finger [Sun, 26 Feb 2012 00:10:21 +0000 (18:10 -0600)]
staging: r8712u: Fix regression caused by commit 8c213fa
In commit 8c213fa "staging: r8712u: Use asynchronous firmware loading",
the command to release the firmware was placed in the wrong routine.
In combination with the bug introduced in commit a5ee652 "staging: r8712u:
Interface-state not fully tracked", the driver attempts to upload firmware
that had already been released. This bug is the source of one of the
problems in https://bugs.archlinux.org/task/27996#comment89833.
Tested-by: Alberto Lago Ballesteros <saniukeokusainaya@gmail.com> Tested-by: Adrian <agib@gmx.de> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Larry Finger [Sun, 26 Feb 2012 00:10:20 +0000 (18:10 -0600)]
staging: r8712u: Fix regression introduced by commit a5ee652
In commit a5ee652 "staging: r8712u: Interface-state not fully tracked",
the private boolean "bup" was set false when the interface was brought down,
as that seemed appropriate. This change has not caused any problems when
using NetworkManager or manual control of the device; however, when wicd
control is used, there is a locking problem in wpa_supplicant, as shown in
https://bugzilla.kernel.org/show_bug.cgi?id=42818.
This fix reverts the only code change in commit a5ee652. My
analysis is that "bup" is badly named. In its present form, it
seems to indicate the up/down state of the device, but its usage
is more consistent with an initialized/uninitialized state. That
problem will be addressed in a later patch.
Note: Commit 8c213fa, which introdued asynchronous firmware loading
for this driver, exposed this bug to a greater extent. That bug
is addressed in the next patch in this series.
This bug is also responsible for the bug in
https://bugzilla.kernel.org/show_bug.cgi?id=42815. and this bug is
also part of the problems discussed at https://bugs.archlinux.org/task/27996#comment89950.
Tested-by: Alberto Lago Ballesteros <saniukeokusainaya@gmail.com> Tested-by: Adrian <agib@gmx.de> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Cc: Stable <stable@vger.kernel.org> [3.2+] Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Tue, 28 Feb 2012 08:21:46 +0000 (11:21 +0300)]
Staging: wlan-ng: still setting wrong data
In commit 1ca1a92cc6 "Staging: wlan-ng: memsetting the wrong amount of
data" I changed the code so we didn't memset() past the end of the
msg1.bssid.data[] array. Walter Harms noticed that it was weird that
we were setting the len to 6 when there were 7 elements in the array.
Pavel Roskin pointed out that the intent of the code was actually to
memset() msg1.bssid.data.data[] which is a 6 character array.
Reported-by: Walter Harms <wharms@bfs.de> Reported-by: Pavel Roskin <proski@gnu.org> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jorgyano Vieira [Sat, 25 Feb 2012 23:55:10 +0000 (21:55 -0200)]
Staging: crystalhd: Add global header
In the crystalhd_cmds.h there was a struct dependence bug:
the struct crystalhd_adp (which is declared on crystalhd_lnx.h)
is used on the crystalhd_cmd struct, however the crystalhd_lnx.h is
never included on crystalhd_cmds.h at all. Including the
crystalhd_lnx.h on crystalhd_cmds.h breaks the build,
many dependencies error occurrs, most of the type
"error: 'struct bar' has no member named 'foo'",
so I decided to reorganize the headers by adding a global header.
The gobal header crystalhd.h includes all the local headers.
The idea is that the crystalhd header will be the only included
by the others files, this will avoid the mess of many #include levels.
The order of the headers included by crystalhd.h considers the
dependencies among the headers.
Peter Huewe [Sat, 25 Feb 2012 00:19:41 +0000 (01:19 +0100)]
staging: asus_oled: Remove superfluous loop
This patch removes a superfluous loop in asus_oled.c
The code is equivalent to do{...} while (0) and thus executes the code
exactly once -> so we can simply remove the loop.
Signed-off-by: Peter Huewe <peterhuewe@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tomas Winkler [Sun, 26 Feb 2012 21:18:57 +0000 (23:18 +0200)]
staging: mei: kill struct hbm_cmd bit fields
Replace bitfield struct hbm_cmd with simple u8 as we
always access the value as whole. This allows us to remove
few ugly type casts
For possible further uses and documentation purposes we add
corresponding bitmask defines
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Seth Jennings [Tue, 28 Feb 2012 22:02:23 +0000 (16:02 -0600)]
staging: zcache: fix memory corruption bug
This patch fixes a bug where the zv code writes before the allocated
buffer, resulting in system memory corruption. This was introduced
during the switch from xvmalloc to zsmalloc.
Seth Jennings [Tue, 28 Feb 2012 22:01:37 +0000 (16:01 -0600)]
staging: zcache: fix length type mismatch
This fixes a type mismatch in the compression code where
a size_t pointer was cast to a unsigned int pointer. On
little endian archs, there is no issue. However on big
endian archs, the value is incorrect, taking the high
order bits and truncating the lower order bits.
Chris Kelly [Wed, 29 Feb 2012 16:39:57 +0000 (16:39 +0000)]
staging: ozwpan: Reduced size of oz_evtlist structure.
This structure is used in an ioctl definition and was causing the
64-bit PowerPC build to fail. The size of the array in the structure
has been reduced to fix this.
Signed-off-by: Chris Kelly <ckelly@ozmodevices.com> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The core must not modify available_scan_mask, because it causes problems
with drivers where multiple instances of the driver share the same mask set.
So make this explicit by marking available scan masks as const.
The max1363 driver needs some minor adjustment to accommodate this change.
Pull scan mask allocation into a separate function.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com> Acked-by: Jonathan Cameron <jic23@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging:iio:dac:ad5064: Add AD5628/AD5648/AD5668 support
The AD5628/AD5648/AD5668 are similar to the AD5024/AD5044/AD5064. The difference
being that they have an internal reference voltage and 8 instead of 4 DAC
channels.
staging:iio:dac:ad5064: Prepare driver for the addition of chip variants
Prepare the driver for the addition of chip variants with a different number of
channels. This is done by not hard-coding the number of channels, but instead
add a field to the chip info struct holding the number of channels. Also do not
embed the channel specs into the chip info, but rather store them independently.
This allows sharing the same channel spec between different chip infos.
staging:iio:dac:ad5064: Convert to extended channel info attributes
Use extended channel info attributes for the powerdown, powerdown_mode and
powerdown_mode_available attributes.
Note that this patch moves the chip info defintion around to avoid having to use
forward declarations for the extended channel info attributes callbacks.
Sometimes devices have per channel properties which either do not map nicely to
the current channel info scheme (e.g. string properties) or are very device
specific, so it does not make sense to add generic support for them.
Currently drivers define these attributes by hand for each channel. Depending on
the number of channels this can amount to quite a few lines of boilerplate code.
Especially if a driver supports multiple variations of a chip with different
numbers of channels. In this case it becomes necessary to have a individual
attribute list per chip variation and also a individual iio_info struct.
This patch introduces a new scheme for handling such per channel attributes
called extended channel info attributes. A extended channel info attribute
consist of a name, a flag whether it is shared and read and write callbacks.
The read and write callbacks are similar to the {read,write}_raw callbacks and
take a IIO device and a channel as their first parameters, but instead of
pre-parsed integer values they directly get passed the raw string value, which
has been written to the sysfs file.
It is possible to assign a list of extended channel info attributes to a
channel. For each extended channel info attribute the IIO core will create a new
sysfs attribute conforming to the IIO channel naming spec for the channels type,
similar as for normal info attributes. Read and write access to this sysfs
attribute will be redirected to the extended channel info attributes read and
write callbacks.
Use dev_pm_ops instead of legacy suspend/resume callbacks for IIO drivers.
Note that this patch introduces a few new #ifdef CONFIG_PM_SLEEP around the
suspend and resume callbacks to avoid warnings of unused functions if
CONFIG_PM_SLEEP is not defined.
Jonathan Cameron [Wed, 15 Feb 2012 19:48:01 +0000 (19:48 +0000)]
staging:iio:core add in kernel interface mapping and getting IIO channels.
Lifted from proposal for in kernel interface built on the out of staging
branch.
Two elements here:
* Map as defined in "inkern.h"
* Matching code to actually get the iio_dev and channel
that we want from the global list of IIO devices.
V4: Everything now built if iio is built (rather than being optional)
Removal race condition prevented by using info pointer as a check
of removal under a lock.
V3: Drop the option of registering / getting channels using dev pointer.
Stick to name only as suggested by Mark Brown (this has caused user
confusion in the regulator framework.)
V2: As per Greg KH suggestion, move over to registration by passing
the tables into the provider drivers (how regulator does it).
This does not prevent us using the original more flexible approach
if at a later date there is a usecase that demands it.
Jonathan Cameron [Wed, 15 Feb 2012 19:48:00 +0000 (19:48 +0000)]
staging:iio:core set the iio_dev.info pointer to null on unregister under lock.
This prevents use of provider callbacks after it has been unregistered.
Note that all code using this that can be called from a consumer *must*
check the pointer before using and hold the info_exist_lock throughout
the usage of the callbacks in info.
In order to detect it at runtime, we need the code handling wdt
clock available at runtime to decide whether to enable or disable
based on the baseimage symbols. Default timeout has been set to 5
seconds.
Downside is that we will lose the option to set a custom timeout
for overflow, but than can be added (if needed) as part of debugfs.
David Daney [Thu, 23 Feb 2012 19:19:31 +0000 (11:19 -0800)]
staging/octeon: Fix PHY binding in octeon-ethernet driver.
Commit d6c25be (mdio-octeon: use an unique MDIO bus name.) changed the
names used to refer to MDIO buses. The ethernet driver must be
changed to match, so that the PHY drivers can be attached.
Rabin Vincent [Wed, 22 Feb 2012 10:28:00 +0000 (15:58 +0530)]
staging: logger: hold mutex while removing reader
The readers list is traversed under the log->mutex lock
(for example from fix_up_readers()), but the deletion of
elements from this list is not being done under this lock.
Cc: Brian Swetland <swetland@google.com> Cc: Dima Zavin <dima@android.com> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chris Ball [Tue, 21 Feb 2012 21:40:41 +0000 (16:40 -0500)]
staging: Update TODO for rts5139 and rts_pstor
These are each >20k LOC drivers that embed an entire SD stack, and present
SD cards as if they were SCSI devices; both drivers should be rewritten to
be small hooks that connect the PCI (for rts_pstor) or USB (for rts5139)
hardware into Linux's MMC/xD/memorystick stacks.
Dan Carpenter [Tue, 21 Feb 2012 14:18:38 +0000 (17:18 +0300)]
Staging: wlan-ng: memsetting the wrong amount of data
p80211item_pstr6_t is the size of "msg1.bssid" (16 bytes) but
msg1.bssid.data is type p80211pstr6_t and it is smaller (7 bytes). We
had just set that memory to zeroes earlier and now we're writing over it
with 0xff because we're writing past the end of the struct.
I don't know if this actually causes a problem. It may be that we
initialize the extra 0xff bytes correctly later. But the current code
is obviously wrong and we should fix it.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Danny Kukawka [Tue, 21 Feb 2012 12:07:51 +0000 (13:07 +0100)]
Staging: et131x: unify return value of .ndo_set_mac_address if address is invalid
Unify return value of .ndo_set_mac_address if the given address
isn't valid. Return -EADDRNOTAVAIL as eth_mac_addr() already does
if is_valid_ether_addr() fails.
Signed-off-by: Danny Kukawka <danny.kukawka@bisect.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mark Einon [Mon, 20 Feb 2012 22:33:24 +0000 (22:33 +0000)]
staging: et131x: use netif_rx_ni() for packet receive
netif_rx is meant to be called from interrupts because it doesn't wake
up ksoftirqd. For calling from outside interrupts, netif_rx_ni exists.
This stops the error "NOHZ: local_softirq_panding 08" that happens on
some machines with NOHZ and plip --- it is caused by the fact that
softirq is pending and ksoftirqd is sleeping.
Signed-off-by: Mark Einon <mark.einon@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Mon, 20 Feb 2012 16:29:31 +0000 (08:29 -0800)]
staging: ramster: Dont build ramster when CONFIGFS_FS=m
Ramster can't be a module (yet) and depends on CONFIGFS_FS=y, but
allmodconfig builds with CONFIGFS_FS=m, which breaks the build.
And forcing CONFIGFS_FS=y with select breaks the build in other ways.
So just don't build ramster unless CONFIGFS_FS=y.
Also, while we're here, add a comment as to why BROKEN is depended.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Mon, 20 Feb 2012 16:29:30 +0000 (08:29 -0800)]
staging: ramster: build ramster properly when CONFIG_OCFS2=m|y
Due to some conflicting debug vars, kernel build will warn when
CONFIG_RAMSTER=y and CONFIG_OCFS2=m and will fail when
CONFIG_RAMSTER=y and CONFIG_OCFS2=y (rare).
Rename ramster mlog vars to avoid the name conflict.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
New kernel developer inspired by the 2010 FOSDEM talk. Running checkpatch on
p80211netdev.c gave the error: p80211netdev.c:153: ERROR: "foo * bar" should
be "foo *bar". Fixed it by doing what was suggested.
Signed-off-by: Sebastiaan de Haan <sebastiaan@sebastiaandehaan.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tomas Winkler [Mon, 20 Feb 2012 12:46:03 +0000 (14:46 +0200)]
staging/mei: mei-amt-version - make all function static and used
This patch eliminates following type of warnings
warning: no previous prototype for '...func...' [-Wmissing-prototypes]
For this is a single file example lets make all API-like functions
be static.
Since all static functions should be used so let's call
to amt_host_if_deinit() even if in this example it's not really necessary
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Acked-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Looks during cleanup we converted type of host_version_supported
member from UINT8 into int instead of u8.
Since we've queried only for boolean value of
this variable the bug wasn't really visible.
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jorgyano Vieira [Fri, 17 Feb 2012 23:38:09 +0000 (21:38 -0200)]
Staging: crystalhd: Get rid of unecessary BCMLOG_ENTER macro
The BCMLOG_ENTER macro is used only in five functions, perhaps
it is remainder of debugging some specific problem,
now, this macro don't seems to be useful, so it should be removed.
Chris Kelly [Mon, 20 Feb 2012 21:12:35 +0000 (21:12 +0000)]
staging: ozwpan: Added debug support
Added tracing facilities and also memory allocation and URB tracking.
This is for debugging purposes and is all optional and can be switched
out at compile time.
Signed-off-by: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chris Kelly [Mon, 20 Feb 2012 21:12:27 +0000 (21:12 +0000)]
staging: ozwpan: Added event logging support
The event logging subsystem allows internal events in the driver to
be logged. This facilitates testing the correct operation of the
driver. This subsystem is optional and can be switched out at
compile time.
Signed-off-by: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chris Kelly [Mon, 20 Feb 2012 21:12:17 +0000 (21:12 +0000)]
staging: ozwpan: Added character device support
The character device provides a management interface to the driver
and also provides an additional service to the protocol for side
band communication with the device.
Signed-off-by: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chris Kelly [Mon, 20 Feb 2012 21:11:17 +0000 (21:11 +0000)]
staging: ozwpan: Added driver entry code
This series of patches adds the Ozmo USB over WiFi driver to the
driver staging directory. This is a driver for a virtual USB HCD
and uses an L2 network protocol to talk to the device.
This patch adds the driver entry code and a README file with more
details.
Signed-off-by: Chris Kelly <ckelly@ozmodevices.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jesper Juhl [Sun, 12 Feb 2012 23:15:02 +0000 (00:15 +0100)]
Staging, rtl8192e, softmac: remove redundant memset and fix mem leak
In drivers/staging/rtl8192e/rtllib_softmac.c::rtllib_rx_assoc_resp()
we allocate memory for 'network' with kzalloc() and then proceed to
zero the already zeroed mem we got from kzalloc() with
memset(). That's redundant, so remove the memset()
We also fail to kfree() the memory we allocated for 'network' if we do not enter
if (ieee->current_network.qos_data.supported == 1) {
and the variable then goes out of scope.
To fix that I simply moved the kfree() that was inside that 'if'
statement to instead be just after it. It then covers both the case
where we take the branch and when we don't.
Dan Magenheimer [Wed, 15 Feb 2012 15:54:19 +0000 (07:54 -0800)]
staging: ramster: ramster-specific new files
RAMster implements peer-to-peer transcendent memory, allowing a "cluster"
of kernels to dynamically pool their RAM.
This patch adds new files necessary for ramster support: The file
ramster.h declares externs and some pampd bitfield manipulation. The
file zcache.h declares some zcache functions that now must be accessed
from the ramster glue code. The file r2net.c is the glue between zcache
and the messaging layer, providing routines called from zcache that
initiate messages, and routines that handle messages by calling zcache.
TODO explains future plans for merging.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Wed, 15 Feb 2012 15:54:18 +0000 (07:54 -0800)]
staging: ramster: ramster-specific changes to zcache/tmem
RAMster implements peer-to-peer transcendent memory, allowing a "cluster"
of kernels to dynamically pool their RAM.
This patch incorporates changes transforming zcache to work with
a remote store.
In tmem.[ch], new "repatriate" (provoke async get) and "localify" (handle
incoming data resulting from an async get) routines combine with a handful
of changes to existing pamops interfaces allow the generic tmem code
to support asynchronous operations. Also, a new tmem_xhandle struct
groups together key information that must be passed to remote tmem stores.
Zcache-main.c is augmented with a large amount of ramster-specific code
to handle remote operations and "foreign" pages on both ends of the
"remotify" protocol. New "foreign" pools are auto-created on demand.
A "selfshrinker" thread periodically repatriates remote persistent pages
when local memory conditions allow. For certain operations, a queue is
necessary to guarantee strict ordering as out-of-order puts/flushes can
cause strange race conditions. Pampd pointers now either point to local
memory OR describe a remote page; to allow the same 64-bits to describe
either, the LSB is used to differentiate. Some acrobatics must be performed
to ensure local memory is available to handle a remote persistent get,
or deal with the data directly anyway if the malloc failed. Lots
of ramster-specific statistics are available via sysfs.
Note: Some debug ifdefs left in for now. Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Wed, 15 Feb 2012 15:54:17 +0000 (07:54 -0800)]
staging: ramster: xvmalloc allocation files
RAMster implements peer-to-peer transcendent memory, allowing a "cluster"
of kernels to dynamically pool their RAM.
Zcache is in the process of converting allocators, from xvmalloc to zsmalloc.
Further, RAMster V5 testing to date has been done only with xvmalloc.
To avoid merging problems, a linux-3.2 copy of xvmalloc is incorporated by
this patch. Later patches will be able to eliminate xvmalloc and use zsmalloc.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Wed, 15 Feb 2012 15:54:16 +0000 (07:54 -0800)]
staging: ramster: local compression + tmem
RAMster implements peer-to-peer transcendent memory, allowing a "cluster"
of kernels to dynamically pool their RAM.
This patch copies files from drivers/staging/zcache. RAMster compresses
pages locally before transmitting them to another node, so we can
leverage the zcache and tmem code directly. Note: there are
no ramster-specific changes yet to these files.
(Why copy? The ramster tmem.c/tmem.h changes are definitely shareable
between zcache and ramster; the eventual destination for tmem.c
is the linux lib directory. Ramster changes to zcache are more substantial
and zcache is currently undergoing some significant unrelated changes
(including a new allocator and breaking zcache-main.c into smaller files),
so it seemed best to branch temporarily and merge later.)
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Magenheimer [Wed, 15 Feb 2012 15:54:15 +0000 (07:54 -0800)]
staging: ramster: cluster/messaging foundation
RAMster implements peer-to-peer transcendent memory, allowing a "cluster"
of kernels to dynamically pool their RAM.
This patch provides the cluster and messaging foundation for RAMster,
implementing the basic cluster discovery, mapping, heartbeat / keepalive,
and messaging ("r2net") that RAMster requires for internode communication.
This code heavily leverages code from the ocfs2 cluster layer but
has been extended, interfaces to userland changed, and external functions
renamed so that RAMster and ocfs2 can co-exist in the kernel and userland.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are two members of pr_ctxt allocated during bridge_open that
are never freed resulting in memory leaks, these are stream_id and
node_id, they are now freed on release of the handle (bridge_release)
right before freeing pr_ctxt.
Error path for bridge_open was also fixed since the same variables
could result in memory leaking due to missing handling of failure
scenarios. While at it, the indentation changes were introduced to
avoid interleaved goto statements inside big if blocks.
Signed-off-by: Omar Ramirez Luna <omar.ramirez@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kevin McKinney [Mon, 13 Feb 2012 03:24:53 +0000 (22:24 -0500)]
Staging: bcm: Move directives for the preprocessor statement to enum value in led_control.h
DRIVER_HALT is a driver state that was originally
defined as a #define statement. This patch moves
it to the LedEvents type as an enumerated
value for the purpose of removing a compile time warning:
drivers/staging/bcm/led_control.c: In function ‘LEDControlThread’:
drivers/staging/bcm/led_control.c:817:3: warning: case value ‘255’ not in enumerated type ‘LedEventInfo_t’ [-Wswitch]
Signed-off-by: Kevin McKinney <klmckinney1@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Nitin Gupta [Mon, 13 Feb 2012 04:04:45 +0000 (23:04 -0500)]
staging: zram: Rename module parameter
zram accepts number of devices to be created
as a module parameter. This was renamed from
num_devices to zram_num_devices (without updating
the documentation!) since num_devices was declared
as a non-static global variable, polluting the global
namespace. Now, we declare it as a static variable
and revert back the name change.
The documentation (zram.txt) already mentions
num_devices as the module parameter name.
staging:iio: Add event monitor example application
Add a small evtest like application to monitor events generated by an IIO
device. The application can be used as an example on how to listen for IIO
events and also is usful for testing and debugging device drivers which
generate IIO events.
Add macros for extracting whether the event is for a differential channel and
the second channel number from the event code. These were the only two fields
which did not have such an macro yet.
Commit 43ba1100 ("staging:iio:events: Use waitqueue lock to protect event
queue") removed the event_list_lock field from the iio_event_interface struct,
but missed to remove the same field from the documentation for that function.
Szymon Janc [Fri, 10 Feb 2012 20:05:43 +0000 (21:05 +0100)]
Staging: quickstart: Fix compilation warning on 64 bit arch
acpi_size is u32 or u64 depending on architecture. Cast it to
unsigned long and use %lu for printing.
This fix following build warning:
drivers/staging/quickstart/quickstart.c: In function ‘quickstart_acpi_ghid’:
drivers/staging/quickstart/quickstart.c:212:5: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 3 has type ‘acpi_size’ [-Wformat]
projects / karo-tx-linux.git / log