Fengguang Wu's build bot detected that if moduleloader.h is included in
a C file (used by ftrace and kprobes to access module_alloc() when
available), that it can fail to build if CONFIG_MODULES and
CONFIG_MODULES_USE_ELF_REL is not defined.
This is because there's a printk() that dereferences struct module to
print the name of the module. But as struct module does not exist when
CONFIG_MODULES is not defined we get this error:
include/linux/moduleloader.h: In function 'apply_relocate':
>> include/linux/moduleloader.h:48:63: error: dereferencing pointer to
>> incomplete type
printk(KERN_ERR "module %s: REL relocation unsupported\n", me->name);
^
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Based-on-the-true-story-by: Steven Rostedt <rostedt@goodmis.org>
Confirms-rustys-story-ends-the-same-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
With the Bluetooth 4.1 specification the Simultaneous LE and BR/EDR
controller option has been deprecated. It shall be set to zero and
ignored otherwise.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This patch fixes the following checkpatch.pl issues in gdm_usb.c:
ERROR: that open brace { should be on the previous line
Signed-off-by: Cihangir Akturk <cakturk@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This protocol is found on Dreambox remotes
[m.chehab@samsung.com: CodingStyle fixes and conflict fix]
Signed-off-by: Marcel Mol <marcel@mesa.nl>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
This adds a device tree node for the infrared receiver connected to a
GPIO pin on the Radxa Rock.
Signed-off-by: Beniamino Galvani <b.galvani@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This adds the necessary nodex and pinctrl settings for the Rockchip PWM-driver.
Signed-off-by: Beniamino Galvani <b.galvani@gmail.com>
Modified to use the new clock defines and added rk3066 pins.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Michel Dänzer and a couple of other people reported inexplicable random
oopses in the scheduler, and the cause turns out to be gcc mis-compiling
the load_balance() function when debugging is enabled. The gcc bug
apparently goes back to gcc-4.5, but slight optimization changes means
that it now showed up as a problem in 4.9.0 and 4.9.1.
The instruction scheduling problem causes gcc to schedule a spill
operation to before the stack frame has been created, which in turn can
corrupt the spilled value if an interrupt comes in. There may be other
effects of this bug too, but that's the code generation problem seen in
Michel's case.
This is fixed in current gcc HEAD, but the workaround as suggested by
Markus Trippelsdorf is pretty simple: use -fno-var-tracking-assignments
when compiling the kernel, which disables the gcc code that causes the
problem. This can result in slightly worse debug information for
variable accesses, but that is infinitely preferable to actual code
generation problems.
Doing this unconditionally (not just for CONFIG_DEBUG_INFO) also allows
non-debug builds to verify that the debug build would be identical: we
can do
export GCC_COMPARE_DEBUG=1
to make gcc internally verify that the result of the build is
independent of the "-g" flag (it will make the compiler build everything
twice, toggling the debug flag, and compare the results).
Without the "-fno-var-tracking-assignments" option, the build would fail
(even with 4.8.3 that didn't show the actual stack frame bug) with a gcc
compare failure.
See also gcc bugzilla:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61801
Reported-by: Michel Dänzer <michel@daenzer.net>
Suggested-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Cc: Jakub Jelinek <jakub@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use the newly ammended dw_8250 clock binding to define both the baudclk as
well as the pclk supplying the ip.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This enables the 2nd i2c bus and adds the act8846 pmic as device.
Signed-off-by: Beniamino Galvani <b.galvani@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The Curie2 uses a tps659102 as its main pmic, so add the i2c1 and tps65910
node as well as define the used voltages and regulator-names according to
the schematics.
Also fix the supply of the sd0 regulator, as it is supplied by the vio reg.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Shortly before 3.16-rc1, Dave Jones reported:
WARNING: CPU: 3 PID: 19721 at fs/xfs/xfs_aops.c:971
xfs_vm_writepage+0x5ce/0x630 [xfs]()
CPU: 3 PID: 19721 Comm: trinity-c61 Not tainted 3.15.0+ #3
Call Trace:
xfs_vm_writepage+0x5ce/0x630 [xfs]
shrink_page_list+0x8f9/0xb90
shrink_inactive_list+0x253/0x510
shrink_lruvec+0x563/0x6c0
shrink_zone+0x3b/0x100
shrink_zones+0x1f1/0x3c0
try_to_free_pages+0x164/0x380
__alloc_pages_nodemask+0x822/0xc90
alloc_pages_vma+0xaf/0x1c0
handle_mm_fault+0xa31/0xc50
etc.
970 if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
971 PF_MEMALLOC))
I did not respond at the time, because a glance at the PageDirty block
in shrink_page_list() quickly shows that this is impossible: we don't do
writeback on file pages (other than tmpfs) from direct reclaim nowadays.
Dave was hallucinating, but it would have been disrespectful to say so.
However, my own /var/log/messages now shows similar complaints
WARNING: CPU: 1 PID: 28814 at fs/ext4/inode.c:1881 ext4_writepage+0xa7/0x38b()
WARNING: CPU: 0 PID: 27347 at fs/ext4/inode.c:1764 ext4_writepage+0xa7/0x38b()
from stressing some mmotm trees during July.
Could a dirty xfs or ext4 file page somehow get marked PageSwapBacked,
so fail shrink_page_list()'s page_is_file_cache() test, and so proceed
to mapping->a_ops->writepage()?
Yes, 3.16-rc1's commit 68711a7463 ("mm, migration: add destination
page freeing callback") has provided such a way to compaction: if
migrating a SwapBacked page fails, its newpage may be put back on the
list for later use with PageSwapBacked still set, and nothing will clear
it.
Whether that can do anything worse than issue WARN_ON_ONCEs, and get
some statistics wrong, is unclear: easier to fix than to think through
the consequences.
Fixing it here, before the put_new_page(), addresses the bug directly,
but is probably the worst place to fix it. Page migration is doing too
many parts of the job on too many levels: fixing it in
move_to_new_page() to complement its SetPageSwapBacked would be
preferable, except why is it (and newpage->mapping and newpage->index)
done there, rather than down in migrate_page_move_mapping(), once we are
sure of success? Not a cleanup to get into right now, especially not
with memcg cleanups coming in 3.17.
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The pincontroller uses the GRF and PMU syscons nowadays, so should not
contain an address in its device node.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Some nodes that are changed in the dtsi hierarchy do not have handles yet.
As it was suggested in the rk3288 submission to do subsequent nodes changes
through such handle-references, add the missing ones.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Comments received from the rk3288 submission indicated that a generic subnode
to group soc components should not be used.
So to keep all rockchip devicetree files similar, remove it from rk3066 and rk3188.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The clock and reset unit is now provided by the rk3188-cru clock driver and thus
the old style definitions of the gate clocks can go away.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-By: Max Schwarz <max.schwarz@online.de>
Tested-By: Max Schwarz <max.schwarz@online.de>
The clock definitions get a lot shorter due to the soc clocks being handled by
rk3188-cru and only the input clock remains. These can now simply live
in the main rk3xxx.dtsi without affecting readability.
At the same time, rename the node to oscillator, adding a clock-output-names
property to match how the rk3288 handles this.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-By: Max Schwarz <max.schwarz@online.de>
Tested-By: Max Schwarz <max.schwarz@online.de>
This adds a node for the clock and reset unit on rk3188 and rk3066 SoCs and
also updates the device nodes retrieve their clocks from there, instead of
the previous gate clock nodes.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-By: Max Schwarz <max.schwarz@online.de>
Tested-By: Max Schwarz <max.schwarz@online.de>
Such as found on the new Toshiba Portégé Z30-A and Z40-A.
Signed-off-by: Yunkang Tang <yunkang.tang@cn.alps.com>
[hdegoede@redhat.com: Remove softbutton handling, this is done in userspace]
[hdegoede@redhat.com: Report INPUT_PROP_BUTTONPAD]
[hdegoede@redhat.com: Do not report fake PRESSURE, reporting BTN_TOUCH is
enough]
[hdegoede@redhat.com: Various cleanups / refactoring]
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
So that decode functions can return a failure when appropriate.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
If we detect more then 2 fingers report 2 touches, rather then only
reporting the upper left corner of the bounding box.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Move all the semi-mt specific handling shared between the v3 and v4
handling code to a common helper function.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
For v3 protocol devices, use the more accurate single touch data when the
mt data contains only one finger. Note the mt data reporting a finger count
of 1 should never happen, but better safe then sorry.
This brings the v3 bitmap handling in line with what the v4 code does,
allowing to factor out the common bits into a helper function.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
When there are 2 fingers on the pad we don't know which one is which, so
use input_mt_assign_slots to make sure the right set of coordinates ends
up in the right slot.
Besides ensuring things end up in the right slot, this also results in a
nice cleanup, since sync_frame also handles non mt position and btn_touch
reporting.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
This is a preparation patch for switching the DIY mt handling to using
input_mt_assign_slots && input_mt_sync_frame.
struct alps_fields is quite large, so while making changes to almost all uses
of it lets put it in our priv data instead of on the stack.
Having it in our priv data also allows using it directly for storing values
which need to be cached, rather then having separate x, y, z, fingers, etc.
copies in our priv data.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
This fixes 2 fingers at the same height or width on the touchpad getting
reported at different y / x coordinates.
Note num_bits is always at least 1.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
alps_process_bitmap was resetting the point bit-count as soon as it saw
2 0 bits in a row. This means that unless the high point actually is at
the end of the bitmap, it would always get its num_bits set to 0.
Instead reset num_bits to 0 on a 0->1 transition, so that with > 2 fingers
we only count the number of bits occupied by the highest finger.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Factor out the identical code for getting the bitmap points for x and y into
a helper function.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Rushmore models don't have the Y-axis data in the bitmap inverted. Since
we now have 2 different Y orientations, make the Y bitmap data processing
use a forward loop like the X bitmap data processing, unifying the 2,
and invert the data later, except on Rushmore.
So far no-one has noticed this because the synaptics driver only uses the
non mt coordinates (except on clickpads, and there are no alps clickpads
using process_bitmap).
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
V3 models only report mt bitmap data when there are 2 or more fingers on
the touchpad. So always generate 2 positions in alps_process_bitmap, and
for v3 models only fall back to st data when there was no mt data in a
mt packet (which should never happen).
This fixes 2 finger scrolling not working when using 2 fingers close to
each other.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Multiple pinctrl states are defined for 8, 16 and 24 data pin groups in PPI peripheral.
The driver should select correct group before set up further PPI parameters.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Acked-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
On NOMMU architecture page fault is not triggered if a deleted list entry is
accessed without reinit.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Acked-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
if the pinctrl driver is enabled.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Acked-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Fix sparse warnings:
drivers/staging/rtl8188eu/core/rtw_cmd.c:52:5: warning:
symbol '_rtw_enqueue_cmd' was not declared. Should it be static?
drivers/staging/rtl8188eu/core/rtw_wlan_util.c:1225:5: warning:
symbol 'wifirate2_ratetbl_inx' was not declared. Should it be static?
drivers/staging/rtl8188eu/hal/HalPhyRf_8188e.c:25:4: warning:
symbol 'ODM_GetRightChnlPlaceforIQK' was not declared. Should it be static?
drivers/staging/rtl8188eu/os_dep/ioctl_linux.c:3109:5: warning:
symbol 'rtw_ioctl' was not declared. Should it be static?
Signed-off-by: Jeff Oczek <jeffoczek@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes an old copy+paste bug that has survived all recent code
changes in this code area.
Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Instead of counting the number of opened file handles, use function
v4l2_fh_is_singular_file() in em28xx_v4l2_open() and em28xx_v4l2_close() to
determine if the file handle is the first/last opened one.
Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Instead of calling
...
struct v4l2_fh *fh = kzalloc(sizeof(*fh), GFP_KERNEL);
filp->private_data = fh;
v4l2_fh_init(fh, vdev);
v4l2_fh_add(fh);
...
simply use function v4l2_fh_open() which does all of these calls for us.
Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
struct em28xx_fh isn't needed anymore because the only used field which is left is struct v4l2_fh fh.
Use struct v4l2_fh directly and remvove struct em28xx_fh.
Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Pointer v4l2" can be used instead of "dev->v4l2, which saves some characters.
Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
This patch fix checkpatch:
WARNING: else is not generally useful after a break or return
WARNING: line over 80 characters
[m.chehab@samsung.com: Fix conflicts and use a better coding style for
passing the parameters to parport_register_device()]
Signed-off-by: Zheng Di <zhengdi05@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Expression ((7bit i2c_addr << 1) & 0x01) can not be right
because it is always 0
Signed-off-by: Raimonds Cicans <ray@apollo.lv>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
The crashed member of struct iss_device is documented to be a bitmask,
but a bool doesn't hold that many (usable) bits. Lines 589 and 659 of
iss.c strongly suggest that "unsigned int" was meant (the same type as
struct iss_pipeline::entities). Currently, any crashed entity will be
blamed on index 0, which is unlikely to be what was intended.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
This patch introduces the use of managed interfaces like
devm_request_mem_region and devm_request_irq and does away with the
calls to free the allocated memory in the probe and remove functions.
The remove function is no longer required and is removed completely.
Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
This patch fix checkpath "WARNING: void function return statements are not generally useful".
The removed return were useless in that case.
Signed-off-by: Raphaël Poggi <poggi.raph@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
This patch fix checkpath "WARNING: Missing a blank line after declarations"
Signed-off-by: Raphaël Poggi <poggi.raph@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
