MIPS: math-emu: <MAX|MAXA|MIN|MINA>.<D|S>: Fix cases of both inputs zero

commit 15560a58bfd4ff82cdd16b2270d4ef9b06d2cc4d upstream.

Fix the value returned by <MAX|MAXA|MIN|MINA>.<D|S>, if both inputs
are zeros. The right behavior in such cases is stated in instruction
reference manual and is as follows:

   fs  ft       MAX     MIN       MAXA    MINA
  ---------------------------------------------
    0   0        0       0         0       0
    0  -0        0      -0         0      -0
   -0   0        0      -0         0      -0
   -0  -0       -0      -0        -0      -0

Prior to this patch, some of the above cases were yielding correct
results. However, for the sake of code consistency, all such cases
are rewritten in this patch.

A relevant example:

MAX.S fd,fs,ft:
  If fs contains +0.0, and ft contains -0.0, fd is going to contain
  +0.0 (without this patch, it used to contain -0.0).

Fixes: a79f5f9ba508 ("MIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction")
Fixes: 4e9561b20e2f ("MIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction")
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Cc: Bo Hu <bohu@google.com>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: Jin Qian <jinqian@google.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/16881/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

MIPS: math-emu: <MAX|MAXA|MIN|MINA>.<D|S>: Fix quiet NaN propagation

commit e78bf0dc4789bdea1453595ae89e8db65918e22e upstream.

Fix the value returned by <MAX|MAXA|MIN|MINA>.<D|S> fd,fs,ft, if both
inputs are quiet NaNs. The <MAX|MAXA|MIN|MINA>.<D|S> specifications
state that the returned value in such cases should be the quiet NaN
contained in register fs.

A relevant example:

MAX.S fd,fs,ft:
  If fs contains qNaN1, and ft contains qNaN2, fd is going to contain
  qNaN1 (without this patch, it used to contain qNaN2).

Fixes: a79f5f9ba508 ("MIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction")
Fixes: 4e9561b20e2f ("MIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction")
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Cc: Bo Hu <bohu@google.com>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: Jin Qian <jinqian@google.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/16880/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Input: i8042 - add Gigabyte P57 to the keyboard reset table

commit 697c5d8a36768b36729533fb44622b35d56d6ad0 upstream.

Similar to other Gigabyte laptops, the touchpad on P57 requires a
keyboard reset to detect Elantech touchpad correctly.

BugLink: https://bugs.launchpad.net/bugs/1594214
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

pinctrl/amd: save pin registers over suspend/resume

commit 79d2c8bede2c93f9432d7da0bc2f76a195c90fc0 upstream.

The touchpad in the Asus laptop models X505BA/BP and X542BA/BP is
unresponsive after suspend/resume. The following error appears during
resume:

  i2c_hid i2c-ELAN1300:00: failed to reset device.

The problem here is that i2c_hid does not notice the interrupt being
generated at this point, because the GPIO is no longer configured
for interrupts.

Fix this by saving pinctrl-amd pin registers during suspend and
restoring them at resume time.

Based on code from pinctrl-intel.

Signed-off-by: Daniel Drake <drake@endlessm.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tty: fix __tty_insert_flip_char regression

commit 8a5a90a2a477b86a3dc2eaa5a706db9bfdd647ca upstream.

Sergey noticed a small but fatal mistake in __tty_insert_flip_char,
leading to an oops in an interrupt handler when using any serial
port.

The problem is that I accidentally took the tty_buffer pointer
before calling __tty_buffer_request_room(), which replaces the
buffer. This moves the pointer lookup to the right place after
allocating the new buffer space.

Fixes: 979990c62848 ("tty: improve tty_insert_flip_char() fast path")
Reported-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tty: improve tty_insert_flip_char() slow path

commit 065ea0a7afd64d6cf3464bdd1d8cd227527e2045 upstream.

While working on improving the fast path of tty_insert_flip_char(),
I noticed that by calling tty_buffer_request_room(), we needlessly
move to the separate flag buffer mode for the tty, even when all
characters use TTY_NORMAL as the flag.

This changes the code to call __tty_buffer_request_room() with the
correct flag, which will then allocate a regular buffer when it rounds
out of space but no special flags have been used. I'm guessing that
this is the behavior that Peter Hurley intended when he introduced
the compacted flip buffers.

Fixes: acc0f67f307f ("tty: Halve flip buffer GFP_ATOMIC memory consumption")
Cc: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

tty: improve tty_insert_flip_char() fast path

commit 979990c6284814617d8f2179d197f72ff62b5d85 upstream.

kernelci.org reports a crazy stack usage for the VT code when CONFIG_KASAN
is enabled:

drivers/tty/vt/keyboard.c: In function 'kbd_keycode':
drivers/tty/vt/keyboard.c:1452:1: error: the frame size of 2240 bytes is larger than 2048 bytes [-Werror=frame-larger-than=]

The problem is that tty_insert_flip_char() gets inlined many times into
kbd_keycode(), and also into other functions, and each copy requires 128
bytes for stack redzone to check for a possible out-of-bounds access on
the 'ch' and 'flags' arguments that are passed into
tty_insert_flip_string_flags as a variable-length string.

This introduces a new __tty_insert_flip_char() function for the slow
path, which receives the two arguments by value. This completely avoids
the problem and the stack usage goes back down to around 100 bytes.

Without KASAN, this is also slightly better, as we don't have to
spill the arguments to the stack but can simply pass 'ch' and 'flag'
in registers, saving a few bytes in .text for each call site.

This should be backported to linux-4.0 or later, which first introduced
the stack sanitizer in the kernel.

Fixes: c420f167db8c ("kasan: enable stack instrumentation")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

IB/addr: Fix setting source address in addr6_resolve()

commit 79e25959403e6a79552db28a87abed34de32a1df upstream.

Commit eea40b8f624f ("infiniband: call ipv6 route lookup via the stub
interface") introduced a regression in address resolution when connecting
to IPv6 destination addresses.  The old code called ip6_route_output(),
while the new code calls ipv6_stub->ipv6_dst_lookup().  The two are almost
the same, except that ipv6_dst_lookup() also calls ip6_route_get_saddr()
if the source address is in6addr_any.

This means that the test of ipv6_addr_any(&fl6.saddr) now never succeeds,
and so we never copy the source address out.  This ends up causing
rdma_resolve_addr() to fail, because without a resolved source address,
cma_acquire_dev() will fail to find an RDMA device to use.  For me, this
causes connecting to an NVMe over Fabrics target via RoCE / IPv6 to fail.

Fix this by copying out fl6.saddr if ipv6_addr_any() is true for the original
source address passed into addr6_resolve().  We can drop our call to
ipv6_dev_get_saddr() because ipv6_dst_lookup() already does that work.

Fixes: eea40b8f624 ("infiniband: call ipv6 route lookup via the stub interface")
Signed-off-by: Roland Dreier <roland@purestorage.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Raju Rangoju <rajur@chelsio.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drm/sun4i: Implement drm_driver lastclose to restore fbdev console

commit 2a596fc9d974bb040eda9ab70bf8756fcaaa6afe upstream.

The drm_driver lastclose callback is called when the last userspace
DRM client has closed. Call drm_fbdev_cma_restore_mode to restore
the fbdev console otherwise the fbdev console will stop working.

Fixes: 9026e0d122ac ("drm: Add Allwinner A10 Display Engine support")
Tested-by: Olliver Schinagl <oliver@schinagl.nl>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
[net147@gmail.com: Backport to 4.9, minor context change]
Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

IB/{qib, hfi1}: Avoid flow control testing for RDMA write operation

commit 5b0ef650bd0f820e922fcc42f1985d4621ae19cf upstream.

Section 9.7.7.2.5 of the 1.3 IBTA spec clearly says that receive
credits should never apply to RDMA write.

qib and hfi1 were doing that.  The following situation will result
in a QP hang:
- A prior SEND or RDMA_WRITE with immmediate consumed the last
  credit for a QP using RC receive buffer credits
- The prior op is acked so there are no more acks
- The peer ULP fails to post receive for some reason
- An RDMA write sees that the credits are exhausted and waits
- The peer ULP posts receive buffers
- The ULP posts a send or RDMA write that will be hung

The fix is to avoid the credit test for the RDMA write operation.

Reviewed-by: Kaike Wan <kaike.wan@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

orangefs: Don't clear SGID when inheriting ACLs

commit b5accbb0dfae36d8d36cd882096943c98d5ede15 upstream.

When new directory 'DIR1' is created in a directory 'DIR0' with SGID bit
set, DIR1 is expected to have SGID bit set (and owning group equal to
the owning group of 'DIR0'). However when 'DIR0' also has some default
ACLs that 'DIR1' inherits, setting these ACLs will result in SGID bit on
'DIR1' to get cleared if user is not member of the owning group.

Fix the problem by creating __orangefs_set_acl() function that does not
call posix_acl_update_mode() and use it when inheriting ACLs. That
prevents SGID bit clearing and the mode has been properly set by
posix_acl_create() anyway.

Fixes: 073931017b49d9458aa351605b43a7e34598caef
CC: stable@vger.kernel.org
CC: Mike Marshall <hubcap@omnibond.com>
CC: pvfs2-developers@beowulf-underground.org
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: prevent double decrease of nr_reserved_highatomic

commit 4855e4a7f29d6d10b0b9c84e189c770c9a94e91e upstream.

There is race between page freeing and unreserved highatomic.

 CPU 0     CPU 1

    free_hot_cold_page
      mt = get_pfnblock_migratetype
      set_pcppage_migratetype(page, mt)
         unreserve_highatomic_pageblock
         spin_lock_irqsave(&zone->lock)
         move_freepages_block
         set_pageblock_migratetype(page)
         spin_unlock_irqrestore(&zone->lock)
      free_pcppages_bulk
        __free_one_page(mt) <- mt is stale

By above race, a page on CPU 0 could go non-highorderatomic free list
since the pageblock's type is changed.  By that, unreserve logic of
highorderatomic can decrease reserved count on a same pageblock severak
times and then it will make mismatch between nr_reserved_highatomic and
the number of reserved pageblock.

So, this patch verifies whether the pageblock is highatomic or not and
decrease the count only if the pageblock is highatomic.

Link: http://lkml.kernel.org/r/1476259429-18279-3-git-send-email-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Sangseok Lee <sangseok.lee@lge.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Miles Chen <miles.chen@mediatek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

vcsm: use dma APIs for cache functions

* Will handle multi-platform builds

vcsm: add macros for cache functions

arm: partially revert 702b94bff3c50542a6e4ab9a4f4cef093262fe65

* Re-expose some dmi APIs for use in VCSM

vc-sm: Fix a printf of a dma_addr_t from %u to %pad

Avoids issues when other build parameters result in
dma_addr_t being a 64 bit value.
See #2196.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

Touchscreen spurious event fix

Touchscreen driver was not checking event type,
which meant bad coords that arrived with the
finger up event were being used, causing this
spurious coordinate event.

Fix is to ignore all finger up events.

Signed-off-by: James Hughes <james.hughes@raspberrypi.org>

vcsm: fix multi-platform build

cgroup: Add kernel param to enable memory group

cgroup support is useful, but enabling the "memory" cgroup has a 32-byte
overhead per page. Arrange that "memory" cgroup is disabled by default
unless the "cgroup_memory" parameter is set to a true value.

See: https://github.com/raspberrypi/issues/1950

Signed-off-by: Phil Elwell <phil@raspberrypi.org>

Revert "config: Enable CONFIG_MEMCG, but leave it disabled (due to memory cost). Enable with cgroup_enable=memory."

This reverts commit 7dbc189967bef8e7f431c982557d0e73eb13f161.

config: Add CONFIG_GPIO_WATCHDOG=m

Enable the gpio_wdt module, to support external watchdogs.

See: https://github.com/raspberrypi/linux/issues/2199

Signed-off-by: Phil Elwell <phil@raspberrypi.org>

vcsm: Handle the clean+invalidate case directly

vcsm: Replace SM_PDE_T with sm_pde_t

Follows the Linux kernel coding style

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Replace SM_PRIV_DATA_T with sm_priv_data_t

Follows the Linux kernel coding style

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Replace SM_STATE_T with sm_state_t

Follow the Linux kernel coding style

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Replace SM_STATS_T with sm_stats_t

Follows the kernel coding style

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Replace SM_RESOURCE_T with sm_resource_t

Follow the Linux kernel coding style.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Remove VC_SM_LOCK_CACHE_MODE_T typedef.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix "Prefer kcalloc over kzalloc with multiply"

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix "void function return statements are not generally useful"

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix "struct vm_operations_struct should normally be const"

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Correct "long unsigned int" to "unsigned long"

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix use of S_IRUGO and use 0444 instead

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix spaces around operators.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix indentation of switch/case statement.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix lots of block quote formatting issues

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix erroneous space in 'if' statement.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix blank lines after declarations.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: "unsigned int" preferred over "unsigned" fixed.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Remove typedef of VC_VCHI_SM_HANDLE_T

Replaced with struct sm_instance *.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Fix block comment formatting

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Correct block comment style on licence headers.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Remove all typedefs from vc_sm_defs.h and calling code

Remove typedefs on the structures that make up the IPC
to the firmware for VCSM. Update all calling code appropriately.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

VCSM: New option to import a DMABUF for VPU use

Takes a dmabuf, and then calls over to the VPU to wrap
it into a suitable handle.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Convert to loading via device tree.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: code-style: Fix comment indentation

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

VCSM: Specify pointers in IPC msgs as uint32_t

Pointers were specified as void* which will fail
rather badly if the kernel is switched to 64 bit.
GPU wants 32 bit addresses, so use uint32_t and
cast whereever necessary.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>

vcsm: Allow both clean and invalidate to be requested

brcmfmac: Disable ARP offloading when promiscuous

This is a test patch for brcmfmac from Franky Lin at Broadcom to disable
ARP offloading when in promiscuous mode, re-enabling the ability to
sniff ARP packets over WiFi.

See: https://github.com/raspberrypi/linux/issues/2171

Signed-off-by: Phil Elwell <phil@raspberrypi.org>

mm: Remove the PFN busy warning

See commit dae803e165a11bc88ca8dbc07a11077caf97bbcb -- the warning is
expected sometimes when using CMA.  However, that commit still spams
my kernel log with these warnings.

Signed-off-by: Eric Anholt <eric@anholt.net>

vcsm: Allow build with multi-platform v6-v7

drm/vc4: Fix refcounting of runtime PM get if it errors out.

We were returning without decrementing if the error happened, meaning
that at the next submit we wouldn't try to bring up the power domain.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: http://patchwork.freedesktop.org/patch/msgid/20170417162603.12726-1-eric@anholt.net
Reviewed-by: Sean Paul <seanpaul@chromium.org>
(cherry picked from commit 925d05e1f825db9490da33afe35bd5383d301e97)

drm/vc4: Demote user-accessible DRM_ERROR paths to DRM_DEBUG.

Userspace shouldn't be able to spam dmesg by passing bad arguments.
This has particularly become an issues since we started using a bad
argument to set_tiling to detect if set_tiling was supported.

Signed-off-by: Eric Anholt <eric@anholt.net>
Fixes: 83753117f1de ("drm/vc4: Add get/set tiling ioctls.")

drm/vc4: Fix pitch setup for T-format scanout.

The documentation said to use src_w here, and I didn't consider that
we actually needed to be using pitch somewhere in our setup.  Fixes
scanout on my DSI panel when X11 does initial setup with 1920x1080
HDMI and 800x480 DSI both at 0,0 of the same framebuffer.

Signed-off-by: Eric Anholt <eric@anholt.net>
Fixes: 98830d91da08 ("drm/vc4: Add T-format scanout support.")

drm/vc4: prepare for CEC support

In order to support CEC the hsm clock needs to be enabled in
vc4_hdmi_bind(), not in vc4_hdmi_encoder_enable(). Otherwise you wouldn't
be able to support CEC when there is no hotplug detect signal, which is
required by some monitors that turn off the HPD when in standby, but keep
the CEC bus alive so they can be woken up.

The HDMI core also has to be enabled in vc4_hdmi_bind() for the same
reason.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20170716104804.48308-3-hverkuil@xs4all.nl
(cherry picked from commit 10ee275cb12f884ce09bed69ea387eae73d7fece)

drm/vc4: Add runtime PM support to the HDMI encoder driver

The HDMI driver is currently enabling all clocks at probe time and
keeps the power-domain connected to the HDMI encoder enabled.

Move all activation code to vc4_hdmi_encoder_enable() and make sure
the clks and power domain are released when the HDMI encoder is not used
by adding deactivation steps in vc4_hdmi_encoder_disable().

Note that the sequencing imposed by the IP requires that we move
vc4_hdmi_encoder_mode_set() code into vc4_hdmi_encoder_enable().

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
(cherry picked from commit 4f6e3d66ac522dec9733d433ad00e8e77747c372)

drm/vc4: Fix misleading name of the continuous flag.

The logic was all right in the end, the name was just backwards.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: http://patchwork.freedesktop.org/patch/msgid/20170627195839.3338-3-eric@anholt.net
Reviewed-by: Andrzej Hajda <a.hajda@samsung.com>
(cherry picked from commit ec878c0756a0c202e86256dca1de307ab1189ab8)

drm/vc4: Fix DSI T_INIT timing.

The DPHY spec requires a much larger T_INIT than I was specifying
before.  In the absence of clear specs from the slave of what their
timing is, just use the value that the firmware was using.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: http://patchwork.freedesktop.org/patch/msgid/20170627195839.3338-2-eric@anholt.net
Reviewed-by: Andrzej Hajda <a.hajda@samsung.com>
(cherry picked from commit e65d51126f89a0d67ee6c5df58363730b1410ab5)

drm/vc4: Convert more lock requirement comments to lockdep assertions.

Since I do my development with lockdep on, this will help make sure I
don't introduce bugs here.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20170725182718.31468-3-eric@anholt.net
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
(cherry picked from commit 4e6b1e91253ebf701b4ec3586cba5596527a4d0a)

drm/vc4: Add an ioctl for labeling GEM BOs for summary stats

This has proven immensely useful for debugging memory leaks and
overallocation (which is a rather serious concern on the platform,
given that we typically run at about 256MB of CMA out of up to 1GB
total memory, with framebuffers that are about 8MB ecah).

The state of the art without this is to dump debug logs from every GL
application, guess as to kernel allocations based on bo_stats, and try
to merge that all together into a global picture of memory allocation
state.  With this, you can add a couple of calls to the debug build of
the 3D driver and get a pretty detailed view of GPU memory usage from
/debug/dri/0/bo_stats (or when we debug print to dmesg on allocation
failure).

The Mesa side currently labels at the gallium resource level (so you
see that a 1920x20 pixmap has been created, presumably for the window
system panel), but we could extend that to be even more useful with
glObjectLabel() names being sent all the way down to the kernel.

(partial) example of sorted debugfs output with Mesa labeling all
resources:

               kernel BO cache:  16392kb BOs (3)
       tiling shadow 1920x1080:   8160kb BOs (1)
       resource 1920x1080@32/0:   8160kb BOs (1)
scanout resource 1920x1080@32/0:   8100kb BOs (1)
                        kernel:   8100kb BOs (1)

v2: Use strndup_user(), use lockdep assertion instead of just a
    comment, fix an array[-1] reference, extend comment about name
    freeing.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20170725182718.31468-2-eric@anholt.net
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
(cherry picked from commit f30994622b2bf8e4fa224237ac65304b27a9cb6a)

drm/vc4: Start using u64_to_user_ptr.

Chris Wilson pointed out this little cleanup in a review of new code,
so let's fix up the code I was copying from.

Signed-off-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20170725182718.31468-1-eric@anholt.net
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
(cherry picked from commit 95d7cbcb6fbe3885d7a968cc86b98a08466742af)

drm/vc4: Don't disable DSI clocks on component unload.

The clocks are enabled/disabled at encoder enable/disable time, not at
component load.  Fixes a WARN_ON at boot if V3D fails to probe.

Fixes: 4078f5757144 ("drm/vc4: Add DSI driver")
Signed-off-by: Eric Anholt <eric@anholt.net>

drm/vc4: Fix double destroy of the BO cache on teardown.

It's also destroyed from the top level vc4_drv.c initialization, which
is where the cache was actually initialized from.

This used to just involve duplicate del_timer() and cancel_work_sync()
being called, but it started causing kmalloc issues once we
double-freed the new BO label array.

Fixes: 1908a876f909 ("drm/vc4: Add an ioctl for labeling GEM BOs for summary stats")
Signed-off-by: Eric Anholt <eric@anholt.net>

cache: Fix up previous commit for v6

vcsm: Provide new ioctl to clean/invalidate a 2D block

vcsm: Provide flag to map pages up front

cache: export clean and invalidate

config: Enable CONFIG_BRCMDBG temporarily for debugging on pi0

BCM270X_DT: Set spidev spi-max-frequency to 125MHz

The BCM2835 SPI controllers have a maximum bus clock of half the system
clock speed, so with the 250MHz system clock found on Raspberry Pis
you get a theoretical maximum bus speed of 125MHz. Note that this
speed is unlikely to be reliable, and the maximum usable bus spee will
depend on both the attached device and the wiring.

See: https://github.com/raspberrypi/linux/issues/2165
Signed-off-by: Phil Elwell <phil@raspberrypi.org>

config: Enable CONFIG_BRCMDBG temporarily for debugging

allo-digione: 192kHz clicking sound fix

See: https://github.com/raspberrypi/linux/pull/2149

allo-piano-dac-plus: Master volume added

Master volume added, which controls both DACs volumes.

See: https://github.com/raspberrypi/linux/pull/2149

overlays: i2c-rtc: add m41t62

Add support for the ST M41T62 real-time clock chip.

Bug fix: unmuted

ASoC: bcm2835: Enforce full symmetry

bcm2835's configuration registers can't be changed when a stream
is running, which means asymmetric configurations aren't supported.

Channel and rate symmetry are already enforced by constraints
but samplebits had been missed.

As hw_params doesn't check for symmetry constraints by itself
and just returns success if a stream is running this led to
situations where asymmetric configurations were seeming to
succeed but of course didn't work because the hardware wasn't
configured at all.

Fix this by adding the missing samplerate symmetry constraint.

Signed-off-by: Matthias Reichl <hias@horus.com>

ASoC: bcm2835: Support additional samplerates up to 384kHz

Sample rates are only restricted by the capabilities of the
clock driver, so use SNDRV_PCM_RATE_CONTINUOUS instead of
SNDRV_PCM_RATE_8000_192000.

Tests (eg with pcm5122) have shown that bcm2835 works fine
in 384kHz/32bit stereo mode, so change the maximum allowed
rate from 192kHz to 384kHz.

Signed-off-by: Matthias Reichl <hias@horus.com>

ASoC: bcm2835: Support left/right justified and DSP modes

DSP modes and left/right justified modes can be supported
on bcm2835 by configuring the frame sync polarity and
frame sync length registers and by adjusting the
channel data position registers.

Clock and frame sync polarity handling in hw_params has
been refactored to make the interaction between logical
rising/falling edge frame start and physical configuration
(changed by normal/inverted polarity modes) clearer.

Modes where the first active data bit is transmitted immediately
after frame start (eg DSP mode B with slot 0 active)
only work reliable if bcm2835 is configured as frame master.
In frame slave mode channel swap (or shift, this isn't quite
clear yet) can occur.

Currently the driver only warns if an unstable configuration
is detected but doensn't prevent using them.

Signed-off-by: Matthias Reichl <hias@horus.com>

ASoC: bcm2835: Add support for TDM modes

bcm2835 supports arbitrary positioning of channel data within
a frame and thus is capable of supporting TDM modes. Since
the driver is limited to 2-channel operations only TDM setups
with exactly 2 active slots are supported.

Logical TDM slot numbering follows the usual convention:

For I2S-like modes, with a 50% duty-cycle frame clock,
slots 0, 2, ... are transmitted in the first half of a frame,
slots 1, 3, ... are transmitted in the second half.

For DSP modes slot numbering is ascending: 0, 1, 2, 3, ...

Channel position calculation has been refactored to use
TDM info and moved out of hw_params.

set_tdm_slot, set_bclk_ratio and hw_params now check more
strictly if the configuration is valid. Illegal configurations
like odd number of slots in I2S mode, data lengths exceeding
slot width or frame sizes larger than the hardware limit of
1024 are rejected. Also hw_params now properly checks for
errors from clk_set_rate.

Allowed PCM formats are already guarded by stream constraints,
thus the formats check in hw_params has been removed and
data_length is now retrieved via params_width().

Also standard functions like snd_soc_params_to_bclk are now
being used instead of manual calculations to make the code
more readable.

Special care has been taken to ensure that set_bclk_ratio works
as before. The bclk ratio is mapped to a 2-channel TDM config
with a slot width of half the ratio. In order to support odd ratios,
which can't be expressed via a TDM config, the ratio (frame length)
is stored and used by hw_params.

Signed-off-by: Matthias Reichl <hias@horus.com>

Sets the BCDC priority to constant 0

This is to workaround for a possible issue in the
wireless chip firmware where some packets with
higher priorities seem to go missing.

See https://github.com/raspberrypi/linux/issues/1342 for
details.

BCM270X_DT: Add PaPiRus overlay

Add Device Tree overlay for the PaPiRus ePaper Screens by Pi Supply.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

config: Enable TINYDRM and drivers

Enable tinydrm library, mi0283qt and repaper drivers.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/tinydrm: Add RePaper e-ink driver

This adds support for the Pervasive Displays RePaper branded displays.
The controller code is taken from the userspace driver available
through repaper.org. Only the V231 film is supported since the others
are EOL.

Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
[rebase on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/tinydrm: Add tinydrm_xrgb8888_to_gray8() helper

Drm has no monochrome or greyscale support so add a conversion
from the common format XR24.

Also reorder includes into the common order.

Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
[rebase on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/tinydrm: Fix drm_driver.fops.owner

drm_driver.fops can't be shared since the owner then becomes tinydrm.ko.
Move the fops declaration to the driver.

v2: Use DEFINE_DRM_GEM_CMA_FOPS

Reported-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20170326142529.16938-1-noralf@tronnes.org

drm: Create DEFINE_DRM_GEM_CMA_FOPS and roll it out to drivers

Less code ftw.

This converts all drivers except the tinydrm helper module. That one
needs more work, since it gets the THIS_MODULE reference from
tinydrm.ko instead of the actual driver module like it should.
Probably needs a similar trick like I used here with generating the
entire struct with a macro.

Cc: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Reviewed-by: Liviu Dudau <Liviu.Dudau@arm.com>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
[rebased on 4.9, only macro no driver changes]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/tinydrm: helpers: Properly fix backlight dependency

BACKLIGHT_CLASS_DEVICE was selected in the last version of the
tinydrm patchset to fix the backlight dependency, but the
ifdef CONFIG_BACKLIGHT_CLASS_DEVICE was forgotten. Fix that.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/tinydrm: mipi-dbi: Fix field width specifier warning

This warning is seen on 64-bit builds in functions:
   'mipi_dbi_typec1_command':
   'mipi_dbi_typec3_command_read':
   'mipi_dbi_typec3_command':

>> drivers/gpu/drm/tinydrm/mipi-dbi.c:65:20: warning: field width specifier '*' expects argument of type 'int', but argument 5 has type 'size_t {aka long unsigned int}' [-Wformat=]
      DRM_DEBUG_DRIVER("cmd=%02x, par=%*ph\n", cmd, len, data); \
                       ^
   include/drm/drmP.h:228:40: note: in definition of macro 'DRM_DEBUG_DRIVER'
     drm_printk(KERN_DEBUG, DRM_UT_DRIVER, fmt, ##__VA_ARGS__)
                                           ^~~
>> drivers/gpu/drm/tinydrm/mipi-dbi.c:671:2: note: in expansion of macro 'MIPI_DBI_DEBUG_COMMAND'
     MIPI_DBI_DEBUG_COMMAND(cmd, parameters, num);
     ^~~~~~~~~~~~~~~~~~~~~~

Fix by casting 'len' to int in the macro MIPI_DBI_DEBUG_COMMAND().
There is no chance of overflow.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/tinydrm: mipi-dbi: Silence: ‘cmd’ may be used uninitialized

Fix this warning:
drivers/gpu/drm/tinydrm/mipi-dbi.c: In function ‘mipi_dbi_debugfs_command_write’:
drivers/gpu/drm/tinydrm/mipi-dbi.c:905:8: warning: ‘cmd’ may be used uninitialized in this function [-Wmaybe-uninitialized]
  ret = mipi_dbi_command_buf(mipi, cmd, parameters, i);
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

cmd can't be used uninitialized, but to satisfy the compiler,
initialize it to zero.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/tinydrm: Add support for Multi-Inno MI0283QT display

Add driver to support the Multi-Inno MI0283QT display panel.
It has an ILI9341 MIPI DBI compatible display controller.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Acked-by: Thierry Reding <treding@nvidia.com>

drm/tinydrm: Add MIPI DBI support

Add support for MIPI DBI compatible controllers.
Interface type C option 1 and 3 are supported (SPI).

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Acked-by: Thierry Reding <treding@nvidia.com>
[rebased on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm/tinydrm: Add helper functions

Add common functionality needed by many tinydrm drivers.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Thierry Reding <treding@nvidia.com>
[rebased on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm: Add DRM support for tiny LCD displays

tinydrm provides helpers for very simple displays that can use
CMA backed framebuffers and need flushing on changes.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Thierry Reding <treding@nvidia.com>
[rebased on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

drm: debugfs: Remove all files automatically on cleanup

Instead of having the drivers call drm_debugfs_remove_files() in
their drm_driver->debugfs_cleanup hook, do it automatically by
traversing minor->debugfs_list.
Also use debugfs_remove_recursive() so drivers who add their own
debugfs files don't have to keep track of them for removal.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20170126225621.12314-2-noralf@tronnes.org
[rebased on 4.9]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

overlays: i2c1-bcm2708: Don't overwrite i2c1 pins node

It is bad practise to overwrite an entire node in an overlay. Instead,
target the node and overwrite any properties that need changing.

See: https://github.com/raspberrypi/linux/pull/2118

Suggested-by: soodvarun78 <soodvarun78@gmail.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.org>

bcm2835-mmc: Prevent DMA race condition

The end of a read operation is triggered by the completion of the DMA
transfer, but writes are complete when the data IRQ is raised. The
bcm2835-mmc driver contains a race between the handling of the DMA
completion interrupt and the submission of the next request. Fix the
race by deferring the completion of the request until the DMA
transfer finishes.

Signed-off-by: Phil Elwell <phil@raspberrypi.org>

config: enhance DualShock3 controller support

Enable rumble support in Sony HID & HID battery strength.

bcm2835-mmc: Fix DMA usage

The previous change ("bcm2835-mmc: Only claim one DMA channel")
used an incorrect variable, the effect of which was to prevent
DMA from being used at all. Fix that bug by using the right
variable.

Signed-off-by: Phil Elwell <phil@raspberrypi.org>

config: enable generic S/PDIF codec drivers (#2104)

These drivers can be used as dummy ADC/DAC drivers for
attaching general codecs that don't need to be configured.

This option will build 2 additional drivers, spdif_receiver
and spdif_transmitter.

Since these drivers have DT bindings they are handy for quick
testing of I2S peripherals with simple-audio-card.

eg:

fragment@0 {
    target-path = "/";
    __overlay__ {
        dummy_receiver: spdif-receiver {
            #address-cells = <0>;
            #size-cells = <0>;
            #sound-dai-cells = <0>;
            compatible = "linux,spdif-dir";
            status = "okay";
        };
    };
};

Signed-off-by: Matthias Reichl <hias@horus.com>

overlays: Add gpio-shutdown overlay (#2103)

This overlay facilitates the addition of a powerbutton by converting
GPIO edges into KEY_POWER keypresses, which can be handled by
systemd-logind to shut down the system.

Signed-off-by: Matthijs Kooijman <matthijs@stdin.nl>

drm/vc4: Add support for setting DPMS in firmwarekms.

This ensures that the screen goes blank during DPMS (screensaver),
including the cursor.  Planes don't necessarily get disabled during
CRTC disable, so we need to be careful to not leave them on or turn
them back on early.

Signed-off-by: Eric Anholt <eric@anholt.net>