This was added in

commit 0a3e67a4
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date:   Tue Sep 30 12:14:26 2008 -0700

    drm: Rework vblank-wait handling to allow interrupt reduction.

to stay backwards-compatible with old UMS code that didn't even tell
the kernel when it did a modeset, so that the kernel could
save/restore vblank counters. At worst this means vblanks will be
somewhat funky on a setup that very likely no one still runs.

So let's just nuke it.

Plan B would be to set it unconditionally in drm_vblank_init for kms
drivers, instead of in each driver separately. So if this patch breaks
anything please only restore the hunks in drmP.h and drm_irq.c, plus
add a check for DRIVER_MODESET in drm_vblank_init.

Stumbled over this in a discussion on irc with Chris.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Liviu Dudau <liviu.dudau@arm.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Eric Anholt <eric@anholt.net>
Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Cc: Inki Dae <inki.dae@samsung.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Mark Yao <mark.yao@rock-chips.com>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Acked-by: NLiviu Dudau <Liviu.Dudau@arm.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This reverts commit e91abf80.

Since

commit 24e4a8c3
Author: John Stultz <john.stultz@linaro.org>
Date:   Wed Jul 16 21:03:53 2014 +0000

    ktime: Kill non-scalar ktime_t implementation for 2038

there is no longer a 32bit version that's unsigned, and we don't have
to jump through ridiculous hoops to make the calculations correct.

I didn't look whether there's more of this pattern in the kernel.

Cc: John Stultz <john.stultz@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Michel Dänzer <michel.daenzer@amd.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1459249942-21589-1-git-send-email-daniel.vetter@ffwll.ch

drm_vblank_offdelay can have three different types of values:

< 0 is to be always treated the same as dev->vblank_disable_immediate
= 0 is to be treated as "never disable vblanks"
> 0 is to be treated as disable immediate if kms driver wants it
    that way via dev->vblank_disable_immediate. Otherwise it is
    a disable timeout in msecs.

This got broken in Linux 3.18+ for the implementation of
drm_vblank_on. If the user specified a value of zero which should
always reenable vblank irqs in this function, a kms driver could
override the users choice by setting vblank_disable_immediate
to true. This patch fixes the regression and keeps the user in
control.

v2: Only reenable vblank if there are clients left or the user
    requested to "never disable vblanks" via offdelay 0. Enabling
    vblanks even in the "delayed disable" case (offdelay > 0) was
    specifically added by Ville in commit cd19e52a
    ("drm: Kick start vblank interrupts at drm_vblank_on()"),
    but after discussion it turns out that this was done by accident.

    Citing Ville: "I think it just ended up as a mess due to changing
    some of the semantics of offdelay<0 vs. offdelay==0 vs.
    disable_immediate during the review of the series. So yeah, given
    how drm_vblank_put() works now, I'd just make this check for
    offdelay==0."
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Cc: <stable@vger.kernel.org> # 3.18+
Cc: michel@daenzer.net
Cc: vbabka@suse.cz
Cc: ville.syrjala@linux.intel.com
Cc: daniel.vetter@ffwll.ch
Cc: dri-devel@lists.freedesktop.org
Cc: alexander.deucher@amd.com
Cc: christian.koenig@amd.com
Signed-off-by: NDave Airlie <airlied@redhat.com>

Changes to drm_update_vblank_count() in Linux 4.4 broke the
behaviour of the pre/post modeset functions as the new update
code doesn't deal with hw vblank counter resets inbetween calls
to drm_vblank_pre_modeset an drm_vblank_post_modeset, as it
should.

This causes mistreatment of such hw counter resets as counter
wraparound, and thereby large forward jumps of the software
vblank counter which in turn cause vblank event dispatching
and vblank waits to fail/hang --> userspace clients hang.

This symptom was reported on radeon-kms to cause a infinite
hang of KDE Plasma 5 shell's login procedure, preventing users
from logging in.

Fix this by detecting when drm_update_vblank_count() is called
inside a pre->post modeset interval. If so, clamp valid vblank
increments to the safe values 0 and 1, pretty much restoring
the update behavior of the old update code of Linux 4.3 and
earlier. Also reset the last recorded hw vblank count at call
to drm_vblank_post_modeset() to be safe against hw that after
modesetting, dpms on etc. only fires its first vblank irq after
drm_vblank_post_modeset() was already called.
Reported-by: NVlastimil Babka <vbabka@suse.cz>
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: NVlastimil Babka <vbabka@suse.cz>

Cc: <stable@vger.kernel.org> # 4.4+
Cc: michel@daenzer.net
Cc: vbabka@suse.cz
Cc: ville.syrjala@linux.intel.com
Cc: daniel.vetter@ffwll.ch
Cc: dri-devel@lists.freedesktop.org
Cc: alexander.deucher@amd.com
Cc: christian.koenig@amd.com
Signed-off-by: NDave Airlie <airlied@redhat.com>

This fixes a regression introduced by the new drm_update_vblank_count()
implementation in Linux 4.4:

Restrict the bump of the software vblank counter in drm_update_vblank_count()
to a safe maximum value of +1 whenever there is the possibility that
concurrent readers of vblank timestamps could be active at the moment,
as the current implementation of the timestamp caching and updating is
not safe against concurrent readers for calls to store_vblank() with a
bump of anything but +1. A bump != 1 would very likely return corrupted
timestamps to userspace, because the same slot in the cache could
be concurrently written by store_vblank() and read by one of those
readers in a non-atomic fashion and without the read-retry logic
detecting this collision.

Concurrent readers can exist while drm_update_vblank_count() is called
from the drm_vblank_off() or drm_vblank_on() functions or other non-vblank-
irq callers. However, all those calls are happening with the vbl_lock
locked thereby preventing a drm_vblank_get(), so the vblank refcount
can't increase while drm_update_vblank_count() is executing. Therefore
a zero vblank refcount during execution of that function signals that
is safe for arbitrary counter bumps if called from outside vblank irq,
whereas a non-zero count is not safe.

Whenever the function is called from vblank irq, we have to assume concurrent
readers could show up any time during its execution, even if the refcount
is currently zero, as vblank irqs are usually only enabled due to the
presence of readers, and because when it is called from vblank irq it
can't hold the vbl_lock to protect it from sudden bumps in vblank refcount.
Therefore also restrict bumps to +1 when the function is called from vblank
irq.

Such bumps of more than +1 can happen at other times than reenabling
vblank irqs, e.g., when regular vblank interrupts get delayed by more
than 1 frame due to long held locks, long irq off periods, realtime
preemption on RT kernels, or system management interrupts.

A better solution would be to rewrite the timestamp caching to use
full seqlocks to allow concurrent writes and reads for arbitrary
vblank counter increments.

v2: Add code comment that this is essentially a hack and should
    be replaced by a full seqlock implementation for caching of
    timestamps.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Cc: <stable@vger.kernel.org> # 4.4+
Cc: michel@daenzer.net
Cc: vbabka@suse.cz
Cc: ville.syrjala@linux.intel.com
Cc: daniel.vetter@ffwll.ch
Cc: dri-devel@lists.freedesktop.org
Cc: alexander.deucher@amd.com
Cc: christian.koenig@amd.com
Signed-off-by: NDave Airlie <airlied@redhat.com>

Otherwise if a kms driver calls into drm_vblank_off() more than once
before calling drm_vblank_on() again, the redundant calls to
vblank_disable_and_save() will call drm_update_vblank_count()
while hw vblank counters and vblank timestamping are in a undefined
state during modesets, dpms off etc.

At least with the legacy drm helpers it is not unusual to
get multiple calls to drm_vblank_off and drm_vblank_on, e.g.,
half a dozen calls to drm_vblank_off and two calls to drm_vblank_on
were observed on radeon-kms during dpms-off -> dpms-on transition.

We don't no-op calls from atomic modesetting drivers, as they
should do a proper job of tracking hw state.

Fixes large jumps of the software maintained vblank counter due to
the hardware vblank counter resetting to zero during dpms off or
modeset, e.g., if radeon-kms is modified to use drm_vblank_off/on
instead of drm_vblank_pre/post_modeset().

This fixes a regression caused by the changes made to
drm_update_vblank_count() in Linux 4.4.

v2: Don't no-op on atomic modesetting drivers, per suggestion
    of Daniel Vetter.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Cc: <stable@vger.kernel.org> # 4.4+
Cc: michel@daenzer.net
Cc: vbabka@suse.cz
Cc: ville.syrjala@linux.intel.com
Cc: alexander.deucher@amd.com
Cc: christian.koenig@amd.com
Signed-off-by: NDave Airlie <airlied@redhat.com>

Well we can't use that directly since that code must hold
dev->event_lock already. Extract an _unlocked version.

Embarrassingly I've totally forgotten about this patch and any kind of
event-based vblank wait totally blew up, killing the kernel.

v2: Pick the right base struct, someone didn't noticed that gcc was
unhappy. No bug since the addresses at least matched (Daniel Stone)

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Daniel Stone <daniels@collabora.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: NDaniel Stone <daniels@collabora.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1453978864-1513-1-git-send-email-daniel.vetter@ffwll.ch

Use them in the core vblank code and exynos/vmwgfx drivers.

Note that the difference between wake_up_all and _interruptible in
vmwgfx doesn't matter since the only waiter is the core code in
drm_fops.c. And that is interruptible.

v2: Adjust existing kerneldoc too.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Acked-by: NDaniel Stone <daniels@collabora.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Inki Dae <inki.dae@samsung.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1452548477-15905-6-git-send-email-daniel.vetter@ffwll.chReviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
[danvet: Squash in compile fixup, spotted by 0-day.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We moved the module options from drm_drv.c to drm_irq.c in 18882995
('drm: Move vblank related module options into drm_irq.c').  Let's move
the MODULE_PARM_DESC()s as well so they're together.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160108110045.GF32195@mwandaSigned-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Apparently pre-nv50 pageflip events happen before the actual vblank
period. Therefore that functionality got semi-disabled in

commit af4870e4
Author: Mario Kleiner <mario.kleiner.de@gmail.com>
Date:   Tue May 13 00:42:08 2014 +0200

    drm/nouveau/kms/nv04-nv40: fix pageflip events via special case.

Unfortunately that hack got uprooted in

commit cc1ef118
Author: Thierry Reding <treding@nvidia.com>
Date:   Wed Aug 12 17:00:31 2015 +0200

    drm/irq: Make pipe unsigned and name consistent

Triggering a warning when trying to sample the vblank timestamp for a
non-existing pipe. There's a few ways to fix this:

- Open-code the old behaviour, which just enshrines this slight
  breakage of the userspace ABI.

- Revert Mario's commit and again inflict broken timestamps, again not
  pretty.

- Fix this for real by delaying the pageflip TS until the next vblank
  interrupt, thereby making it accurate.

This patch implements the third option. Since having a page flip
interrupt that happens when the pageflip gets armed and not when it
completes in the next vblank seems to be fairly common (older i915 hw
works very similarly) create a new helper to arm vblank events for
such drivers.

v2 (Mario Kleiner):
- Fix function prototypes in drmP.h
- Add missing vblank_put() for pageflip completion without
  pageflip event.
- Initialize sequence number for queued pageflip event to avoid
  trouble in drm_handle_vblank_events().
- Remove dead code and spelling fix.

v3 (Mario Kleiner):
- Add a signed-off-by and cc stable tag per Ilja's advice.

v4 (Thierry Reding):
- Fix kerneldoc typo, discovered by Michel Dänzer
- Rearrange tags and changelog

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=106431
Cc: Thierry Reding <treding@nvidia.com>
Cc: Mario Kleiner <mario.kleiner.de@gmail.com>
Acked-by: NBen Skeggs <bskeggs@redhat.com>
Cc: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>
Reviewed-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Cc: stable@vger.kernel.org # v4.3
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Seems the crtc helpers call drm_calc_timestamping_constants()
unconditionally even if the driver didn't initialize vblank support by
calling drm_vblank_init(). That used to be OK since the constants were
stored under drm_crtc.

However I broke this with
commit eba1f35d ("drm: Move timestamping constants into drm_vblank_crtc")
when I moved the constants to live inside the drm_vblank_crtc struct
instead. If drm_vblank_init() isn't called, we don't allocate these
structures, and so drm_calc_timestamping_constants() will oops.

Fix it by adding a check into drm_calc_timestamping_constants() to see
if vblank support was initialized at all. And to keep in line with other
such checks, also toss in a check and warn for the case where vblank
support was initialized, but the wrong number of crtcs was specified.

Fixes the following sort of oops:
 BUG: unable to handle kernel NULL pointer dereference at 00000000000000b0
 IP: [<ffffffffa014b266>] drm_calc_timestamping_constants+0x86/0x130 [drm]
 PGD 0
 Oops: 0002 [#1] SMP
 Modules linked in: sr_mod cdrom mgag200(+) i2c_algo_bit drm_kms_helper ahci syscopyarea sysfillrect sysimgblt libahci fb_sys_fops bnx2x ttm tg3(+) mdio drm ptp sd_mod libata i2c_core pps_core libcrc32c hpsa dm_mirror dm_region_hash dm_log dm_mod
 CPU: 0 PID: 418 Comm: kworker/0:2 Not tainted 4.3.0+ #1
 Hardware name: HP ProLiant DL380 Gen9, BIOS P89 06/09/2015
 Workqueue: events work_for_cpu_fn
 task: ffff88046ca95500 ti: ffff88007830c000 task.ti: ffff88007830c000
 RIP: 0010:[<ffffffffa014b266>]  [<ffffffffa014b266>] drm_calc_timestamping_constants+0x86/0x130 [drm]
 RSP: 0018:ffff88007830f4e8  EFLAGS: 00010246
 RAX: 0000000000fe4c00 RBX: ffff88006a849160 RCX: 0000000000000540
 RDX: 0000000000000000 RSI: 000000000000fde8 RDI: ffff88006a849000
 RBP: ffff88007830f518 R08: ffff88007830c000 R09: 00000001b87e3712
 R10: 00000000000050c4 R11: 0000000000000000 R12: 0000000000fe4c00
 R13: ffff88006a849000 R14: 0000000000000000 R15: 000000000000fde8
 FS:  0000000000000000(0000) GS:ffff88046f800000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00000000000000b0 CR3: 00000000019d6000 CR4: 00000000001406f0
 Stack:
  ffff88007830f518 ffff88006a849000 ffff880c69b90340 ffff880c69b90000
  ffff880c69b90348 ffff880c69b90340 ffff88007830f748 ffffffffa042f7e7
  ffff88006a849090 0000000000000000 ffff88006a849160 0000000000000000
 Call Trace:
  [<ffffffffa042f7e7>] drm_crtc_helper_set_mode+0x3d7/0x4b0 [drm_kms_helper]
  [<ffffffffa04307d4>] drm_crtc_helper_set_config+0x8d4/0xb10 [drm_kms_helper]
  [<ffffffffa01548d4>] drm_mode_set_config_internal+0x64/0x100 [drm]
  [<ffffffffa043c342>] drm_fb_helper_pan_display+0xa2/0x280 [drm_kms_helper]
  [<ffffffff81392c7b>] fb_pan_display+0xbb/0x170
  [<ffffffff8138cf70>] bit_update_start+0x20/0x50
  [<ffffffff8138b81b>] fbcon_switch+0x39b/0x590
  [<ffffffff8140a3d0>] redraw_screen+0x1a0/0x240
  [<ffffffff8140b30e>] do_bind_con_driver+0x2ee/0x310
  [<ffffffff8140b651>] do_take_over_console+0x141/0x1b0
  [<ffffffff81387377>] do_fbcon_takeover+0x57/0xb0
  [<ffffffff8138c98b>] fbcon_event_notify+0x60b/0x750
  [<ffffffff810a5599>] notifier_call_chain+0x49/0x70
  [<ffffffff810a58dd>] __blocking_notifier_call_chain+0x4d/0x70
  [<ffffffff810a5916>] blocking_notifier_call_chain+0x16/0x20
  [<ffffffff8139282b>] fb_notifier_call_chain+0x1b/0x20
  [<ffffffff81394881>] register_framebuffer+0x1f1/0x330
  [<ffffffffa043d9aa>] drm_fb_helper_initial_config+0x27a/0x3d0 [drm_kms_helper]
  [<ffffffffa0469b4d>] mgag200_fbdev_init+0xdd/0xf0 [mgag200]
  [<ffffffffa0468586>] mgag200_modeset_init+0x176/0x1e0 [mgag200]
  [<ffffffffa0464659>] mgag200_driver_load+0x3f9/0x580 [mgag200]
  [<ffffffffa014e067>] drm_dev_register+0xa7/0xb0 [drm]
  [<ffffffffa015054f>] drm_get_pci_dev+0x8f/0x1e0 [drm]
  [<ffffffffa046937b>] mga_pci_probe+0x9b/0xc0 [mgag200]
  [<ffffffff813662d5>] local_pci_probe+0x45/0xa0
  [<ffffffff8109afe4>] work_for_cpu_fn+0x14/0x20
  [<ffffffff8109e13c>] process_one_work+0x14c/0x3c0
  [<ffffffff8109eaa4>] worker_thread+0x244/0x470
  [<ffffffff8168bfba>] ? __schedule+0x2aa/0x760
  [<ffffffff8109e860>] ? rescuer_thread+0x310/0x310
  [<ffffffff810a4438>] kthread+0xd8/0xf0
  [<ffffffff810a4360>] ? kthread_park+0x60/0x60
  [<ffffffff8169030f>] ret_from_fork+0x3f/0x70
  [<ffffffff810a4360>] ? kthread_park+0x60/0x60
 Code: f6 31 d2 41 89 c2 8b 83 b4 00 00 00 0f af c1 48 98 48 69 c0 40 42 0f 00 48 f7 f6 f6 43 74 10 41 89 c4 75 26 f6 05 9a 6f 03 00 01 <45> 89 96 b0 00 00 00 45 89 a6 ac 00 00 00 75 35 48 83 c4 08 5b
 RIP  [<ffffffffa014b266>] drm_calc_timestamping_constants+0x86/0x130 [drm]
  RSP <ffff88007830f4e8>
 CR2: 00000000000000b0

Cc: Jeff Moyer <jmoyer@redhat.com>
Reported-by: NJeff Moyer <jmoyer@redhat.com>
References: http://lists.freedesktop.org/archives/dri-devel/2015-November/094217.html
Fixes: eba1f35d ("drm: Move timestamping constants into drm_vblank_crtc")
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Switch everything to the new and more capable implementation of abs().
Mainly to give the new abs() a bit of a workout.

Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add a new debug class for _verbose_ debug message from the vblank code.
That is message we spew out potentially for every vblank interrupt.
Thierry already got annoyed at the spew, and now I managed to lock up
my box with these debug prints (seems serial console + a few debug
prints every vblank aren't a good combination).

Or should I maybe call it DRM_DEBUG_IRQ?

Cc: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

String literals get concatenated just fine on their own,
no need to use '\'.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

drm_vblank_count() returns the software counter. We should not pretend
it's the hw counter since we use the hw counter to figuere out what the
software counter value should be. So instead provide a new function
drm_vblank_no_hw_counter() for drivers that don't have a real hw
counter. The new function simply returns 0, which is about the only
thing it can do.

Cc: Vincent Abriou <vincent.abriou@st.com>
Cc: Thierry Reding <treding@nvidia.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NVincent Abriou <vincent.abriou@st.com>
[danvet: s/int pipe/unsigned int pipe/ to follow Thierry's interface
change.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This continues the pattern started in commit cc1ef118 ("drm/irq:
Make pipe unsigned and name consistent"). This is applied to the public
APIs and driver callbacks, so pretty much all drivers need to be updated
to match the new prototypes.

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Inki Dae <inki.dae@samsung.com>
Cc: Jianwei Wang <jianwei.wang.chn@gmail.com>
Cc: Alison Wang <alison.wang@freescale.com>
Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Philipp Zabel <p.zabel@pengutronix.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Mark Yao <mark.yao@rock-chips.com>
Cc: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Cc: Vincent Abriou <vincent.abriou@st.com>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If we couldn't get a high precisions vblank timestamp, we currently
store a zeroed timestamp instead and assume the next vblank irq to
get us something better. This makes sense when trying to update the
timestamp from eg. vblank enable. But if we do this from the vblank
irq we will never get a vblank timestamp unless we high precision
timestamps are available and succeeded. This break weston for instance
on drivers lacking high precision timestamps.

To fix this, zero the timestamp only when not called from vbl irq.
When called from the irq, we still want the timestamp, even if not
perfect.

This fixes a regression from
4dfd6486 drm: Use vblank timestamps to guesstimate how many vblanks were missed

Cc: Mario Kleiner <mario.kleiner.de@gmail.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
Reported-by: NThierry Reding <thierry.reding@gmail.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This function is the KMS native variant of drm_vblank_count_and_time().
It takes a struct drm_crtc * instead of a struct drm_device * and an
index of the CRTC.

Eventually the goal is to access vblank data through the CRTC only so
that the per-CRTC data can be moved to struct drm_crtc.
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Since the original crtc parameter was renamed to pipe, there is no
longer a need to artificially prefix the CRTC parameter.
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When lacking am accurate hardware frame counter, we can fall back to
using the vblank timestamps to guesstimagte how many vblanks have
elapsed since the last time the vblank counter was updated.

Take the oppostunity to unify the vblank_disable_and_save() and
drm_handle_vblank_events() to call the same function
(drm_update_vblank_count()) to perform the vblank updates.

If the hardware/driver has an accurate frame counter use it instead of
the timestamp based guesstimate. If the hardware/driver has neither
a frame counter nor acurate vblank timestamps, we fall back to assuming
that each drm_handle_vblank_events() should increment the vblank count
by one.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Remove the NULL 't_vblank' checks from store_vblank() since that will
never happen.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Avoid confusion and don't use 'vbl_status' as both the
.get_scanout_position() return value and the return value from
drm_calc_vbltimestamp_from_scanoutpos().

While at it make 'vbl_status' unsigned and print it as hex in the
debug prints since it's a bitmask.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Pontential infinite loops in the vblank code are a bad idea. Add some
limits.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We'll soon have use for the 'flags' in drm_update_vblank_count() so pass
it in.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

pixeldur_ns is now unsued, so kill it from drm_vblank_crtc. framedur_ns
is also currently unused but we will have use for it in the near future
so leave it be. linedur_ns is still used by nouveau for some internal
delays.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

linedur_ns, and especially pixeldur_ns are becoming rather inaccurate
to be used for the vblank timestamp correction. With 4k@60 the pixel
duration is already below 2ns, so the amount of error due to the
truncation to nanoseconds is introducing quite a bit of error.

We can avoid such problems if we instead calculate the timestamp
delta_ns directly from the dislay timings, avoiding the use of
these intermediate truncated values.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
[danvet: Squash in fixup from Thierry Reding for amdgpu.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Collect the timestamping constants alongside the rest of the relevant
stuff under drm_vblank_crtc.

We can now get rid of the 'refcrtc' parameter to
drm_calc_vbltimestamp_from_scanoutpos().
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Finish the recent replacement of 'int pipe' with 'unsigned int pipe'

Cc: Thierry Reding <treding@nvidia.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some of the functions are documented inconsistently. Add Returns:
sections where missing and use consistent style to describe the return
value.
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Name all references to the pipe number (CRTC index) consistently to make
it easier to distinguish which is a pipe number and which is a pointer
to struct drm_crtc.

While at it also make all references to the pipe number unsigned because
there is no longer any reason why it should ever be negative.
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When accessing the array of per-CRTC VBLANK structures we must always
check that the index into the array is valid before dereferencing to
avoid crashing.
Signed-off-by: NThierry Reding <treding@nvidia.com>
[danvet: Squash in my own whitespace ocd fixup in drm_vblank_count.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The drm_send_vblank_event() function treats negative CRTC indices as
meaning that a driver doesn't have proper VBLANK handling. This is the
only place where DRM needs negative CRTC indices, so in order to enable
subsequent cleanup, remove this special case and replace it by the more
obvious check for whether or not VBLANK support was initialized.
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

In

commit 99264a61
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Apr 15 19:34:43 2015 +0200

    drm/vblank: Fixup and document timestamp update/read barriers

I've switched vblank->count from atomic_t to unsigned long and
accidentally created an integer comparison bug in
drm_vblank_count_and_time since vblanke->count might overflow the u32
local copy and hence the retry loop never succeed.

Fix this by consistently using u32.

Cc: Michel Dänzer <michel@daenzer.net>
Reported-by: NMichel Dänzer <michel@daenzer.net>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

This patch fix following warnings.

Warning(.//drivers/gpu/drm/drm_irq.c:1279): No description
found for parameter drm_crtc'
Warning(.//drivers/gpu/drm/drm_irq.c:1279): Excess function
parameter 'crtc' description in 'drm_crtc_vblank_reset'
Signed-off-by: NMasanari Iida <standby24x7@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

dev->max_vblank_count contains the largest value that can be represented
by the hardware counter. When the hardware counter wraps around, we have
to add that value + 1 to get the same value as if the hardware counter
didn't wrap around.
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Since commit 844b03f2 we make
sure that after vblank irq off, we return the last valid
(vblank count, vblank timestamp) pair to clients, e.g., during
modesets, which is good.

An overlooked side effect of that commit for kms drivers without
support for precise vblank timestamping is that at vblank irq
enable, when we update the vblank counter from the hw counter, we
can't update the corresponding vblank timestamp, so now we have a
totally mismatched timestamp for the new count to confuse clients.

Restore old client visible behaviour from before Linux 3.17, but
zero out the timestamp at vblank counter update (instead of disable
as in original implementation) if we can't generate a meaningful
timestamp immediately for the new vblank counter. This will fix
this regression, so callers know they need to retry again later
if they need a valid timestamp, but at the same time preserves
the improvements made in the commit mentioned above.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Cc: <stable@vger.kernel.org> #v3.17+

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Since commit 844b03f2 we make
sure that after vblank irq off, we return the last valid
(vblank count, vblank timestamp) pair to clients, e.g., during
modesets, which is good.

An overlooked side effect of that commit for kms drivers without
support for precise vblank timestamping is that at vblank irq
enable, when we update the vblank counter from the hw counter, we
can't update the corresponding vblank timestamp, so now we have a
totally mismatched timestamp for the new count to confuse clients.

Restore old client visible behaviour from before Linux 3.18, but
zero out the timestamp at vblank counter update (instead of disable
as in original implementation) if we can't generate a meaningful
timestamp immediately for the new vblank counter. This will fix
this regression, so callers know they need to retry again later
if they need a valid timestamp, but at the same time preserves
the improvements made in the commit mentioned above.

v2: Rebased on top of Daniel Vetter's fixup and documentation
    patch for timestamp updates. Drop request for stable kernel
    backport as this would be more difficult, unless the original
    patch would get applied to stable kernels.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

For a kms driver to support immediate disable of vblank
irq's reliably without introducing off by one errors or
other mayhem for clients, it must not only support a
hardware vblank counter query, but also high precision
vblank timestamping, so vblank count and timestamp can be
instantaneously reinitialzed to valid values. Additionally
the exposed hardware counter must behave as if it is
incrementing at leading edge of vblank to avoid off by
one errors during reinitialization of the counter while
the display happens to be inside or close to vblank.

Check during drm_vblank_init that a driver which claims to
be capable of vblank_disable_immediate at least supports
high precision timestamping and prevent use of instant
disable if that isn't present as a minimum requirement.

v2: Changed from DRM_ERROR to DRM_INFO and made message
    more clear, as suggested by Michel Dänzer.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This was a bit too much cargo-culted, so lets make it solid:
- vblank->count doesn't need to be an atomic, writes are always done
  under the protection of dev->vblank_time_lock. Switch to an unsigned
  long instead and update comments. Note that atomic_read is just a
  normal read of a volatile variable, so no need to audit all the
  read-side access specifically.

- The barriers for the vblank counter seqlock weren't complete: The
  read-side was missing the first barrier between the counter read and
  the timestamp read, it only had a barrier between the ts and the
  counter read. We need both.

- Barriers weren't properly documented. Since barriers only work if
  you have them on boths sides of the transaction it's prudent to
  reference where the other side is. To avoid duplicating the
  write-side comment 3 times extract a little store_vblank() helper.
  In that helper also assert that we do indeed hold
  dev->vblank_time_lock, since in some cases the lock is acquired a
  few functions up in the callchain.

Spotted while reviewing a patch from Chris Wilson to add a fastpath to
the vblank_wait ioctl.

v2: Add comment to better explain how store_vblank works, suggested by
Chris.

v3: Peter noticed that as-is the 2nd smp_wmb is redundant with the
implicit barrier in the spin_unlock. But that can only be proven by
auditing all callers and my point in extracting this little helper was
to localize all the locking into just one place. Hence I think that
additional optimization is too risky.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mario Kleiner <mario.kleiner.de@gmail.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Michel Dänzer <michel@daenzer.net>
Cc: Peter Hurley <peter@hurleysoftware.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-and-tested-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>