Ville Syrjälä [Tue, 25 Sep 2018 19:37:08 +0000 (22:37 +0300)]
drm/i915: Decouple SKL stride units from intel_fb_stride_alignment()
In the future framebuffer stride alignment requirements won't exactly
match the units in which skl+ plane stride is specified. So extract
the code for the skl+ stuff into a separate helper.
Ville Syrjälä [Tue, 23 Oct 2018 16:02:01 +0000 (19:02 +0300)]
drm/i915: Make sure fb gtt offsets stay within 32bits
Let's try to make sure the fb offset computations never hit
an integer overflow by making sure the entire fb stays
below 32bits. framebuffer_check() in the core already does
the same check, but as it doesn't know about tiling some things
can slip through. Repeat the check in the driver with tiling
taken into account.
v2: Use add_overflows() after massaging it to work for me (Chris)
v3: Call it add_overflow_t() to match min_t() & co. (Chris)
Ville Syrjälä [Wed, 14 Nov 2018 21:07:29 +0000 (23:07 +0200)]
drm/i915: Pass the plane to icl_program_input_csc_coeff()
On icl+ the plane state that gets passed to update_slave() is not
the plane state of the plane we're programming. With NV12 the
plane state would be coming from the master (UV) plane whereas
the plane we're programming is the slave (Y) plane. For that reason
we need to explicitly pass around the slave plane (or we'd have to
otherwise deduce it by checking whether we were called via
.update_plane() or .update_slave()).
In the case of icl_program_input_csc_coeff() it's actually OK to
assume that we are always the master plane because the input CSC
only exists on HDR planes which can never be a slave plane. But
for consistency let's pass in the plane explicitly anyway.
While at it drop the "_coeff" from the function name since it's
kinda redundant, and this makes the name a bit shorter :)
Cc: Uma Shankar <uma.shankar@intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181114210729.16185-14-ville.syrjala@linux.intel.com Reviewed-by: Uma Shankar <uma.shankar@intel.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Ville Syrjälä [Wed, 14 Nov 2018 21:07:28 +0000 (23:07 +0200)]
drm/i915: Rename the confusing 'plane_id' to 'color_plane'
A variable whose name is 'plane_id' is expected to be of the
enum plane_id type. In this case we have a raw int, which turns
out to refer to the plane of the framebuffer. Rename the variable
to 'color_plane' in line with the trend started earlier.
Ville Syrjälä [Wed, 14 Nov 2018 21:07:27 +0000 (23:07 +0200)]
drm/i915: Commit skl+ planes in an order that avoids ddb overlaps
skl+ can go belly up if there are overlapping ddb allocations between
planes. If we could absolutely guarantee that we can perform the atomic
update within a single frame we shouldn't have to worry about this. But
we can't rely on that so let's steal the ddb overlap check trick from
skl_update_crtcs() and apply it to the plane updates. Since each step
of the sequence is free from ddb overlaps we don't have to worry about
a vblank sneaking up on us in the middle of the sequence. The partial
state that gets latched by the hardware will be safe. And unlike
skl_update_crtcs() we don't have to intoduce any extra vblank waits
on account of only having to worry about a single pipe.
Ville Syrjälä [Tue, 27 Nov 2018 16:59:00 +0000 (18:59 +0200)]
drm/i915: Move ddb/wm programming into plane update/disable hooks on skl+
On SKL+ the plane WM/BUF_CFG registers are a proper part of each
plane's register set. That means accessing them will cancel any
pending plane update, and we would need a PLANE_SURF register write
to arm the wm/ddb change as well.
To avoid all the problems with that let's just move the wm/ddb
programming into the plane update/disable hooks. Now all plane
registers get written in one (hopefully atomic) operation.
To make that feasible we'll move the plane ddb tracking into
the crtc state. Watermarks were already tracked there.
v2: Rebase due to input CSC
v3: Split out a bunch of junk (Matt)
v4: Add skl_wm_add_affected_planes() to deal with
cursor special case and non-zero wm register reset value
v5: Drop the unrelated for_each_intel_plane_mask() fix (Matt)
Remove the redundant ddb memset() (Matt)
Ville Syrjälä [Wed, 14 Nov 2018 21:07:25 +0000 (23:07 +0200)]
drm/i915: Don't pass dev_priv around so much
Simplify the calling convention of the skl+ watermark functions
by not passing around dev_priv needlessly. The callees have
what they need to dig it out anyway.
Ville Syrjälä [Tue, 27 Nov 2018 16:57:26 +0000 (18:57 +0200)]
drm/i915: Clean up skl+ vs. icl+ watermark computation
Make a cleaner split between the skl+ and icl+ ways of computing
watermarks. This way skl_build_pipe_wm() doesn't have to know any
of the gritty details of icl+ master/slave planes.
We can also simplify a bunch of the lower level code by pulling
the plane visibility checks a bit higher up.
v2: WARN_ON(!visible) for the icl+ master plane case (Matt)
Ville Syrjälä [Wed, 14 Nov 2018 21:07:23 +0000 (23:07 +0200)]
drm/i915: Pass the entire skl_plane_wm to skl_compute_transition_wm()
We have to pass both level 0 watermark struct and the transition
watermark struct to skl_compute_transition_wm(). Make life less
confusing by just passing the entire plane watermark struct that
contains both aforementioned structures.
Ville Syrjälä [Wed, 14 Nov 2018 21:07:20 +0000 (23:07 +0200)]
drm/i915: Pass the new crtc_state to ->disable_plane()
We're going to need access to the new crtc state in ->disable_plane()
for SKL+ wm/ddb programming and pre-skl pipe gamma/csc control. Pass
the crtc state down.
We'll also try to make intel_crtc_disable_planes() do the right
thing as much as it's possible. The fact that we don't have a
separate crtc state for the disabled state when we're going to
re-enable the crtc later means we might end up poking at a few
extra planes in there. But that's harmless. I suppose one might
argue that we wouldn't have to care about proper ddb/wm/csc/gamma
if the pipe is going to permanently disable anyway, but the state
checker probably cares so we should try our best to make sure
everything is programmed correctly even in that case.
Keep track which planes need updating during the commit. For now
we set the bit for any plane that was or will be visible (including
icl+ nv12 slave planes). In the future I'll have need to update
invisible planes as well, for skl plane ddbs and for pre-skl pipe
gamma/csc control (which lives in the primary plane control register).
v2: Pimp the commit message to mention icl+ nv12 slave planes (Matt)
Ville Syrjälä [Wed, 14 Nov 2018 21:07:18 +0000 (23:07 +0200)]
drm/i915: Move single buffered plane register writes to the end
The plane color correction registers are single buffered. So
ideally we would write them at the start of vblank just after the
double buffered plane registers have been latched. Since we have
no convenient way to do that for now let's at least move the
single buffered register writes to happen after the double
buffered registers have been written.
Ville Syrjälä [Wed, 14 Nov 2018 21:07:17 +0000 (23:07 +0200)]
drm/i915: Reorganize plane register writes to make them more atomic
Some observations about the plane registers:
- the control register will self-arm if the plane is not already
enabled, thus we want to write it as close to (or ideally after)
the surface register
- tileoff/linoff/offset/aux_offset are self-arming as well so we want
them close to the surface register as well
- color keying registers we maybe self arming before SKL. Not 100%
sure but we can try to keep them near to the surface register
as well
- chv pipe b csc register are double buffered but self arming so
moving them down a bit
- the rest should be mostly armed by the surface register so we can
safely write them first, and to just for some consistency let's try
to follow keep them in order based on the register offset
None of this will have any effect of course unless the vblank evasion
fails (which it still does sometimes). Another potential future benefit
might be pulling the non-self armings registers outside the vblank
evasion since they won't latch until the arming register has been
written. This would make the critical section a bit lighter and thus
less likely to exceed the deadline.
v2: Rebase due to input CSC
v3: Swap LINOFF/TILEOFF and KEYMSK/KEYMAX to actually follow
the last rule above (Matt)
Add a bit more rationale to the commit message (Matt)
Manasi Navare [Tue, 27 Nov 2018 21:41:08 +0000 (13:41 -0800)]
drm/dsc: Define the DSC 1.1 and 1.2 Line Buffer depth constants
DSC specification defines linebuf_depth which contains the
line buffer bit depth used to generate the bitstream.
These values are defined as per Table 4.1 in DSC 1.2 spec
v2 (From Manasi):
* Rename as MAX_LINEBUF_DEPTH for DSC 1.1 and DSC 1.2
Cc: dri-devel@lists.freedesktop.org Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Gaurav K Singh <gaurav.k.singh@intel.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Acked-by: Sean Paul <seanpaul@chromium.org> (For merging through
drm-intel) Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181127214125.17658-6-manasi.d.navare@intel.com
Manasi Navare [Tue, 27 Nov 2018 21:41:07 +0000 (13:41 -0800)]
drm/dsc: Add helpers for DSC picture parameter set infoframes
According to Display Stream compression spec 1.2, the picture
parameter set metadata is sent from source to sink device
using the DP Secondary data packet. An infoframe is formed
for the PPS SDP header and PPS SDP payload bytes.
This patch adds helpers to fill the PPS SDP header
and PPS SDP payload according to the DSC 1.2 specification.
v7:
* Use BUILD_BUG_ON() to protect changing struct size (Ville)
* Remove typecaseting (Ville)
* Include byteorder.h in drm_dsc.c (Ville)
* Correct kernel doc spacing (Anusha)
v6:
* Use proper sequence points for breaking down the
assignments (Chris Wilson)
* Use SPDX identifier
v5:
Do not use bitfields for DRM structs (Jani N)
v4:
* Use DSC constants for params that dont change across
configurations
v3:
* Add reference to added kernel-docs in
Documentation/gpu/drm-kms-helpers.rst (Daniel Vetter)
v2:
* Add EXPORT_SYMBOL for the drm functions (Manasi)
Cc: dri-devel@lists.freedesktop.org Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Anusha Srivatsa <anusha.srivatsa@intel.com> Cc: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com> Acked-by: Sean Paul <seanpaul@chromium.org> (For merging through
drm-intel) Link: https://patchwork.freedesktop.org/patch/msgid/20181127214125.17658-5-manasi.d.navare@intel.com
Manasi Navare [Tue, 27 Nov 2018 21:41:05 +0000 (13:41 -0800)]
drm/dsc: Define VESA Display Stream Compression Capabilities
This defines all the DSC parameters as per the VESA DSC spec
that will be required for DSC encoder/decoder
v6: (From Manasi)
* Add a bit mask for RANGE_BPG_OFFSET for 6 bits(Manasi)
v5 (From Manasi)
* Add the RC constants as per the spec
v4 (From Manasi)
* Add the DSC_MUX_WORD_SIZE constants (Manasi)
v2: Define this struct in DRM (From Manasi)
* Changed the data types to u8/u16 instead of unsigned longs (Manasi)
* Remove driver specific fields (Manasi)
* Move this struct definition to DRM (Manasi)
* Define DSC 1.2 parameters (Manasi)
* Use DSC_NUM_BUF_RANGES (Manasi)
* Call it drm_dsc_config (Manasi)
Cc: dri-devel@lists.freedesktop.org Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Anusha Srivatsa <anusha.srivatsa@intel.com> Cc: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Gaurav K Singh <gaurav.k.singh@intel.com> Co-developed-by: Gaurav K Singh <gaurav.k.singh@intel.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Acked-by: Sean Paul <seanpaul@chromium.org> (For merging through
drm-intel) Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181127214125.17658-3-manasi.d.navare@intel.com
This patch defines a new header file for all the DSC 1.2 structures
and creates a structure for PPS infoframe which will be used to send
picture parameter set secondary data packet for display stream compression.
All the PPS infoframe syntax elements are taken from DSC 1.2 specification
from VESA.
v4:
* Remove redundant blankline in doc (Ville)
* use drm_dsc namespace for all structs (Ville)
* Use packed struct (Ville)
v3:
* Add the SPDX shorthand (Chris Wilson)
v2:
* Do not use bitfields in the struct (Jani Nikula)
Cc: Gaurav K Singh <gaurav.k.singh@intel.com> Cc: dri-devel@lists.freedesktop.org Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Anusha Srivatsa <anusha.srivatsa@intel.com> Cc: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Acked-by: Sean Paul <seanpaul@chromium.org> (For merging through
drm-intel) Reviewed-by: Harry Wentland <harry.wentland@amd.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181127214125.17658-2-manasi.d.navare@intel.com
Manasi Navare [Tue, 27 Nov 2018 21:41:03 +0000 (13:41 -0800)]
drm/dsc: Modify DRM helper to return complete DSC color depth capabilities
DSC DPCD color depth register advertises its color depth capabilities
by setting each of the bits that corresponding to a specific color
depth. This patch defines those specific color depths and adds
a helper to return an array of color depth capabilities.
v2:
* Simplify the logic (Ville)
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Acked-by: Sean Paul <seanpaul@chromium.org> (For merging through
drm-intel) Reviewed-by: Ville Syrjala <ville.syrjala@linux.intel.com> Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181127214125.17658-1-manasi.d.navare@intel.com
Chris Wilson [Mon, 26 Nov 2018 09:56:10 +0000 (09:56 +0000)]
drm/i915: Skip engine serialisation for no-op seqno reset
If the engine's seqno is already at our target seqno (most likely it
hasn't been used since the last reset), we can skip serialising the
engine and leave it as is.
Chris Wilson [Mon, 26 Nov 2018 12:28:21 +0000 (12:28 +0000)]
drm/i915/ringbuffer: 2-step restart
We may be simply restarting too fast for the culmudgeonly gen3/gen4 as
we still see missing interrupts following a reset. So let's try
restarting a little slower, first wake up the ring empty and then tell
it about the work it has to perform.
Chris Wilson [Fri, 23 Nov 2018 13:23:25 +0000 (13:23 +0000)]
drm/i915: Cache the error string
Currently, we convert the error state into a string every time we read
from sysfs (and sysfs reads in page size (4KiB) chunks). We do try to
window the string and only capture the portion that is being read, but
that means that we must always convert up to the window to find the
start. For a very large error state bordering on EXEC_OBJECT_CAPTURE
abuse, this is noticeable as it degrades to O(N^2)!
As we do not have a convenient hook for sysfs open(), and we would like
to keep the lazy conversion into a string, do the conversion of the
whole string on the first read and keep the string until the error state
is freed.
v2: Don't double advance simple_read_from_buffer
v3: Due to extreme pain of lack of vrealloc, use a scatterlist
v4: Keep the forward iterator loosely cached
v5: Stylistic improvements to reduce patch size
Reported-by: Jason Ekstrand <jason@jlekstrand.net>
Testcase: igt/gem_exec_capture/many* Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181123132325.26541-1-chris@chris-wilson.co.uk
drm/i915: Keep PSR disabled after a driver reload after a PSR error
If a PSR error happened and the driver is reloaded, the EDP_PSR_IIR
will still keep the error set even after the reset done in the
irq_preinstall and irq_uninstall hooks.
And enabling in this situation cause the screen to freeze in the
first time that PSR HW tries to activate so lets keep PSR disabled
to avoid any rendering problems.
v5: rebased: using edp_psr_shift()
v4: Moved handling from intel_psr_compute_config() to
intel_psr_init() to avoid hardware access during compute(Ville)
While PSR is active hardware will do aux transactions by it self to
wakeup sink to receive a new frame when necessary. If that
transaction is not acked by sink, hardware will trigger this
interruption.
So let's disable PSR as it is a hint that there is problem with this
sink.
The removed FIXME was asking to manually train the link but we don't
need to do that as by spec sink should do a short pulse when it is
out of sync with source, we just need to make sure it is awaken and
the SDP header with PSR inactive set it will trigger the short pulse
with a error set in the link status.
v3: added workarround to fix scheduled work starvation cause by
to frequent PSR error interruption
v4: only setting irq_aux_error as we don't care in clear it and
not using dev_priv->irq_lock as consequence.
drm/i915: Do not enable PSR in the next modeset after a error
When we detect a error and disable PSR, it is kept disabled until the
next modeset but as the sink already show signs that it do not
properly work with PSR lets disabled it for good to avoid any
additional flickering.
drm/i915: Check PSR errors instead of retrain while PSR is enabled
When a PSR error happens sink sets the PSR error register and also
set the link status to a error status.
So in the short pulse handling it was returning earlier and doing a
full detection and attempting to retrain but it fails as PSR HW is
in change of the main-link.
Just call intel_psr_short_pulse() before
intel_dp_needs_link_retrain() is not the right fix as
intel_dp_needs_link_retrain() would return true and trigger a full
detection while PSR HW is still in change of main-link.
Check for PSR active is also not safe as it could be inactive due a
frontbuffer invalidate and still doing the PSR exit sequence.
v3: added comment in intel_dp_needs_link_retrain()
drm/i915: Avoid a full port detection in the first eDP short pulse
Some eDP panels do not set a valid sink count value and even for the
ones that sets is should always be one for eDP, that is why it is not
cached in intel_edp_init_dpcd().
But intel_dp_short_pulse() compares the old count with the read one
if there is a mistmatch a full port detection will be executed, what
was happening in the first short pulse interruption of eDP panels
that sets sink count.
Instead of just skip the compasison for eDP panels, lets not read
the sink count at all for eDP.
v2: the previous version of this patch it was caching the sink count
in intel_edp_init_dpcd() but I was pointed out by Ville a patch that
handled a case of a eDP panel that do not set sink count and as sink
count is not used to eDP certification was choosed to just not read
it at all.
Lyude Paul [Wed, 21 Nov 2018 00:37:17 +0000 (19:37 -0500)]
drm/i915: Synchronize hpd work in i915_hpd_storm_ctl_show()
While trying to add a chamelium test for short HPD IRQs, I ran into
issues where a hotplug storm would be triggered, but the point at which
it would be reported by the kernel would be after igt actually finished
checking i915_hpd_storm_ctl's status. So, fix this by simply
synchronizing our IRQ work, dig_port_work, and hotplug_work before
printing out the HPD storm status in i915_hpd_storm_ctl_show().
Ville Syrjälä [Mon, 29 Oct 2018 18:34:53 +0000 (20:34 +0200)]
drm/i915: Eliminate the horrendous format check code
Replace the messy framebuffer format/modifier validation code
with a single call to drm_any_plane_has_format(). The code was
extremely annoying to maintain as you had to have a lot of platform
checks for different formats. The new code requires zero maintenance.
v2: Nuke the modifier checks as well since the core does that too now
v3: Call drm_any_plane_has_format() from the driver code
v4: Rebase
Chris Wilson [Wed, 21 Nov 2018 15:16:53 +0000 (15:16 +0000)]
drm/i915: Show waiter's status on engine dump
When showing the list of waiters, include the task's status so that we
can tell if they have been woken up and are waiting for the CPU, or if
they are still waiting to be woken.
Ville Syrjälä [Tue, 20 Nov 2018 13:54:50 +0000 (15:54 +0200)]
drm/i915: Add rotation readout for plane initial config
If we need to force a full plane update before userspace/fbdev
have given us a proper plane state we should try to maintain the
current plane state as much as possible (apart from the parts
of the state we're trying to fix up with the plane update).
To that end add basic readout for the plane rotation and
maintain it during the initial fb takeover.
Cc: Hans de Goede <hdegoede@redhat.com> Fixes: f9df69ff7d22 ("drm/i915: Fix assert_plane() warning on bootup with external display") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181120135450.3634-2-ville.syrjala@linux.intel.com Tested-by: Hans de Goede <hdegoede@redhat.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Ville Syrjälä [Tue, 20 Nov 2018 13:54:49 +0000 (15:54 +0200)]
drm/i915: Force a LUT update in intel_initial_commit()
If we force a plane update to fix up our half populated plane state
we'll also force on the pipe gamma for the plane (since we always
enable pipe gamma currently). If the BIOS hasn't programmed a sensible
LUT into the hardware this will cause the image to become corrupted.
Typical symptoms are a purple/yellow/etc. flash when the driver loads.
To avoid this let's program something sensible into the LUT when
we do the plane update. In the future I plan to add proper plane
gamma enable readout so this is just a temporary measure.
Cc: Hans de Goede <hdegoede@redhat.com> Fixes: f9df69ff7d22 ("drm/i915: Fix assert_plane() warning on bootup with external display") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181120135450.3634-1-ville.syrjala@linux.intel.com Tested-by: Hans de Goede <hdegoede@redhat.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Imre Deak [Mon, 19 Nov 2018 18:00:21 +0000 (20:00 +0200)]
drm/i915: Make CHICKEN_TRANS reg not depend on enum value
Depending on the transcoder enum values to translate from transcoder to
the corresponding CHICKEN_TRANS register can easily break if we add a
new transcoder. Add an explicit mapping instead, by using helpers to
look up the register instance either by transcoder or port (since
unconveniently the registers have both port and transcoder specific
bits).
While at it also check for the correctness of GEN, port, transcoder. I
wasn't sure if psr2_enabled can only be set for GEN9+, but that seems to
be the case indeed (see setting of sink_psr2_support in
intel_psr_init_dpcd()).
v2 (Ville):
- Make gen9_chicken_trans_reg() internal to intel_psr.c.
- s/trans/cpu_transcoder/
Imre Deak [Tue, 20 Nov 2018 09:23:25 +0000 (11:23 +0200)]
drm/i915: Add code comment on assumption of pipe==transcoder
Add a comment to the pipe and transcoder enum definitions about our
assumption in the code about enum values for pipes and transcoders
with a 1:1 transcoder -> pipe mapping.
v2:
- Clarify more what are the assumptions about the enum values. (Ville)
v3: (Lucas)
- s/->/ -> / so it looks less like pointer dereferencing.
- Use pipe enums as initializers in the transcoder enum definition.
Imre Deak [Tue, 20 Nov 2018 09:23:24 +0000 (11:23 +0200)]
drm/i915: Make EDP PSR flags not depend on enum values
Depending on the transcoder enum values to translate from transcoder
to EDP PSR flags can easily break if we add a new transcoder. So remove
the dependency by using an explicit mapping.
While at it also add a WARN for unexpected trancoders.
v2:
- Simplify things by defining flag shift values instead of indices.
- s/trans/cpu_transcoder/ (Ville)
v3:
- Define flags to look like separate bits instead of the values of
the same bitfield. (Ville)
Imre Deak [Tue, 20 Nov 2018 09:23:23 +0000 (11:23 +0200)]
drm/i915: Make pipe/transcoder offsets not depend on enum values
Depending on the transcoder enum values to translate from transcoder
to pipe/transcoder register addresses can easily break if we add a new
transcoder. So remove the dependency by using named initializers.
Lucas De Marchi [Sat, 17 Nov 2018 00:42:34 +0000 (16:42 -0800)]
drm/i915: Downgrade unknown CSR firmware warnings
Like it was done in commit b017622154c7 ("drm/i915: Downgrade unknown
firmware warnings") for huc and guc: downgrade CSR firmware warnings. If
we have released no firmware yet for a platform, stop scaring the
consumer and merely note its expected absence.
By simply removing the warning and early return we hit the condition
with the appropriate message.
Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181117004234.23437-2-lucas.demarchi@intel.com
Lucas De Marchi [Sat, 17 Nov 2018 00:42:33 +0000 (16:42 -0800)]
drm/i915: allow to load DMC firmware on next gen
Before commit 64c0fcc4a687 ("drm/i915/csr: keep max firmware size together
with firmare name and version") it was possible to load the firmware for
testing purposes via parameter. Let's use the size of the last known
platform to recover that behavior.
Jani Nikula [Fri, 16 Nov 2018 12:07:29 +0000 (14:07 +0200)]
drm/i915/fixed: cosmetic cleanup
Clean up fixed point temp variable initialization, use the more
conventional tmp name for temp variables, add empty lines before
return. No functional changes.
Chris Wilson [Mon, 19 Nov 2018 15:41:53 +0000 (15:41 +0000)]
drm/i915: Write GPU relocs harder with gen3
Under moderate amounts of GPU stress, we can observe on Bearlake and
Pineview (later gen3 models) that we execute the following batch buffer
before the write into the batch is coherent. Adding extra (tested with
upto 32x) MI_FLUSH to either the invalidation, flush or both phases does
not solve the incoherency issue with the relocations, but emitting the
MI_STORE_DWORD_IMM twice does. So be it.
Chris Wilson [Fri, 2 Nov 2018 16:12:12 +0000 (16:12 +0000)]
drm/i915: Prevent machine hang from Broxton's vtd w/a and error capture
Since capturing the error state requires fiddling around with the GGTT
to read arbitrary buffers and is itself run under stop_machine(), it
deadlocks the machine (effectively a hard hang) when run in conjunction
with Broxton's VTd workaround to serialize GGTT access.
v2: Store the ERR_PTR in first_error so that the error can be reported
to the user via sysfs.
v3: Mention the quirk in dmesg (using info as per usual)
Fixes: 0359347efd36 ("drm/i915: Serialize GTT/Aperture accesses on BXT") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: John Harrison <john.C.Harrison@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181102161232.17742-5-chris@chris-wilson.co.uk
Dave Airlie [Mon, 19 Nov 2018 01:07:52 +0000 (11:07 +1000)]
Merge branch 'drm-next-4.21' of git://people.freedesktop.org/~agd5f/linux into drm-next
New features for 4.21:
amdgpu:
- Support for SDMA paging queue on vega
- Put compute EOP buffers into vram for better performance
- Share more code with amdkfd
- Support for scanout with DCC on gfx9
- Initial kerneldoc for DC
- Updated SMU firmware support for gfx8 chips
- Rework CSA handling for eventual support for preemption
- XGMI PSP support
- Clean up RLC handling
- Enable GPU reset by default on VI, SOC15 dGPUs
- Ring and IB test cleanups
amdkfd:
- Share more code with amdgpu
ttm:
- Move global init out of the drivers
scheduler:
- Track if schedulers are ready for work
- Timeout/fault handling changes to facilitate GPU recovery
Dave Airlie [Mon, 19 Nov 2018 00:40:00 +0000 (10:40 +1000)]
Merge tag 'drm-misc-next-2018-11-07' of git://anongit.freedesktop.org/drm/drm-misc into drm-next
drm-misc-next for v4.21, part 1:
UAPI Changes:
- Add syncobj timeline support to drm.
Cross-subsystem Changes:
- Remove shared fence staging in dma-buf's fence object, and allow
reserving more than 1 fence and add more paranoia when debugging.
- Constify infoframe functions in video/hdmi.
Core Changes:
- Add vkms todo, and a lot of assorted doc fixes.
- Drop transitional helpers and convert drivers to use drm_atomic_helper_shutdown().
- Move atomic state helper functions to drm_atomic_state_helper.[ch]
- Refactor drm selftests, and add new tests.
- DP MST atomic state cleanups.
- Drop EXPORT_SYMBOL from drm leases.
- Lease cleanups and fixes.
- Create render node for vgem.
Driver Changes:
- Fix build failure in imx without fbdev emulation.
- Add rotation quirk for GPD win2 panel.
- Add support for various CDTech panels, Banana Pi Panel, DLC1010GIG,
Olimex LCD-O-LinuXino, Samsung S6D16D0, Truly NT35597 WQXGA,
Himax HX8357D, simulated RTSM AEMv8.
- Add dw_hdmi support to rockchip driver.
- Fix YUV support in vc4.
- Fix resource id handling in virtio.
- Make rockchip use dw-mipi-dsi bridge driver, and add dual dsi support.
- Advertise that tinydrm only supports DRM_FORMAT_MOD_LINEAR.
- Convert many drivers to use atomic helpers, and drm_fbdev_generic_setup().
- Add Mali linear tiled formats, and enable them in the Mali-DP driver.
- Add support for H6 DE3 mixer 0, DW HDMI, HDMI PHY and TCON TOP.
- Assorted driver cleanups and fixes.
Linus Torvalds [Sun, 18 Nov 2018 20:21:09 +0000 (12:21 -0800)]
Merge tag 'libnvdimm-fixes-4.20-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm
Pull libnvdimm fixes from Dan Williams:
"A small batch of fixes for v4.20-rc3.
The overflow continuation fix addresses something that has been broken
for several releases. Arguably it could wait even longer, but it's a
one line fix and this finishes the last of the known address range
scrub bug reports. The revert addresses a lockdep regression. The unit
tests are not critical to fix, but no reason to hold this fix back.
Summary:
- Address Range Scrub overflow continuation handling has been broken
since it was initially merged. It was only recently that error
injection and platform-BIOS support enabled this corner case to be
exercised.
- The recent attempt to provide more isolation for the kernel Address
Range Scrub state machine from userapace initiated sessions
triggers a lockdep report. Revert and try again at the next merge
window.
- Fix a kasan reported buffer overflow in libnvdimm unit test
infrastrucutre (nfit_test)"
* tag 'libnvdimm-fixes-4.20-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
Revert "acpi, nfit: Further restrict userspace ARS start requests"
acpi, nfit: Fix ARS overflow continuation
tools/testing/nvdimm: Fix the array size for dimm devices.
Linus Torvalds [Sun, 18 Nov 2018 19:31:26 +0000 (11:31 -0800)]
Merge branch 'akpm' (patches from Andrew)
Merge misc fixes from Andrew Morton:
"16 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
mm/memblock.c: fix a typo in __next_mem_pfn_range() comments
mm, page_alloc: check for max order in hot path
scripts/spdxcheck.py: make python3 compliant
tmpfs: make lseek(SEEK_DATA/SEK_HOLE) return ENXIO with a negative offset
lib/ubsan.c: don't mark __ubsan_handle_builtin_unreachable as noreturn
mm/vmstat.c: fix NUMA statistics updates
mm/gup.c: fix follow_page_mask() kerneldoc comment
ocfs2: free up write context when direct IO failed
scripts/faddr2line: fix location of start_kernel in comment
mm: don't reclaim inodes with many attached pages
mm, memory_hotplug: check zone_movable in has_unmovable_pages
mm/swapfile.c: use kvzalloc for swap_info_struct allocation
MAINTAINERS: update OMAP MMC entry
hugetlbfs: fix kernel BUG at fs/hugetlbfs/inode.c:444!
kernel/sched/psi.c: simplify cgroup_move_task()
z3fold: fix possible reclaim races
Linus Torvalds [Sun, 18 Nov 2018 18:58:20 +0000 (10:58 -0800)]
Merge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fix from Ingo Molnar:
"Fix an exec() related scalability/performance regression, which was
caused by incorrectly calculating load and migrating tasks on exec()
when they shouldn't be"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/fair: Fix cpu_util_wake() for 'execl' type workloads
Linus Torvalds [Sun, 18 Nov 2018 18:54:59 +0000 (10:54 -0800)]
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Ingo Molnar:
"Fix uncore PMU enumeration for CofeeLake CPUs"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel/uncore: Support CoffeeLake 8th CBOX
perf/x86/intel/uncore: Add more IMC PCI IDs for KabyLake and CoffeeLake CPUs
Linus Torvalds [Sun, 18 Nov 2018 18:52:26 +0000 (10:52 -0800)]
Merge branch 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull EFI fixes from Ingo Molnar:
"Misc fixes: two warning splat fixes, a leak fix and persistent memory
allocation fixes for ARM"
* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
efi: Permit calling efi_mem_reserve_persistent() from atomic context
efi/arm: Defer persistent reservations until after paging_init()
efi/arm/libstub: Pack FDT after populating it
efi/arm: Revert deferred unmap of early memmap mapping
efi: Fix debugobjects warning on 'efi_rts_work'
Linus Torvalds [Sun, 18 Nov 2018 18:45:09 +0000 (10:45 -0800)]
Merge branch 'spectre' of git://git.armlinux.org.uk/~rmk/linux-arm
Pull ARM spectre updates from Russell King:
"These are the currently known final bits that resolve the Spectre
issues. big.Little systems used to be sufficiently identical in that
there were no differences between individual CPUs in the system that
mattered to the kernel. With the advent of the Spectre problem, the
CPUs now have differences in how the workaround is applied.
As a result of previous Spectre patches, these systems ended up
reporting quite a lot of:
"CPUx: Spectre v2: incorrect context switching function, system vulnerable"
messages due to the action of the big.Little switcher causing the CPUs
to be re-initialised regularly. This series resolves that issue by
making the CPU vtable unique to each CPU.
However, since this is used very early, before per-cpu is setup,
per-cpu can't be used. We also have a problem that two of the methods
are not called from preempt-safe paths, but thankfully these remain
identical between all CPUs in the system. To make sure, we validate
that these are identical during boot"
* 'spectre' of git://git.armlinux.org.uk/~rmk/linux-arm:
ARM: spectre-v2: per-CPU vtables to work around big.Little systems
ARM: add PROC_VTABLE and PROC_TABLE macros
ARM: clean up per-processor check_bugs method call
ARM: split out processor lookup
ARM: make lookup_processor_type() non-__init
the problem is that we only check for an out of bound order in the slow
path and the node reclaim might happen from the fast path already. This
is fixable by making sure that kvmalloc doesn't ever use kmalloc for
requests that are larger than KMALLOC_MAX_SIZE but this also shows that
the code is rather fragile. A recent UBSAN report just underlines that
by the following report
Note that this is not a kvmalloc path. It is just that the fast path
really depends on having sanitzed order as well. Therefore move the
order check to the fast path.
Link: http://lkml.kernel.org/r/20181113094305.GM15120@dhcp22.suse.cz Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Reported-by: Kyungtae Kim <kt0755@gmail.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Pavel Tatashin <pavel.tatashin@microsoft.com> Cc: Oscar Salvador <osalvador@suse.de> Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Aaron Lu <aaron.lu@intel.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Byoungyoung Lee <lifeasageek@gmail.com> Cc: "Dae R. Jeong" <threeearcat@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Uwe Kleine-König [Fri, 16 Nov 2018 23:08:43 +0000 (15:08 -0800)]
scripts/spdxcheck.py: make python3 compliant
Without this change the following happens when using Python3 (3.6.6):
$ echo "GPL-2.0" | python3 scripts/spdxcheck.py -
FAIL: 'str' object has no attribute 'decode'
Traceback (most recent call last):
File "scripts/spdxcheck.py", line 253, in <module>
parser.parse_lines(sys.stdin, args.maxlines, '-')
File "scripts/spdxcheck.py", line 171, in parse_lines
line = line.decode(locale.getpreferredencoding(False), errors='ignore')
AttributeError: 'str' object has no attribute 'decode'
So as the line is already a string, there is no need to decode it and
the line can be dropped.
/usr/bin/python on Arch is Python 3. So this would indeed be worth
going into 4.19.
Link: http://lkml.kernel.org/r/20181023070802.22558-1-u.kleine-koenig@pengutronix.de Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Joe Perches <joe@perches.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Yufen Yu [Fri, 16 Nov 2018 23:08:39 +0000 (15:08 -0800)]
tmpfs: make lseek(SEEK_DATA/SEK_HOLE) return ENXIO with a negative offset
Other filesystems such as ext4, f2fs and ubifs all return ENXIO when
lseek (SEEK_DATA or SEEK_HOLE) requests a negative offset.
man 2 lseek says
: EINVAL whence is not valid. Or: the resulting file offset would be
: negative, or beyond the end of a seekable device.
:
: ENXIO whence is SEEK_DATA or SEEK_HOLE, and the file offset is beyond
: the end of the file.
Make tmpfs return ENXIO under these circumstances as well. After this,
tmpfs also passes xfstests's generic/448.
[akpm@linux-foundation.org: rewrite changelog] Link: http://lkml.kernel.org/r/1540434176-14349-1-git-send-email-yuyufen@huawei.com Signed-off-by: Yufen Yu <yuyufen@huawei.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Hugh Dickins <hughd@google.com> Cc: William Kucharski <william.kucharski@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Arnd Bergmann [Fri, 16 Nov 2018 23:08:35 +0000 (15:08 -0800)]
lib/ubsan.c: don't mark __ubsan_handle_builtin_unreachable as noreturn
gcc-8 complains about the prototype for this function:
lib/ubsan.c:432:1: error: ignoring attribute 'noreturn' in declaration of a built-in function '__ubsan_handle_builtin_unreachable' because it conflicts with attribute 'const' [-Werror=attributes]
This is actually a GCC's bug. In GCC internals
__ubsan_handle_builtin_unreachable() declared with both 'noreturn' and
'const' attributes instead of only 'noreturn':
Janne Huttunen [Fri, 16 Nov 2018 23:08:32 +0000 (15:08 -0800)]
mm/vmstat.c: fix NUMA statistics updates
Scan through the whole array to see if an update is needed. While we're
at it, use sizeof() to be safe against any possible type changes in the
future.
The bug here is that we wouldn't sync per-cpu counters into global ones
if there was an update of numa_stats for higher cpus. Highly
theoretical one though because it is much more probable that zone_stats
are updated so we would refresh anyway. So I wouldn't bother to mark
this for stable, yet something nice to fix.
[mhocko@suse.com: changelog enhancement] Link: http://lkml.kernel.org/r/1541601517-17282-1-git-send-email-janne.huttunen@nokia.com Fixes: 276a77e43d19 ("mm: update NUMA counter threshold size") Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 3b26eddf2d06 ("mm/gup: cache dev_pagemap while pinning pages")
modified the signature of follow_page_mask() but left the parameter
description behind.
Update the description to make the code and comments agree again.
While at it, update formatting of the return value description to match
Documentation/doc-guide/kernel-doc.rst guidelines.
Wengang Wang [Fri, 16 Nov 2018 23:08:25 +0000 (15:08 -0800)]
ocfs2: free up write context when direct IO failed
The write context should also be freed even when direct IO failed.
Otherwise a memory leak is introduced and entries remain in
oi->ip_unwritten_list causing the following BUG later in unlink path:
Roman Gushchin [Fri, 16 Nov 2018 23:08:18 +0000 (15:08 -0800)]
mm: don't reclaim inodes with many attached pages
Spock reported that commit d11df0151322 ("mm: slowly shrink slabs with a
relatively small number of objects") leads to a regression on his setup:
periodically the majority of the pagecache is evicted without an obvious
reason, while before the change the amount of free memory was balancing
around the watermark.
The reason behind is that the mentioned above change created some
minimal background pressure on the inode cache. The problem is that if
an inode is considered to be reclaimed, all belonging pagecache page are
stripped, no matter how many of them are there. So, if a huge
multi-gigabyte file is cached in the memory, and the goal is to reclaim
only few slab objects (unused inodes), we still can eventually evict all
gigabytes of the pagecache at once.
The workload described by Spock has few large non-mapped files in the
pagecache, so it's especially noticeable.
To solve the problem let's postpone the reclaim of inodes, which have
more than 1 attached page. Let's wait until the pagecache pages will be
evicted naturally by scanning the corresponding LRU lists, and only then
reclaim the inode structure.
Link: http://lkml.kernel.org/r/20181023164302.20436-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Reported-by: Spock <dairinin@gmail.com> Tested-by: Spock <dairinin@gmail.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Rik van Riel <riel@surriel.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: <stable@vger.kernel.org> [4.19.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Michal Hocko [Fri, 16 Nov 2018 23:08:15 +0000 (15:08 -0800)]
mm, memory_hotplug: check zone_movable in has_unmovable_pages
Page state checks are racy. Under a heavy memory workload (e.g. stress
-m 200 -t 2h) it is quite easy to hit a race window when the page is
allocated but its state is not fully populated yet. A debugging patch to
dump the struct page state shows
Note that the state has been checked for both PageLRU and PageSwapBacked
already. Closing this race completely would require some sort of retry
logic. This can be tricky and error prone (think of potential endless
or long taking loops).
Workaround this problem for movable zones at least. Such a zone should
only contain movable pages. Commit 852ae9513e49 ("mm, memory_hotplug:
make has_unmovable_pages more robust") has told us that this is not
strictly true though. Bootmem pages should be marked reserved though so
we can move the original check after the PageReserved check. Pages from
other zones are still prone to races but we even do not pretend that
memory hotremove works for those so pre-mature failure doesn't hurt that
much.
Link: http://lkml.kernel.org/r/20181106095524.14629-1-mhocko@kernel.org Fixes: 852ae9513e49 ("mm, memory_hotplug: make has_unmovable_pages more robust") Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Baoquan He <bhe@redhat.com> Tested-by: Baoquan He <bhe@redhat.com> Acked-by: Baoquan He <bhe@redhat.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Acked-by: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Vasily Averin [Fri, 16 Nov 2018 23:08:11 +0000 (15:08 -0800)]
mm/swapfile.c: use kvzalloc for swap_info_struct allocation
Commit b86aaba4df26 ("swap: choose swap device according to numa node")
changed 'avail_lists' field of 'struct swap_info_struct' to an array.
In popular linux distros it increased size of swap_info_struct up to 40
Kbytes and now swap_info_struct allocation requires order-4 page.
Switch to kvzmalloc allows to avoid unexpected allocation failures.
Link: http://lkml.kernel.org/r/fc23172d-3c75-21e2-d551-8b1808cbe593@virtuozzo.com Fixes: b86aaba4df26 ("swap: choose swap device according to numa node") Signed-off-by: Vasily Averin <vvs@virtuozzo.com> Acked-by: Aaron Lu <aaron.lu@intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Huang Ying <ying.huang@intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Aaro Koskinen [Fri, 16 Nov 2018 23:08:08 +0000 (15:08 -0800)]
MAINTAINERS: update OMAP MMC entry
Jarkko's e-mail address hasn't worked for a long time. We still want to
keep this driver working as it is critical for some of the OMAP boards.
I use and test this driver frequently, so change myself as a maintainer
with "Odd Fixes" status.
Link: http://lkml.kernel.org/r/20181106222750.12939-1-aaro.koskinen@iki.fi Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mike Kravetz [Fri, 16 Nov 2018 23:08:04 +0000 (15:08 -0800)]
hugetlbfs: fix kernel BUG at fs/hugetlbfs/inode.c:444!
This bug has been experienced several times by the Oracle DB team. The
BUG is in remove_inode_hugepages() as follows:
/*
* If page is mapped, it was faulted in after being
* unmapped in caller. Unmap (again) now after taking
* the fault mutex. The mutex will prevent faults
* until we finish removing the page.
*
* This race can only happen in the hole punch case.
* Getting here in a truncate operation is a bug.
*/
if (unlikely(page_mapped(page))) {
BUG_ON(truncate_op);
In this case, the elevated map count is not the result of a race.
Rather it was incorrectly incremented as the result of a bug in the huge
pmd sharing code. Consider the following:
- Process A maps a hugetlbfs file of sufficient size and alignment
(PUD_SIZE) that a pmd page could be shared.
- Process B maps the same hugetlbfs file with the same size and
alignment such that a pmd page is shared.
- Process B then calls mprotect() to change protections for the mapping
with the shared pmd. As a result, the pmd is 'unshared'.
- Process B then calls mprotect() again to chage protections for the
mapping back to their original value. pmd remains unshared.
- Process B then forks and process C is created. During the fork
process, we do dup_mm -> dup_mmap -> copy_page_range to copy page
tables. Copying page tables for hugetlb mappings is done in the
routine copy_hugetlb_page_range.
In copy_hugetlb_page_range(), the destination pte is obtained by:
dst_pte = huge_pte_alloc(dst, addr, sz);
If pmd sharing is possible, the returned pointer will be to a pte in an
existing page table. In the situation above, process C could share with
either process A or process B. Since process A is first in the list,
the returned pte is a pointer to a pte in process A's page table.
However, the check for pmd sharing in copy_hugetlb_page_range is:
/* If the pagetables are shared don't copy or take references */
if (dst_pte == src_pte)
continue;
Since process C is sharing with process A instead of process B, the
above test fails. The code in copy_hugetlb_page_range which follows
assumes dst_pte points to a huge_pte_none pte. It copies the pte entry
from src_pte to dst_pte and increments this map count of the associated
page. This is how we end up with an elevated map count.
To solve, check the dst_pte entry for huge_pte_none. If !none, this
implies PMD sharing so do not copy.
Link: http://lkml.kernel.org/r/20181105212315.14125-1-mike.kravetz@oracle.com Fixes: e6532b8d838f ("fix hugepages leak due to pagetable page sharing") Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com> Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Prakash Sangappa <prakash.sangappa@oracle.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Olof Johansson [Fri, 16 Nov 2018 23:08:00 +0000 (15:08 -0800)]
kernel/sched/psi.c: simplify cgroup_move_task()
The existing code triggered an invalid warning about 'rq' possibly being
used uninitialized. Instead of doing the silly warning suppression by
initializa it to NULL, refactor the code to bail out early instead.
Warning was:
kernel/sched/psi.c: In function `cgroup_move_task':
kernel/sched/psi.c:639:13: warning: `rq' may be used uninitialized in this function [-Wmaybe-uninitialized]
Link: http://lkml.kernel.org/r/20181103183339.8669-1-olof@lixom.net Fixes: 44083edf138ee ("psi: cgroup support") Signed-off-by: Olof Johansson <olof@lixom.net> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Vitaly Wool [Fri, 16 Nov 2018 23:07:56 +0000 (15:07 -0800)]
z3fold: fix possible reclaim races
Reclaim and free can race on an object which is basically fine but in
order for reclaim to be able to map "freed" object we need to encode
object length in the handle. handle_to_chunks() is then introduced to
extract object length from a handle and use it during mapping.
Moreover, to avoid racing on a z3fold "headless" page release, we should
not try to free that page in z3fold_free() if the reclaim bit is set.
Also, in the unlikely case of trying to reclaim a page being freed, we
should not proceed with that page.
While at it, fix the page accounting in reclaim function.
This patch supersedes "[PATCH] z3fold: fix reclaim lock-ups".
Link: http://lkml.kernel.org/r/20181105162225.74e8837d03583a9b707cf559@gmail.com Signed-off-by: Vitaly Wool <vitaly.vul@sony.com> Signed-off-by: Jongseok Kim <ks77sj@gmail.com> Reported-by-by: Jongseok Kim <ks77sj@gmail.com> Reviewed-by: Snild Dolkow <snild@sony.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Fri, 16 Nov 2018 17:38:14 +0000 (11:38 -0600)]
Merge tag 'gfs2-4.20.fixes3' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull bfs2 fixes from Andreas Gruenbacher:
"Fix two bugs leading to leaked buffer head references:
- gfs2: Put bitmap buffers in put_super
- gfs2: Fix iomap buffer head reference counting bug
And one bug leading to significant slow-downs when deleting large
files:
- gfs2: Fix metadata read-ahead during truncate (2)"
* tag 'gfs2-4.20.fixes3' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
gfs2: Fix iomap buffer head reference counting bug
gfs2: Fix metadata read-ahead during truncate (2)
gfs2: Put bitmap buffers in put_super
gfs2: Fix iomap buffer head reference counting bug
GFS2 passes the inode buffer head (dibh) from gfs2_iomap_begin to
gfs2_iomap_end in iomap->private. It sets that private pointer in
gfs2_iomap_get. Users of gfs2_iomap_get other than gfs2_iomap_begin
would have to release iomap->private, but this isn't done correctly,
leading to a leak of buffer head references.
To fix this, move the code for setting iomap->private from
gfs2_iomap_get to gfs2_iomap_begin.
Linus Torvalds [Fri, 16 Nov 2018 16:37:27 +0000 (10:37 -0600)]
Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto fixes from Herbert Xu:
"This fixes the following issues:
- Potential memory overwrite in simd
- Kernel info leaks in crypto_user
- NULL dereference and use-after-free in hisilicon"
* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
crypto: user - Zeroize whole structure given to user space
crypto: user - fix leaking uninitialized memory to userspace
crypto: simd - correctly take reqsize of wrapped skcipher into account
crypto: hisilicon - Fix reference after free of memories on error path
crypto: hisilicon - Fix NULL dereference for same dst and src
Linus Torvalds [Fri, 16 Nov 2018 16:17:29 +0000 (10:17 -0600)]
Merge tag 'drm-fixes-2018-11-16' of git://anongit.freedesktop.org/drm/drm
Pull drm fixes from Dave Airlie:
"Live from Vancouver, SoC maintainer talk, this weeks drm fixes pull
for rc3:
omapdrm:
- regression fixes for the reordering bridge stuff that went into rc1
i915:
- incorrect EU count fix
- HPD storm fix
- MST fix
- relocation fix for gen4/5
amdgpu:
- huge page handling fix
- IH ring setup
- XGMI aperture setup
- watermark setup fix
misc:
- docs and MST fix"
* tag 'drm-fixes-2018-11-16' of git://anongit.freedesktop.org/drm/drm: (23 commits)
drm/i915: Account for scale factor when calculating initial phase
drm/i915: Clean up skl_program_scaler()
drm/i915: Move programming plane scaler to its own function.
drm/i915/icl: Drop spurious register read from icl_dbuf_slices_update
drm/i915: fix broadwell EU computation
drm/amdgpu: fix huge page handling on Vega10
drm/amd/pp: Fix truncated clock value when set watermark
drm/amdgpu: fix bug with IH ring setup
drm/meson: venc: dmt mode must use encp
drm/amdgpu: set system aperture to cover whole FB region
drm/i915: Fix hpd handling for pins with two encoders
drm/i915/execlists: Force write serialisation into context image vs execution
drm/i915/icl: Fix power well 2 wrt. DC-off toggling order
drm/i915: Fix NULL deref when re-enabling HPD IRQs on systems with MST
drm/i915: Fix possible race in intel_dp_add_mst_connector()
drm/i915/ringbuffer: Delay after EMIT_INVALIDATE for gen4/gen5
drm/omap: dsi: Fix missing of_platform_depopulate()
drm/omap: Move DISPC runtime PM handling to omapdrm
drm/omap: dsi: Ensure the device is active during probe
drm/omap: hdmi4: Ensure the device is active during bind
...
Linus Torvalds [Fri, 16 Nov 2018 16:14:54 +0000 (10:14 -0600)]
Merge tag 'powerpc-4.20-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
"Two weeks worth of fixes since rc1.
- I broke 16-byte alignment of the stack when we moved PPR into
pt_regs. Despite being required by the ABI this broke almost
nothing, we eventually hit it in code where GCC does arithmetic on
the stack pointer assuming the bottom 4 bits are clear. Fix it by
padding the in-kernel pt_regs by 8 bytes.
- A couple of commits fixing minor bugs in the recent SLB rewrite.
- A build fix related to tracepoints in KVM in some configurations.
- Our old "IO workarounds" code written for Cell couldn't coexist in
a kernel that runs on Power9 with the Radix MMU, fix that.
- Remove the NPU DMA ops, these just printed a warning and should
never have been called.
- Suppress an overly chatty message triggered by CPU hotplug in some
configs.
- Two small selftest fixes.
Thanks to: Alistair Popple, Gustavo Romero, Nicholas Piggin, Satheesh
Rajendran, Scott Wood"
* tag 'powerpc-4.20-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
selftests/powerpc: Adjust wild_bctr to build with old binutils
powerpc/64: Fix kernel stack 16-byte alignment
powerpc/numa: Suppress "VPHN is not supported" messages
selftests/powerpc: Fix wild_bctr test to work on ppc64
powerpc/io: Fix the IO workarounds code to work with Radix
powerpc/mm/64s: Fix preempt warning in slb_allocate_kernel()
KVM: PPC: Move and undef TRACE_INCLUDE_PATH/FILE
powerpc/mm/64s: Only use slbfee on CPUs that support it
powerpc/mm/64s: Use PPC_SLBFEE macro
powerpc/mm/64s: Consolidate SLB assertions
powerpc/powernv/npu: Remove NPU DMA ops
Linus Torvalds [Fri, 16 Nov 2018 16:10:27 +0000 (10:10 -0600)]
Merge tag 'xtensa-20181115' of git://github.com/jcmvbkbc/linux-xtensa
Pull Xtensa fixes from Max Filippov:
- fix stack alignment for bFLT binaries.
- fix physical-to-virtual address translation for boot parameters in
MMUv3 256+256 and 512+512 virtual memory layouts.
* tag 'xtensa-20181115' of git://github.com/jcmvbkbc/linux-xtensa:
xtensa: fix boot parameters address translation
xtensa: make sure bFLT stack is 16 byte aligned
Linus Torvalds [Fri, 16 Nov 2018 15:31:59 +0000 (09:31 -0600)]
Merge tag 'for-linus-20181115' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
- Discard loop fix, caused by integer overflow (Dave)
- Blacklist of Samsung drive that hangs with power management (Diego)
- Copy bio priority when cloning it (Hannes)
- Fix race condition exposed in floppy (me)
- Fix SCSI queue cleanup regression. While elusive, it caused oopses in
queue running (Ming)
- Fix bad string copy in kyber tracing (Omar)
* tag 'for-linus-20181115' of git://git.kernel.dk/linux-block:
SCSI: fix queue cleanup race before queue initialization is done
block: fix 32 bit overflow in __blkdev_issue_discard()
libata: blacklist SAMSUNG MZ7TD256HAFV-000L9 SSD
block: copy ioprio in __bio_clone_fast() and bounce
kyber: fix wrong strlcpy() size in trace_kyber_latency()
floppy: fix race condition in __floppy_read_block_0()
Ville Syrjälä [Wed, 14 Nov 2018 17:34:40 +0000 (19:34 +0200)]
drm/i915: Disable LP3 watermarks on all SNB machines
I have a Thinkpad X220 Tablet in my hands that is losing vblank
interrupts whenever LP3 watermarks are used.
If I nudge the latency value written to the WM3 register just
by one in either direction the problem disappears. That to me
suggests that the punit will not enter the corrsponding
powersave mode (MPLL shutdown IIRC) unless the latency value
in the register matches exactly what we read from SSKPD. Ie.
it's not really a latency value but rather just a cookie
by which the punit can identify the desired power saving state.
On HSW/BDW this was changed such that we actually just write
the WM level number into those bits, which makes much more
sense given the observed behaviour.
We could try to handle this by disallowing LP3 watermarks
only when vblank interrupts are enabled but we'd first have
to prove that only vblank interrupts are affected, which
seems unlikely. Also we can't grab the wm mutex from the
vblank enable/disable hooks because those are called with
various spinlocks held. Thus we'd have to redesigne the
watermark locking. So to play it safe and keep the code
simple we simply disable LP3 watermarks on all SNB machines.
To do that we simply zero out the latency values for
watermark level 3, and we adjust the watermark computation
to check for that. The behaviour now matches that of the
g4x/vlv/skl wm code in the presence of a zeroed latency
value.
v2: s/USHRT_MAX/U32_MAX/ for consistency with the types (Chris)
Cc: stable@vger.kernel.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101269
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103713 Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181114173440.6730-1-ville.syrjala@linux.intel.com
Lucas De Marchi [Wed, 14 Nov 2018 01:15:08 +0000 (17:15 -0800)]
drm/i915/icl: reverse uninit order
Bspec 21257 says "DDIA PHY is the comp master, so it must
not be un-initialized if other combo PHYs are in use". Here
we are shutting down all phys, so it's not strictly required.
However let's be consistent on deinitializing things in the
reversed order we initialized them.
v2: simplify protection for enum port being unsigned in future
v3: spell out reverse rather than rev
Lucas De Marchi [Wed, 14 Nov 2018 01:15:07 +0000 (17:15 -0800)]
drm/i915/icl: replace check for combo phy
These are the only places that assume ports A and B are the ones with
combo phy. Let's use intel_port_is_combophy() there to make sure
it checks for combo phy ports the same way everywhere.
v2: define for_each_combo_port() helper to check the ports
Dave Airlie [Thu, 15 Nov 2018 22:52:14 +0000 (08:52 +1000)]
Merge tag 'drm-intel-fixes-2018-11-15' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes
- Fix Bugzilla #108712: Fix incorrect EU count report from kernel
- Fix to account for scale factor when calculating initial phase on scaled output
- Avoid too trigger-happy HPD storm detection and fix a race and an OOPS for MST systems.
- Relocation race fix for Gen4/5
- A couple ICL fixes and dependencies for above Fixes:.
projects / kernel.git / log