Ben Widawsky [Tue, 24 Feb 2015 16:22:36 +0000 (16:22 +0000)]
drm/i915: Create page table allocators
As we move toward dynamic page table allocation, it becomes much easier
to manage our data structures if break do things less coarsely by
breaking up all of our actions into individual tasks. This makes the
code easier to write, read, and verify.
Aside from the dissection of the allocation functions, the patch
statically allocates the page table structures without a page directory.
This remains the same for all platforms,
The patch itself should not have much functional difference. The primary
noticeable difference is the fact that page tables are no longer
allocated, but rather statically declared as part of the page directory.
This has non-zero overhead, but things gain additional complexity as a
result.
This patch exists for a few reasons:
1. Splitting out the functions allows easily combining GEN6 and GEN8
code. Page tables have no difference based on GEN8. As we'll see in a
future patch when we add the DMA mappings to the allocations, it
requires only one small change to make work, and error handling should
just fall into place.
2. Unless we always want to allocate all page tables under a given PDE,
we'll have to eventually break this up into an array of pointers (or
pointer to pointer).
3. Having the discrete functions is easier to review, and understand.
All allocations and frees now take place in just a couple of locations.
Reviewing, and catching leaks should be easy.
4. Less important: the GFP flags are confined to one location, which
makes playing around with such things trivial.
v2: Updated commit message to explain why this patch exists
v3: For lrc, s/pdp.page_directory[i].daddr/pdp.page_directory[i]->daddr/
v4: Renamed free_pt/pd_single functions to unmap_and_free_pt/pd (Daniel)
v5: Added additional safety checks in gen8 clear/free/unmap.
v6: Use WARN_ON and return -EINVAL in alloc_pt_range (Mika).
v7: Make err_out loop symmetrical to the way we allocate in
alloc_pt_range. Also s/page_tables/page_table and correct commit
message (Mika)
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v3+) Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 24 Feb 2015 16:22:35 +0000 (16:22 +0000)]
drm/i915: Complete page table structures
Move the remaining members over to the new page table structures.
This can be squashed with the previous commit if desire. The reasoning
is the same as that patch. I simply felt it is easier to review if split.
v2: In lrc: s/ppgtt->pd_dma_addr[i]/ppgtt->pdp.page_directory[i].daddr/
v3: Rebase.
v4: Rebased after s/page_tables/page_table/.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v2+) Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 24 Feb 2015 16:22:34 +0000 (16:22 +0000)]
drm/i915: page table abstractions
When we move to dynamic page allocation, keeping page_directory and pagetabs as
separate structures will help to break actions into simpler tasks.
To help transition the code nicely there is some wasted space in gen6/7.
This will be ameliorated shortly.
Following the x86 pagetable terminology:
PDPE = struct i915_page_directory_pointer_entry.
PDE = struct i915_page_directory_entry [page_directory].
PTE = struct i915_page_table_entry [page_tables].
v2: fixed mismatches after clean-up/rebase.
v3: Clarify the names of the multiple levels of page tables (Daniel)
v4: Addressing Mika's review comments.
s/gen8_free_page_directories/gen8_free_page_directory and free the
page tables for the directory there.
In gen8_ppgtt_allocate_page_directories, do not leak previously allocated
pt in case the page_directory alloc fails.
Update error return handling in gen8_ppgtt_alloc.
v5: Do not leak pt on error in gen6_ppgtt_allocate_page_tables. (Mika)
v6: s/page_tables/page_table/. (Mika)
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v2+) Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Sonika Jindal [Wed, 25 Feb 2015 04:59:12 +0000 (10:29 +0530)]
drm/i915/skl: Add support for edp1.4 low vswing
Based upon vbt's vswing preemph settings value select the appropriate
translations for edp.
v2: Incorporating bspec changes for vswing and preemph levels, adding edp
translation table. Removed HSW from selection 9 which is specific to skl and
correcting the returning of level2 from max pre emph (Damien)
v3: Rebasing on top of renaming patches. Adding level(3,0) since level(2,2) as
mentioned in bspec is invalid as per edp spec. Also changed the determining of
size of the table selected (Satheesh).
v4: Adding level 3 in max voltage selection if low vswing is selected (Satheesh)
v5: Add a comment stating that skl_ddi_translations_edp is for eDP 1.4
low vswing panels.
v6: Updating recommended DDI translation table for edp 1.4
Sonika Jindal [Wed, 25 Feb 2015 04:59:11 +0000 (10:29 +0530)]
drm/i915/skl: Support for edp low_vswing param in vbt
v2: Adding VBT version check for low_vswing field, and correcting parsing
v3: (Damien)
- Restrain the scope of the 'vswing' variable
- Use the more idiomatic "ev_priv->vbt.edp_low_vswing = vswing == 0;"
instead of if (foo) var = true; else var = false;
- Shorten edp_vswing_premph_setting to edp_vswing_premph to fit in 80 chars
- Add the version from which the edp_vswing_premph field is valid in the
struct definition
Reviewed-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v2) Signed-off-by: Sonika Jindal <sonika.jindal@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Dan Carpenter [Wed, 25 Feb 2015 13:17:48 +0000 (16:17 +0300)]
drm/i915: cleanup some indenting
Static checkers complain that we should probably add curly braces
because, from the indenting, it looks like seq_printf() should be inside
the list_for_each_entry() loop. But the code is actually correct, it's
just the indenting which is off.
Besides fixing the indenting on seq_printf(), I did add curly braces,
because generally mult-line indents should have curly braces to make
them more readable.
The unintended indent was left behind and not unindented in
drm/i915: Make pin count per VMA Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Mon, 9 Feb 2015 13:35:10 +0000 (13:35 +0000)]
drm/i915/skl: Make sure to allocate mininum sizes in the DDB
I overlooked the fact that we need to allocate a minimum 8 blocks and
that just allocating the planes depending on how much they need to fetch
from the DDB in proportion of how much memory bw is necessary for the
whole display can lead to cases where we don't respect those minima (and
thus overrun).
So, instead, start by allocating 8 blocks to each active display plane
and then allocate the remaining blocks like before.
v2: Rebase on top of -nightly
Cc: Mahesh Kumar <mahesh1.kumar@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Vandana Kannan [Fri, 13 Feb 2015 10:03:00 +0000 (15:33 +0530)]
drm/i915/bdw: Add support for DRRS to switch RR
For Broadwell, there is one instance of Transcoder MN values per transcoder.
For dynamic switching between multiple refreshr rates, M/N values may be
reprogrammed on the fly. Link N programming triggers update of all data and
link M & N registers and the new M/N values will be used in the next frame
that is output.
V2: [By Ram]: intel_dp_set_m_n() is rewritten to accommodate
gen >= 8 [Rodrigo]
V3: Coding style correction [Ram]
V4: [By Ram] intel_dp_set_m_n modifications are moved into a
separate patch, retaining only DRRS related changes here [Rodrigo]
Signed-off-by: Vandana Kannan <vandana.kannan@intel.com> Signed-off-by: Pradeep Bhat <pradeep.bhat@intel.com> Signed-off-by: Ramalingam C <ramalingam.c@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ramalingam C [Fri, 13 Feb 2015 10:02:59 +0000 (15:32 +0530)]
drm/i915: Add support for DRRS in intel_dp_set_m_n
Till Gen 7 we have two sets of M_N registers, but Gen 8 onwards
we have only one M_N register set. To support DRRS on both scenarios
a input parameter to intel_dp_set_m_n is added.
In case of DRRS, When platform provides two set of M_N registers for dp,
we can program them with two different dividers and switch between them.
But when only one such register set is provided, we have to program
the required divider M_N value on that registers itself.
Two enum members M1_N1 and M2_N2 are defined to represent the above
scenarios.
M1_N1 : Program dp_m_n on M1_N1 registers
dp_m2_n2 on M2_N2 registers (If supported)
M2_N2 : Program dp_m2_n2 on M1_N1 registers
M2_N2 registers are not supported
Signed-off-by: Ramalingam C <ramalingam.c@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Thomas Daniel [Wed, 18 Feb 2015 11:48:21 +0000 (11:48 +0000)]
drm/i915: Shift driver's HWSP usage out of reserved range
As of Gen6, the general purpose area of the hardware status page has shrunk and
now begins at dword 0x30. i915 driver uses dword 0x20 to store the seqno which
is now reserved. So shift our HWSP dwords up into the general purpose range
before this bites us.
Note that all available documentation just says this is reserved
without going into details about what it's used for.
Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> Reviewed-by: Dave Gordon <david.s.gordon@intel.com>
[danvet: Add clarification from Thomas that unfortunately Bspec is
silent on what "reserverd" precisely means.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Thomas Daniel [Mon, 16 Feb 2015 16:12:53 +0000 (16:12 +0000)]
drm/i915: Reset logical ring contexts' head and tail during GPU reset
Work was getting left behind in LRC contexts during reset. This causes a hang
if the GPU is reset when HEAD==TAIL because the context's ringbuffer head and
tail don't get reset and retiring a request doesn't alter them, so the ring
still appears full.
Added a function intel_lr_context_reset() to reset head and tail on a LRC and
its ringbuffer.
Call intel_lr_context_reset() for each context in i915_gem_context_reset() when
in execlists mode.
Testcase: igt/pm_rps --run-subtest reset #bdw
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88096 Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> Reviewed-by: Dave Gordon <david.s.gordon@intel.com>
[danvet: Flatten control flow in the lrc reset code a notch.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Sat, 14 Feb 2015 18:30:29 +0000 (18:30 +0000)]
drm/i915/skl: Tune IZ hashing when subslices are unbalanced
When one EU is disabled in a particular subslice, we can tune how the
work is spread between subslices to improve EU utilization.
v2: - Use a bitfield to record which subslice(s) has(have) 7 EUs. That
will also make the machinery work if several sublices have 7 EUs.
(Jeff Mcgee)
- Only apply the different hashing algorithm if the slice is
effectively unbalanced by checking there's a single subslice with
7 EUs. (Jeff Mcgee)
v3: Fix typo in comment (Jeff Mcgee)
Issue: VIZ-3845 Cc: Jeff Mcgee <jeff.mcgee@intel.com> Reviewed-by: Jeff Mcgee <jeff.mcgee@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Fri, 13 Feb 2015 19:23:47 +0000 (17:23 -0200)]
drm/i915: don't reallocate the compressed FB at every frame
With the current code we just reallocate the compressed FB at every
FBC update: we have X in one frame, then in the other frame we need X
again, but we check "needed < have" instead of "needed <= have".
v2: Rebase after Jani addressed the other problems described in v1.
Cc: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jeff McGee [Fri, 13 Feb 2015 16:27:56 +0000 (10:27 -0600)]
drm/i915: Request full SSEU enablement on Gen9
On Gen9 the render power gating can leave slice/subslice/EU in
a partially enabled state. We must make an explicit request for
full SSEU enablement through the Render Power Clock State
register when resuming render work. This register is save/
restored in the logical ring context image for execlist
submission mode. Initialize its value in each LRC image to
request full enablement according to the device SSEU config.
Thanks to Sharma Ankitprasad and Akash Goel for highlighting the
issue and proposing the initial fix on which this patch is based.
v2: Adjusted the names of the power gating support flags to fit
update of an earlier patch.
Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> Reviewed-by: "Akash Goel <akash.goel@intel.com>" Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jeff McGee [Fri, 13 Feb 2015 16:27:55 +0000 (10:27 -0600)]
drm/i915/skl: Add SKL HW status to SSEU status
Add a new section to the 'i915_sseu_status' debugfs entry to
report the currently enabled counts of slice, subslice, and
execution units on the device. The count of enabled subslice
per slice represents the most enabled subslice on any one
slice for devices where imbalances may exist. Similarly, the
count of enabled EU per subslice represents the most enabled
EU on any one subslice.
Collect this device status for Skylake by reading the Gen9
power gate control ack message registers. Power gate control
operates on EU in pairs, therefore our reported counts of
enabled EU can be overestimated by one for each pair in which
one EU is fused-off.
Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jeff McGee [Fri, 13 Feb 2015 16:27:54 +0000 (10:27 -0600)]
drm/i915/skl: Determine SKL slice/subslice/EU info
Read fuse registers to determine the available slice total,
subslice total, subslice per slice, EU total, and EU per subslice
counts of the SKL device. The EU per subslice attribute is more
precisely defined as the maximum EU available on any one subslice,
since available EU counts may vary across subslices due to fusing.
Set flags indicating the SKL device's slice/subslice/EU (SSEU)
power gating capability. Make all values available via debugfs
entry 'i915_sseu_status'.
v2: Several small clean-ups suggested by Damien. Most notably,
used smaller types for the new device info fields to reduce
memory usage and improved the clarity/readability of the
method used to extract attribute values from the fuse
registers.
Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Wed, 14 Jan 2015 11:20:57 +0000 (11:20 +0000)]
drm/i915: Trim the command parser allocations
Currently, the command parser tries to create a secondary batch exactly
as large as the original, and vmap both. This is open to abuse by
userspace using extremely large batch objects, but only executing very
short batches. For example, this would be if userspace were to implement
a command submission ringbuffer. However, we only need to allocate pages
for just the contents of the command sequence in the batch - all
relocations copied to the secondary batch will reference the original
batch and so there can be no access to the secondary batch outside of
the explicit execution region.
Testcase: igt/gem_exec_big #ivb,byt,hsw
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88308 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: John Harrison <John.C.Harrison@Intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Dave Gordon [Wed, 10 Dec 2014 18:12:12 +0000 (18:12 +0000)]
drm/i915: FIFO space query code refactor
When querying the GTFIFOCTL register to check the FIFO space, the read value
must be masked. The operation is repeated explicitly in several places. This
change refactors the read-and-mask code into a function call.
v2: rebased on top of Mika's forcewake patch set, specifically:
[PATCH 8/8] drm/i915: Enum forcewake domains and domain identifiers
Change-Id: Id1a9f3785cb20b82d4caa330c37b31e4e384a3ef Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Sun, 22 Feb 2015 14:11:20 +0000 (15:11 +0100)]
drm: WARN if drm_handle_vblank is called errornously
KMS drivers are in full control of their irq and vblank handling, if
they get a vblank interrupt before drm_vblank_init or after
drm_vblank_cleanup that's just a driver bug.
For ums driver there's only r128 and radeon which support vblank, and
they call drm_vblank_init in their driver load functions. Which again
means that userspace can do whatever it wants with interrupt, vblank
structures will always be there.
So this should never happen, let's catch driver issues with a WARN_ON.
Motivated by some discussions with Imre.
v2: Use WARN_ON_ONCE as suggested by Imre.
Cc: Imre Deak <imre.deak@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Acked-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Daniel Vetter [Sun, 22 Feb 2015 14:11:19 +0000 (15:11 +0100)]
drm/irq: Don't call ->get_vblank_counter directly from irq_uninstall/cleanup
The pipe might already have been shut down, and then it's not a good
idea to call hw accessor functions. Instead use the same logic as
drm_vblank_off which has all the necessary checks to avoid troubles or
inconsistency.
Noticed by Imre while reviewing my patches to remove some sanity
checks from ->get_vblank_counter.
v2: Try harder. disable_and_save can still access the vblank stuff
when vblank->enabled isn't set. It has to, since vlbank irq could be
disable but the pipe is still on when being called from
drm_vblank_off. But we still want to use that code for more code
sharing. So add a check for vblank->enabled on top - if that's not set
we shouldn't have anyone waiting for the vblank. If we have that's a
pretty serious bug.
The other issue that Imre spotted is drm_vblank_cleanup. That code
again calls disable_and_save and so suffers from the same issues. But
really drm_irq_uninstall should have cleaned that all up, so replace
the code with WARN_ON. Note that we can't delete the timer cleanup
since drivers aren't required to use drm_irq_install/uninstall, but
can do their own irq handling.
v3: Make it clear that all that gunk in drm_irq_uninstall is really
just bandaids for UMS races between the irq/vblank code. In UMS
userspace is in control of enabling/disabling interrupts in general
and vblanks specifically.
v4: Imre observed that KMS drivers all call drm_vblank_cleanup before
drm_irq_uninstall (as they should), so again the code in there is dead
for KMS (due to dev->num_crtcs == 0 after drm_vblank_cleanup). Or
should be, so only WARN for KMS - with UMS userspace could try to do
evil things.
v5: After more discussion on irc we've gone back to v3: the
del_timer_sync is required in all cases in drm_vblank_cleanup, but
let's restrict the WARN_ON to kms drivers only. Imre was also
concerned that bad things could happen without the disable_and_save
call. But we immediately free vblank structures afterwards which makes
the save useless. And drm_handle_vblank has a check for dev->num_crtcs
to avoid surprises with ums.
Cc: Imre Deak <imre.deak@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Acked-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Daniel Vetter [Fri, 13 Feb 2015 20:03:43 +0000 (21:03 +0100)]
drm/i915: Drop pipe_enable checks in vblank funcs
With Ville's rework to use drm_crtc_vblank_on/off the core will take
care of rejecting drm_vblank_get calls when the pipe is off. Also the
core won't call the get_vblank_counter hooks in that case either. And
since we've dropped ums support recently we can now remove these
hacks, yay!
Noticed while trying to answer questions Laurent had about how the new
atomic helpers work.
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Daniel Vetter [Fri, 13 Feb 2015 20:03:42 +0000 (21:03 +0100)]
drm/irq: Add drm_crtc_vblank_reset
At driver load we need to tell the vblank code about the state of the
pipes, so that the logic around reject vblank_get when the pipe is off
works correctly.
Thus far i915 used drm_vblank_off, but one of the side-effects of it
is that it also saves the vblank counter. And for that it calls down
into the ->get_vblank_counter hook. Which isn't really a good idea
when the pipe is off for a few reasons:
- With runtime pm the register might not respond.
- If the pipe is off some datastructures might not be around or
unitialized.
The later is what blew up on gen3: We look at intel_crtc->config to
compute the vblank counter, and for a disabled pipe at boot-up that's
just not there. Thus far this was papered over by a check for
intel_crtc->active, but I want to get rid of that (since it's fairly
race, vblank hooks are called from all kinds of places).
So prep for that by adding a _reset functions which only does what we
really need to be done at driver load: Mark the vblank pipe as off,
but don't do any vblank counter saving or event flushing - neither of
that is required.
v2: Clarify the code flow slightly as suggested by Ville.
v3: Fix kerneldoc spelling, spotted by Laurent.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Imre Deak <imre.deak@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> (v2) Acked-by: Dave Airlie <airlied@redhat.com> Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
drm/i915: Remove references to previously removed UMS config option
Commit 03dae59c72ffffd8 ("drm/i915: Ditch UMS config option") removed
CONFIG_DRM_I915_UMS from the Kconfig file, but i915_drv.c still
references this option in two #ifndef statements.
As an undefined config option will always be 'false', we can drop
the #ifndefs alltogether and adapt the printed error message.
This inconsistency was found with the undertaker tool.
Signed-off-by: Andreas Ruprecht <rupran@einserver.de>
[danvet: Undo logging change as requested by Jani.] Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Mika Kuoppala [Thu, 12 Feb 2015 08:26:02 +0000 (10:26 +0200)]
drm/i915: Add process identifier to requests
We use the pid of the process which opened our device when
we track which was the culprit of the gpu hang. But as that
file descriptor might get inherited, we might blame the
wrong process when we record the error state.
Track process identifiers in requests to always find
the correct offender.
v2: Track only user processes (Chris)
Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
[danvet: drop NULL check before put_pid as suggested by Chris.] Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Tue, 3 Feb 2015 14:25:14 +0000 (14:25 +0000)]
drm/i915: Detect eDRAM with the enabled bit only
At the moment we compare the whole EDRAM_PRESENT/EDRAMCAP register value
to 1 while EDRAM_PRESENT is only bit 0 (the rest may be used to describe
eDRAM capabilities).
To be more future proof, only look at bit 0 to detect eDRAM presence.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Tue, 10 Feb 2015 19:32:21 +0000 (19:32 +0000)]
drm/i915: Make intel_logical_ring_advance_and_submit() static
This function is only used in intel_lrc.c, so restrict it to that file. The
function was moved around to avoid a forward declaration and group it with its
user.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Tue, 10 Feb 2015 19:32:19 +0000 (19:32 +0000)]
drm/i915: Make intel_lr_context_render_state_init() static
This function is only used in intel_lrc.c, so restrict it to that file. The
function was moved around to avoid a forward declaration and group it with its
user.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Damien Lespiau [Tue, 10 Feb 2015 19:32:17 +0000 (19:32 +0000)]
drm/i915: Make intel_ring_setup_status_page() static
This function is only used in intel_ringbuffer.c, so restrict it to that
file. The function was moved around to avoid a forward declaration and
group it with its user.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: Squash in fixup from Wu Fengguang.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:54 +0000 (19:05 +0800)]
drm/i915: Support alias ppgtt in VM if ppgtt is enabled
The current Intel GVT-g only supports alias ppgtt. And the
emulation is done in the host by first trapping PP_DIR_BASE
mmio accesses. Updating PP_DIR_BASE by using instructions such
as MI_LOAD_REGISTER_IMM are hard to detect and are not supported
in current code. Therefore this patch also adds a new callback
routine - vgpu_mm_switch() to set the PP_DIR_BASE by mmio writes.
v2:
take Chris' comments:
- move the code into sanitize_enable_ppgtt()
v4:
take Tvrtko's comments:
- fix the parenthesis alignment warning
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:53 +0000 (19:05 +0800)]
drm/i915: Create vGPU specific MMIO operations to reduce traps
In the virtualized environment, forcewake operations are not
necessary for the driver, because mmio accesses will be trapped
and emulated by the host side, and real forcewake operations are
also done in the host. New mmio access handlers are added to directly
call the __raw_i915_read/write, therefore will reduce many traps and
increase the overall performance for drivers running in the VM with
Intel GVT-g enhancement.
v2:
take Chris' comments:
- register the mmio hooks in intel_uncore_init()
v3:
take Daniel's comments:
- use macros to assign mmio write functions for vGPU
v4:
take Tvrtko's comments:
- also use mmio hooks for read operations
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Kevin Tian <kevin.tian@intel.com>k Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:52 +0000 (19:05 +0800)]
drm/i915: Disable power management for i915 driver in VM
With Intel GVT-g, GPU power management is controlled by
host driver, so there is no need to provide virtualized
GPU PM support. In the future it might be useful to gather
VM input for freq boost, but now let's disable it simply.
v2:
take Chris' comments:
- do not special case this to gen6+
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:51 +0000 (19:05 +0800)]
drm/i915: Add the display switch logic for vGPU in i915 driver
Display switch logic is added to notify the host side that
current vGPU have a valid surface to show. It does so by
writing the display_ready field in PV INFO page, and then
will be handled in the host side. This is useful to avoid
trickiness when the VM's framebuffer is being accessed in
the middle of VM modesetting, e.g. compositing the framebuffer
in the host side.
v2:
- move the notification code outside the 'else' in load sequence
- remove the notification code in intel_crtc_set_config()
v4:
- code rebase, no need to define another dev_priv
- use #define instead of enum for display readiness
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:50 +0000 (19:05 +0800)]
drm/i915: Disable framebuffer compression for i915 driver in VM
Framebuffer compression is disabled when driver detects it's
running in a Intel GVT-g enlightened VM, because FBC is not
emulated and there is no stolen memory for a vGPU.
v2:
take Chris' comments:
- move the code into intel_update_fbc()
v4:
take Tvrtko's comments:
- rebase the code into intel_fbc_update()
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:49 +0000 (19:05 +0800)]
drm/i915: Partition the fence registers for vGPU in i915 driver
With Intel GVT-g, the fence registers are partitioned by multiple
vGPU instances in different VMs. Routine i915_gem_load() is modified
to reset the num_fence_regs, when the driver detects it's running in
a VM. Accesses to the fence registers from vGPU will be trapped and
remapped by the host side. And the allocated fence number is provided
in PV INFO page structure. By now, the value of fence number is fixed,
but in the future we can relax this limitation, to allocate the fence
registers dynamically from host side.
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Eddie Dong <eddie.dong@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:48 +0000 (19:05 +0800)]
drm/i915: Adds graphic address space ballooning logic
With Intel GVT-g, the global graphic memory space is partitioned by
multiple vGPU instances in different VMs. The ballooning code is called
in i915_gem_setup_global_gtt(), utilizing the drm mm allocator APIs to
mark the graphic address space which are partitioned out to other vGPUs
as reserved. With ballooning, host side does not need to translate a
grahpic address from guest view to host view. By now, current implementation
only support the static ballooning, but in the future, with more cooperation
from guest driver, the same interfaces can be extended to grow/shrink the
guest graphic memory dynamically.
v2:
take Chris and Daniel's comments:
- no guard page between different VMs
- use drm_mm_reserve_node() to do the reservation for ballooning,
instead of the previous drm_mm_insert_node_in_range_generic()
v3:
take Daniel's comments:
- move ballooning functions into i915_vgpu.c
- add kerneldoc to ballooning functions
v4:
take Tvrtko's comments:
- more accurate comments and commit message
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Zhi Wang <zhi.a.wang@intel.com> Signed-off-by: Eddie Dong <eddie.dong@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Yu Zhang [Tue, 10 Feb 2015 11:05:47 +0000 (19:05 +0800)]
drm/i915: Introduce a PV INFO page structure for Intel GVT-g.
Introduce a PV INFO structure, to facilitate the Intel GVT-g
technology, which is a GPU virtualization solution with mediated
pass-through. This page contains the shared information between
i915 driver and the host emulator. For now, this structure utilizes
an area of 4K bytes on HSW GPU's unused MMIO space. Future hardware
will have the reserved window architecturally defined, and layout
of the page will be added in future BSpec.
The i915 driver load routine detects if it is running in a VM by
reading the contents of this PV INFO page. Thereafter a flag,
vgpu.active is set, and intel_vgpu_active() is used by checking
this flag to conclude if GPU is virtualized with Intel GVT-g. By
now, intel_vgpu_active() will return true, only when the driver
is running as a guest in the Intel GVT-g enhanced environment on
HSW platform.
v2:
take Chris' comments:
- call the i915_check_vgpu() in intel_uncore_init()
- sanitize i915_check_vgpu() by adding BUILD_BUG_ON() and debug info
take Daniel's comments:
- put the definition of PV INFO into a new header - i915_vgt_if.h
other changes:
- access mmio regs by readq/readw in i915_check_vgpu()
v3:
take Daniel's comments:
- move the i915/vgt interfaces into a new i915_vgpu.c
- update makefile
- add kerneldoc to functions which are non-static
- add a DOC: section describing some of the high-level design
- update drm docbook
other changes:
- rename i915_vgt_if.h to i915_vgpu.h
v4:
take Tvrtko's comments:
- fix a typo in commit message
- add debug message when vgt version mismatches
- rename low_gmadr/high_gmadr to mappable/non-mappable in PV INFO
structure
Signed-off-by: Yu Zhang <yu.c.zhang@linux.intel.com> Signed-off-by: Jike Song <jike.song@intel.com> Signed-off-by: Eddie Dong <eddie.dong@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Zhi Wang [Tue, 10 Feb 2015 09:11:36 +0000 (17:11 +0800)]
drm/i915: Introduce bit definitions of CTXT_SR_CTRL register.
This patch introduces 2 bit definitions of context save/restore
control register.
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com> Suggested-by: Dave Gordon <david.s.gordon@intel.com> Cc: Dave Gordon <david.s.gordon@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tvrtko Ursulin [Tue, 10 Feb 2015 17:16:14 +0000 (17:16 +0000)]
drm/i915: Use fb modifiers in intel_check_cursor_plane
Also drop the mutex since with universal planes there is always a
proper framebuffer around which wraps the underlying bo. Which means
tiling is locked down. This was different in the old code which
directly took gem handles. The looking though was always cargo-cult
since races where not prevented in any way.
v2: Unconditionally enforce untiled, because cursors are always
untiled. The check for physical or gtt cursor is irrelevant. Also
clarify the commit message a bit
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> (v1) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tvrtko Ursulin [Tue, 10 Feb 2015 17:16:13 +0000 (17:16 +0000)]
drm/i915/skl: Use fb modifiers for sprites
While at it just outright remove the tiling check in
intel_check_sprite_plane because it's impossible: We only allow
untiled and X-tiled. This essentially reverts
drm/i915: Error out when trying to set a y-tiled as a sprite
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
[danvet: Drop the hunk in check_sprite, it's impossible.] Cc: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 10 Feb 2015 17:16:11 +0000 (17:16 +0000)]
drm/i915: Use fb format modifiers in skylake_update_primary_plane
Just a little demo really. We probably need to introduce skl specific
functions for a lot of the format validation stuff, or at least
helpers. Specifically I think intel_framebuffer_init and
intel_fb_align_height must be adjusted to have an i915_ and a skl_
variant. And only shared code should be converted to fb modifiers,
platform code (like the plane config readout can keep on using old
tiling_mode defines to avoid some churn).
Daniel Vetter [Tue, 10 Feb 2015 17:16:10 +0000 (17:16 +0000)]
drm/i915: Switch intel_fb_align_height to fb format modifiers
With this we can treat the fb format modifier completely independently
from the fencing mode in obj->tiling_mode in the initial plane code.
Which means new tiling modes without any gtt fence are now fully
support in the core i915 driver code.
v2: Also add pixel_format while at it, we need this to compute the
height for the new tiling formats.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Daniel Vetter [Tue, 10 Feb 2015 17:16:09 +0000 (17:16 +0000)]
drm/i915: Set up fb format modifier for initial plane config
No functional changes yet since intel_framebuffer_init would have
fixed this up for us. But this is prep work to be able to handle new
tiling layouts in the initial plane config code.
Follow-up patches will start to make use of this and switch over to fb
modifiers where needed.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Daniel Vetter [Tue, 10 Feb 2015 17:16:06 +0000 (17:16 +0000)]
drm/i915: Add fb format modifier support
Currently we don't support anything but X tiled. And for an easier
transition it makes a lot of sense to just keep requiring that X tiled
is properly fenced.
Which means we need to do absolutely nothing in old code to support fb
modifiers, yay!
v2: Fix the Y tiling check, noticed by Tvrtko.
v3: Catch Y-tiled fb for legacy addfb again (Tvrtko) and explain why
we want X tiling to match in the comment.
Tvrtko Ursulin [Tue, 10 Feb 2015 17:16:05 +0000 (17:16 +0000)]
drm/i915: Add tiled framebuffer modifiers
To be used from the new addfb2 extension.
v2:
- Drop Intel-specific untiled modfier.
- Move to drm_fourcc.h.
- Document layouts a bit and denote them as platform-specific and not
useable for cross-driver sharing.
- Add Y-tiling for completeness.
- Drop special docstring markers to avoid confusing kerneldoc.
v3: Give Y-tiling a unique idea, noticed by Tvrtko.
Paulo Zanoni [Mon, 9 Feb 2015 16:46:29 +0000 (14:46 -0200)]
drm/i915: change dev_priv->fbc.plane to dev_priv->fbc.crtc
Since the mapping from CRTCs to planes is fixed, looking at the CRTC
is essentially the same as looking at the plane. Also, the next
patches wil start using the frontbuffer_bits macros, and they take the
pipe as the parameter instead of the plane, and this could differ on
gens 2 and 3.
Another nice thing is that we don't risk accidentally initializing
things to PLANE_A if we don't set the value before it is used for the
first time. But this shouldn't be a problem with the current code.
V2: Rebase.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> (v1) Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Mon, 9 Feb 2015 16:46:27 +0000 (14:46 -0200)]
drm/i915: don't try to find crtcs for FBC if it's disabled
.. because it would be a waste of time, so move the place where the
check is done. Also, with this we won't risk printing "more than one
pipe active, disabling compression" or "no output, disabling" when FBC
is actually disabled.
This patch also represents a small behavior difference when using
i915.powersave=0: it is now exactly the same as i915.enable_fbc=0 on
this part of the code.
V2: Rebase.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> (v1) Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Rob Clark [Thu, 5 Feb 2015 14:41:52 +0000 (14:41 +0000)]
drm: add support for tiled/compressed/etc modifier in addfb2
In DRM/KMS we are lacking a good way to deal with tiled/compressed
formats. Especially in the case of dmabuf/prime buffer sharing, where
we cannot always rely on under-the-hood flags passed to driver specific
gem-create ioctl to pass around these extra flags.
The proposal is to add a per-plane format modifier. This allows to, if
necessary, use different tiling patters for sub-sampled planes, etc.
The format modifiers are added at the end of the ioctl struct, so for
legacy userspace it will be zero padded.
v1: original
v1.5: increase modifier to 64b
v2: Incorporate review comments from the big thread, plus a few more.
- Add a getcap so that userspace doesn't have to jump through hoops.
- Allow modifiers only when a flag is set. That way drivers know when
they're dealing with old userspace and need to fish out e.g. tiling
from other information.
- After rolling out checks for ->modifier to all drivers I've decided
that this is way too fragile and needs an explicit opt-in flag. So
do that instead.
- Add a define (just for documentation really) for the "NONE"
modifier. Imo we don't need to add mask #defines since drivers
really should only do exact matches against values defined with
fourcc_mod_code.
- Drop the Samsung tiling modifier on Rob's request since he's not yet
sure whether that one is accurate.
v3:
- Also add a new ->modifier[] array to struct drm_framebuffer and fill
it in drm_helper_mode_fill_fb_struct. Requested by Tvrkto Uruslin.
- Remove TODO in comment and add code comment that modifiers should be
properly documented, requested by Rob.
Cc: Rob Clark <robdclark@gmail.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Daniel Stone <daniel@fooishbar.org> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Michel Dänzer <michel@daenzer.net> Signed-off-by: Rob Clark <robdclark@gmail.com> (v1.5) Reviewed-by: Rob Clark <robdclark@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Acked-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Ben Widawsky [Thu, 22 Jan 2015 17:01:26 +0000 (17:01 +0000)]
drm/i915/gen8: Un-hardcode number of page directories
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Thu, 22 Jan 2015 17:01:25 +0000 (17:01 +0000)]
drm/i915: Setup less PPGTT on failed page_directory
The current code will both potentially print a WARN, and setup part of
the PPGTT structure. Neither of these harm the current code, it is
simply for clarity, and to perhaps prevent later bugs, or weird
debug messages.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Thu, 22 Jan 2015 17:01:24 +0000 (17:01 +0000)]
drm/i915: Rename to GEN8_LEGACY_PDPES
In gen8, 32b PPGTT has always had one "pdp" (it doesn't actually have
one, but it resembles having one). The #define was confusing as is, and
using "PDPE" is a much better description.
sed -i 's/GEN8_LEGACY_PDPS/GEN8_LEGACY_PDPES/' drivers/gpu/drm/i915/*.[ch]
It also matches the x86 pagetable terminology:
PTE = Page Table Entry - pagetable level 1 page
PDE = Page Directory Entry - pagetable level 2 page
PDPE = Page Directory Pointer Entry - pagetable level 3 page
And in the near future (for 48b addressing):
PML4E = Page Map Level 4 Entry
v2: Expanded information about Page Directory/Table nomenclature.
Cc: Daniel Vetter <daniel@ffwll.ch> CC: Dave Gordon <david.s.gordon@intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v2) Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Thu, 22 Jan 2015 17:01:23 +0000 (17:01 +0000)]
drm/i915/trace: Fix offsets for 64b
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
projects / linux-beck.git / log