Chris Wilson [Wed, 16 May 2018 18:33:48 +0000 (19:33 +0100)]
drm/i915: Remove tasklet flush before disable
The idea was to try and let the existing tasklet run to completion
before we began the reset, but it involves a racy check against anything
else that tries to run the tasklet. Rather than acknowledge and ignore
the race, let it be and don't try and be too clever.
The tasklet will resume execution after reset (after spinning a bit
during reset), but before we allow it to resume we will have cleared all
the pending state.
Chris Wilson [Tue, 15 May 2018 14:31:49 +0000 (15:31 +0100)]
drm/i915: Shrink search list for active timelines
When switching to the kernel context, we force the switch to occur after
all currently active requests (so that we know the GPU won't switch
immediately away and the kernel context remains current as we work). To
do so we have to inspect all the timelines and add a fence from the
active work to queue our switch afterwards. We can use the tracked set
of active rings to shrink our search for active timelines.
v2: Use a local to shrink the list_for_each_entry()
Chris Wilson [Sat, 12 May 2018 08:49:57 +0000 (09:49 +0100)]
drm/i915: Mark up nested spinlocks
When we process the outstanding requests upon banning a context, we need
to acquire both the engine and the client's timeline, nesting the locks.
This requires explicit markup as the two timelines are now of the same
class, since commit a89d1f921c15 ("drm/i915: Split i915_gem_timeline into
individual timelines").
Testcase: igt/gem_eio/banned Fixes: a89d1f921c15 ("drm/i915: Split i915_gem_timeline into individual timelines") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180512084957.9829-1-chris@chris-wilson.co.uk
Chris Wilson [Fri, 11 May 2018 12:11:47 +0000 (13:11 +0100)]
drm/i915/execlists: Relax CSB force-mmio for VT-d
The original switch to use CSB from the HWSP was plagued by the effect
of read ordering on VT-d; we would read the WRITE pointer from the HWSP
before it had completed writing the CSB contents. The mystery comes down
to the lack of rmb() for correct ordering with respect to the writes
from HW, and with that resolved we can remove the VT-d special casing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180511121147.31915-3-chris@chris-wilson.co.uk Tested-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Jani Nikula [Mon, 14 May 2018 08:18:06 +0000 (11:18 +0300)]
Merge tag 'gvt-next-2018-05-14' of https://github.com/intel/gvt-linux into drm-intel-next-queued
- Improve the emulation of virtual non-priv register. (Yan)
- Reverse the hack of host of preeption of GVT-g. (Weinan)
- Improve untracked warning message.(Changbin)
There are issues which will block the host preemption before, instead of
disabling it use one workaround "setting max priority for gvt context"
to avoid the gvt context be preempted by the host. Now the issues have been
cleared, so revert this patch to enable host preemption.
v2:
- refine description(Zhenyu)
Signed-off-by: Weinan Li <weinan.z.li@intel.com> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Zhao Yan [Tue, 8 May 2018 06:52:50 +0000 (14:52 +0800)]
drm/i915/gvt: let force_to_nonpriv cmd handler only valid for LRI cmd
the cmd_reg_handler() is called by cmds LRM, PIPE_CTRL, SRM...
for LRM, SRM, we cannot get write data in a simple way.
On other side, the force_to_nonpriv reigsters will only be written in LRI
in current drivers. so we don't want to bother the handler to handle those
memory access cmds, just leave a print message here.
Signed-off-by: Zhao Yan <yan.y.zhao@intel.com> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Zhao Yan [Tue, 8 May 2018 06:52:42 +0000 (14:52 +0800)]
drm/i915/gvt: do not return error on handling force_to_nonpriv registers
Return error will cause vm hang and enter failsafe mode.
However, we don't want that happen on detecting an wrong force_to_nonpriv
register write.
Therefore, we just omit the wrong write or patch it to default value.
v2: only return 0 on detecting lri write of registers outside whitelist,
but still return error on other error conditions. (zhenyu wang)
Signed-off-by: Zhao Yan <yan.y.zhao@intel.com> Reviewed-by: Zhang Yulei <yulei.zhang@intel.com> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Changbin Du [Thu, 19 Apr 2018 04:12:37 +0000 (12:12 +0800)]
drm/i915/gvt: Remove disable_warn_untrack and print untracked mmio with debug level
The disable_warn_untrack never prevent gvt from printing untracked
mmio errors. We were disturbed by this error storm and the fix is
just adding them to the list with no essential new change.
This message is only useful for enabling new platform during
developing process. So lower the message level to debug and then
remove disable_warn_untrack.
Signed-off-by: Changbin Du <changbin.du@intel.com> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Michel Thierry [Thu, 10 May 2018 20:07:08 +0000 (13:07 -0700)]
drm/i915/gen9: Add WaClearHIZ_WM_CHICKEN3 for bxt and glk
Factor in clear values wherever required while updating destination
min/max.
References: HSDES#1604444184 Signed-off-by: Michel Thierry <michel.thierry@intel.com> Cc: mesa-dev@lists.freedesktop.org Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Oscar Mateo <oscar.mateo@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20180510200708.18097-1-michel.thierry@intel.com Cc: stable@vger.kernel.org
Matthew Auld [Fri, 11 May 2018 09:51:40 +0000 (10:51 +0100)]
drm/i915/selftests: scrub 64K
We write all 4K page entries, even when using 64K pages. In order to
verify that the HW isn't cheating by using the 4K PTE instead of the 64K
PTE, we want to remove all the surplus entries. If the HW skipped the
64K PTE, it will read/write into the scratch page instead - which we
detect as missing results during selftests.
Chris Wilson [Fri, 11 May 2018 13:52:07 +0000 (14:52 +0100)]
drm/i915/oa: Check that OA is disabled before unpinning
Before we unpin the buffer used for OA reports and return it to the
system, we need to be sure that the HW has finished writing into it.
For lack of a better idea, poll OACONTROL to check it is switched off.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106379 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Tested-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180511135207.12880-1-chris@chris-wilson.co.uk
Chris Wilson [Fri, 11 May 2018 12:11:46 +0000 (13:11 +0100)]
Revert "drm/i915/cnl: Use mmio access to context status buffer"
In the previous patch (to include a rmb() after readig the CSB WRITE
pointer from the HWSP) we believe we have fixed the underlying bug, and
so can re-enable using the HWSP on Cannolake.
This reverts commit 61bf9719fa17 ("drm/i915/cnl: Use mmio access to
context status buffer").
References: https://bugs.freedesktop.org/show_bug.cgi?id=105888
References: https://bugs.freedesktop.org/show_bug.cgi?id=106185
References: 61bf9719fa17 ("drm/i915/cnl: Use mmio access to context status buffer") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Rafael Antognolli <rafael.antognolli@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Cc: Timo Aaltonen <tjaalton@ubuntu.com> Tested-by: Timo Aaltonen <tjaalton@ubuntu.com> Acked-by: Michel Thierry <michel.thierry@intel.com> Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180511121147.31915-2-chris@chris-wilson.co.uk
Chris Wilson [Fri, 11 May 2018 12:11:45 +0000 (13:11 +0100)]
drm/i915/execlists: Use rmb() to order CSB reads
We assume that the CSB is written using the normal ringbuffer
coherency protocols, as outlined in kernel/events/ring_buffer.c:
* (HW) (DRIVER)
*
* if (LOAD ->data_tail) { LOAD ->data_head
* (A) smp_rmb() (C)
* STORE $data LOAD $data
* smp_wmb() (B) smp_mb() (D)
* STORE ->data_head STORE ->data_tail
* }
So we assume that the HW fulfils its ordering requirements (B), and so
we should use a complimentary rmb (C) to ensure that our read of its
WRITE pointer is completed before we start accessing the data.
The final mb (D) is implied by the uncached mmio we perform to inform
the HW of our READ pointer.
References: https://bugs.freedesktop.org/show_bug.cgi?id=105064
References: https://bugs.freedesktop.org/show_bug.cgi?id=105888
References: https://bugs.freedesktop.org/show_bug.cgi?id=106185 Fixes: 767a983ab255 ("drm/i915/execlists: Read the context-status HEAD from the HWSP")
References: 61bf9719fa17 ("drm/i915/cnl: Use mmio access to context status buffer") Suggested-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Rafael Antognolli <rafael.antognolli@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Cc: Timo Aaltonen <tjaalton@ubuntu.com> Tested-by: Timo Aaltonen <tjaalton@ubuntu.com> Acked-by: Michel Thierry <michel.thierry@intel.com> Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180511121147.31915-1-chris@chris-wilson.co.uk
Oscar Mateo [Tue, 8 May 2018 21:29:35 +0000 (14:29 -0700)]
drm/i915/icl: WaForwardProgressSoftReset
Avoids a hang during soft reset.
v2: Rebased on top of the WA refactoring
v3: Added References (Mika)
v4:
- Rebased
- C, not lisp (Chris)
- Which steppings affected by this are not clear.
For the moment, apply unconditionally as per the
BSpec (Mika)
- Add reference to another HSD also related
Oscar Mateo [Tue, 8 May 2018 21:29:34 +0000 (14:29 -0700)]
drm/i915/icl: Wa_1406838659
Disable CGPSF unit clock gating to prevent an issue.
v2: Rebased on top of the WA refactoring
v3: Added References (Mika)
v4:
- Rebased
- C, not lisp (Chris)
- Remove unintentional whitespaces (Mika)
- Fixed in C0 (Mika)
Oscar Mateo [Tue, 8 May 2018 21:29:33 +0000 (14:29 -0700)]
drm/i915/icl: Wa_1604302699
Disable I2M Write for performance reasons.
v2: Rebased on top of the WA refactoring
v3: Added References (Mika)
v4:
- Rebased
- C, not lisp (Chris)
- GEN7 chicken bit in the wrong side of the fence (Mika)
- Use two spaces to align bit macros
Oscar Mateo [Tue, 8 May 2018 21:29:32 +0000 (14:29 -0700)]
drm/i915/icl: Wa_1406680159
Disable GWL clock gating to prevent an issue that might
cause hangs.
v2: Rebased on top of the WA refactoring
v3: Wa_2201832410 officially merged with Wa_1406680159
v4: Added References (Mika)
v5:
- Rebased
- C, not lisp (Chris)
- Add reference where WA is better explained (Rodrigo)
- Add reference to WA that got merged with this
Oscar Mateo [Tue, 8 May 2018 21:29:30 +0000 (14:29 -0700)]
drm/i915/icl: WaDisCtxReload
Revert to the legacy implementation to avoid a system hang.
v2: Correct the address for GAMW_ECO_DEV_RW_IA_REG
v3: Renamed to Wa_220166154
v4: Rebased on top of the WA refactoring
v5: Added References (Mika)
v6:
- Rebased
- C, not lisp (Chris)
Oscar Mateo [Tue, 8 May 2018 21:29:29 +0000 (14:29 -0700)]
drm/i915/icl: WaCL2SFHalfMaxAlloc
This workarounds an issue with insufficient storage for the
CL2 and SF units.
v2: Renamed to Wa_1405766107
v3: Wrapped the commit message
v4: Rebased on top of the WA refactoring
v5: Added References (Mika)
v6:
- Rebased
- s/MACALLOC/MAXALLOC (Mika)
- C, not lisp (Chris)
Oscar Mateo [Tue, 8 May 2018 21:29:28 +0000 (14:29 -0700)]
drm/i915/icl: WaDisableCleanEvicts
Avoids an undefined LLC behavior.
BSpec: 9613
v2: Renamed to Wa_1405733216
v3: Spaces around '<<' and fix surrounding code
v4: Rebased on top of the WA refactoring
v5: Added References (Mika)
v6:
- Rebased
- C, not lisp (Chris)
Oscar Mateo [Tue, 8 May 2018 21:29:27 +0000 (14:29 -0700)]
drm/i915/icl: WaModifyGamTlbPartitioning
Adjust default GAM TLB partitioning for performance reasons.
v2: Only touch the bits that we really need
v3: Rebased on top of the WA refactoring
v4:
- Added References (Mika)
- Rebased
v5:
- Rebased
- C, not lisp (Chris)
- Correct reference number (Mika)
Oscar Mateo [Tue, 8 May 2018 21:29:26 +0000 (14:29 -0700)]
drm/i915/icl: WaL3BankAddressHashing
Revert to an L3 non-hash model, for performance reasons.
v2:
- Place the WA name above the actual change
- Improve the register naming
v3:
- Rebased
- Renamed to Wa_1604223664
v4: Rebased on top of the WA refactoring
v5:
- Added References (Mika)
- Fixed wrong mask and value (Mika)
- Do not apply together with another WA for the same
register (not worth the hassle)
v6:
- Rebased
- C, not lisp (Chris)
Oscar Mateo [Tue, 8 May 2018 21:29:24 +0000 (14:29 -0700)]
drm/i915/icl: Enable Sampler DFR
Sampler Dynamic Frequency Rebalancing (DFR) aims to reduce Sampler
power by dynamically changing its clock frequency in low-throughput
conditions. This patches enables it by default on Gen11.
v2: Wrong operation to clear the bit (Praveen)
v3: Rebased on top of the WA refactoring
v4: Move to icl_init_clock_gating, since it's not a WA (Rodrigo)
v5: C, not lisp (Chris)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> Cc: Praveen Paneri <praveen.paneri@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/1525814984-20039-3-git-send-email-oscar.mateo@intel.com
Inherit workarounds from previous platforms that are still valid for
Icelake.
v2: GEN7_ROW_CHICKEN2 is masked
v3:
- Since it has been fixed already in upstream, removed the TODO
comment about WA_SET_BIT for WaInPlaceDecompressionHang.
- Squashed with this patch:
drm/i915/icl: add icelake_init_clock_gating()
from Paulo Zanoni <paulo.r.zanoni@intel.com>
- Squashed with this patch:
drm/i915/icl: WaForceEnableNonCoherent
from Oscar Mateo <oscar.mateo@intel.com>
- WaPushConstantDereferenceHoldDisable is now Wa_1604370585 and
applies to B0 as well.
- WaPipeControlBefore3DStateSamplePattern WABB was being applied
to ICL incorrectly.
v4:
- Wrap the commit message
- s/dev_priv/p to please checkpatch
v5: Rebased on top of the WA refactoring
v6: Rebased on top of further whitelist registers refactoring (Michel)
v7: Added WaRsForcewakeAddDelayForAck
v8: s/ICL_HDC_CHICKEN0/ICL_HDC_MODE (Mika)
v9:
- C, not lisp (Chris)
- WaIncreaseDefaultTLBEntries is the same for GEN > 9_LP (Tvrtko)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Tomasz Lis <tomasz.lis@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/1525814984-20039-2-git-send-email-oscar.mateo@intel.com
Mika Kuoppala [Tue, 8 May 2018 12:41:54 +0000 (15:41 +0300)]
drm/i915/gtt: Trust the uncached store to flush wcb
Not all architectures guarantee that uncached read will
flush the write combining buffer. So marking it explicitly
is recommended [1].
However we know the architecture we are operating on
and can avoid wmb as the UC store will flush the wcb [2].
Omit the wmb() before invalidate as redudant.
v2: squash combining and removal (Chris)
v3: remove obsolete comments about posting reads (Chris)
References: http://yarchive.net/comp/linux/write_combining.html [1]
References: http://download.intel.com/design/PentiumII/applnots/24442201.pdf [2] Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20180508124154.14586-1-mika.kuoppala@linux.intel.com
Chandra Konduru [Fri, 11 May 2018 21:33:17 +0000 (03:03 +0530)]
drm/i915: Add NV12 as supported format for sprite plane
This patch adds NV12 to list of supported formats for sprite plane.
v2: Rebased (me)
v3: Review comments by Ville addressed
- Removed skl_plane_formats_with_nv12 and added
NV12 case in existing skl_plane_formats
- Added the 10bpc RGB formats
v4: Addressed review comments from Clinton A Taylor
"Why are we adding 10 bit RGB formats with the NV12 series patches?
Trying to set XR30 or AB30 results in error returned even though
the modes are advertised for the planes"
- Removed 10bit RGB formats added previously with NV12 series
v5: Missed the Tested-by/Reviewed-by in the previous series
Adding the same to commit message in this version.
Addressed review comments from Clinton A Taylor
"Why are we adding 10 bit RGB formats with the NV12 series patches?
Trying to set XR30 or AB30 results in error returned even though
the modes are advertised for the planes"
- Previous version has 10bit RGB format removed from VLV formats
by mistake. Fixing that in this version.
Removed 10bit RGB formats added previously with NV12 series
for SKL.
v6: Addressed review comments by Ville
Restricting the NV12 to BXT and PIPE A and B
v7: Rebased (me)
v8: Rebased (me)
Restricting NV12 changes to BXT and KBL
Restricting NV12 changes for plane 0 (overlay)
v9: Rebased (me)
v10: Addressed review comments from Maarten.
Adding NV12 to skl_plane_formats itself.
v11: Addressed review comments from Shashank Sharma
v12: Addressed review comments from Shashank Sharma
Made the condition in intel_sprite_plane_create
simple and easy to read as suggested.
v13: Adding reviewed by tag from Shashank Sharma
Addressed review comments from Juha-Pekka Heikkila
"NV12 not to be supported by SKL"
v14: Addressed review comments from Ville
Added skl_planar_formats to include NV12
and a check skl_plane_has_planar in sprite create
Added NV12 format to skl_mod_supported. These were
review comments from Kristian Høgsberg <hoegsberg@gmail.com>
v15: Added reviewed by from Juha-Pekka Heikkila
v16: Rebased the series
v17: Added all tiling under mod supported for NV12
Credits to Megha Aggarwal
v18: Added RB by Maarten and Kristian
v19: Addressed review comments from Maarten
Made modification to skl_mod_supported
Credits-to: Megha Aggarwal <megha.aggarwal@intel.com> Credits-to: Kristian Høgsberg <hoegsberg@gmail.com> Reviewed-by: Kristian Høgsberg <hoegsberg@gmail.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Tested-by: Clinton Taylor <clinton.a.taylor@intel.com> Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com> Reviewed-by: Shashank Sharma <shashank.sharma@intel.com> Reviewed-by: Clinton Taylor <clinton.a.taylor@intel.com> Signed-off-by: Chandra Konduru <chandra.konduru@intel.com> Signed-off-by: Nabendu Maiti <nabendu.bikash.maiti@intel.com> Signed-off-by: Vidya Srinivas <vidya.srinivas@intel.com> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/1526074397-10457-7-git-send-email-vidya.srinivas@intel.com
v12: Addressed review comments from Juha-Pekka Heikkila
"NV12 not to be supported by SKL"
v13: Addressed review comments from Ville
Added skl_pri_planar_formats to include NV12
and skl_plane_has_planar function to check for
NV12 support on plane. Added NV12 format to
skl_mod_supported. These were review comments
from Kristian Høgsberg <hoegsberg@gmail.com>
v14: Added reviewed by from Juha-Pekka Heikkila
v15: Rebased the series
v16: Added all tiling support under mod supported
for NV12. Credits to Megha Aggarwal
v17: Added RB by Maarten and Kristian
v18: Review comments from Maarten addressed -
Removing BROXTON support for NV12 due to WA826
v19: Addressed review comments from Maarten
Make changes to skl_mod_supported
Credits-to: Megha Aggarwal megha.aggarwal@intel.com Credits-to: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Tested-by: Clinton Taylor <clinton.a.taylor@intel.com> Reviewed-by: Kristian Høgsberg <hoegsberg@gmail.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com> Reviewed-by: Clinton Taylor <clinton.a.taylor@intel.com> Reviewed-by: Shashank Sharma <shashank.sharma@intel.com> Signed-off-by: Chandra Konduru <chandra.konduru@intel.com> Signed-off-by: Nabendu Maiti <nabendu.bikash.maiti@intel.com> Signed-off-by: Vidya Srinivas <vidya.srinivas@intel.com> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/1526074397-10457-6-git-send-email-vidya.srinivas@intel.com
Chandra Konduru [Fri, 11 May 2018 21:33:15 +0000 (03:03 +0530)]
drm/i915: Add NV12 support to intel_framebuffer_init
This patch adds NV12 as supported format
to intel_framebuffer_init and performs various checks.
v2:
-Fix an issue in checks added (Chandra Konduru)
v3: rebased (me)
v4: Review comments by Ville addressed
Added platform check for NV12 in intel_framebuffer_init
Removed offset checks for NV12 case
v5: Addressed review comments by Clinton A Taylor
This NV12 support only correctly works on SKL.
Plane color space conversion is different on GLK and later platforms
causing the colors to display incorrectly.
Ville's plane color space property patch series
in review will fix this issue.
- Restricted the NV12 case in intel_framebuffer_init to
SKL and BXT only.
v6: Rebased (me)
v7: Addressed review comments by Ville
Restricting the NV12 to BXT for now.
v8: Rebased (me)
Restricting the NV12 changes to BXT and KBL for now.
v9: Rebased (me)
v10: NV12 supported by all GEN >= 9.
Making this change in intel_framebuffer_init. This is
part of addressing Maarten's review comments.
Comment under v8 no longer applicable
v11: Addressed review comments from Shashank Sharma
v12: Adding Reviewed By from Shashank Sharma
v13: Addressed review comments from Juha-Pekka Heikkila
"NV12 not to be supported by SKL"
v14: Addressed review comments from Maarten.
Add checks for fb width height for NV12 and fail the fb
creation if check fails. Added reviewed by from
Juha-Pekka Heikkila
v15: Rebased the series
v16: Setting the minimum value during fb creating to 16
as per Bspec for NV12. Earlier minimum was expected
to be > 16. Now changed it to >=16.
v17: Adding restriction to framebuffer_init - the fb
width and height should be a multiplier of 4
v18: Added RB from Maarten. Included Maarten's review comments
Dont allow CCS formats for fb creation of NV12
v19: Review comments from Maarten addressed -
Removing BROXTON support for NV12 due to WA826
We skip src trunction/adjustments for
NV12 case and handle the sizes directly.
Without this, pipe fifo underruns are seen on APL/KBL.
v2: For NV12, making the src coordinates multiplier of 4
v3: Moving all the src coords handling code for NV12
to skl_check_nv12_surface
v4: Added RB from Mika
v5: Rebased the series. Removed checks of mult of 4 in
skl_update_scaler, Added NV12 condition in intel_check_sprite_plane
where src x/w is being checked for mult of 2 for yuv planes.
v6: Made changes to skl_check_nv12_surface as per WA#1106
Vidya Srinivas [Fri, 11 May 2018 21:33:13 +0000 (03:03 +0530)]
drm/i915: Enable Display WA 0528
Possible hang with NV12 plane surface formats.
WA: When the plane source pixel format is NV12,
the CHICKEN_PIPESL_* register bit 22 must be set to 1
and the render decompression must not be enabled
on any of the planes in that pipe.
drm/i915: Enable display workaround 827 for all planes, v2.
The workaround was applied only to the primary plane, but is required
on all planes. Iterate over all planes in the crtc atomic check to see
if the workaround is enabled, and only perform the actual toggling in
the pre/post plane update functions.
Changes since v1:
- Track active NV12 planes in a nv12_planes bitmask. (Ville)
v2: Removing BROXTON support for NV12 due to WA826
drm/i915/psr: Check if VBT says PSR can be enabled.
Driver features data block has a boolean flag for PSR, use this to decide
whether PSR should be enabled on a platform. The module parameter can
still be used to override this.
Note: The feature currently remains disabled by default for all platforms
irrespective of what VBT says.
Chris Wilson [Tue, 8 May 2018 21:03:18 +0000 (22:03 +0100)]
drm/i915/guc: Make submission tasklet hardirq safe
Prepare to allow the GuC submission to be run from underneath a
hardirq timer context (and not just the current softirq context) as is
required for fast preemption resets and context switches.
Chris Wilson [Tue, 8 May 2018 21:03:17 +0000 (22:03 +0100)]
drm/i915/execlists: Make submission tasklet hardirq safe
Prepare to allow the execlists submission to be run from underneath a
hardirq timer context (and not just the current softirq context) as is
required for fast preemption resets and context switches.
Chris Wilson [Tue, 8 May 2018 21:10:56 +0000 (22:10 +0100)]
drm/i915/selftests: Create mock_engine() under struct_mutex
Calling mock_engine() calls i915_timeline_init() and that requires
struct_mutex to be held as it adds itself to the global list of
timelines. This error was introduced by commit a89d1f921c15 ("drm/i915:
Split i915_gem_timeline into individual timelines") but the issue was
masked in CI by the earlier lockdep spam.
Fixes: a89d1f921c15 ("drm/i915: Split i915_gem_timeline into individual timelines") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Michel Thierry <michel.thierry@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180508211056.17151-1-chris@chris-wilson.co.uk
but did not report it as an issue as it only occurred during the first
module on boot. This is due to the removal of the distinct global
timeline, and its separate lock class. So instead mark up the expected
nesting. An alternative would be to define a separate lock class for the
engine, but since we only expect to have a single point of nesting, we
can avoid having multiple lock classes for the struct.
Fixes: a89d1f921c15 ("drm/i915: Split i915_gem_timeline into individual timelines") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Tested-by: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180508153514.20251-1-chris@chris-wilson.co.uk
Chris Wilson [Tue, 8 May 2018 15:15:52 +0000 (16:15 +0100)]
drm/i915: Remove unused i915_flip tracepoints
The i915_flip* tracepoints are no longer in use since the removal of CS
flip in commit 8b5d27b911d7 ("drm/i915: Remove intel_flip_work
infrastructure")
References: 8b5d27b911d7 ("drm/i915: Remove intel_flip_work infrastructure") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180508151552.31024-1-chris@chris-wilson.co.uk
Chris Wilson [Mon, 7 May 2018 13:57:26 +0000 (14:57 +0100)]
drm/i915: Disable tasklet scheduling across initial scheduling
During request submission, we call the engine->schedule() function so
that we may reorder the active requests as required for inheriting the
new request's priority. This may schedule several tasklets to run on the
local CPU, but we will need to schedule the tasklets again for the new
request. Delay all the local tasklets until the end, so that we only
have to process the queue just once.
v2: Beware PREEMPT_RCU, as then local_bh_disable() is then not a
superset of rcu_read_lock().
Chris Wilson [Mon, 7 May 2018 13:57:25 +0000 (14:57 +0100)]
drm/i915: Flush submission tasklet after bumping priority
When called from process context tasklet_schedule() defers itself to
ksoftirqd. From experience this may cause unacceptable latencies of over
200ms in executing the submission tasklet, our goal is to reprioritise
the HW execution queue and trigger HW preemption immediately, so disable
bh over the call to schedule and force the tasklet to run afterwards if
scheduled.
Chris Wilson [Tue, 8 May 2018 11:53:12 +0000 (12:53 +0100)]
drm/i915/selftests: Return to kernel context after each test
As we flush each test and wait for idle before the next, also switch
back to the kernel context. This helps limit the amount of collateral
damage a test may cause by resetting to the default state each time (and
also helps clean up temporaries used by the test).
Matthew Auld [Thu, 1 Mar 2018 11:46:39 +0000 (11:46 +0000)]
drm/i915: don't leak the pin_map on error
Add some onion to populate_lr_context.
v2: prefer err_unpin_ctx
drop the fixes tag, worst case we just spew a warn before everything
is cleaned up and balance is restored
Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20180301114639.510-1-matthew.auld@intel.com
Chris Wilson [Sat, 5 May 2018 09:10:14 +0000 (10:10 +0100)]
drm/i915/selftests: Flush GPU activity before completing live_contexts
igt_ctx_exec() expects that we retire all active requests/objects before
completing, so that when we clean up the files afterwards they are ready
to be freed. Before we do so, it is then prudent to ensure that we have
indeed retired the GPU activity, raising an error if it fails. If we do
not, we run the risk of triggering an assertion when freeing the object:
Matthew Auld [Wed, 2 May 2018 19:50:21 +0000 (20:50 +0100)]
drm/i915/userptr: reject zero user_size
Operating on a zero sized GEM userptr object will lead to explosions.
Fixes: 5cc9ed4b9a7a ("drm/i915: Introduce mapping of user pages into video memory (userptr) ioctl")
Testcase: igt/gem_userptr_blits/input-checking Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20180502195021.30900-1-matthew.auld@intel.com
Chris Wilson [Tue, 8 May 2018 00:30:46 +0000 (01:30 +0100)]
drm/i915/execlists: Cache the priolist when rescheduling
When rescheduling a change of dependencies, they all need to be added to
the same priolist (at least the ones on the same engine!). Since we
likely want to move a batch of requests, keep the priolist around.
v2: Throw in an assert to catch trivial errors quickly.
Chris Wilson [Tue, 8 May 2018 00:30:45 +0000 (01:30 +0100)]
drm/i915/execlists: Drop unused parameter to lookup_priolist()
lookup_priolist() no longer attaches the request into the priolist, it
just returns the priolist for the given priority instead. Drop the
unused parameter.
Chris Wilson [Sun, 6 May 2018 18:31:47 +0000 (19:31 +0100)]
drm/i915: Don't request a bug report for unsafe module parameters
Unsafe module parameters are just that, unsafe. If the user is foolish
enough to try them and the kernel breaks, they get to keep both pieces.
Don't ask them to file a bug report if they broke it themselves.
References: https://bugs.freedesktop.org/show_bug.cgi?id=106423 Fixes: d15d7538c6d2 ("drm/i915: Tune down init error message due to failure injection") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Acked-by: Jani Nikula <jani.nikula@intel.com> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180506183147.2690-1-chris@chris-wilson.co.uk
Paulo Zanoni [Wed, 28 Mar 2018 21:58:01 +0000 (14:58 -0700)]
drm/i915/icl: compute the MG PLL registers
This implements the "MG PLL Programming" sequence from our spec. The
biggest problem was that the spec assumes real numbers, so we had to
adjust some numbers and calculations due to the fact that the Kernel
prefers to deal with integers.
I recommend grabbing some coffee, a pen and paper before reviewing
this patch.
v2:
- Correctly identify DP encoders after upstream change.
- Small checkpatch issues.
- Rebase.
v3:
- Try to impove the comment on the tdc_targetcnt calculation based on
Manasi's feedback (Manasi).
- Rebase.
Paulo Zanoni [Wed, 28 Mar 2018 21:57:59 +0000 (14:57 -0700)]
drm/i915/icl: compute the combo PHY (DPLL) HDMI registers
HDMI mode DPLL programming on ICL is the same as CNL, so just reuse
the CNL code.
v2:
- Properly detect HDMI crtcs.
- Rebase after changes to the cnl function (clock * 1000).
v3:
- Add a comment to clarify why we treat 38.4 as 19.2 (James).
Paulo Zanoni [Fri, 27 Apr 2018 23:14:36 +0000 (16:14 -0700)]
drm/i915/icl: add basic support for the ICL clocks
This commit introduces the definitions for the ICL clocks and adds the
basic functions to the shared DPLL framework. It adds code for the
Enable and Disable sequences for some PLLs, but it does not have the
code to compute the actual PLL values, which are marked as TODO
comments and should be introduced as separate commits.
Special thanks to James Ausmus for investigating and fixing a bug with
the placement of icl_unmap_plls_to_ports() function.
v2:
- Rebase around dpll_lock changes.
v3:
- The spec now says what the timeouts should be.
- Touch DPCLKA_CFGCR0_ICL at the appropriate time so we don't freeze
the machine.
- Checkpatch found a white space problem.
- Small adjustments before upstreaming.
v4:
- Move the ICL checks out of the *map_plls_to_ports() functions
(James)
- Add extra encoder check (James)
- Call icl_unmap_plls_to_ports() later (James)
v5:
- Rebase after the pll struct changes.
v6:
- Properly make the unmap function based on encoders_post_disable()
with regarding to checks and iterators.
- Address checkpatch comment on "min = max = x()".
drm/i915: Fix drm:intel_enable_lvds ERROR message in kernel log
Fix `[drm:intel_enable_lvds] *ERROR* timed out waiting for panel to
power on` in kernel log at boot time.
Toshiba Satellite Z930 laptops needs between 1 and 2 seconds to power
on its screen during Intel i915 DRM initialization. This currently
results in a `[drm:intel_enable_lvds] *ERROR* timed out waiting for
panel to power on` message appearing in the kernel log during boot
time and when stopping the machine.
This change increases the timeout of the `intel_enable_lvds` function
from 1 to 5 seconds, letting enough time for the Satellite 930 LCD
screen to power on, and suppressing the error message from the kernel
log.
This patch has been successfully tested on Linux 4.14 running on a
Toshiba Satellite Z930.
[vsyrjala: bump the timeout from 2 to 5 seconds to match the DP
code and properly cover the max hw timeout of ~4 seconds, and
drop the comment about the specific machine since this is not
a particulary surprising issue, nor specific to that one machine]
Signed-off-by: Florent Flament <contact@florentflament.com> Cc: stable@vger.kernel.org Cc: Pavel Petrovic <ppetrovic@acm.org> Cc: Sérgio M. Basto <sergio@serjux.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103414
References: https://bugzilla.kernel.org/show_bug.cgi?id=57591 Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180419160700.19828-1-ville.syrjala@linux.intel.com Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Chris Wilson [Fri, 4 May 2018 10:11:47 +0000 (11:11 +0100)]
drm/i915: Remove assertion of active_rings must be non-empty if active_requests
"An outstanding request must still be on an active ring somewhere" is
only true if we haven't just been interrupted by the shrinker in the
middle of allocating the request itself. (At the start of
i915_request_alloc() we pin the context and prepare the GT for activity,
marking it as active, and then try to allocate the request. If this
allocation invokes the shrinker, we try to reclaim some space by calling
i915_retire_requests() which may then be confused by the pre-reservation
of active_requests.)
Chris Wilson [Thu, 3 May 2018 21:29:56 +0000 (22:29 +0100)]
drm/i915/gtt: Tidy up duplicate branches in gen8_gmch_probe()
Following commit f773568b6ff8 ("drm/i915: nuke the duplicated stolen
discovery"), the if-else-chain for determining the GTT size is redundant
with the !chv branches all being the same.
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
References: f773568b6ff8 ("drm/i915: nuke the duplicated stolen discovery") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180503212956.3948-1-chris@chris-wilson.co.uk
Chris Wilson [Thu, 3 May 2018 19:54:16 +0000 (20:54 +0100)]
drm/i915/execlists: Drop preemption arbitrations points along the ring
Limit the arbitration (where preemption may occur) to inside the batch,
and prevent it from happening on the pipecontrols/flushes we use to
write the breadcrumb seqno. Once the user batch is complete, we have
nothing left to do but serialise and emit the breadcrumb; switching
contexts at this point is futile so don't.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180503195416.22498-1-chris@chris-wilson.co.uk
Chris Wilson [Thu, 3 May 2018 19:51:15 +0000 (20:51 +0100)]
drm/i915: Keep one request in our ring_list
Don't pre-emptively retire the oldest request in our ring's list if it
is the only request. We keep various bits of state alive using the
active reference from the request and would rather transfer that state
over to a new request rather than the more involved process of retiring
and reacquiring it.
Chris Wilson [Thu, 3 May 2018 19:51:14 +0000 (20:51 +0100)]
drm/i915: Lazily unbind vma on close
When userspace is passing around swapbuffers using DRI, we frequently
have to open and close the same object in the foreign address space.
This shows itself as the same object being rebound at roughly 30fps
(with a second object also being rebound at 30fps), which involves us
having to rewrite the page tables and maintain the drm_mm range manager
every time.
However, since the object still exists and it is only the local handle
that disappears, if we are lazy and do not unbind the VMA immediately
when the local user closes the object but defer it until the GPU is
idle, then we can reuse the same VMA binding. We still have to be
careful to mark the handle and lookup tables as closed to maintain the
uABI, just allowing the underlying VMA to be resurrected if the user is
able to access the same object from the same context again.
If the object itself is destroyed (neither userspace keeping a handle to
it), the VMA will be reaped immediately as usual.
In the future, this will be even more useful as instantiating a new VMA
for use on the GPU will become heavier. A nuisance indeed, so nip it in
the bud.
v2: s/__i915_vma_final_close/i915_vma_destroy/ etc.
v3: Leave a hint as to why we deferred the unbind on close.
Tomasz Lis [Wed, 2 May 2018 22:31:42 +0000 (15:31 -0700)]
drm/i915/icl: Add configuring MOCS in new Icelake engines
In Icelake, there are more engines on which Memory Object Control
States need to be configured. Besides adding Icelake under Skylake
config, the patch makes sure MOCS register addresses for the new
engines are properly defined.
Additional patch might be need later, in case the specification will
propose different MOCS config values for Icelake than in previous
gens.
v2: Restricted comments to gen11, updated description, renamed
defines.
Ville Syrjälä [Thu, 26 Apr 2018 16:30:15 +0000 (19:30 +0300)]
drm/i915: Correctly populate user mode h/vdisplay with pipe src size during readout
During state readout we first read out the pipe src size, store
that information in the user mode h/vdisplay, but later on we overwrite
that with the actual crtc timings. That makes our read out crtc state
inconsistent with itself when the BIOS has enabled the panel fitter to
scale the pipe contents. Let's preserve the pipe src size based
information in the user mode to make things consistent again.
This fixes a problem introduced by commit a2936e3d9a9c ("drm/i915:
Use drm_mode_get_hv_timing() to populate plane clip rectangle")
where the inconsistent state is now leading the plane clipping code
to report a failure on account the plane dst coordinates not matching
the user mode size. Previously we did the plane clipping based on
the pipe src size instead and thus never noticed the inconsistency.
The failure manifests as a WARN:
[ 0.762117] [drm:intel_dump_pipe_config [i915]] requested mode:
[ 0.762142] [drm:drm_mode_debug_printmodeline [drm]] Modeline 0:"1366x768" 60 72143 1366 1414 1446 1526 768 771 777 784 0x40 0xa
...
[ 0.762327] [drm:intel_dump_pipe_config [i915]] port clock: 72143, pipe src size: 1024x768, pixel rate 72143
...
[ 0.764666] [drm:drm_atomic_helper_check_plane_state [drm_kms_helper]] Plane must cover entire CRTC
[ 0.764690] [drm:drm_rect_debug_print [drm]] dst: 1024x768+0+0
[ 0.764711] [drm:drm_rect_debug_print [drm]] clip: 1366x768+0+0
[ 0.764713] ------------[ cut here ]------------
[ 0.764714] Could not determine valid watermarks for inherited state
[ 0.764792] WARNING: CPU: 4 PID: 159 at drivers/gpu/drm/i915/intel_display.c:14584 intel_modeset_init+0x3ce/0x19d0 [i915]
...
Cc: FadeMind <fademind@gmail.com> Cc: Dave Jones <davej@codemonkey.org.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reported-by: FadeMind <fademind@gmail.com> Reported-by: Dave Jones <davej@codemonkey.org.uk> Tested-by: Dave Jones <davej@codemonkey.org.uk>
References: https://lists.freedesktop.org/archives/intel-gfx/2018-April/163186.html
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105992 Fixes: a2936e3d9a9c ("drm/i915: Use drm_mode_get_hv_timing() to populate plane clip rectangle") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180426163015.14232-1-ville.syrjala@linux.intel.com Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Tested-by: FadeMind <fademind@gmail.com>
Tarun [Wed, 2 May 2018 23:33:00 +0000 (16:33 -0700)]
drm/i915: Remove redundant check for negative timeout while doing an atomic pipe update
No functional changes, just a minor knit. Stumbled across the kernel doc for
schedule_timeout() which quotes "In all cases the return value is guaranteed
to be non-negative". Also, the return code of schedule_timeout() already checks
for negative values "return timeout < 0 ? 0 : timeout;" and returns 0
in such cases. Furthermore, the msec_to_jiffies returns an ungined long
value. So, let's do away with the redundant check for an atomic
pipe update.
Rodrigo Vivi [Wed, 2 May 2018 17:52:55 +0000 (10:52 -0700)]
drm/i915: Adjust eDP's logical vco in a reliable place.
On intel_dp_compute_config() we were calculating the needed vco
for eDP on gen9 and we stashing it in
intel_atomic_state.cdclk.logical.vco
However few moments later on intel_modeset_checks() we fully
replace entire intel_atomic_state.cdclk.logical with
dev_priv->cdclk.logical fully overwriting the logical desired
vco for eDP on gen9.
So, with wrong VCO value we end up with wrong desired cdclk, but
also it will raise a lot of WARNs: On gen9, when we read
CDCLK_CTL to verify if we configured properly the desired
frequency the CD Frequency Select bits [27:26] == 10b can mean
337.5 or 308.57 MHz depending on the VCO. So if we have wrong
VCO value stashed we will believe the frequency selection didn't
stick and start to raise WARNs of cdclk mismatch.
v2: Move the entire eDP's vco logical adjustment to inside
the skl_modeset_calc_cdclk as suggested by Ville.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Fixes: bb0f4aab0e76 ("drm/i915: Track full cdclk state for the logical and actual cdclk frequencies") Cc: <stable@vger.kernel.org> # v4.12+ Link: https://patchwork.freedesktop.org/patch/msgid/20180502175255.5344-1-rodrigo.vivi@intel.com
Chris Wilson [Wed, 2 May 2018 22:03:13 +0000 (23:03 +0100)]
drm/i915: Mark the hangcheck as idle when unparking the engines
As we unpark the engines and are about to begin a new cycle of activity,
mark the current status of the hangceck as idle so that we avoid
carrying over a stale timestamp/action into the next cycle.
Chris Wilson [Wed, 2 May 2018 22:03:12 +0000 (23:03 +0100)]
drm/i915: Reset the hangcheck timestamp before repeating a seqno
In the unusual circumstance where we reuse a seqno (for example, in
igt), make sure that we reset the hangcheck timestamp before it sees the
same seqno again.
Chris Wilson [Mon, 9 Apr 2018 09:49:05 +0000 (10:49 +0100)]
drm/i915: Silence debugging DRM_ERROR for failing to suspend vlv powerwells
If we try to suspend a wedged device following a GPU reset failure, we
will also fail to turn off the rc6 powerwells (on vlv), leading to a
*ERROR*. This is quite expected in this case, so the best we can do is
shake our heads and reduce the *ERROR* to a debug so CI stops
complaining.
Testcase: igt/gem_eio/in-flight-suspend #vlv
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105583 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180409094905.4516-1-chris@chris-wilson.co.uk
Chris Wilson [Wed, 2 May 2018 23:02:02 +0000 (00:02 +0100)]
drm/i915/execlists: Emit i915_trace_request_out for preemption
Move the tracepoint into the common execlists_context_schedule_out() and
call it from preemption completion as well. A small bit of refactoring
code should help with when tracing, or else we end up with requests
mysteriously disappearing and some being emitted to HW multiple times.
Chris Wilson [Wed, 2 May 2018 16:38:39 +0000 (17:38 +0100)]
drm/i915: Split i915_gem_timeline into individual timelines
We need to move to a more flexible timeline that doesn't assume one
fence context per engine, and so allow for a single timeline to be used
across a combination of engines. This means that preallocating a fence
context per engine is now a hindrance, and so we want to introduce the
singular timeline. From the code perspective, this has the notable
advantage of clearing up a lot of mirky semantics and some clumsy
pointer chasing.
By splitting the timeline up into a single entity rather than an array
of per-engine timelines, we can realise the goal of the previous patch
of tracking the timeline alongside the ring.
v2: Tweak wait_for_idle to stop the compiling thinking that ret may be
uninitialised.
Chris Wilson [Wed, 2 May 2018 16:38:38 +0000 (17:38 +0100)]
drm/i915: Move timeline from GTT to ring
In the future, we want to move a request between engines. To achieve
this, we first realise that we have two timelines in effect here. The
first runs through the GTT is required for ordering vma access, which is
tracked currently by engine. The second is implied by sequential
execution of commands inside the ringbuffer. This timeline is one that
maps to userspace's expectations when submitting requests (i.e. given the
same context, batch A is executed before batch B). As the rings's
timelines map to userspace and the GTT timeline an implementation
detail, move the timeline from the GTT into the ring itself (per-context
in logical-ring-contexts/execlists, or a global per-engine timeline for
the shared ringbuffers in legacy submission.
The two timelines are still assumed to be equivalent at the moment (no
migrating requests between engines yet) and so we can simply move from
one to the other without adding extra ordering.
v2: Reinforce that one isn't allowed to mix the engine execution
timeline with the client timeline from userspace (on the ring).
Replace 01.org URL with upstream linux-firmware repo URL.
We no longer release firmware to 01.org.
linux-firmware.git is the ultimate place to find
the i915 firmwares.
Chris Wilson [Tue, 1 May 2018 07:52:03 +0000 (08:52 +0100)]
drm/i915/guc: Assert we have the doorbell before setting it up
As our early doorbell is split between early allocation and a late setup
after we have a channel to the GuC, it may happen due to a lapse of
programmer judgement that we try to setup an invalid doorbell. Make use
of our has_doorbell() function to check the doorbell does exist for the
client before we try and tell the guc about it. In doing so, we prevent
the compiler from warning about the otherwise unused function in some
configurations.
Reported-by: Matthias Kaehlcke <mka@chromium.org> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Reviewed-by: Michel Thierry <michel.thierry@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180501075203.12458-1-chris@chris-wilson.co.uk
Commit 39bf4de89ff7 ("drm/i915: Add -Wall -Wextra to our build, set
warnings to full") enabled extra warnings for i915 to spot possible
bugs in new code, and then disabled a subset of these warnings to keep
the current code building without warnings (with gcc). Enabling the
extra warnings also enabled some additional clang-only warnings, as a
result building i915 with clang currently is extremely noisy. For now
also disable the clang warnings sign-compare, sometimes-uninitialized,
unneeded-internal-declaration and initializer-overrides. If desired
they can be re-enabled after the code has been fixed.
Chris Wilson [Wed, 2 May 2018 10:41:50 +0000 (11:41 +0100)]
drm/i915: Show ring->start for the ELSP context/request queue
Since the advent of execlists, the HW no longer executes from a single
statically assigned ring, but instead switches to a different ring for
each context (logical ringbuffer contexts as it is called). So a good way
to tally the executing context against what we have queued is by
comparing the RING_START register against our requests. Make it so.
Chris Wilson [Tue, 17 Apr 2018 17:06:38 +0000 (18:06 +0100)]
drm/i915/selftests: Fix error checking for wait_var_timeout
The old wait_on_atomic_t used a custom callback to perform the
schedule(), which used my return semantics of reporting an error code on
timeout. wait_var_event_timeout() uses the schedule() return semantics
of reporting the remaining jiffies (1 if it timed out with 0 jiffies
remaining!) and 0 on failure. This semantic mismatch lead to us falsely
claiming a time out occurred.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106085 Fixes: d224985a5e31 ("sched/wait, drivers/drm: Convert wait_on_atomic_t() usage to the new wait_var_event() API") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180417170638.20550-1-chris@chris-wilson.co.uk
Jani Nikula [Wed, 2 May 2018 09:20:32 +0000 (12:20 +0300)]
Merge drm/drm-next into drm-intel-next-queued
Need d224985a5e31 ("sched/wait, drivers/drm: Convert wait_on_atomic_t()
usage to the new wait_var_event() API") in dinq to be able to fix
https://bugs.freedesktop.org/show_bug.cgi?id=106085.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Mika Kuoppala [Mon, 30 Apr 2018 07:52:59 +0000 (10:52 +0300)]
drm/i915: Print error state times relative to capture
Using plain jiffies in error state output makes the output
time differences relative to the current system time. This
is wrong as it makes output time differences dependent
of when the error state is printed rather than when it is
captured.
Store capture jiffies into error state and use it
when outputting the state to fix time differences output.
v2: use engine timestamp as epoch, output formatting (Chris)
v3: pass epoch to print_engine/request (Chris)
Chris Wilson [Tue, 1 May 2018 12:21:31 +0000 (13:21 +0100)]
drm/i915/execlists: Don't trigger preemption if complete
Due to the latency of the tasklet running from ksoftirqd, by the time we
process the execlist dequeue may be a long time behind the GPU. If the
request was completed when we ran reschedule, we will not have tweaked
its priority, but if it is still listed as being in-flight for dequeue
we will use it as a reference for the rest of the queue, including
requests from its own context which will now be at higher priority. This
can cause us to issue a preempt-to-idle request, even though the request
we want to preempt is already complete.
Manasi Navare [Wed, 28 Mar 2018 21:58:03 +0000 (14:58 -0700)]
drm/i915/icl: Fix the DP Max Voltage for ICL
On clock recovery this function is called to find out
the max voltage swing level that we could go.
However gen 9 functions use the old buffer translation tables
to figure that out. ICL uses different set of tables for eDP
and DP for both Combo and MG PHY ports. This patch adds the hook
for ICL for getting this information from appropriate buf trans tables.
v5 (from Paulo):
* New rebase after changes to earlier patches.
v4:
* Rebase.
v3:
* Follow the coding conventions here
(https://cgit.freedesktop.org/drm-intel/tree/Documentation/process/codin
g-style.rst#n191) (Paulo)
v2:
* Rebase after patch that adds voltage check inside buf trans
function (Rodrigo)
Manasi Navare [Wed, 28 Mar 2018 21:58:02 +0000 (14:58 -0700)]
drm/i915/icl: Implement voltage swing programming sequence for Combo PHY DDI
This is an important part of the DDI initalization as well as
for changing the voltage during DisplayPort link training.
The Voltage swing seqeuence is similar to Cannonlake.
However it has different register definitions and hence
it makes sense to create a separate vswing sequence and
program functions for ICL to leave room for more changes
in case the Bspec changes later and deviates from CNL sequence.
v2:
Use ~TAP3_DISABLE for enbaling that bit (Jani Nikula)
v3:
* Use dw4_scaling column for PORT_TX_DW4 values (Rodrigo)
v4:
* Call it combo_vswing, use switch statement (Paulo)
v5 (from Paulo):
* Fix a typo.
* s/rate < 600000/rate <= 600000/.
* Don't remove blank lines that should be there.
v6:
* Rebased by Rodrigo on top of Cannonlake changes
where non vswing sequences are not aligned with iboost
anymore.
v7: Another rebase after an upstream rework.
v8 (from Paulo):
* Adjust the code to the upstream output type changes.
* Squash the patch that moved some functions up.
* Merge both get_combo_buf_trans functions in order to simplify the
code.
* Change the changelog format.
v9 (from Paulo):
* Use RTERM_SELECT instead of SCALING_MODE_SEL.
* Adjust the output type handling according to how the other platforms
do it now.
v10 (from Paulo):
* Fix comment left out from v9 changes (Rodrigo).
Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: James Ausmus <james.ausmus@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180328215803.13835-8-paulo.r.zanoni@intel.com
Chris Wilson [Mon, 30 Apr 2018 13:15:03 +0000 (14:15 +0100)]
drm/i915: Only track live rings for retiring
We don't need to track every ring for its lifetime as they are managed
by the contexts/engines. What we do want to track are the live rings so
that we can sporadically clean up requests if userspace falls behind. We
can simply restrict the gt->rings list to being only gt->live_rings.
v2: s/live/active/ for consistency with gt.active_requests
Chris Wilson [Mon, 30 Apr 2018 13:15:02 +0000 (14:15 +0100)]
drm/i915: Retire requests along rings
In the next patch, rings are the central timeline as requests may jump
between engines. Therefore in the future as we retire in order along the
engine timeline, we may retire out-of-order within a ring (as the ring now
occurs along multiple engines), leading to much hilarity in miscomputing
the position of ring->head.
As an added bonus, retiring along the ring reduces the penalty of having
one execlists client do cleanup for another (old legacy submission
shares a ring between all clients). The downside is that slow and
irregular (off the critical path) process of cleaning up stale requests
after userspace becomes a modicum less efficient.
In the long run, it will become apparent that the ordered
ring->request_list matches the ring->timeline, a fun challenge for the
future will be unifying the two lists to avoid duplication!
v2: We need both engine-order and ring-order processing to maintain our
knowledge of where individual rings have completed upto as well as
knowing what was last executing on any engine. And finally by decoupling
retiring the contexts on the engine and the timelines along the rings,
we do have to keep a reference to the context on each request
(previously it was guaranteed by the context being pinned).
v3: Not just a reference to the context, but we need to keep it pinned
as we manipulate the rings; i.e. we need a pin for both the manipulation
of the engine state during its retirements, and a separate pin for the
manipulation of the ring state.
Chris Wilson [Mon, 30 Apr 2018 13:15:00 +0000 (14:15 +0100)]
drm/i915: Stop tracking timeline->inflight_seqnos
In commit 9b6586ae9f6b ("drm/i915: Keep a global seqno per-engine"), we
moved from a global inflight counter to per-engine counters in the
hope that will be easy to run concurrently in future. However, with the
advent of the desire to move requests between engines, we do need a
global counter to preserve the semantics that no engine wraps in the
middle of a submit. (Although this semantic is now only required for gen7
semaphore support, which only supports greater-then comparisons!)
v2: Keep a global counter of all requests ever submitted and force the
reset when it wraps.
Chris Wilson [Sat, 28 Apr 2018 11:15:32 +0000 (12:15 +0100)]
drm/i915/lrc: Scrub the GPU state of the guilty hanging request
Previously, we just reset the ring register in the context image such
that we could skip over the broken batch and emit the closing
breadcrumb. However, on resume the context image and GPU state would be
reloaded, which may have been left in an inconsistent state by the
reset. The presumption was that at worst it would just cause another
reset and skip again until it recovered, however it seems just as likely
to cause an unrecoverable hang. Instead of risking loading an incomplete
context image, restore it back to the default state.
v2: Fix up off-by-one from including the ppHSWP in with the register
state.
v3: Use a ring local to compact a few lines.
v4: Beware setting the ring local before checking for a NULL request.
References: https://bugs.freedesktop.org/show_bug.cgi?id=105304 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Michel Thierry <michel.thierry@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Michel Thierry <michel.thierry@intel.com> #v2 Link: https://patchwork.freedesktop.org/patch/msgid/20180428111532.15819-1-chris@chris-wilson.co.uk
Dave Airlie [Sun, 29 Apr 2018 23:32:34 +0000 (09:32 +1000)]
Merge tag 'drm-misc-next-2018-04-26' of git://anongit.freedesktop.org/drm/drm-misc into drm-next
drm-misc-next for v4.18:
UAPI Changes:
- Add support for a generic plane alpha property to sun4i, rcar-du and atmel-hclcdc. (Maxime)
Core Changes:
- Stop looking at legacy plane->fb and crtc members in atomic drivers. (Ville)
- mode_valid return type fixes. (Luc)
- Handle zpos normalization in the core. (Peter)
Driver Changes:
- Implement CTM, plane alpha and generic async cursor support in vc4. (Stefan)
- Various fixes for HPD and aux chan in drm_bridge/analogix_dp. (Lin, Zain, Douglas)
- Add support for MIPI DSI to sun4i. (Maxime)
Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Thomas Gleixner:
"Another set of x86 related updates:
- Fix the long broken x32 version of the IPC user space headers which
was noticed by Arnd Bergman in course of his ongoing y2038 work.
GLIBC seems to have non broken private copies of these headers so
this went unnoticed.
- Two microcode fixlets which address some more fallout from the
recent modifications in that area:
- Unconditionally save the microcode patch, which was only saved
when CPU_HOTPLUG was enabled causing failures in the late
loading mechanism
- Make the later loader synchronization finally work under all
circumstances. It was exiting early and causing timeout failures
due to a missing synchronization point.
- Do not use mwait_play_dead() on AMD systems to prevent excessive
power consumption as the CPU cannot go into deep power states from
there.
- Address an annoying sparse warning due to lost type qualifiers of
the vmemmap and vmalloc base address constants.
- Prevent reserving crash kernel region on Xen PV as this leads to
the wrong perception that crash kernels actually work there which
is not the case. Xen PV has its own crash mechanism handled by the
hypervisor.
- Add missing TLB cpuid values to the table to make the printout on
certain machines correct.
- Enumerate the new CLDEMOTE instruction
- Fix an incorrect SPDX identifier
- Remove stale macros"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/ipc: Fix x32 version of shmid64_ds and msqid64_ds
x86/setup: Do not reserve a crash kernel region if booted on Xen PV
x86/cpu/intel: Add missing TLB cpuid values
x86/smpboot: Don't use mwait_play_dead() on AMD systems
x86/mm: Make vmemmap and vmalloc base address constants unsigned long
x86/vector: Remove the unused macro FPU_IRQ
x86/vector: Remove the macro VECTOR_OFFSET_START
x86/cpufeatures: Enumerate cldemote instruction
x86/microcode: Do not exit early from __reload_late()
x86/microcode/intel: Save microcode patch unconditionally
x86/jailhouse: Fix incorrect SPDX identifier
Merge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 pti fixes from Thomas Gleixner:
"A set of updates for the x86/pti related code:
- Preserve r8-r11 in int $0x80. r8-r11 need to be preserved, but the
int$80 entry code removed that quite some time ago. Make it correct
again.
- A set of fixes for the Global Bit work which went into 4.17 and
caused a bunch of interesting regressions:
- Triggering a BUG in the page attribute code due to a missing
check for early boot stage
- Warnings in the page attribute code about holes in the kernel
text mapping which are caused by the freeing of the init code.
Handle such holes gracefully.
- Reduce the amount of kernel memory which is set global to the
actual text and do not incidentally overlap with data.
- Disable the global bit when RANDSTRUCT is enabled as it
partially defeats the hardening.
- Make the page protection setup correct for vma->page_prot
population again. The adjustment of the protections fell through
the crack during the Global bit rework and triggers warnings on
machines which do not support certain features, e.g. NX"
* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/entry/64/compat: Preserve r8-r11 in int $0x80
x86/pti: Filter at vma->vm_page_prot population
x86/pti: Disallow global kernel text with RANDSTRUCT
x86/pti: Reduce amount of kernel text allowed to be Global
x86/pti: Fix boot warning from Global-bit setting
x86/pti: Fix boot problems from Global-bit setting
projects / mirror_ubuntu-hirsute-kernel.git / log