- 19 Oct, 2018 1 commit
-
-
Xiong Zhang authored
Currently the guest couldn't boot up under GVT-g environment as the following call trace exists: [ 272.504762] BUG: unable to handle kernel NULL pointer dereference at 0000000000000100 [ 272.504834] Call Trace: [ 272.504852] execlists_context_pin+0x2b2/0x520 [i915] [ 272.504869] intel_gvt_scan_and_shadow_workload+0x50/0x4d0 [i915] [ 272.504887] intel_vgpu_create_workload+0x3e2/0x570 [i915] [ 272.504901] intel_vgpu_submit_execlist+0xc0/0x2a0 [i915] [ 272.504916] elsp_mmio_write+0xc7/0x130 [i915] [ 272.504930] intel_vgpu_mmio_reg_rw+0x24a/0x4c0 [i915] [ 272.504944] intel_vgpu_emulate_mmio_write+0xac/0x240 [i915] [ 272.504947] intel_vgpu_rw+0x22d/0x270 [kvmgt] [ 272.504949] intel_vgpu_write+0x164/0x1f0 [kvmgt] GVT GEM context is created by i915_gem_context_create_gvt() which doesn't allocate ppgtt. So GVT GEM context structure doesn't have a valid i915_hw_ppgtt. This patch create ppgtt table at GVT GEM context creation, then assign shadow ppgtt's root table address to this ppgtt when shadow ppgtt will be used on GPU. So GVT GEM context has valid ppgtt address. But note that this ppgtt only contain valid ppgtt root table address, the table entry in this ppgtt structure are invalid. Fixes:4a3d3f67("drm/i915: Match code to comment and enforce ppgtt for execlists") Signed-off-by:
Xiong Zhang <xiong.y.zhang@intel.com> Reviewed-by:
Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/1539841231-3157-1-git-send-email-xiong.y.zhang@intel.com
-
- 18 Oct, 2018 10 commits
-
-
Ville Syrjälä authored
Replace the kvmalloc_array() with i915_gem_object_get_dma_address() when populating rotated vmas. One random access mechanism ought to be enough for everyone? To calculate the size of the radix tree I think we can do something like this (assuming 64bit pointers): num_pages = obj_size / 4096 tree_height = ceil(log64(num_pages)) num_nodes = sum(64^n, n, 0, tree_height-1) tree_size = num_nodes * 576 If we compare that with the object size we should get a relative overhead of around .2% to 1% for reasonable sized objects, which framebuffers tend to be. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Suggested-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181016150413.11577-1-ville.syrjala@linux.intel.comReviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com>
-
Joonas Lahtinen authored
Use single exit point to drop rpm wakeref in case of an error. Fixes: 9d3eb2c3 ("drm/i915: Hold rpm wakeref for debugfs/i915_drop_caches_set") Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Daniele Ceraolo Spurio authored
The test requires driver tweaks to avoid causing error messages on intentionally-triggered errors and to stop accessing non existing register. However, this is a pure GuC FW interface test and should be covered by FW validation, so it isn't really worth tweaking the driver for it and we're better off dropping it instead. Testing the driver running out of doorbells is already covered by igt_guc_doorbells Suggested-by:
Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by:
Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Reviewed-by:
-
Jani Nikula authored
intel_dsi_vbt_init() has grown too unwieldy, and it's about to be modified due to ICL DSI. Abstract out the VLV specific dphy param init. No functional changes. Intentionally no stylistic changes during code movement. Cc: Madhav Chauhan <madhav.chauhan@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Reviewed-by:
Madhav Chauhan <madhav.chauhan@intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/96d15760db027a137f298ec330520ef8ec6474b0.1539613303.git.jani.nikula@intel.com
-
Jani Nikula authored
Abstract bitrate calculation to a newly resurrected intel_dsi.c file that will contain common code for VLV and ICL DSI. No functional changes. Cc: Madhav Chauhan <madhav.chauhan@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Encoders are not alike, make enable and disable hooks optional like other hooks. Utilize this in DSI code, and remove the silly nop hook. v2: Add the check also to intel_sanitize_encoder() (Madhav) Reviewed-by:
-
Jani Nikula authored
Pass the type we want to simplify. No functional changes. v2: s/dev_priv/i915/g (Chris) Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by:
-
Michal Wajdeczko authored
In response for I915_PARAM_HAS_HUC we are returning value that indicates if HuC firmware was loaded and verified. However, our previously used positive value was based on specific register bit which is about to change on future platform. Let's normalize our return values to 0 and 1 before clients will start to use Gen9 value. v2: use bool for implicit conversion (Chris) Signed-off-by:
-
Tvrtko Ursulin authored
GEM_WARN_ON currently has dangerous semantics where it is completely compiled out on !GEM_DEBUG builds. This can leave users who expect it to be more like a WARN_ON, just without a warning in non-debug builds, in complete ignorance. Another gotcha with it is that it cannot be used as a statement. Which is again different from a standard kernel WARN_ON. This patch fixes both problems by making it behave as one would expect. It can now be used both as an expression and as statement, and also the condition evaluates properly in all builds - code under the conditional will therefore not unexpectedly disappear. To satisfy call sites which really want the code under the conditional to completely disappear, we add GEM_DEBUG_WARN_ON and convert some of the callers to it. This one can also be used as both expression and statement. >From the above it follows GEM_DEBUG_WARN_ON should be used in situations where we are certain the condition will be hit during development, but at a place in code where error can be handled to the benefit of not crashing the machine. GEM_WARN_ON on the other hand should be used where condition may happen in production and we just want to distinguish the level of debugging output emitted between the production and debug build. v2: * Dropped BUG_ON hunk. Signed-off-by:
Tomasz Lis <tomasz.lis@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181012063142.16080-1-tvrtko.ursulin@linux.intel.com
-
Rodrigo Vivi authored
Since when it was introduced we forgot to add this case so ICL was using a wrong signal_levels as reference. Fixes: fb5c8e9d ("drm/i915/icl: Implement voltage swing programming sequence for Combo PHY DDI") Cc: José Roberto de Souza <jose.souza@intel.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by:
Manasi Navare <manasi.d.navare@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181017215652.26841-1-rodrigo.vivi@intel.com
-
- 17 Oct, 2018 9 commits
-
-
Lyude Paul authored
Unfortunately, it appears our fix in: commit b5d29843 ("drm/atomic_helper: Allow DPMS On<->Off changes for unregistered connectors") Which attempted to work around the problems introduced by: commit 4d802739 ("drm/atomic_helper: Disallow new modesets on unregistered connectors") Is still not the right solution, as modesets can still be triggered outside of drm_atomic_set_crtc_for_connector(). So in order to fix this, while still being careful that we don't break modesets that a driver may perform before being registered with userspace, we replace connector->registered with a tristate member, connector->registration_state. This allows us to keep track of whether or not a connector is still initializing and hasn't been exposed to userspace, is currently registered and exposed to userspace, or has been legitimately removed from the system after having once been present. Using this info, we can prevent userspace from performing new modesets on unregistered connectors while still allowing the driver to perform modesets on unregistered connectors before the driver has finished being registered. Changes since v1: - Fix WARN_ON() in drm_connector_cleanup() that CI caught with this patchset in igt@drv_module_reload@basic-reload-inject and igt@drv_module_reload@basic-reload by checking if the connector is registered instead of unregistered, as calling drm_connector_cleanup() on a connector that hasn't been registered with userspace yet should stay valid. - Remove unregistered_connector_check(), and just go back to what we were doing before in commit 4d802739 ("drm/atomic_helper: Disallow new modesets on unregistered connectors") except replacing READ_ONCE(connector->registered) with drm_connector_is_unregistered(). This gets rid of the behavior of allowing DPMS On<->Off, but that should be fine as it's more consistent with the UAPI we had before - danvet - s/drm_connector_unregistered/drm_connector_is_unregistered/ - danvet - Update documentation, fix some typos. Fixes: b5d29843 ("drm/atomic_helper: Allow DPMS On<->Off changes for unregistered connectors") Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: stable@vger.kernel.org Cc: David Airlie <airlied@linux.ie> Signed-off-by:
Lyude Paul <lyude@redhat.com> Reviewed-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20181016203946.9601-1-lyude@redhat.com
-
Daniele Ceraolo Spurio authored
The ENTER/EXIT_S_STATE actions queue the save/restore operation in GuC FW and then return, so waiting on the H2G is not enough to guarantee GuC is done. When all the processing is done, GuC writes 0 to scratch register 14, so we can poll on that. Note that GuC does not ensure that the value in the register is different from 0 while the action is in progress so we need to take care of that ourselves as well. v2: improve comment, return early on GuC error and improve error message (Michal) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by: -
Maarten Lankhorst authored
Wait for idle, and iterate over connectors instead of encoders. With this information we know crtc->state is the actual state, and we can enable/disable drrs safely. Signed-off-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181011100457.8776-2-maarten.lankhorst@linux.intel.comReviewed-by:
-
Imre Deak authored
If BIOS configured a Y tiled FB we failed to set up the backing object tiling accordingly, leading to a lack of GT fence installed and a garbled console. The problem was bisected to commit 011f22eb ("drm/i915: Do NOT skip the first 4k of stolen memory for pre-allocated buffers v2") but it just revealed a pre-existing issue. Kudos to Ville who suspected a missing fence looking at the corruption on the screen. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Mika Westerberg <mika.westerberg@linux.intel.com> Cc: Hans de Goede <hdegoede@redhat.com> Cc: <ronald@innovation.ch> Cc: <stable@vger.kernel.org> Reported-by:
Mika Westerberg <mika.westerberg@linux.intel.com> Reported-by: <ronald@innovation.ch> Tested-by: <ronald@innovation.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108264 Fixes: bc8d7dff ("drm/i915/skl: Provide a Skylake version of get_plane_config()") Signed-off-by:
Imre Deak <imre.deak@intel.com> Reviewed-by:
-
Michal Wajdeczko authored
In commit 4502e9ec ("drm/i915/uc: Unify firmware loading") we stopped converting errors detected during firmware transfer into -EAGAIN and this indirectly killed our workarounds for Gen9 GuC. Reactivate those workarounds by looking for actual -ETIMEDOUT error. Testcase: igt@drv_selftest@live_hangcheck Reported-by:
-
Jani Nikula authored
Reduce intel_display.c by splitting out intel_quirks.c. No functional changes. Reviewed-by:
-
Jani Nikula authored
Clang build with UBSAN enabled leads to the following build error: drivers/gpu/drm/i915/intel_engine_cs.o: In function `intel_engine_init_execlist': drivers/gpu/drm/i915/intel_engine_cs.c:411: undefined reference to `__compiletime_assert_411' Again, for this to work the code would first need to be inlined and then constant folded, which doesn't work for Clang because semantic analysis happens before optimization/inlining. Use GEM_BUG_ON() instead of BUILD_BUG_ON(). v2: Use is_power_of_2() from log2.h (Chris) References: http://mid.mail-archive.com/20181015203410.155997-1-swboyd@chromium.orgReported-by:
Stephen Boyd <swboyd@chromium.org> Cc: Stephen Boyd <swboyd@chromium.org> Cc: Chris Wilson <chris@chris-wilson.co.uk> Tested-by:
Nathan Chancellor <natechancellor@gmail.com> Tested-by:
Nick Desaulniers <ndesaulniers@google.com> Signed-off-by:
-
Jani Nikula authored
When building the kernel with Clang with defconfig and CONFIG_64BIT disabled, vmlinux fails to link because of the BUILD_BUG in _print_param. ld: drivers/gpu/drm/i915/i915_params.o: in function `i915_params_dump': i915_params.c:(.text+0x56): undefined reference to `__compiletime_assert_191' This function is semantically invalid unless the code is first inlined then constant folded, which doesn't work for Clang because semantic analysis happens before optimization/inlining. [The above written by Nathan Chancellor <natechancellor@gmail.com>] Use WARN_ONCE() instead of BUILD_BUG() to avoid the problem. The WARN_ONCE() should get optimized away unless there's a type that's not handled by _print_param(). References: https://github.com/ClangBuiltLinux/linux/issues/191 References: http://mid.mail-archive.com/20181009171401.14980-1-natechancellor@gmail.com Cc: Nick Desaulniers <ndesaulniers@google.com> Cc: Nathan Chancellor <natechancellor@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reported-by:
-
Jani Nikula authored
Rename intel_get_pipe_from_connector() to intel_connector_get_pipe() and move it near its connector function friends in intel_connector.c. No functional changes. Reviewed-by:
-
- 16 Oct, 2018 20 commits
-
-
Mahesh Kumar authored
DDI/TC clock-off bits are not equally distanced. TC1-3 bits are from offset 12 & TC4 is at offset 21. Create a function to choose correct clk-off bit. v2: Add fixes tag (Lucas) Fixes: c27e917e ("drm/i915/icl: add basic support for the ICL clocks") Signed-off-by:
Mahesh Kumar <mahesh1.kumar@intel.com> Signed-off-by:
Vandita Kulkarni <vandita.kulkarni@intel.com> Reviewed-by:
Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
-
Lucas De Marchi authored
combo-phy register instances are at same offset from base for each combo-phy port, i.e. Port A base offset: 0x16200 Port B base offset: 0x6C000 All the other addresses for both ports can be derived by calculating offset to these base addresses. PORT_CL_DW_OFFSET 0x0 PORT_CL_DW<x> 0 + x * 4 PORT_COMP_OFFSET 0x100 PORT_COMP_DW<x> 0x100 + x * 4 PORT_PCS_AUX_OFFSET 0x300 PORT_PCS_GRP_OFFSET 0x600 PORT_PCS_LN<y>_OFFSET 0x800 + y * 0x100 PORT_TX_AUX_OFFSET 0x380 PORT_TX_GRP_OFFSET 0x680 PORT_TX_LN<y>_OFFSET 0x880 + y * 0x100 And inside each PORT_TX_[AUX|GRP|LN] we add `dw * 4`. Based on original patch by Mahesh Kumar <mahesh1.kumar@intel.com>. v2: make port, dw and ln arguments follow the order in register's name Signed-off-by: -
Mahesh Kumar authored
This patch combines CNL/ICL specific port/combophy macros together at one location. This is prework for patches later in series where new macros to find port/combophy register will be introduced. v2: remove wrong empty line Signed-off-by:
-
Vandita Kulkarni authored
This patch adds helper function for identifying whether the given PLL is combo PHY PLL or not. This helper function is used inside various ICL functions to make them scalable. Signed-off-by:
-
Vandita Kulkarni authored
Use intel_port_is_tc and intel_port_is_combophy functions to replace the individual port checks from port C to F and port A to B respectively. Signed-off-by:
-
Vandita Kulkarni authored
Use the existing port-to-id helper function, to refactor hence making it scalable. Signed-off-by:
-
Mahesh Kumar authored
Instead of directly comparing HPD pins use intel_port_is_combophy/tc helper functions to distinguish between combophy/TC ports. Signed-off-by:
-
Mahesh Kumar authored
This patch creates a function/wrapper to check if port is combophy port instead of explicitly comparing ports. Changes since V1: - keep all intel_port_is_* helper together (Lucas) Signed-off-by:
-
Lucas De Marchi authored
Display WA #1178 is meant to fix Aux channel voltage swing too low with some type C dongles. It applies to external ports on combo phy. On Icelake this is port A and B when those are not eDP. v2: follow the spec to the letter: include Aux A and just check if it's not eDP instead of checking only for Aux B. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Imre Deak <imre.deak@intel.com> Signed-off-by: -
Maarten Lankhorst authored
We have to look at crtc_state, so pass that instead. Also cleanup the use of dev vs dev_priv, we really want to pass along dev_priv. Signed-off-by:
-
Maarten Lankhorst authored
Instead of derferencing crtc->config, look at crtc_state. Signed-off-by:
-
Maarten Lankhorst authored
The CRTC is idle at this point, so we can dereference crtc->state safely. Signed-off-by:
-
Maarten Lankhorst authored
We know the crtc is idle because we're at the beginning of sanitization, so just dereference crtc->state instead. Signed-off-by:
-
Maarten Lankhorst authored
No need to look at crtc->config when we have crtc_state in the caller. Signed-off-by:
-
Maarten Lankhorst authored
has_drrs is a flag we can't read out. We set it when seamless DRRS is enabled in pipe_config, so intel_dump_pipe_config() and intel_pipe_config_compare() will continue to do the right thing when has_drrs is set on the real state. This removes one more dereference of crtc->config. While at it, fixup the comment and also read out M2_N2 for CHV, since we program it in the set_m_n function. Changes since v1: - Only read out M2/N2 on platforms that support DRRS. Signed-off-by:
-
Maarten Lankhorst authored
We already have a perfectly nice pipe_config, use that instead. Signed-off-by:
-
Maarten Lankhorst authored
Another user of crtc->config gone. The functions it calls also needed crtc->config, so convert those as well. Signed-off-by:
-
Mika Kuoppala authored
Disable master interrupt before reading level indications. This will close a race where we get a level indication between reading and disabling, generating an extra interrupt where we could have avoided one. Further, as the reading acts also as a post, replace the write/post on the irq reset with the helper. On enabling side, posting doesn't serve any purpose so it can also be replaced with helper. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by:
Mika Kuoppala <mika.kuoppala@linux.intel.com> Acked-by:
-
Mika Kuoppala authored
All other master control register bits, except the enable, are read only and they are level indications of the second level interrupt status. Only touch enable bit and rectify the comment. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181015141440.21845-2-mika.kuoppala@linux.intel.com
-
Mika Kuoppala authored
Disable master interrupt before reading level indications. This will close a race where we get a level indication between reading and disabling, generating an extra interrupt where we could have avoided one. Further, as the reading acts also as a post, replace the write/post on the irq reset with the helper. On enabling side, posting doesn't serve any purpose so it can also be replaced with helper. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
