- 11 Dec, 2019 25 commits
-
-
Ville Syrjälä authored
In order to eliminate intel_pipe_to_cpu_transcoder() (and its crtc->config usage) let's pass the cpu transcoder to assert_pipe() so we don't have to do the pipe->cpu transcoder lookup on HSW+. On VLV/CHV this can get called during eDP init, which happens before crtc->config->cpu_transcoder is even populated. So currently we're always reading PIPECONF(A) there even if we're trying to check the state of some other pipe. Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191112163812.22075-4-ville.syrjala@linux.intel.comReviewed-by:
José Roberto de Souza <jose.souza@intel.com>
-
Ville Syrjälä authored
Let's start to eliminate intel_pipe_to_cpu_transcoder() so that we can get rid of one more crtc->config usage (which we will want to nuke as well). In the case of assert_fdi_tx() we know that we're never dealing with the EDP transcoder so we can simply replace this with a cast. v2: Fix poor English in comment Signed-off-by:
-
Chris Wilson authored
Print the i915_active state on selftest failure, with a hope it helps illuminate the cause of the failure. References: https://gitlab.freedesktop.org/drm/intel/issues/765Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Acked-by:
Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191210115502.3767070-1-chris@chris-wilson.co.uk
-
Chris Wilson authored
When we are told we have to drop all references to userptr, wait for any barriers required for unbinding. <4> [2055.808787] WARNING: CPU: 3 PID: 6239 at mm/mmu_notifier.c:472 __mmu_notifier_invalidate_range_start+0x1f2/0x250 <4> [2055.808792] Modules linked in: vgem mei_hdcp snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic x86_pkg_temp_thermal coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel r8169 lpc_ich realtek i915 snd_hda_intel snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core pinctrl_broxton snd_pcm pinctrl_intel mei_me intel_lpss_pci mei prime_numbers [last unloaded: vgem] <4> [2055.808834] CPU: 3 PID: 6239 Comm: gem_userptr_bli Tainted: G U 5.5.0-rc1-CI-CI_DRM_7522+ #1 <4> [2055.808839] Hardware name: /NUC6CAYB, BIOS AYAPLCEL.86A.0049.2018.0508.1356 05/08/2018 <4> [2055.808847] RIP: 0010:__mmu_notifier_invalidate_range_start+0x1f2/0x250 <4> [2055.808853] Code: c2 48 c7 c7 70 17 2e 82 44 89 45 d4 48 8b 70 28 e8 ec 01 ef ff 41 f6 46 20 01 44 8b 45 d4 75 0a 41 83 f8 f5 44 89 7d d4 74 89 <0f> 0b 44 89 45 d4 eb 81 0f 0b 49 8b 46 18 49 8b 76 10 4c 89 ff 48 <4> [2055.808858] RSP: 0018:ffffc90002937d40 EFLAGS: 00010202 <4> [2055.808865] RAX: 0000000000000061 RBX: ffff8882703a33e0 RCX: 0000000000000001 <4> [2055.808870] RDX: 0000000000000000 RSI: ffff888277da8cb8 RDI: 00000000ffffffff <4> [2055.808874] RBP: ffffc90002937d70 R08: 00000000fffffff5 R09: 0000000000000000 <4> [2055.808879] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000001 <4> [2055.808884] R13: ffffffff822e1716 R14: ffffc90002937d80 R15: 00000000fffffff5 <4> [2055.808890] FS: 00007fda75004e40(0000) GS:ffff888277d80000(0000) knlGS:0000000000000000 <4> [2055.808895] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 <4> [2055.808900] CR2: 000055ad72ec3000 CR3: 00000002697b2000 CR4: 00000000003406e0 <4> [2055.808904] Call Trace: <4> [2055.808920] unmap_vmas+0x13e/0x150 <4> [2055.808937] unmap_region+0xa3/0x100 <4> [2055.808964] __do_munmap+0x26d/0x490 <4> [2055.808980] __vm_munmap+0x66/0xc0 <4> [2055.808994] __x64_sys_munmap+0x12/0x20 <4> [2055.809001] do_syscall_64+0x4f/0x220 Closes: https://gitlab.freedesktop.org/drm/intel/issues/771Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191210133719.3874455-1-chris@chris-wilson.co.uk
-
Chris Wilson authored
Check that we own the global pointer before deregistering. Reported-by:
Venkata Sandeep Dhanalakota <venkata.s.dhanalakota@intel.com> Signed-off-by:
-
Tvrtko Ursulin authored
Convert i915_gem_check_execbuffer to return the error code instead of a boolean so our neat EINVAL debugging trick works within this function. Signed-off-by:
-
Chris Wilson authored
We want the bonded request to have the same scheduler properties as its master so that it is placed at the same depth in the queue. For example, consider we have requests A, B and B', where B & B' are a bonded pair to run in parallel on two engines. A -> B \- B' B will run after A and so may be scheduled on an idle engine and wait on A using a semaphore. B' sees B being executed and so enters the queue on the same engine as A. As B' did not inherit the semaphore-chain from B, it may have higher precedence than A and so preempts execution. However, B' then sits on a semaphore waiting for B, who is waiting for A, who is blocked by B. Ergo B' needs to inherit the scheduler properties from B (i.e. the semaphore chain) so that it is scheduled with the same priority as B and will not be executed ahead of Bs dependencies. Furthermore, to prevent the priorities changing via the expose fence on B', we need to couple in the dependencies for PI. This requires us to relax our sanity-checks that dependencies are strictly in order. v2: Synchronise (B, B') execution on all platforms, regardless of using a scheduler, any no-op syncs should be elided. Fixes: ee113690 ("drm/i915/execlists: Virtual engine bonding") Closes: https://gitlab.freedesktop.org/drm/intel/issues/464 Testcase: igt/gem_exec_balancer/bonded-chain Testcase: igt/gem_exec_balancer/bonded-semaphore Signed-off-by: -
Animesh Manna authored
As the current usage is restricted to first DSB instance per pipe, so existing code could not catch the issue to calculate the mmio offset of different DSB instance per pipe. Corrected the offset calculation. Fixes: a6e58d9a ("drm/i915/dsb: Check DSB engine status.") Signed-off-by:
Animesh Manna <animesh.manna@intel.com> Reviewed-by:
Anshuman Gupta <anshuman.gupta@intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191205123513.22603-1-animesh.manna@intel.com
-
Jani Nikula authored
Sync up with v5.5-rc1 to get the updated lock_release() API among other things. Fix the conflict reported by Stephen Rothwell [1]. [1] http://lore.kernel.org/r/20191210093957.5120f717@canb.auug.org.auSigned-off-by:
-
Jani Nikula authored
Enable DSC for DSI, if specified in VBT. This still lacks DSC aware get config implementation, and therefore state checker will fail. Also mode valid is not there yet. v5: - add dsc get config call v4: - convert_rgb = true (Vandita) - ignore max cdclock check (Vandita) - rename pipe_config to crtc_state v3: - take compressed bpp into account v2: - Nuke conn_state->max_requested_bpc, it's not used on DSI Bspec: 49263 Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
Vandita Kulkarni <vandita.kulkarni@intel.com> Signed-off-by:
-
Vandita Kulkarni authored
When DSC is enabled consider the compression ratio that was used during horizontal timing calculations. This may still lead to warns due to rounding errors in the round-trip. v2 by Jani: - rebase on top of the more generic dsc state readout Signed-off-by:
-
Jani Nikula authored
When DSC is enabled, we need to adjust the horizontal timings to account for the compressed (and therefore reduced) link speed. The compressed frequency ratio simplifies down to the ratio between compressed and non-compressed bpp. Bspec: 49263 Suggested-by:
-
Jani Nikula authored
When compression is enabled, configure the DSI transcoder to use compressed format. Suggested-by:
-
Jani Nikula authored
Pass crtc_state to afe_clk() to be able to take compression into account in the computation. Once we enable compression, that is. Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
We'll be expanding afe_clk() to take DSC into account. Switch to using it where DSC matters. Which is really everywhere that intel_dsi_bitrate() is currently used in ICL DSI code. The functional difference is that we round the result closest instead of down. Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
We'll make more use of it in the future. Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
The ICL DSI pipe_bpp currently comes from compute_baseline_pipe_bpp(). Fix it. Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Add basic hardware state readout for DSC, and check the most relevant details in the state checker. v2: - check for DSC power before reading its state - check if source supports DSC at all As a side effect, this should also get the power domains for the enabled DSC on takeover, and subsequently disable DSC if it's not needed. Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Reviewed-by:
Manasi Navare <manasi.d.navare@intel.com> Signed-off-by:
-
Jani Nikula authored
Move intel_dp_source_supports_dsc() from intel_dp.c as intel_dsc_source_support() in intel_vdsc.c. The DSC source support is more about DSC than about DP, and will be needed for DP independent code. Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Reviewed-by:
-
Jani Nikula authored
Add DSI specific computation and transmission to display of PPS. With hopes that this approach will work for both DP and DSI encoders. Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Turns out this isn't compatible with DSI, where we use the value from VBT. No functional changes. Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Turns out future DSI specific parameters aren't workable with the approach of having the encoder specific functions in intel_vdsc.c. Make intel_dsc_compute_params() a helper that does the encoder independent parts, and have encoder code call it. Move intel_dsc_dp_compute_params() to intel_dp.c as intel_dp_dsc_compute_params(). No functional changes. v2: Rename pipe_config to crtc_state while at it. Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Add function for retrieving the DSC data for an encoder. Initially, this is DSI specific, as DP does not use VBT settings for DSC at all. It's also not very pretty. In the future we might have a pointer from encoder to the child device, which would make the child device list query here so much more sensible. v3: - use crtc_state instead of pipe_config - return true by default from intel_bios_get_dsc_params() - expand the comment about rc_buffer_block_size and rc_buffer_size v2: - make more robust, debug log errors better Bspec: 29885 Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Check for child devices that specify compression, and store the device specific compression parameters in the display device data struct for later use. Warn if compression is requested but not available. Use fairly rigid checks for compression data for starters. These can be made more dynamic later. Log about DSC presence in DDI port parse, though this is not universal across platforms or port types (DSI). v2: amended debug logging Bspec: 29885 Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Jani Nikula authored
Allow accessing the parent structure later on. Drop const for allowing future modification as well. Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
- 10 Dec, 2019 3 commits
-
-
Matt Roper authored
Gen12 can improve bandwidth efficiency by pairing up memory requests with similar addresses. We need to program the BW_BUDDY1 and BW_BUDDY2 registers according to the memory configuration during display initialization to take advantage of this capability. The magic numbers we program here feel like something that could definitely change on future platforms, so let's use a table-based programming scheme to make this easy to extend in the future. v2: - Add separate table for Wa_1409767108. (Stan) - Reorder structure reduce size by a word. Page mask can still be up to 28 bits (even though current values are small) so we should keep it as a u32, but just using a u8 for DRAM type instead of the actual enum type saves space. (Lucas, Ville) - Rename function to tgl_bw_buddy_init() to be more precise about what it does. (Lucas) Bspec: 49189 Bspec: 49218 Bspec: 52890 Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
Matt Roper <matthew.d.roper@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191205224848.76712-1-matthew.d.roper@intel.comReviewed-by:
Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
-
Colin Ian King authored
Currently the variable sum is not uninitialized and hence will cause an incorrect result in the summation values. Fix this by initializing sum to the first item in the summation. Addresses-Coverity: ("Uninitialized scalar variable") Fixes: 3c7a44bb ("drm/i915/selftests: Perform some basic cycle counting of MI ops") Signed-off-by:Colin Ian King <colin.king@canonical.com> Reviewed-by:
-
Chris Wilson authored
In order to avoid confusing the HW, we must never submit an empty ring during lite-restore, that is we should always advance the RING_TAIL before submitting to stay ahead of the RING_HEAD. Normally this is prevented by keeping a couple of spare NOPs in the request->wa_tail so that on resubmission we can advance the tail. This relies on the request only being resubmitted once, which is the normal condition as it is seen once for ELSP[1] and then later in ELSP[0]. On preemption, the requests are unwound and the tail reset back to the normal end point (as we know the request is incomplete and therefore its RING_HEAD is even earlier). However, if this w/a should fail we would try and resubmit the request with the RING_TAIL already set to the location of this request's wa_tail potentially causing a GPU hang. We can spot when we do try and incorrectly resubmit without advancing the RING_TAIL and spare any embarrassment by forcing the context restore. In the case of preempt-to-busy, we leave the requests running on the HW while we unwind. As the ring is still live, we cannot rewind our rq->tail without forcing a reload so leave it set to rq->wa_tail and only force a reload if we resubmit after a lite-restore. (Normally, the forced reload will be a part of the preemption event.) Fixes: 22b7a426 ("drm/i915/execlists: Preempt-to-busy") Closes: https://gitlab.freedesktop.org/drm/intel/issues/673Signed-off-by:
-
- 09 Dec, 2019 12 commits
-
-
Daniele Ceraolo Spurio authored
We now only use 1 client without any plan to add more. The client is also only holding information about the WQ and the process desc, so we can just move those in the intel_guc structure and always use stage_id 0. v2: fix comment (John) v3: fix the comment for real, fix kerneldoc Signed-off-by:
Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: John Harrison <John.C.Harrison@Intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Reviewed-by:
John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191205220243.27403-4-daniele.ceraolospurio@intel.com
-
Daniele Ceraolo Spurio authored
Instead of relying on the workqueue, the upcoming reworked GuC submission flow will offer the host driver indipendent control over the execution status of each context submitted to GuC. As part of this, the doorbell usage model has been reworked, with each doorbell being paired to a single lrc and a doorbell ring representing new work available for that specific context. This mechanism, however, limits the number of contexts that can be registered with GuC to the number of doorbells, which is an undesired limitation. To avoid this limitation, we requested the GuC team to also provide a H2G that will allow the host to notify the GuC of work available for a specified lrc, so we can use that mechanism instead of relying on the doorbells. We can therefore drop the doorbell code we currently have, also given the fact that in the unlikely case we'd want to switch back to using doorbells we'd have to heavily rework it. The workqueue will still have a use in the new interface to pass special commands, so that code has been retained for now. With the doorbells gone and the GuC client becoming even simpler, the existing GuC selftests don't give us any meaningful coverage so we can remove them as well. Some selftests might come with the new code, but they will look different from what we have now so if doesn't seem worth it to keep the file around in the meantime. v2: fix comments and commit message (John) Signed-off-by:
-
Daniele Ceraolo Spurio authored
We already have a couple of use-cases in the code and another one will come in one of the later patches in the series. v2: use the new function for the CT object as well Signed-off-by:
-
Daniele Ceraolo Spurio authored
Remove unused enums and ctx_save_restore_disabled() function, leftover from the legacy preemption removal. Signed-off-by:
-
Ville Syrjälä authored
intel_hdcp_transcoder_config() is clobbering some globally visible state in .compute_config(). That is a big no no as .compute_config() is supposed to have no visible side effects when either the commit fails or it's just a TEST_ONLY commit. Inline this stuff into intel_hdcp_enable() so that the state only gets modified when we actually commit the state to the hardware. Cc: Ramalingam C <ramalingam.c@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Uma Shankar <uma.shankar@intel.com> Fixes: 39e2df09 ("drm/i915/hdcp: update current transcoder into intel_hdcp") Signed-off-by:
Ramalingam C <ramalingam.c@intel.com>
-
Ville Syrjälä authored
The code assumes we can omit the cfb allocation once fbc has been enabled once. That's nonsense. Let's try to reallocate it if we need to. The code is still a mess, but maybe this is enough to get fbc going in some cases where it initially underallocates the cfb and there's no full modeset to fix it up. Cc: Daniel Drake <drake@endlessm.com> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Jian-Hong Pan <jian-hong@endlessm.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by:
Daniel Drake <drake@endlessm.com> Reviewed-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
-
Ville Syrjälä authored
Now that we have the glk+ w/a for back to back fbc disable + plane update in place we can once more enable fbc on glk+ by default. Cc: Daniel Drake <drake@endlessm.com> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Jian-Hong Pan <jian-hong@endlessm.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by:
-
Ville Syrjälä authored
On glk+ the hardware gets confused if we disable FBC while it's recompressing and we perform a plane update during the same frame. The result is that top of the screen gets corrupted. We can avoid that by giving the hardware enough time to finish the FBC disable before we touch the plane registers. Ie. we need an extra vblank wait after FBC disable. v2: Don't do the vblank wait if we never activated FBC in hw Signed-off-by:
-
Ville Syrjälä authored
The hardware automagically nukes the cfb on flip. We can use that whenever the plane/crtc configuration doesn't change too much. Let's hook that up. We'll need this for glk+ since we need to introduce an extra vblank wait after FBC disable. As we're currently disabling FBC around all plane updates we'd slow them down by an extra frame. Not a great user experience when your fps is always capped at vrefres/2. With flip nuke we don't need the extra vblank wait. Signed-off-by:
-
Ville Syrjälä authored
fbc.enabled == (fbc.crtc != NULL), so let's just nuke fbc.enabled. Signed-off-by:
-
Ville Syrjälä authored
Replace the 'gen9 && !glk' with the slightly more obvious 'gen9_bc || bxt'. Signed-off-by:
-
Ville Syrjälä authored
i965gm no longer needs the fence for scanout so we should be do what we do for ctg+ and only configure a fence for FBC when we have one. In theory this should do nothing atm on account of intel_fbc_can_activate() requiring the fence, but since we do this for g4x+ let's do it for i965gm as well. We may want to relax the requirements at some point and allow FBC without a fence. Signed-off-by:
-
