- 21 Aug, 2020 1 commit
-
-
Mika Kuoppala authored
Our variety of defined gpu commands have the actual command id field and possibly length and flags applied. We did start to apply the mask during initialization of the cmd descriptors but forgot to also apply it on comparisons. Fix comparisons in order to properly deny access with associated commands. v2: fix lri with correct mask (Chris) References: 926abff2 ("drm/i915/cmdparser: Ignore Length operands during command matching") Reported-by:
Nicolai Stange <nstange@suse.de> Cc: stable@vger.kernel.org # v5.4+ Cc: Miroslav Benes <mbenes@suse.cz> Cc: Takashi Iwai <tiwai@suse.de> Cc: Tyler Hicks <tyhicks@canonical.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris.p.wilson@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/20200817195926.12671-1-mika.kuoppala@linux.intel.com
-
- 19 Aug, 2020 1 commit
-
-
Matt Atwood authored
Add minimum width to planes, variable with specific formats for gen11+ to reflect recent bspec changes. Signed-off-by:
Matt Atwood <matthew.s.atwood@intel.com> Reviewed-by:
José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
- 17 Aug, 2020 38 commits
-
-
José Roberto de Souza authored
Major upgrade of HuC firmware with fixes and new features. Signed-off-by:
Matt Roper <matthew.d.roper@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200813200029.25307-2-jose.souza@intel.comSigned-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com>
-
José Roberto de Souza authored
Fixes around DC5, DC6 and DC3CO in those new firmwares. Signed-off-by:
-
José Roberto de Souza authored
From the 3 WAs for PSR2 man track/selective fetch this is only one needed when doing single full frames at every flip. Reviewed-by:
Gwan-gyeong Mun <gwan-gyeong.mun@intel.com> Signed-off-by:
-
José Roberto de Souza authored
All GEN12 platforms supports PSR2 selective fetch but not all GEN12 platforms supports PSR2 hardware tracking(aka RKL). This feature consists in software programming registers with the damaged area of each plane this way hardware will only fetch from memory those areas and sent the PSR2 selective update blocks to panel, saving even more power. But as initial step it is only enabling the full frame fetch at every flip, the actual selective fetch part will come in a future patch. Also this is only handling the page flip side, it is still completely missing frontbuffer modifications, that is why the enable_psr2_sel_fetch parameter was added. v3: - calling intel_psr2_sel_fetch_update() during the atomic check phase (Ville) BSpec: 55229 Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Reviewed-by:
-
Matt Roper authored
We usually assume that increasing PCI device revision ID's translates to newer steppings; macros like IS_KBL_REVID() that we use rely on this behavior. Unfortunately this turns out to not be true on KBL; the newer device 2 revision ID's sometimes go backward to older steppings. The situation is further complicated by different GT and display steppings associated with each revision ID. Let's work around this by providing a table to map the revision ID to specific GT and display steppings, and then perform our comparisons on the mapped values. v2: - Move the kbl_revids[] array to intel_workarounds.c to avoid compiler warnings about an unused variable in files that don't call the macros (kernel test robot). Bspec: 18329 Signed-off-by:
Swathi Dhanavanthri <swathi.dhanavanthri@intel.com> Signed-off-by:
-
José Roberto de Souza authored
This new HBR2 table for TGL-U and TGL-Y is required to pass DisplayPort compliance. BSpec: 49291 Cc: Khaled Almahallawy <khaled.almahallawy@intel.com> Reviewed-by: Khaled Almahallawy<khaled.almahallawy@intel.com> Signed-off-by:
-
José Roberto de Souza authored
There is no way to differentiate TGL-U from TGL-Y by the PCI ids as some ids are available in both SKUs. So here using the root device id in the PCI bus that iGPU is in to differentiate between U and Y. BSpec: 44455 Reviewed-by:
-
Imre Deak authored
The command register is the PCODE MBOX low register not the high one as described by the spec. This left the system with the TC-cold power state being blocked all the time. Fix things by using the correct register. Also to make sure we retry a request for at least 600usec, when the PCODE MBOX command itself succeeded, but the TC-cold block command failed, sleep for 1msec unconditionally after any fail. The change was tested with JTAG register read of the HW/FW's actual TC-cold state, which reported the expected states after this change. Tested-by:
Nivedita Swaminathan <nivedita.swaminathan@intel.com> Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
Imre Deak <imre.deak@intel.com> Reviewed-by:
-
Imre Deak authored
The dependency between power wells is determined by the ordering of the power well list: when enabling the power wells for a domain, this happens walking the power well list forward, while disabling them happens in the reverse direction. Accordingly a power well on the list must follow any other power well it depends on. Since the TC AUX power wells depend on TC-cold being blocked, move the TC-cold off power well before all AUX power wells. Fixes: 3c02934b ("drm/i915/tc/tgl: Implement TC cold sequences") Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
George Spelvin authored
igt_mm_config() calls ilog2() on the (pseudo)random 21-bit number s>>12. Once in 2 million seeds, this is zero and ilog2 summons the nasal demons. There was an attempt to handle this case with a max(), but that's too late; ms could already be something bizarre. Given that the low 12 bits of s and ms are always zero, it's a lot simpler just to divide them by 4096, then everything fits into 32 bits, and we can easily generate a random number 1 <= s <= 0x1fffff. Fixes: 14d1b9a6 ("drm/i915: buddy allocator") Signed-off-by:
George Spelvin <lkml@sdf.org> Cc: Matthew Auld <matthew.auld@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> Cc: intel-gfx@lists.freedesktop.org Reviewed-by:
Matthew Auld <matthew.auld@intel.com> Signed-off-by:
-
Lionel Landwerlin authored
Introduces a new parameters to execbuf so that we can specify syncobj handles as well as timeline points. v2: Reuse i915_user_extension_fn v3: Check that the chained extension is only present once (Chris) v4: Check that dma_fence_chain_find_seqno returns a non NULL fence (Lionel) v5: Use BIT_ULL (Chris) v6: Fix issue with already signaled timeline points, dma_fence_chain_find_seqno() setting fence to NULL (Chris) v7: Report ENOENT with invalid syncobj handle (Lionel) v8: Check for out of order timeline point insertion (Chris) v9: After explanations on https://lists.freedesktop.org/archives/dri-devel/2019-August/229287.html drop the ordering check from v8 (Lionel) v10: Set first extension enum item to 1 (Jason) v11: Rebase v12: Allow multiple extension nodes of timeline syncobj (Chris) Signed-off-by:Lionel Landwerlin <lionel.g.landwerlin@intel.com> Co-authored-by:
-
Lionel Landwerlin authored
We're planning to use this for a couple of new feature where we need to provide additional parameters to execbuf. v2: Check for invalid flags in execbuffer2 (Lionel) v3: Rename I915_EXEC_EXT -> I915_EXEC_USE_EXTENSIONS (Chris) v4: Rebase Move array fence parsing in i915_gem_do_execbuffer() Signed-off-by:Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by:
-
Matt Roper authored
The hardware team has dropped this workaround from the bspec; it is no longer needed. This reverts commit 111822b21be995a3a4a731066db3d820523c57f7. Bspec: 49291 Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
Tianjia Zhang authored
In the case of calling check_digital_port_conflicts() failed, a negative error code -EINVAL should be returned. Fixes: bf5da83e ("drm/i915: Move check_digital_port_conflicts() earier") Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Tianjia Zhang <tianjia.zhang@linux.alibaba.com> Reviewed-by:
-
Tianjia Zhang authored
In function i915_active_acquire_preallocate_barrier(), not all paths have the return value set correctly, and in case of memory allocation failure, a negative error code should be returned. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
-
Matt Roper authored
A recent bspec update removed the LPDDR4 single channel entry from the buddy register table, but added a new four-channel entry. Workaround 1409767108 hasn't been updated with any guidance for four channel configurations, so we leave that alternate table unchanged for now. Bspec 49218 Fixes: 3fa01d64 ("drm/i915/tgl: Program BW_BUDDY registers during display init") Signed-off-by:
Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
-
Chris Wilson authored
We use i915_active_fini() as a debug check on the i915_active state before freeing. If we forget to call it, we may end up angering the debugobjects contained within. Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by:
Thomas Hellström <thomas.hellstrom@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731085015.32368-1-chris@chris-wilson.co.ukSigned-off-by:
-
Chris Wilson authored
Avoid a GPF at <1>[ 20.177320] BUG: kernel NULL pointer dereference, address: 000000000000007c <1>[ 20.177322] #PF: supervisor read access in kernel mode <1>[ 20.177323] #PF: error_code(0x0000) - not-present page <6>[ 20.177324] PGD 0 P4D 0 <4>[ 20.177327] Oops: 0000 [#1] PREEMPT SMP PTI <4>[ 20.177328] CPU: 1 PID: 944 Comm: debugfs_test Not tainted 5.8.0-rc7-CI-CI_DRM_8814+ #1 <4>[ 20.177330] Hardware name: Dell Inc. XPS 13 9360/0823VW, BIOS 2.9.0 07/09/2018 <4>[ 20.177372] RIP: 0010:i915_lpsp_capability_show+0x44/0xc0 [i915] <4>[ 20.177374] Code: 0f b6 81 ca 0d 00 00 3c 0b 74 77 76 19 3c 0c 75 44 83 7e 7c 01 7e 2f 48 c7 c6 d7 b9 47 a0 e8 43 df 06 e1 31 c0 c3 3c 09 72 2b <8b> 46 7c 85 c0 75 e6 8b 82 e4 00 00 00 89 c2 83 e2 fb 83 fa 0a 74 <4>[ 20.177376] RSP: 0018:ffffc90000cebe38 EFLAGS: 00010246 <4>[ 20.177377] RAX: 0000000000000009 RBX: ffff888267fe6a58 RCX: ffff888252d10000 <4>[ 20.177378] RDX: ffff88824a9a4000 RSI: 0000000000000000 RDI: ffff888267fe6a30 <4>[ 20.177379] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000001 <4>[ 20.177380] R10: 0000000000000001 R11: 0000000000000000 R12: ffffc90000cebf08 <4>[ 20.177381] R13: 00000000ffffffff R14: 0000000000000001 R15: ffff888267fe6a30 <4>[ 20.177383] FS: 00007f6f9c6b5e40(0000) GS:ffff888276480000(0000) knlGS:0000000000000000 <4>[ 20.177384] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 <4>[ 20.177385] CR2: 000000000000007c CR3: 0000000255f04006 CR4: 00000000003606e0 <4>[ 20.177386] Call Trace: <4>[ 20.177390] seq_read+0xcb/0x420 which is presumably from having no encoder attached at that time. Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2175 Fixes: 8806211f ("drm/i915: Add i915_lpsp_capability debugfs") Signed-off-by:
Anshuman Gupta <anshuman.gupta@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200729130912.30093-1-chris@chris-wilson.co.ukSigned-off-by:
-
Chris Wilson authored
It's been a while since gen6_rps_boost() [that only worked on gen6+] was replaced by intel_rps_boost() that understood itself when rps was active. Since the intel_rps_boost() is gen-agnostic, just call it. Signed-off-by:
-
Chris Wilson authored
Since we want to read the values from the HWSP as written to by the GPU, warn the compiler that the values are volatile. Signed-off-by:
-
Chris Wilson authored
Since we use the module parameters stored inside the drm_i915_device itself, we need to ensure the mock i915_device also sets up the right defaults. Fixes: 8a25c4be ("drm/i915/params: switch to device specific parameters") Signed-off-by:
-
Uma Shankar authored
Fbc is causing random underruns in CI execution on TGL platforms. Disabling the same while the problem is being debugged and analyzed. v2: Moved the check below the module param check (Ville) Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Uma Shankar <uma.shankar@intel.com> Acked-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200716145857.6911-1-uma.shankar@intel.comSigned-off-by:
-
José Roberto de Souza authored
Although the WA description targets the platforms it is a workaround for the affected PCHs, that is why it is being checked. v2: excluding DG1 fake PCH from WA BSpec: 52890 BSpec: 53273 BSpec: 52888 Reviewed-by:
-
Chris Wilson authored
Gracefully skip over the failures in the frequency scaling for the moment, the results are under review. References: https://gitlab.freedesktop.org/drm/intel/-/issues/1754Signed-off-by:
Sujaritha Sundaresan <sujaritha.sundaresan@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200723153805.8076-1-chris@chris-wilson.co.ukSigned-off-by:
-
Matt Roper authored
After doing normal PHY-B initialization on Rocket Lake, we need to manually copy some additional PHY-A register values into PHY-B registers. Note that the bspec's combo phy page doesn't specify that this workaround is restricted to specific platform steppings (and doesn't even do a very good job of specifying that RKL is the only platform this is needed on), but the RKL workaround page lists this as relevant only for A and B steppings, so I'm trusting that information for now. v2: Make rkl_combo_phy_b_init_wa() static v3: - Minimize variables in WA function. (Jose) - Fix timeout duration (usec vs msec). (Jose) - Add verification of workaround. (Jose) - Fix stepping bounds in comment. Bspec: 49291 Bspec: 53273 Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
Matt Roper authored
If HTI (also sometimes called HDPORT) is enabled at startup, it may be using some of the PHYs and DPLLs making them unavailable for general usage. Let's read out the HDPORT_STATE register and avoid making use of resources that HTI is already using. v2: - Fix minor checkpatch warnings v3: - Just readout HDPORT_STATE register once during init and then parse it later as needed. - Add a 'has_hti' device info flag to track whether we should readout HDPORT_STATE or not. We can skip the platform/flag tests later since the hti_state in dev_priv will remain 0 for platforms it does not apply to. - Move PLL masking into icl_get_combo_phy_dpll() since at the moment RKL is the only platform that has HTI. (Jose) Bspec: 49189 Bspec: 53707 Cc: Lucas De Marchi <lucas.demarchi@intel.com> Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
Matt Roper authored
Rocket Lake has a third DPLL (called 'DPLL4') that must be used to enable a third display. Unlike EHL's variant of DPLL4, the RKL variant behaves the same as DPLL0/1. And despite its name, the DPLL4 registers are offset as if it were DPLL2. v2: - Add new .update_ref_clks() hook. v3: - Renumber TBT PLL to '3' and switch _MMIO_PLL3 to _MMIO_PLL (Lucas) v4: - Don't drop _MMIO_PLL3; although it's now unused, we're going to need it very soon again for upcoming DG1 patches. (Lucas) v5: - Don't re-number TBT PLL and beyond, just use new RKL_DPLL_CFGCR macros to lookup the proper registers instead. Although renumbering the PLLs might be something we want to consider down the road, it opens a big can of worms right now since a bunch of places in the code have an assumption that the PLL table has idx==id and no holes. Renumbering creates a hole for TGL, so we'd either need to allow holes in the table or break the idx==id invariant, both of which are somewhat invasive changes to the design. Bspec: 49202 Bspec: 49443 Bspec: 50288 Bspec: 50289 Cc: Lucas De Marchi <lucas.demarchi@intel.com> Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
Matt Roper authored
RKL and TGL share some general gen12 workarounds, but each platform also has its own platform-specific workarounds. v2: - Add Wa_1604555607 for RKL. This makes RKL's ctx WA list identical to TGL's, so we'll have both functions call the tgl_ function for now; this workaround isn't listed for DG1 so we don't want to add it to the general gen12_ function. Cc: Matt Atwood <matthew.s.atwood@intel.com> Signed-off-by:
-
Matt Roper authored
RKL uses a slightly different bit layout for the DPCLKA_CFGCR0 register. v2: - Fix inverted mask application when updating ICL_DPCLKA_CFGCR0 - Checkpatch style fixes Bspec: 50287 Cc: Aditya Swarup <aditya.swarup@intel.com> Signed-off-by:
-
Ville Syrjälä authored
It's silly to have if(SKL) checks in gen9_init_clock_gating() when we can just move those bits into skl_init_clock_gating(). I'm not entirely convinced we even need this w/a, or if we do then maybe we want it for kbl/cfl as well. IIRC it was only listed in the wadb, but that is now dead so can't double check anymore. Bspec doesn't seem to have any purely skl specific DOP clock gating workarounds listed. Signed-off-by:
-
José Roberto de Souza authored
Hours Of Battery Life is a new GEN12+ power-saving feature that allows supported motherboards to use a special voltage swing table for eDP panels that uses less power. So here if supported by HW, OEM will set it in VBT and i915 will try to train link with HOBL vswing table if link training fails it fall back to the original table. intel_ddi_dp_preemph_max() was optimized to only check the HOBL flag instead of do something like is done in intel_ddi_dp_voltage_max() because it is only called after the first entry of the voltage swing table was loaded so the HOBL flag is valid at that point. v3: - removed a few parameters of icl_ddi_combo_vswing_program() that can be taken from encoder v4: - using the HOBL vswing table until training fails completely (Ville) v5: - not reducing lane or link rate when link training fails with HOBL active - duplicated the HOBL voltage swing entry to match DP spec requirement v6: - removed the optional VS 3 & pre-emp 0 from HOBL table - changed from u8:1 to bool to store hobl_failed/active BSpec: 49291 BSpec: 49399 Reviewed-by:
-
Alexei Podtelezhnikov authored
0x0155 is rather Ivy Bridge PCI-E Root Port. 0x0157 from the same commit ff049b6c ("drm/i915: bind driver to ValleyView chipsets") is likely wrong too. Nowhere is it independetly confirmed or mentioned. Signed-off-by:
Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Chris Wilson authored
Rather than manually implement our own module reference counting for perf pmu events, finally realise that there is a module parameter to struct pmu for this very purpose. Signed-off-by:
-
Imre Deak authored
The value we program to DDI_BUF_CTL changes at the following places: - At enabling/disabling the output to configure the port width etc, and to enable/disable the DDI BUF function. - At the beginning/end of link re-training to disable/re-enable the DDI BUF function. - On HSW/BDW/SKL to change the voltage swing/pre-emph levels. Except of the above the value we program to the DDI_BUF_CTL register (intel_dp->DP) doesn't change, so no need to reprogram the register when changing the link training patterns (which is programmed via the DP_TP_CTL register on DDI platforms). v2: - Fix the commit message wrt. voltage/pre-emph level values in intel_dp->DP. (Ville) Signed-off-by:
-
Imre Deak authored
According to BSpec this flag should not be changed while the DDI function is enabled. On BDW+ the DP_TP_CTL register spec also states it explicitly that the HW takes care of enabling/disabling the scrambling for training patterns (and it must stay enabled for normal pixel output). Assume that this HW automatic handling of scrambling is also true for HSW. BSpec: 8013, 7557, 50484 Signed-off-by:
-
Ville Syrjälä authored
Explicitly check for i830 when assigning the .get_cdclk() vfunc, and then deal with the case of not having assigned the vfunc separately. Less confusing, and gets rid of the checkpatch complaint about using {} on one branch but not the others. Signed-off-by: -
Ville Syrjälä authored
Some spaces have snuck in where we want tabs. Fix it. Signed-off-by:
-
Ville Syrjälä authored
There's a pointless hole in struct intel_cdclk_vals, get rid of it. Fortunately we already use named initializers so the order does not matter. Signed-off-by:
-
