diff options
author | 2023-11-01 06:28:35 -1000 | |
---|---|---|
committer | 2023-11-01 06:28:35 -1000 | |
commit | 7d461b291e65938f15f56fe58da2303b07578a76 (patch) | |
tree | 015dd7c2f1743dd70be52787dd9aff33822bc938 /drivers/gpu/drm/i915/gt/intel_ggtt.c | |
parent | Merge tag 'devicetree-for-6.7' of git://git.kernel.org/pub/scm/linux/kernel/g... (diff) | |
parent | Merge tag 'amd-drm-next-6.7-2023-10-27' of https://gitlab.freedesktop.org/agd... (diff) | |
download | linux-7d461b291e65938f15f56fe58da2303b07578a76.tar.gz linux-7d461b291e65938f15f56fe58da2303b07578a76.tar.bz2 linux-7d461b291e65938f15f56fe58da2303b07578a76.zip |
Merge tag 'drm-next-2023-10-31-1' of git://anongit.freedesktop.org/drm/drm
Pull drm updates from Dave Airlie:
"Highlights:
- AMD adds some more upcoming HW platforms
- Intel made Meteorlake stable and started adding Lunarlake
- nouveau has a bunch of display rework in prepartion for the NVIDIA
GSP firmware support
- msm adds a7xx support
- habanalabs has finished migration to accel subsystem
Detail summary:
kernel:
- add initial vmemdup-user-array
core:
- fix platform remove() to return void
- drm_file owner updated to reflect owner
- move size calcs to drm buddy allocator
- let GPUVM build as a module
- allow variable number of run-queues in scheduler
edid:
- handle bad h/v sync_end in EDIDs
panfrost:
- add Boris as maintainer
fbdev:
- use fb_ops helpers more
- only allow logo use from fbcon
- rename fb_pgproto to pgprot_framebuffer
- add HPD state to drm_connector_oob_hotplug_event
- convert to fbdev i/o mem helpers
i915:
- Enable meteorlake by default
- Early Xe2 LPD/Lunarlake display enablement
- Rework subplatforms into IP version checks
- GuC based TLB invalidation for Meteorlake
- Display rework for future Xe driver integration
- LNL FBC features
- LNL display feature capability reads
- update recommended fw versions for DG2+
- drop fastboot module parameter
- added deviceid for Arrowlake-S
- drop preproduction workarounds
- don't disable preemption for resets
- cleanup inlines in headers
- PXP firmware loading fix
- Fix sg list lengths
- DSC PPS state readout/verification
- Add more RPL P/U PCI IDs
- Add new DG2-G12 stepping
- DP enhanced framing support to state checker
- Improve shared link bandwidth management
- stop using GEM macros in display code
- refactor related code into display code
- locally enable W=1 warnings
- remove PSR watchdog timers on LNL
amdgpu:
- RAS/FRU EEPROM updatse
- IP discovery updatses
- GC 11.5 support
- DCN 3.5 support
- VPE 6.1 support
- NBIO 7.11 support
- DML2 support
- lots of IP updates
- use flexible arrays for bo list handling
- W=1 fixes
- Enable seamless boot in more cases
- Enable context type property for HDMI
- Rework GPUVM TLB flushing
- VCN IB start/size alignment fixes
amdkfd:
- GC 10/11 fixes
- GC 11.5 support
- use partial migration in GPU faults
radeon:
- W=1 Fixes
- fix some possible buffer overflow/NULL derefs
nouveau:
- update uapi for NO_PREFETCH
- scheduler/fence fixes
- rework suspend/resume for GSP-RM
- rework display in preparation for GSP-RM
habanalabs:
- uapi: expose tsc clock
- uapi: block access to eventfd through control device
- uapi: force dma-buf export to PAGE_SIZE alignments
- complete move to accel subsystem
- move firmware interface include files
- perform hard reset on PCIe AXI drain event
- optimise user interrupt handling
msm:
- DP: use existing helpers for DPCD
- DPU: interrupts reworked
- gpu: a7xx (a730/a740) support
- decouple msm_drv from kms for headless devices
mediatek:
- MT8188 dsi/dp/edp support
- DDP GAMMA - 12 bit LUT support
- connector dynamic selection capability
rockchip:
- rv1126 mipi-dsi/vop support
- add planar formats
ast:
- rename constants
panels:
- Mitsubishi AA084XE01
- JDI LPM102A188A
- LTK050H3148W-CTA6
ivpu:
- power management fixes
qaic:
- add detach slice bo api
komeda:
- add NV12 writeback
tegra:
- support NVSYNC/NHSYNC
- host1x suspend fixes
ili9882t:
- separate into own driver"
* tag 'drm-next-2023-10-31-1' of git://anongit.freedesktop.org/drm/drm: (1803 commits)
drm/amdgpu: Remove unused variables from amdgpu_show_fdinfo
drm/amdgpu: Remove duplicate fdinfo fields
drm/amd/amdgpu: avoid to disable gfxhub interrupt when driver is unloaded
drm/amdgpu: Add EXT_COHERENT support for APU and NUMA systems
drm/amdgpu: Retrieve CE count from ce_count_lo_chip in EccInfo table
drm/amdgpu: Identify data parity error corrected in replay mode
drm/amdgpu: Fix typo in IP discovery parsing
drm/amd/display: fix S/G display enablement
drm/amdxcp: fix amdxcp unloads incompletely
drm/amd/amdgpu: fix the GPU power print error in pm info
drm/amdgpu: Use pcie domain of xcc acpi objects
drm/amd: check num of link levels when update pcie param
drm/amdgpu: Add a read to GFX v9.4.3 ring test
drm/amd/pm: call smu_cmn_get_smc_version in is_mode1_reset_supported.
drm/amdgpu: get RAS poison status from DF v4_6_2
drm/amdgpu: Use discovery table's subrevision
drm/amd/display: 3.2.256
drm/amd/display: add interface to query SubVP status
drm/amd/display: Read before writing Backlight Mode Set Register
drm/amd/display: Disable SYMCLK32_SE RCO on DCN314
...
Diffstat (limited to 'drivers/gpu/drm/i915/gt/intel_ggtt.c')
-rw-r--r-- | drivers/gpu/drm/i915/gt/intel_ggtt.c | 265 |
1 files changed, 257 insertions, 8 deletions
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c index da21f2786b5d..1c93e84278a0 100644 --- a/drivers/gpu/drm/i915/gt/intel_ggtt.c +++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c @@ -15,18 +15,23 @@ #include "display/intel_display.h" #include "gem/i915_gem_lmem.h" +#include "intel_context.h" #include "intel_ggtt_gmch.h" +#include "intel_gpu_commands.h" #include "intel_gt.h" #include "intel_gt_regs.h" #include "intel_pci_config.h" +#include "intel_ring.h" #include "i915_drv.h" #include "i915_pci.h" +#include "i915_request.h" #include "i915_scatterlist.h" #include "i915_utils.h" #include "i915_vgpu.h" #include "intel_gtt.h" #include "gen8_ppgtt.h" +#include "intel_engine_pm.h" static void i915_ggtt_color_adjust(const struct drm_mm_node *node, unsigned long color, @@ -201,22 +206,36 @@ static void gen8_ggtt_invalidate(struct i915_ggtt *ggtt) intel_uncore_write_fw(uncore, GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); } +static void guc_ggtt_ct_invalidate(struct intel_gt *gt) +{ + struct intel_uncore *uncore = gt->uncore; + intel_wakeref_t wakeref; + + with_intel_runtime_pm_if_active(uncore->rpm, wakeref) { + struct intel_guc *guc = >->uc.guc; + + intel_guc_invalidate_tlb_guc(guc); + } +} + static void guc_ggtt_invalidate(struct i915_ggtt *ggtt) { struct drm_i915_private *i915 = ggtt->vm.i915; + struct intel_gt *gt; gen8_ggtt_invalidate(ggtt); - if (GRAPHICS_VER(i915) >= 12) { - struct intel_gt *gt; - - list_for_each_entry(gt, &ggtt->gt_list, ggtt_link) + list_for_each_entry(gt, &ggtt->gt_list, ggtt_link) { + if (intel_guc_tlb_invalidation_is_available(>->uc.guc)) { + guc_ggtt_ct_invalidate(gt); + } else if (GRAPHICS_VER(i915) >= 12) { intel_uncore_write_fw(gt->uncore, GEN12_GUC_TLB_INV_CR, GEN12_GUC_TLB_INV_CR_INVALIDATE); - } else { - intel_uncore_write_fw(ggtt->vm.gt->uncore, - GEN8_GTCR, GEN8_GTCR_INVALIDATE); + } else { + intel_uncore_write_fw(gt->uncore, + GEN8_GTCR, GEN8_GTCR_INVALIDATE); + } } } @@ -252,6 +271,145 @@ u64 gen8_ggtt_pte_encode(dma_addr_t addr, return pte; } +static bool should_update_ggtt_with_bind(struct i915_ggtt *ggtt) +{ + struct intel_gt *gt = ggtt->vm.gt; + + return intel_gt_is_bind_context_ready(gt); +} + +static struct intel_context *gen8_ggtt_bind_get_ce(struct i915_ggtt *ggtt) +{ + struct intel_context *ce; + struct intel_gt *gt = ggtt->vm.gt; + + if (intel_gt_is_wedged(gt)) + return NULL; + + ce = gt->engine[BCS0]->bind_context; + GEM_BUG_ON(!ce); + + /* + * If the GT is not awake already at this stage then fallback + * to pci based GGTT update otherwise __intel_wakeref_get_first() + * would conflict with fs_reclaim trying to allocate memory while + * doing rpm_resume(). + */ + if (!intel_gt_pm_get_if_awake(gt)) + return NULL; + + intel_engine_pm_get(ce->engine); + + return ce; +} + +static void gen8_ggtt_bind_put_ce(struct intel_context *ce) +{ + intel_engine_pm_put(ce->engine); + intel_gt_pm_put(ce->engine->gt); +} + +static bool gen8_ggtt_bind_ptes(struct i915_ggtt *ggtt, u32 offset, + struct sg_table *pages, u32 num_entries, + const gen8_pte_t pte) +{ + struct i915_sched_attr attr = {}; + struct intel_gt *gt = ggtt->vm.gt; + const gen8_pte_t scratch_pte = ggtt->vm.scratch[0]->encode; + struct sgt_iter iter; + struct i915_request *rq; + struct intel_context *ce; + u32 *cs; + + if (!num_entries) + return true; + + ce = gen8_ggtt_bind_get_ce(ggtt); + if (!ce) + return false; + + if (pages) + iter = __sgt_iter(pages->sgl, true); + + while (num_entries) { + int count = 0; + dma_addr_t addr; + /* + * MI_UPDATE_GTT can update 512 entries in a single command but + * that end up with engine reset, 511 works. + */ + u32 n_ptes = min_t(u32, 511, num_entries); + + if (mutex_lock_interruptible(&ce->timeline->mutex)) + goto put_ce; + + intel_context_enter(ce); + rq = __i915_request_create(ce, GFP_NOWAIT | GFP_ATOMIC); + intel_context_exit(ce); + if (IS_ERR(rq)) { + GT_TRACE(gt, "Failed to get bind request\n"); + mutex_unlock(&ce->timeline->mutex); + goto put_ce; + } + + cs = intel_ring_begin(rq, 2 * n_ptes + 2); + if (IS_ERR(cs)) { + GT_TRACE(gt, "Failed to ring space for GGTT bind\n"); + i915_request_set_error_once(rq, PTR_ERR(cs)); + /* once a request is created, it must be queued */ + goto queue_err_rq; + } + + *cs++ = MI_UPDATE_GTT | (2 * n_ptes); + *cs++ = offset << 12; + + if (pages) { + for_each_sgt_daddr_next(addr, iter) { + if (count == n_ptes) + break; + *cs++ = lower_32_bits(pte | addr); + *cs++ = upper_32_bits(pte | addr); + count++; + } + /* fill remaining with scratch pte, if any */ + if (count < n_ptes) { + memset64((u64 *)cs, scratch_pte, + n_ptes - count); + cs += (n_ptes - count) * 2; + } + } else { + memset64((u64 *)cs, pte, n_ptes); + cs += n_ptes * 2; + } + + intel_ring_advance(rq, cs); +queue_err_rq: + i915_request_get(rq); + __i915_request_commit(rq); + __i915_request_queue(rq, &attr); + + mutex_unlock(&ce->timeline->mutex); + /* This will break if the request is complete or after engine reset */ + i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT); + if (rq->fence.error) + goto err_rq; + + i915_request_put(rq); + + num_entries -= n_ptes; + offset += n_ptes; + } + + gen8_ggtt_bind_put_ce(ce); + return true; + +err_rq: + i915_request_put(rq); +put_ce: + gen8_ggtt_bind_put_ce(ce); + return false; +} + static void gen8_set_pte(void __iomem *addr, gen8_pte_t pte) { writeq(pte, addr); @@ -272,6 +430,21 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm, ggtt->invalidate(ggtt); } +static void gen8_ggtt_insert_page_bind(struct i915_address_space *vm, + dma_addr_t addr, u64 offset, + unsigned int pat_index, u32 flags) +{ + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); + gen8_pte_t pte; + + pte = ggtt->vm.pte_encode(addr, pat_index, flags); + if (should_update_ggtt_with_bind(i915_vm_to_ggtt(vm)) && + gen8_ggtt_bind_ptes(ggtt, offset, NULL, 1, pte)) + return ggtt->invalidate(ggtt); + + gen8_ggtt_insert_page(vm, addr, offset, pat_index, flags); +} + static void gen8_ggtt_insert_entries(struct i915_address_space *vm, struct i915_vma_resource *vma_res, unsigned int pat_index, @@ -311,6 +484,50 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, ggtt->invalidate(ggtt); } +static bool __gen8_ggtt_insert_entries_bind(struct i915_address_space *vm, + struct i915_vma_resource *vma_res, + unsigned int pat_index, u32 flags) +{ + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); + gen8_pte_t scratch_pte = vm->scratch[0]->encode; + gen8_pte_t pte_encode; + u64 start, end; + + pte_encode = ggtt->vm.pte_encode(0, pat_index, flags); + start = (vma_res->start - vma_res->guard) / I915_GTT_PAGE_SIZE; + end = start + vma_res->guard / I915_GTT_PAGE_SIZE; + if (!gen8_ggtt_bind_ptes(ggtt, start, NULL, end - start, scratch_pte)) + goto err; + + start = end; + end += (vma_res->node_size + vma_res->guard) / I915_GTT_PAGE_SIZE; + if (!gen8_ggtt_bind_ptes(ggtt, start, vma_res->bi.pages, + vma_res->node_size / I915_GTT_PAGE_SIZE, pte_encode)) + goto err; + + start += vma_res->node_size / I915_GTT_PAGE_SIZE; + if (!gen8_ggtt_bind_ptes(ggtt, start, NULL, end - start, scratch_pte)) + goto err; + + return true; + +err: + return false; +} + +static void gen8_ggtt_insert_entries_bind(struct i915_address_space *vm, + struct i915_vma_resource *vma_res, + unsigned int pat_index, u32 flags) +{ + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); + + if (should_update_ggtt_with_bind(i915_vm_to_ggtt(vm)) && + __gen8_ggtt_insert_entries_bind(vm, vma_res, pat_index, flags)) + return ggtt->invalidate(ggtt); + + gen8_ggtt_insert_entries(vm, vma_res, pat_index, flags); +} + static void gen8_ggtt_clear_range(struct i915_address_space *vm, u64 start, u64 length) { @@ -332,6 +549,27 @@ static void gen8_ggtt_clear_range(struct i915_address_space *vm, gen8_set_pte(>t_base[i], scratch_pte); } +static void gen8_ggtt_scratch_range_bind(struct i915_address_space *vm, + u64 start, u64 length) +{ + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); + unsigned int first_entry = start / I915_GTT_PAGE_SIZE; + unsigned int num_entries = length / I915_GTT_PAGE_SIZE; + const gen8_pte_t scratch_pte = vm->scratch[0]->encode; + const int max_entries = ggtt_total_entries(ggtt) - first_entry; + + if (WARN(num_entries > max_entries, + "First entry = %d; Num entries = %d (max=%d)\n", + first_entry, num_entries, max_entries)) + num_entries = max_entries; + + if (should_update_ggtt_with_bind(ggtt) && gen8_ggtt_bind_ptes(ggtt, first_entry, + NULL, num_entries, scratch_pte)) + return ggtt->invalidate(ggtt); + + gen8_ggtt_clear_range(vm, start, length); +} + static void gen6_ggtt_insert_page(struct i915_address_space *vm, dma_addr_t addr, u64 offset, @@ -1008,7 +1246,18 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND; } - if (intel_uc_wants_guc(&ggtt->vm.gt->uc)) + if (i915_ggtt_require_binder(i915)) { + ggtt->vm.scratch_range = gen8_ggtt_scratch_range_bind; + ggtt->vm.insert_page = gen8_ggtt_insert_page_bind; + ggtt->vm.insert_entries = gen8_ggtt_insert_entries_bind; + /* + * On GPU is hung, we might bind VMAs for error capture. + * Fallback to CPU GGTT updates in that case. + */ + ggtt->vm.raw_insert_page = gen8_ggtt_insert_page; + } + + if (intel_uc_wants_guc_submission(&ggtt->vm.gt->uc)) ggtt->invalidate = guc_ggtt_invalidate; else ggtt->invalidate = gen8_ggtt_invalidate; |