From 2de60af4a4633be12e5559ed2d3c27613f715d53 Mon Sep 17 00:00:00 2001 From: Colin Xu Date: Mon, 1 Jun 2020 11:04:57 +0800 Subject: drm/i915/gvt: Add one missing MMIO handler for D_SKL_PLUS _PLANE_CTL_3_A, _PLANE_CTL_3_B and _PLANE_SURF_3_A are handled, but miss _PLANE_SURF_3_B. Reviewed-by: Zhenyu Wang Signed-off-by: Colin Xu Signed-off-by: Zhenyu Wang Link: http://patchwork.freedesktop.org/patch/msgid/20200601030457.14002-1-colin.xu@intel.com --- drivers/gpu/drm/i915/gvt/handlers.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 3e88e3b5c43a..89873a7b7f32 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -3055,6 +3055,7 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) MMIO_D(_MMIO(0x72380), D_SKL_PLUS); MMIO_D(_MMIO(0x7239c), D_SKL_PLUS); MMIO_D(_MMIO(_PLANE_SURF_3_A), D_SKL_PLUS); + MMIO_D(_MMIO(_PLANE_SURF_3_B), D_SKL_PLUS); MMIO_D(CSR_SSP_BASE, D_SKL_PLUS); MMIO_D(CSR_HTP_SKL, D_SKL_PLUS); -- cgit v1.2.3 From fccd0f7cf4d532674d727c7f204f038456675dee Mon Sep 17 00:00:00 2001 From: Colin Xu Date: Mon, 1 Jun 2020 11:06:38 +0800 Subject: drm/i915/gvt: Fix two CFL MMIO handling caused by regression. D_CFL was incorrectly removed for: GAMT_CHKN_BIT_REG GEN9_CTX_PREEMPT_REG V2: Update commit message. V3: Rebase and split Fixes and mis-handled MMIO. Fixes: 43226e6fe798 (drm/i915/gvt: replaced register address with name) Reviewed-by: Zhenyu Wang Signed-off-by: Colin Xu Signed-off-by: Zhenyu Wang Link: http://patchwork.freedesktop.org/patch/msgid/20200601030638.16002-1-colin.xu@intel.com --- drivers/gpu/drm/i915/gvt/handlers.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 89873a7b7f32..f39a6b20bbaf 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -3132,8 +3132,8 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL); - MMIO_D(GAMT_CHKN_BIT_REG, D_KBL); - MMIO_D(GEN9_CTX_PREEMPT_REG, D_KBL | D_SKL); + MMIO_D(GAMT_CHKN_BIT_REG, D_KBL | D_CFL); + MMIO_D(GEN9_CTX_PREEMPT_REG, D_SKL_PLUS); return 0; } -- cgit v1.2.3 From fc1e3aa0337c76b620726e205372073c234616ab Mon Sep 17 00:00:00 2001 From: Colin Xu Date: Mon, 1 Jun 2020 11:07:21 +0800 Subject: drm/i915/gvt: Fix incorrect check of enabled bits in mask registers Using _MASKED_BIT_ENABLE macro to set mask register bits is straight forward and not likely to go wrong. However when checking which bit(s) is(are) enabled, simply bitwise AND value and _MASKED_BIT_ENABLE() won't output expected result. Suppose the register write is disabling bit 1 by setting 0xFFFF0000, however "& _MASKED_BIT_ENABLE(1)" outputs 0x00010000, and the non-zero check will pass which cause the old code consider the new value set as an enabling operation. We found guest set 0x80008000 on boot, and set 0xffff8000 during resume. Both are legal settings but old code will block latter and force vgpu enter fail-safe mode. Introduce two new macro and make proper masked bit check in mmio handler: IS_MASKED_BITS_ENABLED() IS_MASKED_BITS_DISABLED() V2: Rebase. Reviewed-by: Zhenyu Wang Signed-off-by: Colin Xu Signed-off-by: Zhenyu Wang Link: http://patchwork.freedesktop.org/patch/msgid/20200601030721.17129-1-colin.xu@intel.com --- drivers/gpu/drm/i915/gvt/handlers.c | 19 ++++++++++--------- drivers/gpu/drm/i915/gvt/mmio_context.h | 6 +++--- drivers/gpu/drm/i915/gvt/reg.h | 5 +++++ 3 files changed, 18 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index f39a6b20bbaf..fadd2adb8030 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -1726,13 +1726,13 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, (*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(2); write_vreg(vgpu, offset, p_data, bytes); - if (data & _MASKED_BIT_ENABLE(1)) { + if (IS_MASKED_BITS_ENABLED(data, 1)) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } if (IS_COFFEELAKE(vgpu->gvt->gt->i915) && - data & _MASKED_BIT_ENABLE(2)) { + IS_MASKED_BITS_ENABLED(data, 2)) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } @@ -1741,14 +1741,14 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, * pvinfo, if not, we will treat this guest as non-gvtg-aware * guest, and stop emulating its cfg space, mmio, gtt, etc. */ - if (((data & _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE)) || - (data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE))) - && !vgpu->pv_notified) { + if ((IS_MASKED_BITS_ENABLED(data, GFX_PPGTT_ENABLE) || + IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE)) && + !vgpu->pv_notified) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } - if ((data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE)) - || (data & _MASKED_BIT_DISABLE(GFX_RUN_LIST_ENABLE))) { + if (IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE) || + IS_MASKED_BITS_DISABLED(data, GFX_RUN_LIST_ENABLE)) { enable_execlist = !!(data & GFX_RUN_LIST_ENABLE); gvt_dbg_core("EXECLIST %s on ring %s\n", @@ -1809,7 +1809,7 @@ static int ring_reset_ctl_write(struct intel_vgpu *vgpu, write_vreg(vgpu, offset, p_data, bytes); data = vgpu_vreg(vgpu, offset); - if (data & _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)) + if (IS_MASKED_BITS_ENABLED(data, RESET_CTL_REQUEST_RESET)) data |= RESET_CTL_READY_TO_RESET; else if (data & _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET)) data &= ~RESET_CTL_READY_TO_RESET; @@ -1827,7 +1827,8 @@ static int csfe_chicken1_mmio_write(struct intel_vgpu *vgpu, (*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(0x18); write_vreg(vgpu, offset, p_data, bytes); - if (data & _MASKED_BIT_ENABLE(0x10) || data & _MASKED_BIT_ENABLE(0x8)) + if (IS_MASKED_BITS_ENABLED(data, 0x10) || + IS_MASKED_BITS_ENABLED(data, 0x8)) enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.h b/drivers/gpu/drm/i915/gvt/mmio_context.h index 970704b18f23..3b25e7fe32f6 100644 --- a/drivers/gpu/drm/i915/gvt/mmio_context.h +++ b/drivers/gpu/drm/i915/gvt/mmio_context.h @@ -54,8 +54,8 @@ bool is_inhibit_context(struct intel_context *ce); int intel_vgpu_restore_inhibit_context(struct intel_vgpu *vgpu, struct i915_request *req); -#define IS_RESTORE_INHIBIT(a) \ - (_MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) == \ - ((a) & _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT))) + +#define IS_RESTORE_INHIBIT(a) \ + IS_MASKED_BITS_ENABLED(a, CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) #endif diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h index 5b66e14c5b7b..b88e033cbed4 100644 --- a/drivers/gpu/drm/i915/gvt/reg.h +++ b/drivers/gpu/drm/i915/gvt/reg.h @@ -94,6 +94,11 @@ #define GFX_MODE_BIT_SET_IN_MASK(val, bit) \ ((((bit) & 0xffff0000) == 0) && !!((val) & (((bit) << 16)))) +#define IS_MASKED_BITS_ENABLED(_val, _b) \ + (((_val) & _MASKED_BIT_ENABLE(_b)) == _MASKED_BIT_ENABLE(_b)) +#define IS_MASKED_BITS_DISABLED(_val, _b) \ + ((_val) & _MASKED_BIT_DISABLE(_b)) + #define FORCEWAKE_RENDER_GEN9_REG 0xa278 #define FORCEWAKE_ACK_RENDER_GEN9_REG 0x0D84 #define FORCEWAKE_BLITTER_GEN9_REG 0xa188 -- cgit v1.2.3 From a291e4fba259a56a6a274c1989997acb6f0bb03a Mon Sep 17 00:00:00 2001 From: Colin Xu Date: Mon, 1 Jun 2020 11:55:56 +0800 Subject: drm/i915/gvt: Use GFP_ATOMIC instead of GFP_KERNEL in atomic context GFP_KERNEL flag specifies a normal kernel allocation in which executing in process context without any locks and can sleep. mmio_diff takes sometime to finish all the diff compare and it has locks, continue using GFP_KERNEL will output below trace if LOCKDEP enabled. Use GFP_ATOMIC instead. V2: Rebase. ===================================================== WARNING: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected 5.7.0-rc2 #400 Not tainted ----------------------------------------------------- is trying to acquire: ffffffffb47bea20 (fs_reclaim){+.+.}-{0:0}, at: fs_reclaim_acquire.part.0+0x0/0x30 and this task is already holding: ffff88845b85cc90 (&gvt->scheduler.mmio_context_lock){+.-.}-{2:2}, at: vgpu_mmio_diff_show+0xcf/0x2e0 which would create a new lock dependency: (&gvt->scheduler.mmio_context_lock){+.-.}-{2:2} -> (fs_reclaim){+.+.}-{0:0} but this new dependency connects a SOFTIRQ-irq-safe lock: (&gvt->scheduler.mmio_context_lock){+.-.}-{2:2} ... which became SOFTIRQ-irq-safe at: lock_acquire+0x175/0x4e0 _raw_spin_lock_irqsave+0x2b/0x40 shadow_context_status_change+0xfe/0x2f0 notifier_call_chain+0x6a/0xa0 __atomic_notifier_call_chain+0x5f/0xf0 execlists_schedule_out+0x42a/0x820 process_csb+0xe7/0x3e0 execlists_submission_tasklet+0x5c/0x1d0 tasklet_action_common.isra.0+0xeb/0x260 __do_softirq+0x11d/0x56f irq_exit+0xf6/0x100 do_IRQ+0x7f/0x160 ret_from_intr+0x0/0x2a cpuidle_enter_state+0xcd/0x5b0 cpuidle_enter+0x37/0x60 do_idle+0x337/0x3f0 cpu_startup_entry+0x14/0x20 start_kernel+0x58b/0x5c5 secondary_startup_64+0xa4/0xb0 to a SOFTIRQ-irq-unsafe lock: (fs_reclaim){+.+.}-{0:0} ... which became SOFTIRQ-irq-unsafe at: ... lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_node_trace+0x2e/0x290 alloc_worker+0x2b/0xb0 init_rescuer.part.0+0x17/0xe0 workqueue_init+0x293/0x3bb kernel_init_freeable+0x149/0x325 kernel_init+0x8/0x116 ret_from_fork+0x3a/0x50 other info that might help us debug this: Possible interrupt unsafe locking scenario: CPU0 CPU1 ---- ---- lock(fs_reclaim); local_irq_disable(); lock(&gvt->scheduler.mmio_context_lock); lock(fs_reclaim); lock(&gvt->scheduler.mmio_context_lock); *** DEADLOCK *** 3 locks held by cat/1439: #0: ffff888444a23698 (&p->lock){+.+.}-{3:3}, at: seq_read+0x49/0x680 #1: ffff88845b858068 (&gvt->lock){+.+.}-{3:3}, at: vgpu_mmio_diff_show+0xc7/0x2e0 #2: ffff88845b85cc90 (&gvt->scheduler.mmio_context_lock){+.-.}-{2:2}, at: vgpu_mmio_diff_show+0xcf/0x2e0 the dependencies between SOFTIRQ-irq-safe lock and the holding lock: -> (&gvt->scheduler.mmio_context_lock){+.-.}-{2:2} ops: 31 { HARDIRQ-ON-W at: lock_acquire+0x175/0x4e0 _raw_spin_lock_bh+0x2f/0x40 vgpu_mmio_diff_show+0xcf/0x2e0 seq_read+0x242/0x680 full_proxy_read+0x95/0xc0 vfs_read+0xc2/0x1b0 ksys_read+0xc4/0x160 do_syscall_64+0x63/0x290 entry_SYSCALL_64_after_hwframe+0x49/0xb3 IN-SOFTIRQ-W at: lock_acquire+0x175/0x4e0 _raw_spin_lock_irqsave+0x2b/0x40 shadow_context_status_change+0xfe/0x2f0 notifier_call_chain+0x6a/0xa0 __atomic_notifier_call_chain+0x5f/0xf0 execlists_schedule_out+0x42a/0x820 process_csb+0xe7/0x3e0 execlists_submission_tasklet+0x5c/0x1d0 tasklet_action_common.isra.0+0xeb/0x260 __do_softirq+0x11d/0x56f irq_exit+0xf6/0x100 do_IRQ+0x7f/0x160 ret_from_intr+0x0/0x2a cpuidle_enter_state+0xcd/0x5b0 cpuidle_enter+0x37/0x60 do_idle+0x337/0x3f0 cpu_startup_entry+0x14/0x20 start_kernel+0x58b/0x5c5 secondary_startup_64+0xa4/0xb0 INITIAL USE at: lock_acquire+0x175/0x4e0 _raw_spin_lock_irqsave+0x2b/0x40 shadow_context_status_change+0xfe/0x2f0 notifier_call_chain+0x6a/0xa0 __atomic_notifier_call_chain+0x5f/0xf0 execlists_schedule_in+0x2c8/0x690 __execlists_submission_tasklet+0x1303/0x1930 execlists_submit_request+0x1e7/0x230 submit_notify+0x105/0x2a4 __i915_sw_fence_complete+0xaa/0x380 __engine_park+0x313/0x5a0 ____intel_wakeref_put_last+0x3e/0x90 intel_gt_resume+0x41e/0x440 intel_gt_init+0x283/0xbc0 i915_gem_init+0x197/0x240 i915_driver_probe+0xc2d/0x12e0 i915_pci_probe+0xa2/0x1e0 local_pci_probe+0x6f/0xb0 pci_device_probe+0x171/0x230 really_probe+0x17a/0x380 driver_probe_device+0x70/0xf0 device_driver_attach+0x82/0x90 __driver_attach+0x60/0x100 bus_for_each_dev+0xe4/0x140 bus_add_driver+0x257/0x2a0 driver_register+0xd3/0x150 i915_init+0x6d/0x80 do_one_initcall+0xb8/0x3a0 kernel_init_freeable+0x2b4/0x325 kernel_init+0x8/0x116 ret_from_fork+0x3a/0x50 } __key.77812+0x0/0x40 ... acquired at: lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_trace+0x2e/0x260 mmio_diff_handler+0xc0/0x150 intel_gvt_for_each_tracked_mmio+0x7b/0x140 vgpu_mmio_diff_show+0x111/0x2e0 seq_read+0x242/0x680 full_proxy_read+0x95/0xc0 vfs_read+0xc2/0x1b0 ksys_read+0xc4/0x160 do_syscall_64+0x63/0x290 entry_SYSCALL_64_after_hwframe+0x49/0xb3 the dependencies between the lock to be acquired and SOFTIRQ-irq-unsafe lock: -> (fs_reclaim){+.+.}-{0:0} ops: 1999031 { HARDIRQ-ON-W at: lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_node_trace+0x2e/0x290 alloc_worker+0x2b/0xb0 init_rescuer.part.0+0x17/0xe0 workqueue_init+0x293/0x3bb kernel_init_freeable+0x149/0x325 kernel_init+0x8/0x116 ret_from_fork+0x3a/0x50 SOFTIRQ-ON-W at: lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_node_trace+0x2e/0x290 alloc_worker+0x2b/0xb0 init_rescuer.part.0+0x17/0xe0 workqueue_init+0x293/0x3bb kernel_init_freeable+0x149/0x325 kernel_init+0x8/0x116 ret_from_fork+0x3a/0x50 INITIAL USE at: lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_node_trace+0x2e/0x290 alloc_worker+0x2b/0xb0 init_rescuer.part.0+0x17/0xe0 workqueue_init+0x293/0x3bb kernel_init_freeable+0x149/0x325 kernel_init+0x8/0x116 ret_from_fork+0x3a/0x50 } __fs_reclaim_map+0x0/0x60 ... acquired at: lock_acquire+0x175/0x4e0 fs_reclaim_acquire.part.0+0x20/0x30 kmem_cache_alloc_trace+0x2e/0x260 mmio_diff_handler+0xc0/0x150 intel_gvt_for_each_tracked_mmio+0x7b/0x140 vgpu_mmio_diff_show+0x111/0x2e0 seq_read+0x242/0x680 full_proxy_read+0x95/0xc0 vfs_read+0xc2/0x1b0 ksys_read+0xc4/0x160 do_syscall_64+0x63/0x290 entry_SYSCALL_64_after_hwframe+0x49/0xb3 stack backtrace: CPU: 5 PID: 1439 Comm: cat Not tainted 5.7.0-rc2 #400 Hardware name: Intel(R) Client Systems NUC8i7BEH/NUC8BEB, BIOS BECFL357.86A.0056.2018.1128.1717 11/28/2018 Call Trace: dump_stack+0x97/0xe0 check_irq_usage.cold+0x428/0x434 ? check_usage_forwards+0x2c0/0x2c0 ? class_equal+0x11/0x20 ? __bfs+0xd2/0x2d0 ? in_any_class_list+0xa0/0xa0 ? check_path+0x22/0x40 ? check_noncircular+0x150/0x2b0 ? print_circular_bug.isra.0+0x1b0/0x1b0 ? mark_lock+0x13d/0xc50 ? __lock_acquire+0x1e32/0x39b0 __lock_acquire+0x1e32/0x39b0 ? timerqueue_add+0xc1/0x130 ? register_lock_class+0xa60/0xa60 ? mark_lock+0x13d/0xc50 lock_acquire+0x175/0x4e0 ? __zone_pcp_update+0x80/0x80 ? check_flags.part.0+0x210/0x210 ? mark_held_locks+0x65/0x90 ? _raw_spin_unlock_irqrestore+0x32/0x40 ? lockdep_hardirqs_on+0x190/0x290 ? fwtable_read32+0x163/0x480 ? mmio_diff_handler+0xc0/0x150 fs_reclaim_acquire.part.0+0x20/0x30 ? __zone_pcp_update+0x80/0x80 kmem_cache_alloc_trace+0x2e/0x260 mmio_diff_handler+0xc0/0x150 ? vgpu_mmio_diff_open+0x30/0x30 intel_gvt_for_each_tracked_mmio+0x7b/0x140 vgpu_mmio_diff_show+0x111/0x2e0 ? mmio_diff_handler+0x150/0x150 ? rcu_read_lock_sched_held+0xa0/0xb0 ? rcu_read_lock_bh_held+0xc0/0xc0 ? kasan_unpoison_shadow+0x33/0x40 ? __kasan_kmalloc.constprop.0+0xc2/0xd0 seq_read+0x242/0x680 ? debugfs_locked_down.isra.0+0x70/0x70 full_proxy_read+0x95/0xc0 vfs_read+0xc2/0x1b0 ksys_read+0xc4/0x160 ? kernel_write+0xb0/0xb0 ? mark_held_locks+0x24/0x90 do_syscall_64+0x63/0x290 entry_SYSCALL_64_after_hwframe+0x49/0xb3 RIP: 0033:0x7ffbe3e6efb2 Code: c0 e9 c2 fe ff ff 50 48 8d 3d ca cb 0a 00 e8 f5 19 02 00 0f 1f 44 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 0f 05 <48> 3d 00 f0 ff ff 77 56 c3 0f 1f 44 00 00 48 83 ec 28 48 89 54 24 RSP: 002b:00007ffd021c08a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 RAX: ffffffffffffffda RBX: 0000000000020000 RCX: 00007ffbe3e6efb2 RDX: 0000000000020000 RSI: 00007ffbe34cd000 RDI: 0000000000000003 RBP: 00007ffbe34cd000 R08: 00007ffbe34cc010 R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: 0000562b6f0a11f0 R13: 0000000000000003 R14: 0000000000020000 R15: 0000000000020000 ------------[ cut here ]------------ Acked-by: Zhenyu Wang Signed-off-by: Colin Xu Signed-off-by: Zhenyu Wang Link: http://patchwork.freedesktop.org/patch/msgid/20200601035556.19999-1-colin.xu@intel.com --- drivers/gpu/drm/i915/gvt/debugfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/debugfs.c b/drivers/gpu/drm/i915/gvt/debugfs.c index ec47d4114554..62e6a14ad58e 100644 --- a/drivers/gpu/drm/i915/gvt/debugfs.c +++ b/drivers/gpu/drm/i915/gvt/debugfs.c @@ -66,7 +66,7 @@ static inline int mmio_diff_handler(struct intel_gvt *gvt, vreg = vgpu_vreg(param->vgpu, offset); if (preg != vreg) { - node = kmalloc(sizeof(*node), GFP_KERNEL); + node = kmalloc(sizeof(*node), GFP_ATOMIC); if (!node) return -ENOMEM; -- cgit v1.2.3 From 55fd7e0222ea01246ef3e6aae28b5721fdfb790f Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Wed, 10 Jun 2020 13:18:07 -0700 Subject: drm/i915: Include asm sources for {ivb, hsw}_clear_kernel.c Alexandre Oliva has recently removed these files from Linux Libre with concerns that the sources weren't available. The sources are available on IGT repository, and only open source tools are used to generate the {ivb,hsw}_clear_kernel.c files. However, the remaining concern from Alexandre Oliva was around GPL license and the source not been present when distributing the code. So, it looks like 2 alternatives are possible, the use of linux-firmware.git repository to store the blob or making sure that the source is also present in our tree. Since the goal is to limit the i915 firmware to only the micro-controller blobs let's make sure that we do include the asm sources here in our tree. Btw, I tried to have some diligence here and make sure that the asms that these commits are adding are truly the source for the mentioned files: igt$ ./scripts/generate_clear_kernel.sh -g ivb \ -m ~/mesa/build/src/intel/tools/i965_asm Output file not specified - using default file "ivb-cb_assembled" Generating gen7 CB Kernel assembled file "ivb_clear_kernel.c" for i915 driver... igt$ diff ~/i915/drm-tip/drivers/gpu/drm/i915/gt/ivb_clear_kernel.c \ ivb_clear_kernel.c < * Generated by: IGT Gpu Tools on Fri 21 Feb 2020 05:29:32 AM UTC > * Generated by: IGT Gpu Tools on Mon 08 Jun 2020 10:00:54 AM PDT 61c61 < }; > }; \ No newline at end of file igt$ ./scripts/generate_clear_kernel.sh -g hsw \ -m ~/mesa/build/src/intel/tools/i965_asm Output file not specified - using default file "hsw-cb_assembled" Generating gen7.5 CB Kernel assembled file "hsw_clear_kernel.c" for i915 driver... igt$ diff ~/i915/drm-tip/drivers/gpu/drm/i915/gt/hsw_clear_kernel.c \ hsw_clear_kernel.c 5c5 < * Generated by: IGT Gpu Tools on Fri 21 Feb 2020 05:30:13 AM UTC > * Generated by: IGT Gpu Tools on Mon 08 Jun 2020 10:01:42 AM PDT 61c61 < }; > }; \ No newline at end of file Used IGT and Mesa master repositories from Fri Jun 5 2020) IGT: 53e8c878a6fb ("tests/kms_chamelium: Force reprobe after replugging the connector") Mesa: 5d13c7477eb1 ("radv: set keep_statistic_info with RADV_DEBUG=shaderstats") Mesa built with: meson build -D platforms=drm,x11 -D dri-drivers=i965 \ -D gallium-drivers=iris -D prefix=/usr \ -D libdir=/usr/lib64/ -Dtools=intel \ -Dkulkan-drivers=intel && ninja -C build v2: Header clean-up and include build instructions in a readme (Chris) Modified commit message to respect check-patch Reference: http://www.fsfla.org/pipermail/linux-libre/2020-June/003374.html Reference: http://www.fsfla.org/pipermail/linux-libre/2020-June/003375.html Fixes: 47f8253d2b89 ("drm/i915/gen7: Clear all EU/L3 residual contexts") Cc: # v5.7+ Cc: Alexandre Oliva Cc: Prathap Kumar Valsan Cc: Akeem G Abodunrin Cc: Mika Kuoppala Cc: Chris Wilson Cc: Jani Nikula Cc: Joonas Lahtinen Signed-off-by: Rodrigo Vivi Reviewed-by: Jon Bloomfield Link: https://patchwork.freedesktop.org/patch/msgid/20200610201807.191440-1-rodrigo.vivi@intel.com (cherry picked from commit 5a7eeb8ba143d860050ecea924a8f074f02d8023) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/gt/shaders/README | 46 ++++++++ .../gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm | 119 +++++++++++++++++++++ .../gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm | 117 ++++++++++++++++++++ 3 files changed, 282 insertions(+) create mode 100644 drivers/gpu/drm/i915/gt/shaders/README create mode 100644 drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm create mode 100644 drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm diff --git a/drivers/gpu/drm/i915/gt/shaders/README b/drivers/gpu/drm/i915/gt/shaders/README new file mode 100644 index 000000000000..e7e96d7073c7 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/README @@ -0,0 +1,46 @@ +ASM sources for auto generated shaders +====================================== + +The i915/gt/hsw_clear_kernel.c and i915/gt/ivb_clear_kernel.c files contain +pre-compiled batch chunks that will clear any residual render cache during +context switch. + +They are generated from their respective platform ASM files present on +i915/gt/shaders/clear_kernel directory. + +The generated .c files should never be modified directly. Instead, any modification +needs to be done on the on their respective ASM files and build instructions below +needes to be followed. + +Building +======== + +Environment +----------- + +IGT GPU tool scripts and the Mesa's i965 instruction assembler tool are used +on building. + +Please make sure your Mesa tool is compiled with "-Dtools=intel" and +"-Ddri-drivers=i965", and run this script from IGT source root directory" + +The instructions bellow assume: + * IGT gpu tools source code is located on your home directory (~) as ~/igt + * Mesa source code is located on your home directory (~) as ~/mesa + and built under the ~/mesa/build directory + * Linux kernel source code is under your home directory (~) as ~/linux + +Instructions +------------ + +~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm \ + ~/igt/lib/i915/shaders/clear_kernel/ivb.asm +~ $ cd ~/igt +igt $ ./scripts/generate_clear_kernel.sh -g ivb \ + -m ~/mesa/build/src/intel/tools/i965_asm + +~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm \ + ~/igt/lib/i915/shaders/clear_kernel/hsw.asm +~ $ cd ~/igt +igt $ ./scripts/generate_clear_kernel.sh -g hsw \ + -m ~/mesa/build/src/intel/tools/i965_asm \ No newline at end of file diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm new file mode 100644 index 000000000000..5fdf384bb621 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm @@ -0,0 +1,119 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright © 2020 Intel Corporation + */ + +/* + * Kernel for PAVP buffer clear. + * + * 1. Clear all 64 GRF registers assigned to the kernel with designated value; + * 2. Write 32x16 block of all "0" to render target buffer which indirectly clears + * 512 bytes of Render Cache. + */ + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/** + * Curbe Format + * + * DW 1.0 - Block Offset to write Render Cache + * DW 1.1 [15:0] - Clear Word + * DW 1.2 - Delay iterations + * DW 1.3 - Enable Instrumentation (only for debug) + * DW 1.4 - Rsvd (intended for context ID) + * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount + * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count) + * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count) + * + * Binding Table + * + * BTI 0: 2D Surface to help clear L3 (Render/Data Cache) + * BTI 1: Wait/Instrumentation Buffer + * Size : (SliceCount * SubSliceCount * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT) + * Expected to be initialized to 0 by driver/another kernel + * Layout: + * RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS] + * Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N + */ +add(1) g1.2<1>UD g1.2<0,1,0>UD 0x00000001UD { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */ +cmp.z.f0.0(1) null<1>UD g1.3<0,1,0>UD 0x00000000UD { align1 1N }; +(+f0.0) jmpi(1) 352D { align1 WE_all 1N }; + +/** + * State Register has info on where this thread is running + * IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + * HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + */ +mov(8) g3<1>UD 0x00000000UD { align1 1Q }; +shr(1) g3<1>D sr0<0,1,0>D 12D { align1 1N }; +and(1) g3<1>D g3<0,1,0>D 1D { align1 1N }; /* g3 has HSID */ +shr(1) g3.1<1>D sr0<0,1,0>D 13D { align1 1N }; +and(1) g3.1<1>D g3.1<0,1,0>D 3D { align1 1N }; /* g3.1 has sliceID */ +mul(1) g3.5<1>D g3.1<0,1,0>D g1.10<0,1,0>UW { align1 1N }; +add(1) g3<1>D g3<0,1,0>D g3.5<0,1,0>D { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */ +shr(1) g3.2<1>D sr0<0,1,0>D 8D { align1 1N }; +and(1) g3.2<1>D g3.2<0,1,0>D 15D { align1 1N }; /* g3.2 = EUID */ +mul(1) g3.4<1>D g3<0,1,0>D 16D { align1 1N }; +add(1) g3.2<1>D g3.2<0,1,0>D g3.4<0,1,0>D { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address ) in instrumentation surf */ + +mov(8) g5<1>UD 0x00000000UD { align1 1Q }; +and(1) g3.3<1>D sr0<0,1,0>D 7D { align1 1N }; +mul(1) g3.3<1>D g3.3<0,1,0>D 4D { align1 1N }; + +mov(8) g4<1>UD g0<8,8,1>UD { align1 1Q }; /* Initialize message header with g0 */ +mov(1) g4<1>UD g3.3<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.1<1>UD g3.2<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.2<1>UD 0x00000003UD { align1 1N }; /* Block size (1 row x 4 bytes) */ +and(1) g4.3<1>UD g4.3<0,1,0>UW 0xffffffffUD { align1 1N }; + +/* Media block read to fetch current value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x02190001 + + render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q }; +add(1) g5<1>D g5<0,1,0>D 1D { align1 1N }; + +/* Media block write for updated value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x040a8001 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q }; + +/* Delay thread for specified parameter */ +add.nz.f0.0(1) g1.2<1>UD g1.2<0,1,0>UD -1D { align1 1N }; +(+f0.0) jmpi(1) -32D { align1 WE_all 1N }; + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/* Initialize looping parameters */ +mov(1) a0<1>D 0D { align1 1N }; /* Initialize a0.0:w=0 */ +mov(1) a0.4<1>W 127W { align1 1N }; /* Loop count. Each loop contains 16 GRF's */ + +/* Write 32x16 all "0" block */ +mov(8) g2<1>UD g0<8,8,1>UD { align1 1Q }; +mov(8) g127<1>UD g0<8,8,1>UD { align1 1Q }; +mov(2) g2<1>UD g1<2,2,1>UW { align1 1N }; +mov(1) g2.2<1>UD 0x000f000fUD { align1 1N }; /* Block size (16x16) */ +and(1) g2.3<1>UD g2.3<0,1,0>UW 0xffffffefUD { align1 1N }; +mov(16) g3<1>UD 0x00000000UD { align1 1H }; +mov(16) g4<1>UD 0x00000000UD { align1 1H }; +mov(16) g5<1>UD 0x00000000UD { align1 1H }; +mov(16) g6<1>UD 0x00000000UD { align1 1H }; +mov(16) g7<1>UD 0x00000000UD { align1 1H }; +mov(16) g8<1>UD 0x00000000UD { align1 1H }; +mov(16) g9<1>UD 0x00000000UD { align1 1H }; +mov(16) g10<1>UD 0x00000000UD { align1 1H }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; +add(1) g2<1>UD g1<0,1,0>UW 0x0010UW { align1 1N }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; + +/* Now, clear all GRF registers */ +add.nz.f0.0(1) a0.4<1>W a0.4<0,1,0>W -1W { align1 1N }; +mov(16) g[a0]<1>UW f0.1<0,1,0>UW { align1 1H }; +add(1) a0<1>D a0<0,1,0>D 32D { align1 1N }; +(+f0.0) jmpi(1) -64D { align1 WE_all 1N }; + +/* Terminante the thread */ +sendc(8) null<1>UD g127<8,8,1>F 0x82000010 + thread_spawner MsgDesc: mlen 1 rlen 0 { align1 1Q EOT }; diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm new file mode 100644 index 000000000000..97c7ac9e3854 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm @@ -0,0 +1,117 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright © 2020 Intel Corporation + */ + +/* + * Kernel for PAVP buffer clear. + * + * 1. Clear all 64 GRF registers assigned to the kernel with designated value; + * 2. Write 32x16 block of all "0" to render target buffer which indirectly clears + * 512 bytes of Render Cache. + */ + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/** + * Curbe Format + * + * DW 1.0 - Block Offset to write Render Cache + * DW 1.1 [15:0] - Clear Word + * DW 1.2 - Delay iterations + * DW 1.3 - Enable Instrumentation (only for debug) + * DW 1.4 - Rsvd (intended for context ID) + * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount + * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count) + * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count) + * + * Binding Table + * + * BTI 0: 2D Surface to help clear L3 (Render/Data Cache) + * BTI 1: Wait/Instrumentation Buffer + * Size : (SliceCount * SubSliceCount * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT) + * Expected to be initialized to 0 by driver/another kernel + * Layout : + * RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS] + * Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N + */ +add(1) g1.2<1>UD g1.2<0,1,0>UD 0x00000001UD { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */ +cmp.z.f0.0(1) null<1>UD g1.3<0,1,0>UD 0x00000000UD { align1 1N }; +(+f0.0) jmpi(1) 44D { align1 WE_all 1N }; + +/** + * State Register has info on where this thread is running + * IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + * HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + */ +mov(8) g3<1>UD 0x00000000UD { align1 1Q }; +shr(1) g3<1>D sr0<0,1,0>D 12D { align1 1N }; +and(1) g3<1>D g3<0,1,0>D 1D { align1 1N }; /* g3 has HSID */ +shr(1) g3.1<1>D sr0<0,1,0>D 13D { align1 1N }; +and(1) g3.1<1>D g3.1<0,1,0>D 3D { align1 1N }; /* g3.1 has sliceID */ +mul(1) g3.5<1>D g3.1<0,1,0>D g1.10<0,1,0>UW { align1 1N }; +add(1) g3<1>D g3<0,1,0>D g3.5<0,1,0>D { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */ +shr(1) g3.2<1>D sr0<0,1,0>D 8D { align1 1N }; +and(1) g3.2<1>D g3.2<0,1,0>D 15D { align1 1N }; /* g3.2 = EUID */ +mul(1) g3.4<1>D g3<0,1,0>D 16D { align1 1N }; +add(1) g3.2<1>D g3.2<0,1,0>D g3.4<0,1,0>D { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address ) in instrumentation surf */ + +mov(8) g5<1>UD 0x00000000UD { align1 1Q }; +and(1) g3.3<1>D sr0<0,1,0>D 7D { align1 1N }; +mul(1) g3.3<1>D g3.3<0,1,0>D 4D { align1 1N }; + +mov(8) g4<1>UD g0<8,8,1>UD { align1 1Q }; /* Initialize message header with g0 */ +mov(1) g4<1>UD g3.3<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.1<1>UD g3.2<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.2<1>UD 0x00000003UD { align1 1N }; /* Block size (1 row x 4 bytes) */ +and(1) g4.3<1>UD g4.3<0,1,0>UW 0xffffffffUD { align1 1N }; + +/* Media block read to fetch current value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x02190001 + render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q }; +add(1) g5<1>D g5<0,1,0>D 1D { align1 1N }; + +/* Media block write for updated value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x040a8001 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q }; +/* Delay thread for specified parameter */ +add.nz.f0.0(1) g1.2<1>UD g1.2<0,1,0>UD -1D { align1 1N }; +(+f0.0) jmpi(1) -4D { align1 WE_all 1N }; + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/* Initialize looping parameters */ +mov(1) a0<1>D 0D { align1 1N }; /* Initialize a0.0:w=0 */ +mov(1) a0.4<1>W 127W { align1 1N }; /* Loop count. Each loop contains 16 GRF's */ + +/* Write 32x16 all "0" block */ +mov(8) g2<1>UD g0<8,8,1>UD { align1 1Q }; +mov(8) g127<1>UD g0<8,8,1>UD { align1 1Q }; +mov(2) g2<1>UD g1<2,2,1>UW { align1 1N }; +mov(1) g2.2<1>UD 0x000f000fUD { align1 1N }; /* Block size (16x16) */ +and(1) g2.3<1>UD g2.3<0,1,0>UW 0xffffffefUD { align1 1N }; +mov(16) g3<1>UD 0x00000000UD { align1 1H }; +mov(16) g4<1>UD 0x00000000UD { align1 1H }; +mov(16) g5<1>UD 0x00000000UD { align1 1H }; +mov(16) g6<1>UD 0x00000000UD { align1 1H }; +mov(16) g7<1>UD 0x00000000UD { align1 1H }; +mov(16) g8<1>UD 0x00000000UD { align1 1H }; +mov(16) g9<1>UD 0x00000000UD { align1 1H }; +mov(16) g10<1>UD 0x00000000UD { align1 1H }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; +add(1) g2<1>UD g1<0,1,0>UW 0x0010UW { align1 1N }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; + +/* Now, clear all GRF registers */ +add.nz.f0.0(1) a0.4<1>W a0.4<0,1,0>W -1W { align1 1N }; +mov(16) g[a0]<1>UW f0.1<0,1,0>UW { align1 1H }; +add(1) a0<1>D a0<0,1,0>D 32D { align1 1N }; +(+f0.0) jmpi(1) -8D { align1 WE_all 1N }; + +/* Terminante the thread */ +sendc(8) null<1>UD g127<8,8,1>F 0x82000010 + thread_spawner MsgDesc: mlen 1 rlen 0 { align1 1Q EOT }; -- cgit v1.2.3