The patch below does not apply to the 6.6-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to stable@vger.kernel.org.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y git checkout FETCH_HEAD git cherry-pick -x 9c33e5fd4fb63b793d9a92bf35d190630d9bada4 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to 'stable@vger.kernel.org' --in-reply-to '2024080738-fasting-raving-635a@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^..

Possible dependencies:

9c33e5fd4fb6 ("drm/amdgpu: fix locking scope when flushing tlb") 08abccc9a7a7 ("drm/amdgpu: further move TLB hw workarounds a layer up") e2e3788850b9 ("drm/amdgpu: rework lock handling for flush_tlb v2") 3983c9fd2d8b ("drm/amdgpu: drop error return from flush_gpu_tlb_pasid") 041a5743883d ("drm/amdgpu: fix and cleanup gmc_v11_0_flush_gpu_tlb_pasid") 72cc99205c0b ("drm/amdgpu: cleanup gmc_v10_0_flush_gpu_tlb_pasid") e7b90e99fa8f ("drm/amdgpu: fix and cleanup gmc_v9_0_flush_gpu_tlb_pasid") 0c525aa40649 ("drm/amdgpu: fix and cleanup gmc_v8_0_flush_gpu_tlb_pasid") fb4c52db6974 ("drm/amdgpu: fix and cleanup gmc_v7_0_flush_gpu_tlb_pasid") a70cb2176f7e ("drm/amdgpu: rework gmc_v10_0_flush_gpu_tlb v2") 24a6eb92b7f6 ("drm/amdgpu: fix and cleanup gmc_v9_0_flush_gpu_tlb") 4e8303cf2c4d ("drm/amdgpu: Use function for IP version check") 6b7d211740da ("drm/amdgpu: Fix refclk reporting for SMU v13.0.6") 1b8e56b99459 ("drm/amdgpu: Restrict bootloader wait to SMUv13.0.6") 983ac45a06ae ("drm/amdgpu: update SET_HW_RESOURCES definition for UMSCH") 822f7808291f ("drm/amdgpu/discovery: enable UMSCH 4.0 in IP discovery") 3488c79beafa ("drm/amdgpu: add initial support for UMSCH") 2da1b04a2096 ("drm/amdgpu: add UMSCH 4.0 api definition") 3ee8fb7005ef ("drm/amdgpu: enable VPE for VPE 6.1.0") 9d4346bdbc64 ("drm/amdgpu: add VPE 6.1.0 support")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 9c33e5fd4fb63b793d9a92bf35d190630d9bada4 Mon Sep 17 00:00:00 2001 From: Yunxiang Li Yunxiang.Li@amd.com Date: Thu, 23 May 2024 07:48:19 -0400 Subject: [PATCH] drm/amdgpu: fix locking scope when flushing tlb MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit

Which method is used to flush tlb does not depend on whether a reset is in progress or not. We should skip flush altogether if the GPU will get reset. So put both path under reset_domain read lock.

Signed-off-by: Yunxiang Li Yunxiang.Li@amd.com Reviewed-by: Christian König christian.koenig@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com CC: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 660599823050..322b8ff67cde 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -682,12 +682,17 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, struct amdgpu_ring *ring = &adev->gfx.kiq[inst].ring; struct amdgpu_kiq *kiq = &adev->gfx.kiq[inst]; unsigned int ndw; - signed long r; + int r; uint32_t seq;

- if (!adev->gmc.flush_pasid_uses_kiq || !ring->sched.ready || - !down_read_trylock(&adev->reset_domain->sem)) { + /* + * A GPU reset should flush all TLBs anyway, so no need to do + * this while one is ongoing. + */ + if (!down_read_trylock(&adev->reset_domain->sem)) + return 0;

+ if (!adev->gmc.flush_pasid_uses_kiq || !ring->sched.ready) { if (adev->gmc.flush_tlb_needs_extra_type_2) adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid, 2, all_hub, @@ -701,44 +706,41 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub, inst); - return 0; - } + r = 0; + } else { + /* 2 dwords flush + 8 dwords fence */ + ndw = kiq->pmf->invalidate_tlbs_size + 8;

- /* 2 dwords flush + 8 dwords fence */ - ndw = kiq->pmf->invalidate_tlbs_size + 8; + if (adev->gmc.flush_tlb_needs_extra_type_2) + ndw += kiq->pmf->invalidate_tlbs_size;

- if (adev->gmc.flush_tlb_needs_extra_type_2) - ndw += kiq->pmf->invalidate_tlbs_size; + if (adev->gmc.flush_tlb_needs_extra_type_0) + ndw += kiq->pmf->invalidate_tlbs_size;

- if (adev->gmc.flush_tlb_needs_extra_type_0) - ndw += kiq->pmf->invalidate_tlbs_size; + spin_lock(&adev->gfx.kiq[inst].ring_lock); + amdgpu_ring_alloc(ring, ndw); + if (adev->gmc.flush_tlb_needs_extra_type_2) + kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 2, all_hub);

- spin_lock(&adev->gfx.kiq[inst].ring_lock); - amdgpu_ring_alloc(ring, ndw); - if (adev->gmc.flush_tlb_needs_extra_type_2) - kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 2, all_hub); + if (flush_type == 2 && adev->gmc.flush_tlb_needs_extra_type_0) + kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 0, all_hub);

- if (flush_type == 2 && adev->gmc.flush_tlb_needs_extra_type_0) - kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 0, all_hub); + kiq->pmf->kiq_invalidate_tlbs(ring, pasid, flush_type, all_hub); + r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT); + if (r) { + amdgpu_ring_undo(ring); + spin_unlock(&adev->gfx.kiq[inst].ring_lock); + goto error_unlock_reset; + }

- kiq->pmf->kiq_invalidate_tlbs(ring, pasid, flush_type, all_hub); - r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT); - if (r) { - amdgpu_ring_undo(ring); + amdgpu_ring_commit(ring); spin_unlock(&adev->gfx.kiq[inst].ring_lock); - goto error_unlock_reset; + if (amdgpu_fence_wait_polling(ring, seq, usec_timeout) < 1) { + dev_err(adev->dev, "timeout waiting for kiq fence\n"); + r = -ETIME; + } }

- amdgpu_ring_commit(ring); - spin_unlock(&adev->gfx.kiq[inst].ring_lock); - r = amdgpu_fence_wait_polling(ring, seq, usec_timeout); - if (r < 1) { - dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r); - r = -ETIME; - goto error_unlock_reset; - } - r = 0; - error_unlock_reset: up_read(&adev->reset_domain->sem); return r;