From: Hawking Zhang Hawking.Zhang@amd.com
Initialize vcn ras block only when vcn ip block supports ras features. Driver queries ras capabilities after early_init, ras block init needs to be moved to sw_int.
Signed-off-by: Hawking Zhang Hawking.Zhang@amd.com Reviewed-by: Stanley Yang Stanley.Yang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit f81c31d975b463c24506d817a48390621f057a57) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 29 ++++++++++++++++--------- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 6 +++-- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 6 +++-- 4 files changed, 28 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index e7974de8b035..e63fcc58e8e0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -1181,19 +1181,28 @@ int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, return 0; }
-void amdgpu_vcn_set_ras_funcs(struct amdgpu_device *adev) +int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev) { + int err; + struct amdgpu_vcn_ras *ras; + if (!adev->vcn.ras) - return; + return 0;
- amdgpu_ras_register_ras_block(adev, &adev->vcn.ras->ras_block); + ras = adev->vcn.ras; + err = amdgpu_ras_register_ras_block(adev, &ras->ras_block); + if (err) { + dev_err(adev->dev, "Failed to register vcn ras block!\n"); + return err; + }
- strcpy(adev->vcn.ras->ras_block.ras_comm.name, "vcn"); - adev->vcn.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__VCN; - adev->vcn.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON; - adev->vcn.ras_if = &adev->vcn.ras->ras_block.ras_comm; + strcpy(ras->ras_block.ras_comm.name, "vcn"); + ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__VCN; + ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON; + adev->vcn.ras_if = &ras->ras_block.ras_comm;
- /* If don't define special ras_late_init function, use default ras_late_init */ - if (!adev->vcn.ras->ras_block.ras_late_init) - adev->vcn.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + if (!ras->ras_block.ras_late_init) + ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index d3e2af902907..c730949ece7d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -400,6 +400,6 @@ void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); -void amdgpu_vcn_set_ras_funcs(struct amdgpu_device *adev); +int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev);
#endif diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index b0b0e69c6a94..223e7dfe4618 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -225,6 +225,10 @@ static int vcn_v2_5_sw_init(void *handle) if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) adev->vcn.pause_dpg_mode = vcn_v2_5_pause_dpg_mode;
+ r = amdgpu_vcn_ras_sw_init(adev); + if (r) + return r; + return 0; }
@@ -2031,6 +2035,4 @@ static void vcn_v2_5_set_ras_funcs(struct amdgpu_device *adev) default: break; } - - amdgpu_vcn_set_ras_funcs(adev); } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 43d587404c3e..720ab36f9c92 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -181,6 +181,10 @@ static int vcn_v4_0_sw_init(void *handle) if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) adev->vcn.pause_dpg_mode = vcn_v4_0_pause_dpg_mode;
+ r = amdgpu_vcn_ras_sw_init(adev); + if (r) + return r; + return 0; }
@@ -2123,6 +2127,4 @@ static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev) default: break; } - - amdgpu_vcn_set_ras_funcs(adev); }
From: Horatio Zhang Hongkun.Zhang@amd.com
Separate vcn RAS poison consumption handling from the instance irq, and register dedicated ras_poison_irq src and funcs for UVD_POISON.
v2: - Separate ras irq from vcn instance irq - Improve the subject and code comments
v3: - Split the patch into three parts - Improve the code comments
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit ac1d8e2f074d9bffc2d368ad0720cdbb4c938fa5) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 27 ++++++++++++++++++++++++- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 3 +++ 2 files changed, 29 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index e63fcc58e8e0..2d94f1b63bd6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -1181,6 +1181,31 @@ int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, return 0; }
+int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block) +{ + int r, i; + + r = amdgpu_ras_block_late_init(adev, ras_block); + if (r) + return r; + + if (amdgpu_ras_is_supported(adev, ras_block->block)) { + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + + r = amdgpu_irq_get(adev, &adev->vcn.inst[i].ras_poison_irq, 0); + if (r) + goto late_fini; + } + } + return 0; + +late_fini: + amdgpu_ras_block_late_fini(adev, ras_block); + return r; +} + int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev) { int err; @@ -1202,7 +1227,7 @@ int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev) adev->vcn.ras_if = &ras->ras_block.ras_comm;
if (!ras->ras_block.ras_late_init) - ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + ras->ras_block.ras_late_init = amdgpu_vcn_ras_late_init;
return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index c730949ece7d..f1397ef66fd7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -234,6 +234,7 @@ struct amdgpu_vcn_inst { struct amdgpu_ring ring_enc[AMDGPU_VCN_MAX_ENC_RINGS]; atomic_t sched_score; struct amdgpu_irq_src irq; + struct amdgpu_irq_src ras_poison_irq; struct amdgpu_vcn_reg external; struct amdgpu_bo *dpg_sram_bo; struct dpg_pause_state pause_state; @@ -400,6 +401,8 @@ void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); +int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev, + struct ras_common_if *ras_block); int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev);
#endif
From: Horatio Zhang Hongkun.Zhang@amd.com
Add ras_poison_irq and functions.
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit 6889f28c736c357700f5755fed852a2badc15a7b) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index 223e7dfe4618..a878fd2c3133 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -143,7 +143,7 @@ static int vcn_v2_5_sw_init(void *handle)
/* VCN POISON TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j], - VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].irq); + VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].ras_poison_irq); if (r) return r; } @@ -343,6 +343,9 @@ static int vcn_v2_5_hw_fini(void *handle) (adev->vcn.cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, i, mmUVD_STATUS))) vcn_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE); + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) + amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0); }
return 0; @@ -1865,6 +1868,14 @@ static int vcn_v2_5_set_interrupt_state(struct amdgpu_device *adev, return 0; }
+static int vcn_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -1895,9 +1906,6 @@ static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev, case VCN_2_0__SRCID__UVD_ENC_LOW_LATENCY: amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[1]); break; - case VCN_2_6__SRCID_UVD_POISON: - amdgpu_vcn_process_poison_irq(adev, source, entry); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -1912,6 +1920,11 @@ static const struct amdgpu_irq_src_funcs vcn_v2_5_irq_funcs = { .process = vcn_v2_5_process_interrupt, };
+static const struct amdgpu_irq_src_funcs vcn_v2_6_ras_irq_funcs = { + .set = vcn_v2_6_set_ras_interrupt_state, + .process = amdgpu_vcn_process_poison_irq, +}; + static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev) { int i; @@ -1921,6 +1934,9 @@ static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev) continue; adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v2_5_irq_funcs; + + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v2_6_ras_irq_funcs; } }
@@ -2023,6 +2039,7 @@ const struct amdgpu_ras_block_hw_ops vcn_v2_6_ras_hw_ops = { static struct amdgpu_vcn_ras vcn_v2_6_ras = { .ras_block = { .hw_ops = &vcn_v2_6_ras_hw_ops, + .ras_late_init = amdgpu_vcn_ras_late_init, }, };
From: Horatio Zhang Hongkun.Zhang@amd.com
Add ras_poison_irq and functions. And fix the amdgpu_irq_put call trace in vcn_v4_0_hw_fini.
[ 44.563572] RIP: 0010:amdgpu_irq_put+0xa4/0xc0 [amdgpu] [ 44.563629] RSP: 0018:ffffb36740edfc90 EFLAGS: 00010246 [ 44.563630] RAX: 0000000000000000 RBX: 0000000000000001 RCX: 0000000000000000 [ 44.563630] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [ 44.563631] RBP: ffffb36740edfcb0 R08: 0000000000000000 R09: 0000000000000000 [ 44.563631] R10: 0000000000000000 R11: 0000000000000000 R12: ffff954c568e2ea8 [ 44.563631] R13: 0000000000000000 R14: ffff954c568c0000 R15: ffff954c568e2ea8 [ 44.563632] FS: 0000000000000000(0000) GS:ffff954f584c0000(0000) knlGS:0000000000000000 [ 44.563632] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 44.563633] CR2: 00007f028741ba70 CR3: 000000026ca10000 CR4: 0000000000750ee0 [ 44.563633] PKRU: 55555554 [ 44.563633] Call Trace: [ 44.563634] <TASK> [ 44.563634] vcn_v4_0_hw_fini+0x62/0x160 [amdgpu] [ 44.563700] vcn_v4_0_suspend+0x13/0x30 [amdgpu] [ 44.563755] amdgpu_device_ip_suspend_phase2+0x240/0x470 [amdgpu] [ 44.563806] amdgpu_device_ip_suspend+0x41/0x80 [amdgpu] [ 44.563858] amdgpu_device_pre_asic_reset+0xd9/0x4a0 [amdgpu] [ 44.563909] amdgpu_device_gpu_recover.cold+0x548/0xcf1 [amdgpu] [ 44.564006] amdgpu_debugfs_reset_work+0x4c/0x80 [amdgpu] [ 44.564061] process_one_work+0x21f/0x400 [ 44.564062] worker_thread+0x200/0x3f0 [ 44.564063] ? process_one_work+0x400/0x400 [ 44.564064] kthread+0xee/0x120 [ 44.564065] ? kthread_complete_and_exit+0x20/0x20 [ 44.564066] ret_from_fork+0x22/0x30
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit 020c76d983151f6f6c9493a3bbe83c1ec927617a) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 36 ++++++++++++++++++++++----- 1 file changed, 30 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 720ab36f9c92..6c62e47a95ff 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -139,7 +139,7 @@ static int vcn_v4_0_sw_init(void *handle)
/* VCN POISON TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i], - VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq); + VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].ras_poison_irq); if (r) return r;
@@ -305,8 +305,8 @@ static int vcn_v4_0_hw_fini(void *handle) vcn_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE); } } - - amdgpu_irq_put(adev, &adev->vcn.inst[i].irq, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) + amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0); }
return 0; @@ -1976,6 +1976,24 @@ static int vcn_v4_0_set_interrupt_state(struct amdgpu_device *adev, struct amdgp return 0; }
+/** + * vcn_v4_0_set_ras_interrupt_state - set VCN block RAS interrupt state + * + * @adev: amdgpu_device pointer + * @source: interrupt sources + * @type: interrupt types + * @state: interrupt states + * + * Set VCN block RAS interrupt state + */ +static int vcn_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + /** * vcn_v4_0_process_interrupt - process VCN block interrupt * @@ -2008,9 +2026,6 @@ static int vcn_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_ case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE: amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]); break; - case VCN_4_0__SRCID_UVD_POISON: - amdgpu_vcn_process_poison_irq(adev, source, entry); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -2025,6 +2040,11 @@ static const struct amdgpu_irq_src_funcs vcn_v4_0_irq_funcs = { .process = vcn_v4_0_process_interrupt, };
+static const struct amdgpu_irq_src_funcs vcn_v4_0_ras_irq_funcs = { + .set = vcn_v4_0_set_ras_interrupt_state, + .process = amdgpu_vcn_process_poison_irq, +}; + /** * vcn_v4_0_set_irq_funcs - set VCN block interrupt irq functions * @@ -2042,6 +2062,9 @@ static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev)
adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v4_0_irq_funcs; + + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v4_0_ras_irq_funcs; } }
@@ -2115,6 +2138,7 @@ const struct amdgpu_ras_block_hw_ops vcn_v4_0_ras_hw_ops = { static struct amdgpu_vcn_ras vcn_v4_0_ras = { .ras_block = { .hw_ops = &vcn_v4_0_ras_hw_ops, + .ras_late_init = amdgpu_vcn_ras_late_init, }, };
From: Hawking Zhang Hawking.Zhang@amd.com
Initialize jpeg ras block only when jpeg ip block supports ras features. Driver queries ras capabilities after early_init, ras block init needs to be moved to sw_int.
Signed-off-by: Hawking Zhang Hawking.Zhang@amd.com Reviewed-by: Stanley Yang Stanley.Yang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit 5640e06e60198d9abdf6c618c54d982d8ec9cc0a) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 29 ++++++++++++++++-------- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 2 +- drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c | 6 +++-- drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c | 6 +++-- 4 files changed, 28 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index 6f81ed4fb0d9..479d9bcc99ee 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -236,19 +236,28 @@ int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, return 0; }
-void jpeg_set_ras_funcs(struct amdgpu_device *adev) +int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev) { + int err; + struct amdgpu_jpeg_ras *ras; + if (!adev->jpeg.ras) - return; + return 0;
- amdgpu_ras_register_ras_block(adev, &adev->jpeg.ras->ras_block); + ras = adev->jpeg.ras; + err = amdgpu_ras_register_ras_block(adev, &ras->ras_block); + if (err) { + dev_err(adev->dev, "Failed to register jpeg ras block!\n"); + return err; + }
- strcpy(adev->jpeg.ras->ras_block.ras_comm.name, "jpeg"); - adev->jpeg.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__JPEG; - adev->jpeg.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON; - adev->jpeg.ras_if = &adev->jpeg.ras->ras_block.ras_comm; + strcpy(ras->ras_block.ras_comm.name, "jpeg"); + ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__JPEG; + ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON; + adev->jpeg.ras_if = &ras->ras_block.ras_comm;
- /* If don't define special ras_late_init function, use default ras_late_init */ - if (!adev->jpeg.ras->ras_block.ras_late_init) - adev->jpeg.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + if (!ras->ras_block.ras_late_init) + ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index e8ca3e32ad52..0ca76f0f23e9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -72,6 +72,6 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); -void jpeg_set_ras_funcs(struct amdgpu_device *adev); +int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev);
#endif /*__AMDGPU_JPEG_H__*/ diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c index f2b743a93915..6b1887808782 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -138,6 +138,10 @@ static int jpeg_v2_5_sw_init(void *handle) adev->jpeg.inst[i].external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH); }
+ r = amdgpu_jpeg_ras_sw_init(adev); + if (r) + return r; + return 0; }
@@ -806,6 +810,4 @@ static void jpeg_v2_5_set_ras_funcs(struct amdgpu_device *adev) default: break; } - - jpeg_set_ras_funcs(adev); } diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c index 3beb731b2ce5..3129094baccc 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c @@ -113,6 +113,10 @@ static int jpeg_v4_0_sw_init(void *handle) adev->jpeg.internal.jpeg_pitch = regUVD_JPEG_PITCH_INTERNAL_OFFSET; adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH);
+ r = amdgpu_jpeg_ras_sw_init(adev); + if (r) + return r; + return 0; }
@@ -685,6 +689,4 @@ static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev) default: break; } - - jpeg_set_ras_funcs(adev); }
From: Horatio Zhang Hongkun.Zhang@amd.com
Separate jpegbRAS poison consumption handling from the instance irq, and register dedicated ras_poison_irq src and funcs for UVD_POISON.
v2: - Separate ras irq from jpeg instance irq - Improve the subject and code comments
v3: - Split the patch into three parts - Improve the code comments
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit ce784421a3e15fd89d5fc1b9da7d846dd8309661) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 27 +++++++++++++++++++++++- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 3 +++ 2 files changed, 29 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index 479d9bcc99ee..6e3858bde271 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -236,6 +236,31 @@ int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, return 0; }
+int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block) +{ + int r, i; + + r = amdgpu_ras_block_late_init(adev, ras_block); + if (r) + return r; + + if (amdgpu_ras_is_supported(adev, ras_block->block)) { + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + r = amdgpu_irq_get(adev, &adev->jpeg.inst[i].ras_poison_irq, 0); + if (r) + goto late_fini; + } + } + return 0; + +late_fini: + amdgpu_ras_block_late_fini(adev, ras_block); + return r; +} + int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev) { int err; @@ -257,7 +282,7 @@ int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev) adev->jpeg.ras_if = &ras->ras_block.ras_comm;
if (!ras->ras_block.ras_late_init) - ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + ras->ras_block.ras_late_init = amdgpu_jpeg_ras_late_init;
return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index 0ca76f0f23e9..1471a1ebb034 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -38,6 +38,7 @@ struct amdgpu_jpeg_reg{ struct amdgpu_jpeg_inst { struct amdgpu_ring ring_dec; struct amdgpu_irq_src irq; + struct amdgpu_irq_src ras_poison_irq; struct amdgpu_jpeg_reg external; };
@@ -72,6 +73,8 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); +int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev, + struct ras_common_if *ras_block); int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev);
#endif /*__AMDGPU_JPEG_H__*/
From: Horatio Zhang Hongkun.Zhang@amd.com
Add ras_poison_irq and functions.
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit 30b2d778f629d51e2ff30beb6d060a0bd7f70104) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c | 28 ++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c index 6b1887808782..0df8b9dbf13f 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -102,13 +102,13 @@ static int jpeg_v2_5_sw_init(void *handle)
/* JPEG DJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i], - VCN_2_6__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].irq); + VCN_2_6__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq); if (r) return r;
/* JPEG EJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i], - VCN_2_6__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].irq); + VCN_2_6__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq); if (r) return r; } @@ -217,6 +217,9 @@ static int jpeg_v2_5_hw_fini(void *handle) if (adev->jpeg.cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS)) jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE); + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) + amdgpu_irq_put(adev, &adev->jpeg.inst[i].ras_poison_irq, 0); }
return 0; @@ -565,6 +568,14 @@ static int jpeg_v2_5_set_interrupt_state(struct amdgpu_device *adev, return 0; }
+static int jpeg_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -589,10 +600,6 @@ static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, case VCN_2_0__SRCID__JPEG_DECODE: amdgpu_fence_process(&adev->jpeg.inst[ip_instance].ring_dec); break; - case VCN_2_6__SRCID_DJPEG0_POISON: - case VCN_2_6__SRCID_EJPEG0_POISON: - amdgpu_jpeg_process_poison_irq(adev, source, entry); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -723,6 +730,11 @@ static const struct amdgpu_irq_src_funcs jpeg_v2_5_irq_funcs = { .process = jpeg_v2_5_process_interrupt, };
+static const struct amdgpu_irq_src_funcs jpeg_v2_6_ras_irq_funcs = { + .set = jpeg_v2_6_set_ras_interrupt_state, + .process = amdgpu_jpeg_process_poison_irq, +}; + static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev) { int i; @@ -733,6 +745,9 @@ static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev)
adev->jpeg.inst[i].irq.num_types = 1; adev->jpeg.inst[i].irq.funcs = &jpeg_v2_5_irq_funcs; + + adev->jpeg.inst[i].ras_poison_irq.num_types = 1; + adev->jpeg.inst[i].ras_poison_irq.funcs = &jpeg_v2_6_ras_irq_funcs; } }
@@ -798,6 +813,7 @@ const struct amdgpu_ras_block_hw_ops jpeg_v2_6_ras_hw_ops = { static struct amdgpu_jpeg_ras jpeg_v2_6_ras = { .ras_block = { .hw_ops = &jpeg_v2_6_ras_hw_ops, + .ras_late_init = amdgpu_jpeg_ras_late_init, }, };
From: Horatio Zhang Hongkun.Zhang@amd.com
Add ras_poison_irq and functions. And fix the amdgpu_irq_put call trace in jpeg_v4_0_hw_fini.
[ 50.497562] RIP: 0010:amdgpu_irq_put+0xa4/0xc0 [amdgpu] [ 50.497619] RSP: 0018:ffffaa2400fcfcb0 EFLAGS: 00010246 [ 50.497620] RAX: 0000000000000000 RBX: 0000000000000001 RCX: 0000000000000000 [ 50.497621] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [ 50.497621] RBP: ffffaa2400fcfcd0 R08: 0000000000000000 R09: 0000000000000000 [ 50.497622] R10: 0000000000000000 R11: 0000000000000000 R12: ffff99b2105242d8 [ 50.497622] R13: 0000000000000000 R14: ffff99b210500000 R15: ffff99b210500000 [ 50.497623] FS: 0000000000000000(0000) GS:ffff99b518480000(0000) knlGS:0000000000000000 [ 50.497623] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 50.497624] CR2: 00007f9d32aa91e8 CR3: 00000001ba210000 CR4: 0000000000750ee0 [ 50.497624] PKRU: 55555554 [ 50.497625] Call Trace: [ 50.497625] <TASK> [ 50.497627] jpeg_v4_0_hw_fini+0x43/0xc0 [amdgpu] [ 50.497693] jpeg_v4_0_suspend+0x13/0x30 [amdgpu] [ 50.497751] amdgpu_device_ip_suspend_phase2+0x240/0x470 [amdgpu] [ 50.497802] amdgpu_device_ip_suspend+0x41/0x80 [amdgpu] [ 50.497854] amdgpu_device_pre_asic_reset+0xd9/0x4a0 [amdgpu] [ 50.497905] amdgpu_device_gpu_recover.cold+0x548/0xcf1 [amdgpu] [ 50.498005] amdgpu_debugfs_reset_work+0x4c/0x80 [amdgpu] [ 50.498060] process_one_work+0x21f/0x400 [ 50.498063] worker_thread+0x200/0x3f0 [ 50.498064] ? process_one_work+0x400/0x400 [ 50.498065] kthread+0xee/0x120 [ 50.498067] ? kthread_complete_and_exit+0x20/0x20 [ 50.498068] ret_from_fork+0x22/0x30
Suggested-by: Hawking Zhang Hawking.Zhang@amd.com Signed-off-by: Horatio Zhang Hongkun.Zhang@amd.com Reviewed-by: Tao Zhou tao.zhou1@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit bc3e1d60f933f823599376f830eb99451afb995a) Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2612 Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c | 28 +++++++++++++++++++------- 1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c index 3129094baccc..230435d30deb 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c @@ -83,13 +83,13 @@ static int jpeg_v4_0_sw_init(void *handle)
/* JPEG DJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, - VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->irq); + VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq); if (r) return r;
/* JPEG EJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, - VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->irq); + VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq); if (r) return r;
@@ -186,7 +186,8 @@ static int jpeg_v4_0_hw_fini(void *handle) RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS)) jpeg_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
- amdgpu_irq_put(adev, &adev->jpeg.inst->irq, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) + amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0);
return 0; } @@ -535,6 +536,14 @@ static int jpeg_v4_0_set_interrupt_state(struct amdgpu_device *adev, return 0; }
+static int jpeg_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -545,10 +554,6 @@ static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev, case VCN_4_0__SRCID__JPEG_DECODE: amdgpu_fence_process(&adev->jpeg.inst->ring_dec); break; - case VCN_4_0__SRCID_DJPEG0_POISON: - case VCN_4_0__SRCID_EJPEG0_POISON: - amdgpu_jpeg_process_poison_irq(adev, source, entry); - break; default: DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -619,10 +624,18 @@ static const struct amdgpu_irq_src_funcs jpeg_v4_0_irq_funcs = { .process = jpeg_v4_0_process_interrupt, };
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_ras_irq_funcs = { + .set = jpeg_v4_0_set_ras_interrupt_state, + .process = amdgpu_jpeg_process_poison_irq, +}; + static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev) { adev->jpeg.inst->irq.num_types = 1; adev->jpeg.inst->irq.funcs = &jpeg_v4_0_irq_funcs; + + adev->jpeg.inst->ras_poison_irq.num_types = 1; + adev->jpeg.inst->ras_poison_irq.funcs = &jpeg_v4_0_ras_irq_funcs; }
const struct amdgpu_ip_block_version jpeg_v4_0_ip_block = { @@ -677,6 +690,7 @@ const struct amdgpu_ras_block_hw_ops jpeg_v4_0_ras_hw_ops = { static struct amdgpu_jpeg_ras jpeg_v4_0_ras = { .ras_block = { .hw_ops = &jpeg_v4_0_ras_hw_ops, + .ras_late_init = amdgpu_jpeg_ras_late_init, }, };
linux-stable-mirror@lists.linaro.org
-
Alex Deucher