dri-devel April 2021

dri-devel@lists.freedesktop.org

317 participants
715 discussions

[PATCH] drm/panel: panel-dsi-cm: Convert sysfs sprintf/snprintf family to sysfs_emit
by Tian Tao 12 Apr '21

12 Apr '21

Fix the following coccicheck warning: drivers/gpu/drm/panel/panel-dsi-cm.c:271:8-16: WARNING: use scnprintf or sprintf drivers/gpu/drm/panel/panel-dsi-cm.c:251:8-16: WARNING: use scnprintf or sprintf Signed-off-by: Tian Tao <tiantao6(a)hisilicon.com> --- drivers/gpu/drm/panel/panel-dsi-cm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/panel/panel-dsi-cm.c b/drivers/gpu/drm/panel/panel-dsi-cm.c index af381d7..422d689 100644 --- a/drivers/gpu/drm/panel/panel-dsi-cm.c +++ b/drivers/gpu/drm/panel/panel-dsi-cm.c @@ -247,7 +247,7 @@ static ssize_t num_dsi_errors_show(struct device *dev, if (r) return r; - return snprintf(buf, PAGE_SIZE, "%d\n", errors); + return sysfs_emit(buf, "%d\n", errors); } static ssize_t hw_revision_show(struct device *dev, @@ -267,7 +267,7 @@ static ssize_t hw_revision_show(struct device *dev, if (r) return r; - return snprintf(buf, PAGE_SIZE, "%02x.%02x.%02x\n", id1, id2, id3); + return sysfs_emit(buf, "%02x.%02x.%02x\n", id1, id2, id3); } static DEVICE_ATTR_RO(num_dsi_errors); -- 2.7.4

1 0

Request for assistance
by o1bigtenor 12 Apr '21

12 Apr '21

Greetings I'm running a Debian testing (11) system using Nouveau as a driver for 2 graphics cards: 1. Nvidia 1050 Ti (GP107) and a Nvidia 570 (GF110) driving 5 monitors 1 - 3840x2160 and 4 - 1920x1080s. The 5th monitor was added some about 8 weeks ago and since life got interesting. Previously I would use an uptime that would last anywhere from 4 to 6 months but after adding the 5th monitor - - - well the best has been some few days and the worst - - - a few hours. I starting digging to try and find possible issues. First thing I found was the idea of adding firmware which was an interesting exercise but was successful and now somewhere between 400 and 700 seconds after reboot I'm seeing this: [ 534.790587] nouveau 0000:02:00.0: firmware: direct-loading firmware nouveau/nvc8_fuc084 I was cheering when I got this far - - - - except this flaw was hiding another one that I hope you might be able to help with. The error I see (using dmesg) looks something like this: [25375.252874] perf: interrupt took too long (3168 > 3150), lowering kernel.perf_event_max_sample_rate to 63000 [35577.509444] perf: interrupt took too long (3963 > 3960), lowering kernel.perf_event_max_sample_rate to 50250 [54648.710595] perf: interrupt took too long (4991 > 4953), lowering kernel.perf_event_max_sample_rate to 40000 [77975.516742] nouveau 0000:01:00.0: fifo: FB_FLUSH_TIMEOUT [85039.583604] nouveau 0000:01:00.0: DRM: core notifier timeout [85041.583597] nouveau 0000:01:00.0: DRM: base-0: timeout [85041.585023] nouveau 0000:01:00.0: bus: MMIO read of 00000000 FAULT at 690400 [ IBUS ] [85041.585259] nouveau 0000:01:00.0: bus: MMIO read of 00000000 FAULT at 616d48 [ IBUS ] [85043.587213] nouveau 0000:01:00.0: DRM: core notifier timeout [85045.587202] nouveau 0000:01:00.0: DRM: base-1: timeout [85047.587302] nouveau 0000:01:00.0: DRM: core notifier timeout [85049.587289] nouveau 0000:01:00.0: DRM: base-2: timeout [85051.628464] nouveau 0000:01:00.0: DRM: core notifier timeout [85053.628538] nouveau 0000:01:00.0: DRM: core notifier timeout [85085.181271] nouveau 0000:01:00.0: DRM: core notifier timeout [85087.181829] nouveau 0000:01:00.0: DRM: core notifier timeout [85089.181909] nouveau 0000:01:00.0: DRM: core notifier timeout That's the error but here's another data point from early in the boot cycle: [ 1.989397] nouveau 0000:01:00.0: firmware: direct-loading firmware nvidia/gp107/acr/ucode_unload.bin [ 1.989401] nouveau 0000:01:00.0: pmu: firmware unavailable [ 1.989535] nouveau 0000:01:00.0: firmware: direct-loading firmware nvidia/gp107/gr/fecs_bl.bin In digging for what I might be able to do I found this final communication from late February with a subject line of: Re: [PATCH v2] drm/nouveau/pmu: fix timeout on GP108 Is what is discussed in this thread the same issue that I'm having on a GP107? If it is the same issue do I need to apply both of the patches? (If the answer to the second question is yes how do I apply the patch? I've found this technique at 'stackoverflow': 1) git clone <path_to_kernel_sources> 2) git checkout 13fac179aa50556ba3c60790a9beb6ca9d0b1b8b 3) git apply <patch_file> is this the right way to patch the kernel? (Never done this and really don't need my main computer down for even hours nevermind weeks.) Hopefully my explanation is clear enough - - - - if more information from dmesg files are needed - - - well I have 5 or 6 to choose from (called from the second m/c on the network that I reboot the main m/c from). TIA

1 0

general protection fault in drm_client_buffer_vunmap
by syzbot 11 Apr '21

11 Apr '21

Hello, syzbot found the following issue on: HEAD commit: 6147c83f Add linux-next specific files for 20201126 git tree: linux-next console output: https://syzkaller.appspot.com/x/log.txt?x=109130a5500000 kernel config: https://syzkaller.appspot.com/x/.config?x=9b91566da897c24f dashboard link: https://syzkaller.appspot.com/bug?extid=10328e8428a896b65119 compiler: gcc (GCC) 10.1.0-syz 20200507 syz repro: https://syzkaller.appspot.com/x/repro.syz?x=12541369500000 C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1365c663500000 IMPORTANT: if you fix the issue, please add the following tag to the commit: Reported-by: syzbot+10328e8428a896b65119(a)syzkaller.appspotmail.com [drm] Initialized udl on minor 2 [drm:udl_get_edid_block] *ERROR* Read EDID byte 0 failed err ffffffb9 udl 1-1:0.0: [drm] Cannot find any crtc or sizes usb 1-1: USB disconnect, device number 2 general protection fault, probably for non-canonical address 0xdffffc0000000002: 0000 [#1] PREEMPT SMP KASAN KASAN: null-ptr-deref in range [0x0000000000000010-0x0000000000000017] CPU: 0 PID: 7 Comm: kworker/0:1 Not tainted 5.10.0-rc5-next-20201126-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Workqueue: usb_hub_wq hub_event RIP: 0010:drm_client_buffer_vunmap+0x26/0x50 drivers/gpu/drm/drm_client.c:347 Code: 00 00 00 00 53 48 89 fb 48 83 ec 08 e8 83 8b 3f fd 48 8d 7b 10 48 b8 00 00 00 00 00 fc ff df 48 8d 73 18 48 89 fa 48 c1 ea 03 <80> 3c 02 00 75 0e 48 8b 7b 10 48 83 c4 08 5b e9 56 61 f7 ff 48 89 RSP: 0018:ffffc90000cc7710 EFLAGS: 00010202 RAX: dffffc0000000000 RBX: 0000000000000000 RCX: ffffffff815524de RDX: 0000000000000002 RSI: 0000000000000018 RDI: 0000000000000010 RBP: ffff88801db748c8 R08: 0000000000000000 R09: ffff88801db74a17 R10: ffffed1003b6e942 R11: 0000000000000000 R12: 0000000000000000 R13: ffff88801db748d8 R14: ffff88801db74810 R15: dead000000000100 FS: 0000000000000000(0000) GS:ffff8880b9e00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 000055f43f9d79e0 CR3: 000000000b08e000 CR4: 00000000001506f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: drm_fbdev_cleanup+0x380/0x440 drivers/gpu/drm/drm_fb_helper.c:2042 drm_fbdev_release drivers/gpu/drm/drm_fb_helper.c:2049 [inline] drm_fbdev_client_unregister+0x61/0x80 drivers/gpu/drm/drm_fb_helper.c:2376 drm_client_dev_unregister+0x239/0x3b0 drivers/gpu/drm/drm_client.c:175 drm_dev_unregister+0xe9/0x2b0 drivers/gpu/drm/drm_drv.c:942 usb_unbind_interface+0x1d8/0x8d0 drivers/usb/core/driver.c:458 __device_release_driver+0x3bd/0x6f0 drivers/base/dd.c:1154 device_release_driver_internal drivers/base/dd.c:1185 [inline] device_release_driver+0x26/0x40 drivers/base/dd.c:1208 bus_remove_device+0x2eb/0x5a0 drivers/base/bus.c:533 device_del+0x502/0xec0 drivers/base/core.c:3113 usb_disable_device+0x35b/0x7b0 drivers/usb/core/message.c:1417 usb_disconnect.cold+0x27d/0x780 drivers/usb/core/hub.c:2218 hub_port_connect drivers/usb/core/hub.c:5074 [inline] hub_port_connect_change drivers/usb/core/hub.c:5363 [inline] port_event drivers/usb/core/hub.c:5509 [inline] hub_event+0x1c8a/0x42d0 drivers/usb/core/hub.c:5591 process_one_work+0x98d/0x15f0 kernel/workqueue.c:2272 process_scheduled_works kernel/workqueue.c:2334 [inline] worker_thread+0x82b/0x1120 kernel/workqueue.c:2420 kthread+0x3b1/0x4a0 kernel/kthread.c:292 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:296 Modules linked in: ---[ end trace 5e45793b7de819bc ]--- RIP: 0010:drm_client_buffer_vunmap+0x26/0x50 drivers/gpu/drm/drm_client.c:347 Code: 00 00 00 00 53 48 89 fb 48 83 ec 08 e8 83 8b 3f fd 48 8d 7b 10 48 b8 00 00 00 00 00 fc ff df 48 8d 73 18 48 --- This report is generated by a bot. It may contain errors. See https://goo.gl/tpsmEJ for more information about syzbot. syzbot engineers can be reached at syzkaller(a)googlegroups.com. syzbot will keep track of this issue. See: https://goo.gl/tpsmEJ#status for how to communicate with syzbot. syzbot can test patches for this issue, for details see: https://goo.gl/tpsmEJ#testing-patches

2 2

[PATCH 1/2] drm/vmwgfx: Fix the lockdep breakage
by Zack Rusin 10 Apr '21

10 Apr '21

Thomas has noticed that the lockdep was broken in vmwgfx. It was broken during the pci initialization rework. This fixes the breakage by making sure we initialize the locking code before doing anything else. This was independently spotted and fixed by Tetsuo Handa as well. Reviewed-by: Martin Krastev <krastevm(a)vmware.com> Reviewed-by: Roland Scheidegger <sroland(a)vmware.com> Cc: Tetsuo Handa <penguin-kernel(a)I-love.SAKURA.ne.jp> Cc: dri-devel(a)lists.freedesktop.org Signed-off-by: Zack Rusin <zackr(a)vmware.com> Fixes: 8772c0bb58bbf98a ("drm/vmwgfx: Cleanup pci resource allocation") --- drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 17 ++++++++--------- drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 2 -- 2 files changed, 8 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index dd69b51c40e4..6fa24645fbbf 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -712,17 +712,8 @@ static int vmw_driver_load(struct vmw_private *dev_priv, u32 pci_id) dev_priv->last_read_seqno = (uint32_t) -100; dev_priv->drm.dev_private = dev_priv; - ret = vmw_setup_pci_resources(dev_priv, pci_id); - if (ret) - return ret; - ret = vmw_detect_version(dev_priv); - if (ret) - goto out_no_pci_or_version; - mutex_init(&dev_priv->cmdbuf_mutex); - mutex_init(&dev_priv->release_mutex); mutex_init(&dev_priv->binding_mutex); - mutex_init(&dev_priv->global_kms_state_mutex); ttm_lock_init(&dev_priv->reservation_sem); spin_lock_init(&dev_priv->resource_lock); spin_lock_init(&dev_priv->hw_lock); @@ -730,6 +721,14 @@ static int vmw_driver_load(struct vmw_private *dev_priv, u32 pci_id) spin_lock_init(&dev_priv->cap_lock); spin_lock_init(&dev_priv->cursor_lock); + ret = vmw_setup_pci_resources(dev_priv, pci_id); + if (ret) + return ret; + ret = vmw_detect_version(dev_priv); + if (ret) + goto out_no_pci_or_version; + + for (i = vmw_res_context; i < vmw_res_max; ++i) { idr_init(&dev_priv->res_idr[i]); INIT_LIST_HEAD(&dev_priv->res_lru[i]); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 336f5086ca26..8087a9013455 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -529,7 +529,6 @@ struct vmw_private { struct vmw_overlay *overlay_priv; struct drm_property *hotplug_mode_update_property; struct drm_property *implicit_placement_property; - struct mutex global_kms_state_mutex; spinlock_t cursor_lock; struct drm_atomic_state *suspend_state; @@ -592,7 +591,6 @@ struct vmw_private { bool refuse_hibernation; bool suspend_locked; - struct mutex release_mutex; atomic_t num_fifo_resources; /* -- 2.27.0

3 5

[PATCH V4 0/2] Add virtual hardware module
by Sumera Priyadarsini 10 Apr '21

10 Apr '21

This patchset adds support for emulating virtual hardware with VKMS. The virtual hardware mode can be enabled by using the following command while loading the module: sudo modprobe vkms enable_virtual_hw=1 The first patch is prep work for adding virtual_hw mode and refactors the plane composition in vkms by adding a helper function vkms_composer_common() which can be used for both vblank mode and virtual mode. The second patch adds virtual hardware support as a module option. It adds new atomic helper functions for the virtual mode and modifies the existing atomic helpers for usage by the vblank mode This gives us two sets of drm_crtc_helper_funcs struct for both modes, making the code flow cleaner and easier to debug. This patchset has been tested with the igt tests- kms_writeback, kms_atomic, kms_lease, kms_flip, kms_pipe_get_crc and preserves results except for subtests related to crc reads and skips tests that rely on vertical blanking. This patchset must be tested after incorporating the igt-tests patch: https://lists.freedesktop.org/archives/igt-dev/2021-February/029355.html Sumera Priyadarsini (2): drm/vkms: Refactor vkms_composer_worker() to prep for virtual_hw mode drm/vkms: Add support for virtual hardware mode drivers/gpu/drm/vkms/vkms_composer.c | 88 +++++++++++++++++----------- drivers/gpu/drm/vkms/vkms_crtc.c | 51 +++++++++++----- drivers/gpu/drm/vkms/vkms_drv.c | 18 ++++-- drivers/gpu/drm/vkms/vkms_drv.h | 4 ++ 4 files changed, 109 insertions(+), 52 deletions(-) -- 2.25.1

3 6

[PATCH] drm/nouveau/nvkm: Remove unnecessary assignment
by Wan Jiabing 10 Apr '21

10 Apr '21

The declaration with assignment here is confusing. Remove the unnecessary assignment. Signed-off-by: Wan Jiabing <wanjiabing(a)vivo.com> --- drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c index 96aca0edfa3c..c51bac76174c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c @@ -313,7 +313,7 @@ nv50_instobj_dtor(struct nvkm_memory *memory) struct nv50_instobj *iobj = nv50_instobj(memory); struct nvkm_instmem *imem = &iobj->imem->base; struct nvkm_vma *bar; - void *map = map; + void *map; mutex_lock(&imem->mutex); if (likely(iobj->lru.next)) -- 2.25.1

1 0

[PATCH] drm/etnaviv: Remove useless error message
by Tian Tao 10 Apr '21

10 Apr '21

Fix the following coccicheck report: drivers/gpu/drm/etnaviv/etnaviv_gpu.c:1775:2-9: line 1775 is redundant because platform_get_irq() already prints an error Remove dev_err() messages after platform_get_irq() failures. Signed-off-by: Tian Tao <tiantao6(a)hisilicon.com> Signed-off-by: Zihao Tang <tangzihao1(a)hisilicon.com> Signed-off-by: Jay Fang <f.fangjian(a)huawei.com> --- drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c index c6404b8..4645349 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c @@ -1771,10 +1771,8 @@ static int etnaviv_gpu_platform_probe(struct platform_device *pdev) /* Get Interrupt: */ gpu->irq = platform_get_irq(pdev, 0); - if (gpu->irq < 0) { - dev_err(dev, "failed to get irq: %d\n", gpu->irq); + if (gpu->irq < 0) return gpu->irq; - } err = devm_request_irq(&pdev->dev, gpu->irq, irq_handler, 0, dev_name(gpu->dev), gpu); -- 2.7.4

2 1

[PATCH v1] drm/radeon: Fix a missing check bug in radeon_dp_mst_detect()
by wangyingjie55＠126.com 10 Apr '21

10 Apr '21

From: Yingjie Wang <wangyingjie55(a)126.com> In radeon_dp_mst_detect(), We should check whether or not @connector has been unregistered from userspace. If the connector is unregistered, we should return disconnected status. Fixes: 9843ead08f18 ("drm/radeon: add DisplayPort MST support (v2)") Signed-off-by: Yingjie Wang <wangyingjie55(a)126.com> --- drivers/gpu/drm/radeon/radeon_dp_mst.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_dp_mst.c b/drivers/gpu/drm/radeon/radeon_dp_mst.c index 2c32186c4acd..4e4c937c36c6 100644 --- a/drivers/gpu/drm/radeon/radeon_dp_mst.c +++ b/drivers/gpu/drm/radeon/radeon_dp_mst.c @@ -242,6 +242,9 @@ radeon_dp_mst_detect(struct drm_connector *connector, to_radeon_connector(connector); struct radeon_connector *master = radeon_connector->mst_port; + if (drm_connector_is_unregistered(connector)) + return connector_status_disconnected; + return drm_dp_mst_detect_port(connector, ctx, &master->mst_mgr, radeon_connector->port); } -- 2.7.4

3 2

[RFC 1/3] drm/msm/dpu: merge dpu_hw_intr_get_interrupt_statuses into dpu_hw_intr_dispatch_irqs
by Dmitry Baryshkov 10 Apr '21

10 Apr '21

There is little sense in reading interrupt statuses and right after that going after the array of statuses to dispatch them. Merge both loops into single function doing read and dispatch. Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov(a)linaro.org> --- drivers/gpu/drm/msm/disp/dpu1/dpu_core_irq.c | 10 +-- .../gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c | 66 ++++++------------- .../gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.h | 8 --- 3 files changed, 20 insertions(+), 64 deletions(-) diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_core_irq.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_core_irq.c index 84ea09d9692f..c977e8484174 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_core_irq.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_core_irq.c @@ -376,15 +376,6 @@ void dpu_core_irq_uninstall(struct dpu_kms *dpu_kms) irqreturn_t dpu_core_irq(struct dpu_kms *dpu_kms) { - /* - * Read interrupt status from all sources. Interrupt status are - * stored within hw_intr. - * Function will also clear the interrupt status after reading. - * Individual interrupt status bit will only get stored if it - * is enabled. - */ - dpu_kms->hw_intr->ops.get_interrupt_statuses(dpu_kms->hw_intr); - /* * Dispatch to HW driver to handle interrupt lookup that is being * fired. When matching interrupt is located, HW driver will call to @@ -392,6 +383,7 @@ irqreturn_t dpu_core_irq(struct dpu_kms *dpu_kms) * dpu_core_irq_callback_handler will perform the registered function * callback, and do the interrupt status clearing once the registered * callback is finished. + * Function will also clear the interrupt status after reading. */ dpu_kms->hw_intr->ops.dispatch_irqs( dpu_kms->hw_intr, diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c index 5c521de71567..a8d463a8e8fe 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c @@ -780,6 +780,7 @@ static void dpu_hw_intr_dispatch_irq(struct dpu_hw_intr *intr, int start_idx; int end_idx; u32 irq_status; + u32 enable_mask; unsigned long irq_flags; if (!intr) @@ -792,8 +793,6 @@ static void dpu_hw_intr_dispatch_irq(struct dpu_hw_intr *intr, */ spin_lock_irqsave(&intr->irq_lock, irq_flags); for (reg_idx = 0; reg_idx < ARRAY_SIZE(dpu_intr_set); reg_idx++) { - irq_status = intr->save_irq_status[reg_idx]; - /* * Each Interrupt register has a range of 32 indexes, and * that is static for dpu_irq_map. @@ -805,6 +804,20 @@ static void dpu_hw_intr_dispatch_irq(struct dpu_hw_intr *intr, start_idx >= ARRAY_SIZE(dpu_irq_map)) continue; + /* Read interrupt status */ + irq_status = DPU_REG_READ(&intr->hw, dpu_intr_set[reg_idx].status_off); + + /* Read enable mask */ + enable_mask = DPU_REG_READ(&intr->hw, dpu_intr_set[reg_idx].en_off); + + /* and clear the interrupt */ + if (irq_status) + DPU_REG_WRITE(&intr->hw, dpu_intr_set[reg_idx].clr_off, + irq_status); + + /* Finally update IRQ status based on enable mask */ + irq_status &= enable_mask; + /* * Search through matching intr status from irq map. * start_idx and end_idx defined the search range in @@ -836,6 +849,10 @@ static void dpu_hw_intr_dispatch_irq(struct dpu_hw_intr *intr, irq_status &= ~dpu_irq_map[irq_idx].irq_mask; } } + + /* ensure register writes go through */ + wmb(); + spin_unlock_irqrestore(&intr->irq_lock, irq_flags); } @@ -987,41 +1004,6 @@ static int dpu_hw_intr_disable_irqs(struct dpu_hw_intr *intr) return 0; } -static void dpu_hw_intr_get_interrupt_statuses(struct dpu_hw_intr *intr) -{ - int i; - u32 enable_mask; - unsigned long irq_flags; - - if (!intr) - return; - - spin_lock_irqsave(&intr->irq_lock, irq_flags); - for (i = 0; i < ARRAY_SIZE(dpu_intr_set); i++) { - if (!test_bit(i, &intr->irq_mask)) - continue; - - /* Read interrupt status */ - intr->save_irq_status[i] = DPU_REG_READ(&intr->hw, - dpu_intr_set[i].status_off); - - /* Read enable mask */ - enable_mask = DPU_REG_READ(&intr->hw, dpu_intr_set[i].en_off); - - /* and clear the interrupt */ - if (intr->save_irq_status[i]) - DPU_REG_WRITE(&intr->hw, dpu_intr_set[i].clr_off, - intr->save_irq_status[i]); - - /* Finally update IRQ status based on enable mask */ - intr->save_irq_status[i] &= enable_mask; - } - - /* ensure register writes go through */ - wmb(); - - spin_unlock_irqrestore(&intr->irq_lock, irq_flags); -} static void dpu_hw_intr_clear_intr_status_nolock(struct dpu_hw_intr *intr, int irq_idx) @@ -1080,7 +1062,6 @@ static void __setup_intr_ops(struct dpu_hw_intr_ops *ops) ops->dispatch_irqs = dpu_hw_intr_dispatch_irq; ops->clear_all_irqs = dpu_hw_intr_clear_irqs; ops->disable_all_irqs = dpu_hw_intr_disable_irqs; - ops->get_interrupt_statuses = dpu_hw_intr_get_interrupt_statuses; ops->clear_intr_status_nolock = dpu_hw_intr_clear_intr_status_nolock; ops->get_interrupt_status = dpu_hw_intr_get_interrupt_status; } @@ -1117,14 +1098,6 @@ struct dpu_hw_intr *dpu_hw_intr_init(void __iomem *addr, return ERR_PTR(-ENOMEM); } - intr->save_irq_status = kcalloc(ARRAY_SIZE(dpu_intr_set), sizeof(u32), - GFP_KERNEL); - if (intr->save_irq_status == NULL) { - kfree(intr->cache_irq_mask); - kfree(intr); - return ERR_PTR(-ENOMEM); - } - intr->irq_mask = m->mdss_irqs; spin_lock_init(&intr->irq_lock); @@ -1135,7 +1108,6 @@ void dpu_hw_intr_destroy(struct dpu_hw_intr *intr) { if (intr) { kfree(intr->cache_irq_mask); - kfree(intr->save_irq_status); kfree(intr); } } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.h index fc9c98617281..d8b9d5fe6b8c 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.h @@ -141,14 +141,6 @@ struct dpu_hw_intr_ops { void (*cbfunc)(void *arg, int irq_idx), void *arg); - /** - * get_interrupt_statuses - Gets and store value from all interrupt - * status registers that are currently fired. - * @intr: HW interrupt handle - */ - void (*get_interrupt_statuses)( - struct dpu_hw_intr *intr); - /** * clear_intr_status_nolock() - clears the HW interrupts without lock * @intr: HW interrupt handle -- 2.30.2

1 2

[PATCH v3 0/3] Add devcoredump support for DPU
by Abhinav Kumar 09 Apr '21

09 Apr '21

This series adds support to use devcoredump for DPU driver. It introduces the dpu_dbg module which assists in the capturing of register dumps during error scenarios. When a display related error happens, the dpu_dbg module captures all the relevant register dumps along with the snapshot of the drm atomic state and triggers a devcoredump. changes in v3: - Get rid of registration mechanism for sub-modules and instead get this information from the dpu catalog itself - Get rid of global dpu_dbg struct and instead store it in dpu_kms - delegate the power management of the sub-modules to the resp drivers - refactor and remove the linked list logic and simplify it to have just an array Abhinav Kumar (3): drm: allow drm_atomic_print_state() to accept any drm_printer drm/msm/dpu: add support to dump dpu registers drm/msm/dpu: add dpu_dbg points across dpu driver drivers/gpu/drm/drm_atomic.c | 28 ++- drivers/gpu/drm/drm_atomic_uapi.c | 4 +- drivers/gpu/drm/drm_crtc_internal.h | 4 +- drivers/gpu/drm/msm/Makefile | 2 + drivers/gpu/drm/msm/disp/dpu1/dpu_dbg.c | 221 ++++++++++++++++++ drivers/gpu/drm/msm/disp/dpu1/dpu_dbg.h | 200 ++++++++++++++++ drivers/gpu/drm/msm/disp/dpu1/dpu_dbg_util.c | 257 +++++++++++++++++++++ drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 18 +- .../gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c | 14 +- .../gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c | 8 +- drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h | 2 +- drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 86 +++++++ drivers/gpu/drm/msm/disp/dpu1/dpu_kms.h | 5 + drivers/gpu/drm/msm/dp/dp_catalog.c | 10 + drivers/gpu/drm/msm/dp/dp_catalog.h | 5 + drivers/gpu/drm/msm/dp/dp_display.c | 37 +++ drivers/gpu/drm/msm/dp/dp_display.h | 1 + drivers/gpu/drm/msm/dsi/dsi.c | 5 + drivers/gpu/drm/msm/dsi/dsi.h | 4 + drivers/gpu/drm/msm/dsi/dsi_host.c | 25 ++ drivers/gpu/drm/msm/msm_drv.c | 29 ++- drivers/gpu/drm/msm/msm_drv.h | 2 + drivers/gpu/drm/selftests/test-drm_framebuffer.c | 1 + 23 files changed, 950 insertions(+), 18 deletions(-) create mode 100644 drivers/gpu/drm/msm/disp/dpu1/dpu_dbg.c create mode 100644 drivers/gpu/drm/msm/disp/dpu1/dpu_dbg.h create mode 100644 drivers/gpu/drm/msm/disp/dpu1/dpu_dbg_util.c -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project

4 6

← Newer
1
...
51
52
53
54
55
56
57
...
72
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel April 2021