Re: Warnings in DRM code when removing/unbinding a driver

[Daniel Vetter <daniel@ffwll.ch>]

On Tue, Dec 17, 2019 at 09:20:43AM +0000, John Garry wrote:
> On 16/12/2019 17:23, John Garry wrote:
> 
> +, -
> 
> > Hi all,
> 
> xinliang <z.liuxinliang@hisilicon.com> is bouncing. We need to get his new
> mail address.
> 
> John
> 
> > 
> > Enabling CONFIG_DEBUG_TEST_DRIVER_REMOVE causes many warns on a system
> > with the HIBMC hw:
> > 
> > [   27.788806] WARNING: CPU: 24 PID: 1 at
> > drivers/gpu/drm/drm_gem_vram_helper.c:564
> > bo_driver_move_notify+0x8c/0x98
> > [   27.798969] Modules linked in:
> > [   27.802018] CPU: 24 PID: 1 Comm: swapper/0 Tainted: G    B
> >  5.5.0-rc1-dirty #565
> > [   27.810358] Hardware name: Huawei D06 /D06, BIOS Hisilicon D06 UEFI
> > RC0 - V1.16.01 03/15/2019
> > [   27.818872] pstate: 20c00009 (nzCv daif +PAN +UAO)
> > [   27.823654] pc : bo_driver_move_notify+0x8c/0x98
> > [   27.828262] lr : bo_driver_move_notify+0x40/0x98
> > [   27.832868] sp : ffff00236f0677e0
> > [   27.836173] x29: ffff00236f0677e0 x28: ffffa0001454e5e0
> > [   27.841476] x27: ffff002366e52128 x26: ffffa000149e67b0
> > [   27.846779] x25: ffff002366e523e0 x24: ffff002336936120
> > [   27.852082] x23: ffff0023346f4010 x22: ffff002336936128
> > [   27.857385] x21: ffffa000149c15c0 x20: ffff0023369361f8
> > [   27.862687] x19: ffff002336936000 x18: 0000000000001258
> > [   27.867989] x17: 0000000000001190 x16: 00000000000011d0
> > [   27.873292] x15: 0000000000001348 x14: ffffa00012d68190
> > [   27.878595] x13: 0000000000000006 x12: 1ffff40003241f91
> > [   27.883897] x11: ffff940003241f91 x10: dfffa00000000000
> > [   27.889200] x9 : ffff940003241f92 x8 : 0000000000000001
> > [   27.894502] x7 : ffffa0001920fc88 x6 : ffff940003241f92
> > [   27.899804] x5 : ffff940003241f92 x4 : ffff0023369363a0
> > [   27.905107] x3 : ffffa00010c104b8 x2 : dfffa00000000000
> > [   27.910409] x1 : 0000000000000003 x0 : 0000000000000001
> > [   27.915712] Call trace:
> > [   27.918151]  bo_driver_move_notify+0x8c/0x98
> > [   27.922412]  ttm_bo_cleanup_memtype_use+0x54/0x100
> > [   27.927194]  ttm_bo_put+0x3a0/0x5d0
> > [   27.930673]  drm_gem_vram_object_free+0xc/0x18
> > [   27.935109]  drm_gem_object_free+0x34/0xd0
> > [   27.939196]  drm_gem_object_put_unlocked+0xc8/0xf0
> > [   27.943978]  hibmc_user_framebuffer_destroy+0x20/0x40
> > [   27.949020]  drm_framebuffer_free+0x48/0x58
> > [   27.953194]  drm_mode_object_put.part.1+0x90/0xe8
> > [   27.957889]  drm_mode_object_put+0x28/0x38
> > [   27.961976]  hibmc_fbdev_fini+0x54/0x78
> > [   27.965802]  hibmc_unload+0x2c/0xd0
> > [   27.969281]  hibmc_pci_remove+0x2c/0x40
> > [   27.973109]  pci_device_remove+0x6c/0x140
> > [   27.977110]  really_probe+0x174/0x548
> > [   27.980763]  driver_probe_device+0x7c/0x148
> > [   27.984936]  device_driver_attach+0x94/0xa0
> > [   27.989109]  __driver_attach+0xa8/0x110
> > [   27.992935]  bus_for_each_dev+0xe8/0x158
> > [   27.996849]  driver_attach+0x30/0x40
> > [   28.000415]  bus_add_driver+0x234/0x2f0
> > [   28.004241]  driver_register+0xbc/0x1d0
> > [   28.008067]  __pci_register_driver+0xbc/0xd0
> > [   28.012329]  hibmc_pci_driver_init+0x20/0x28
> > [   28.016590]  do_one_initcall+0xb4/0x254
> > [   28.020417]  kernel_init_freeable+0x27c/0x328
> > [   28.024765]  kernel_init+0x10/0x118
> > [   28.028245]  ret_from_fork+0x10/0x18
> > [   28.031813] ---[ end trace 35a83b71b657878d ]---
> > [   28.036503] ------------[ cut here ]------------
> > [   28.041115] WARNING: CPU: 24 PID: 1 at
> > drivers/gpu/drm/drm_gem_vram_helper.c:40
> > ttm_buffer_object_destroy+0x4c/0x80
> > [   28.051537] Modules linked in:
> > [   28.054585] CPU: 24 PID: 1 Comm: swapper/0 Tainted: G    B   W
> >  5.5.0-rc1-dirty #565
> > [   28.062924] Hardware name: Huawei D06 /D06, BIOS Hisilicon D06 UEFI
> > RC0 - V1.16.01 03/15/2019
> > 
> > [snip]
> > 
> > Indeed, simply unbinding the device from the driver causes the same sort
> > of issue:
> > 
> > root@(none)$ cd ./bus/pci/drivers/hibmc-drm/
> > root@(none)$ ls
> > 0000:05:00.0  bind          new_id        remove_id     uevent
> > unbind
> > root@(none)$ echo 0000\:05\:00.0 > unbind
> > [  116.074352] ------------[ cut here ]------------
> > [  116.078978] WARNING: CPU: 17 PID: 1178 at
> > drivers/gpu/drm/drm_gem_vram_helper.c:40
> > ttm_buffer_object_destroy+0x4c/0x80
> > [  116.089661] Modules linked in:
> > [  116.092711] CPU: 17 PID: 1178 Comm: sh Tainted: G    B   W
> > 5.5.0-rc1-dirty #565
> > [  116.100704] Hardware name: Huawei D06 /D06, BIOS Hisilicon D06 UEFI
> > RC0 - V1.16.01 03/15/2019
> > [  116.109218] pstate: 20400009 (nzCv daif +PAN -UAO)
> > [  116.114001] pc : ttm_buffer_object_destroy+0x4c/0x80
> > [  116.118956] lr : ttm_buffer_object_destroy+0x18/0x80
> > [  116.123910] sp : ffff0022e6cef8e0
> > [  116.127215] x29: ffff0022e6cef8e0 x28: ffff00231b1fb000
> > [  116.132519] x27: 0000000000000000 x26: ffff00231b1fb000
> > [  116.137821] x25: ffff0022e6cefdc0 x24: 0000000000002480
> > [  116.143124] x23: ffff0023682b6ab0 x22: ffff0023682b6800
> > [  116.148427] x21: ffff0023682b6800 x20: 0000000000000000
> > [  116.153730] x19: ffff0023682b6800 x18: 0000000000000000
> > [  116.159032] x17: 000000000000000000000000001
> > [  116.185545] x7 : ffff0023682b6b07 x6 : ffff80046d056d61
> > [  116.190848] x5 : ffff80046d056d61 x4 : ffff0023682b6ba0
> > [  116.196151] x3 : ffffa00010197338 x2 : dfffa00000000000
> > [  116.201453] x1 : 0000000000000003 x0 : 0000000000000001
> > [  116.206756] Call trace:
> > [  116.209195]  ttm_buffer_object_destroy+0x4c/0x80
> > [  116.213803]  ttm_bo_release_list+0x184/0x220
> > [  116.218064]  ttm_bo_put+0x410/0x5d0
> > [  116.221544]  drm_gem_vram_object_free+0xc/0x18
> > [  116.225979]  drm_gem_object_free+0x34/0xd0
> > [  116.230066]  drm_gem_object_put_unlocked+0xc8/0xf0
> > [  116.234848]  hibmc_user_framebuffer_destroy+0x20/0x40
> > [  116.239890]  drm_framebuffer_free+0x48/0x58
> > [  116.244064]  drm_mode_object_put.part.1+0x90/0xe8
> > [  116.248759]  drm_mode_object_put+0x28/0x38
> > [  116.252846]  hibmc_fbdev_fini+0x54/0x78
> > [  116.256672]  hibmc_unload+0x2c/0xd0
> > [  116.260151]  hibmc_pci_remove+0x2c/0x40
> > [  116.263979]  pci_device_remove+0x6c/0x140
> > [  116.267980]  device_release_driver_internal+0x134/0x250
> > [  116.273196]  device_driver_detach+0x28/0x38
> > [  116.277369]  unbind_store+0xfc/0x150
> > [  116.280934]  drv_attr_store+0x48/0x60
> > [  116.284589]  sysfs_kf_write+0x80/0xb0
> > [  116.288241]  kernfs_fop_write+0x1d4/0x320
> > [  116.292243]  __vfs_write+0x54/0x98
> > [  116.295635]  vfs_write+0xe8/0x270
> > [  116.298940]  ksys_write+0xc8/0x180
> > [  116.302333]  __arm64_sys_write+0x40/0x50
> > [  116.306248]  el0_svc_common.constprop.0+0xa4/0x1f8
> > [  116.311029]  el0_svc_handler+0x34/0xb0
> > [  116.314770]  el0_sync_handler+0x10c/0x1c8
> > [  116.318769]  el0_sync+0x140/0x180
> > [  116.322074] ---[ end trace e60e43d0e316b5c8 ]---
> > [  116.326868] ------------[ cut here ]------------
> > 
> > 
> > dmesg and .config is here:
> > https://pastebin.com/4P5yaZBS
> > 
> > I'm not sure if this is a HIBMC driver issue or issue with the framework.

Display-only drivers shouldn't go boom like this, the drm framework is
fixed for those. Unfortunately there's still many drivers that get their
unload sequence and resource refcounting totally wrong. For a start see
devm_drm_dev_init() and related documentation for recommendations for
current best practices:

https://dri.freedesktop.org/docs/drm/gpu/drm-internals.html#display-driver-example

Cheers, Daniel

> > 
> > john
> > 
> > 
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch