search for: drm_global_mutex

After switching all drivers' fbdev emulation to DRM client, the old fbdev hooks are now obsolete. Only amgdpu and nouveau still use them in a several places. Remove the hooks from the drivers and the DRM core. The series would ideally be merged at once via drm-misc-next. Thomas Zimmermann (8): drm/fbdev-helper: Do delayed switcheroo in lastclose helper drm/amdgpu: Do not set struct

...} else memset(kdata, 0, usize); + if (dev->driver->ioctls_need_rwsem) + down_read(&dev->ioctls_rwsem); + if (ioctl->flags & DRM_UNLOCKED) retcode = func(dev, kdata, file_priv); else { @@ -472,6 +475,9 @@ long drm_ioctl(struct file *filp, mutex_unlock(&drm_global_mutex); } + if (dev->driver->ioctls_need_rwsem) + up_read(&dev->ioctls_rwsem); + if (cmd & IOC_OUT) { if (copy_to_user((void __user *)arg, kdata, usize) != 0) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index aa454f8..4af3227 100644 --- a/dr...

...locks held by swapper/0/1: > [ 0.633632] #0: (&__lockdep_no_validate__){......}, at: [<ffffffff8143375b>] __driver_attach+0x5b/0xb0 > [ 0.633633] #1: (&__lockdep_no_validate__){......}, at: [<ffffffff81433769>] __driver_attach+0x69/0xb0 > [ 0.633636] #2: (drm_global_mutex){+.+.+.}, at: [<ffffffff8135a8f6>] drm_get_pci_dev+0xc6/0x2d0 > [ 0.633640] #3: (registration_lock){+.+.+.}, at: [<ffffffff812c8e75>] register_framebuffer+0x25/0x310 > [ 0.633642] #4: (&fb_info->lock){+.+.+.}, at: [<ffffffff812c7d86>] lock_fb_info+0x26/0x60...

After switching all drivers' fbdev emulation to DRM client, the old fbdev hooks are now obsolete. Only amgdpu and nouveau still use them in a several places. Remove the hooks from the drivers and the DRM core. The series would ideally be merged at once via drm-misc-next. v2: - call vga_switcheroo_process_delayed_switch() from drm_lastclose() (Sima) - documentation updates Thomas

...40.864179] 4 locks held by modprobe/524: [ 40.864179] #0: (&__lockdep_no_validate__){......}, at: [<ffffffff81410ea3>] __driver_attach+0x53/0xb0 [ 40.864179] #1: (&__lockdep_no_validate__){......}, at: [<ffffffff81410eb1>] __driver_attach+0x61/0xb0 [ 40.864179] #2: (drm_global_mutex){+.+.+.}, at: [<ffffffffa0094437>] drm_get_pci_dev+0xb7/0x2a0 [drm] [ 40.864179] #3: (&subdev->mutex){+.+.+.}, at: [<ffffffffa03467f4>] nv50_disp_data_ctor+0x94/0x160 [nouveau] [ 40.864179] [ 40.864179] stack backtrace: [ 40.864179] Pid: 524, comm: modprobe Tainted: G...

...DEADLOCK *** May be due to missing lock nesting notation 4 locks held by modprobe/585: #0: (&__lockdep_no_validate__){......}, at: [<ffffffff813075f3>] __driver_attach+0x53/0xb0 #1: (&__lockdep_no_validate__){......}, at: [<ffffffff81307601>] __driver_attach+0x61/0xb0 #2: (drm_global_mutex){+.+.+.}, at: [<ffffffff812ee59c>] drm_get_pci_dev+0xbc/0x2b0 #3: (&subdev->mutex){+.+.+.}, at: [<ffffffffa017672d>] nv50_disp_data_ctor+0x5d/0xd0 [nouveau] stack backtrace: Pid: 585, comm: modprobe Not tainted 3.8.0-rc6-expert+ #1 Call Trace: [<ffffffff8108fde2>] valida...

...DEADLOCK *** May be due to missing lock nesting notation 4 locks held by modprobe/585: #0: (&__lockdep_no_validate__){......}, at: [<ffffffff813075f3>] __driver_attach+0x53/0xb0 #1: (&__lockdep_no_validate__){......}, at: [<ffffffff81307601>] __driver_attach+0x61/0xb0 #2: (drm_global_mutex){+.+.+.}, at: [<ffffffff812ee59c>] drm_get_pci_dev+0xbc/0x2b0 #3: (&subdev->mutex){+.+.+.}, at: [<ffffffffa017672d>] nv50_disp_data_ctor+0x5d/0xd0 [nouveau] stack backtrace: Pid: 585, comm: modprobe Not tainted 3.8.0-rc6-expert+ #1 Call Trace: [<ffffffff8108fde2>] valida...

On 3.8-rc2 with lockdep enabled and dual-GPU setup (Macbook Pro Retina), I see two releated lock inversion issues with the EFI framebuffer, leading to possible deadlock: when X takes over from the EFI framebuffer [1] and when nouveau releases the framebuffer when being vgaswitcherood [2]. Let me know if you'd like any testing or analysis when I can get the time. Many thanks, Daniel ---

Nvidia released some documentation on how CRC support works on their GPUs, hooray! So: this patch series implements said CRC support in nouveau, along with adding some special debugfs interfaces for some relevant igt-gpu-tools tests (already on the ML). First - we add some new functionality to kthread_work in the kernel, and then use this to add a new feature to DRM that Ville Syrj?l? came up