search for: nouveau_fbcon_accel_save_disable

...le on, and then + * wait to turn it off. It must be synchronous as we are + * about to suspend or unload the driver. + * + * Note that from within the work-handler, we cannot flush + * ourselves, so only flush outstanding work upon suspend! + */ if (state != FBINFO_STATE_RUNNING) - nouveau_fbcon_accel_save_disable(dev); - console_unlock(); + flush_work(&drm->fbdev_suspend_work); + console_lock(); + } else { + /* + * The console lock can be pretty contented on resume due + * to all the printk activity. Try to keep it out of the hot + * path of resume if possible. This also prevents a deadl...

...it off. It must be synchronous as we are > + * about to suspend or unload the driver. > + * > + * Note that from within the work-handler, we cannot flush > + * ourselves, so only flush outstanding work upon suspend! > + */ > if (state != FBINFO_STATE_RUNNING) > - nouveau_fbcon_accel_save_disable(dev); > - console_unlock(); > + flush_work(&drm->fbdev_suspend_work); > + console_lock(); > + } else { > + /* > + * The console lock can be pretty contented on resume due > + * to all the printk activity. Try to keep it out of the hot > + * path of resume...

As it turns out, on cards that actually have CRTCs on them we're already calling drm_kms_helper_poll_enable(drm_dev) from nouveau_display_resume() before we call it in nouveau_pmops_runtime_resume(). This leads us to accidentally trying to enable polling twice, which results in a potential deadlock between the RPM locks and drm_dev->mode_config.mutex if we end up trying to enable polling

...it off. It must be synchronous as we are > + * about to suspend or unload the driver. > + * > + * Note that from within the work-handler, we cannot flush > + * ourselves, so only flush outstanding work upon suspend! > + */ > if (state != FBINFO_STATE_RUNNING) > - nouveau_fbcon_accel_save_disable(dev); > - console_unlock(); > + flush_work(&drm->fbdev_suspend_work); > + console_lock(); > + } else { > + /* > + * The console lock can be pretty contented on resume due > + * to all the printk activity. Try to keep it out of the hot > + * path of resume...

...state; + + if (state == FBINFO_STATE_RUNNING) + pm_runtime_get_sync(drm->dev->dev); + + console_lock(); + if (state == FBINFO_STATE_RUNNING) + nouveau_fbcon_accel_restore(drm->dev); + drm_fb_helper_set_suspend(&drm->fbcon->helper, state); + if (state != FBINFO_STATE_RUNNING) + nouveau_fbcon_accel_save_disable(drm->dev); + console_unlock(); + + if (state == FBINFO_STATE_RUNNING) { + pm_runtime_mark_last_busy(drm->dev->dev); + pm_runtime_put_sync(drm->dev->dev); + } +} + void nouveau_fbcon_set_suspend(struct drm_device *dev, int state) { struct nouveau_drm *drm = nouveau_drm(dev); -...

Signed-off-by: Maarten Lankhorst <maarten.lankhorst at ubuntu.com> --- diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index 3ed12a8cfc91..a4a586807903 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -370,6 +370,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper, ret = -ENOMEM; goto

...e are > > + * about to suspend or unload the driver. > > + * > > + * Note that from within the work-handler, we cannot flush > > + * ourselves, so only flush outstanding work upon suspend! > > + */ > > if (state != FBINFO_STATE_RUNNING) > > - nouveau_fbcon_accel_save_disable(dev); > > - console_unlock(); > > + flush_work(&drm->fbdev_suspend_work); > > + console_lock(); > > + } else { > > + /* > > + * The console lock can be pretty contented on resume due > > + * to all the printk activity. Try to keep it out of...

...it off. It must be synchronous as we are > + * about to suspend or unload the driver. > + * > + * Note that from within the work-handler, we cannot flush > + * ourselves, so only flush outstanding work upon suspend! > + */ > if (state != FBINFO_STATE_RUNNING) > - nouveau_fbcon_accel_save_disable(dev); > - console_unlock(); > + flush_work(&drm->fbdev_suspend_work); > + console_lock(); > + } else { > + /* > + * The console lock can be pretty contented on resume due > + * to all the printk activity. Try to keep it out of the hot > + * path of resume...

...e are > > + * about to suspend or unload the driver. > > + * > > + * Note that from within the work-handler, we cannot flush > > + * ourselves, so only flush outstanding work upon suspend! > > + */ > > if (state != FBINFO_STATE_RUNNING) > > - nouveau_fbcon_accel_save_disable(dev); > > - console_unlock(); > > + flush_work(&drm->fbdev_suspend_work); > > + console_lock(); > > + } else { > > + /* > > + * The console lock can be pretty contented on resume due > > + * to all the printk activity. Try to keep it out of...

...nouveau_fbcon_sw_ops = { - .owner = THIS_MODULE, - DRM_FB_HELPER_DEFAULT_OPS, - .fb_open = nouveau_fbcon_open, - .fb_release = nouveau_fbcon_release, - .fb_fillrect = drm_fb_helper_cfb_fillrect, - .fb_copyarea = drm_fb_helper_cfb_copyarea, - .fb_imageblit = drm_fb_helper_cfb_imageblit, -}; - -void -nouveau_fbcon_accel_save_disable(struct drm_device *dev) -{ - struct nouveau_drm *drm = nouveau_drm(dev); - if (drm->fbcon && drm->fbcon->helper.info) { - drm->fbcon->saved_flags = drm->fbcon->helper.info->flags; - drm->fbcon->helper.info->flags |= FBINFO_HWACCEL_DISABLED; - } -} - -void...

...S, > - .fb_open = nouveau_fbcon_open, > - .fb_release = nouveau_fbcon_release, > - .fb_fillrect = drm_fb_helper_cfb_fillrect, > - .fb_copyarea = drm_fb_helper_cfb_copyarea, > - .fb_imageblit = drm_fb_helper_cfb_imageblit, > -}; > - > -void > -nouveau_fbcon_accel_save_disable(struct drm_device *dev) > -{ > - struct nouveau_drm *drm = nouveau_drm(dev); > - if (drm->fbcon && drm->fbcon->helper.info) { > - drm->fbcon->saved_flags = drm->fbcon->helper.info->flags; > - drm->fbcon->he...

...au_fbcon_open, >> - .fb_release = nouveau_fbcon_release, >> - .fb_fillrect = drm_fb_helper_cfb_fillrect, >> - .fb_copyarea = drm_fb_helper_cfb_copyarea, >> - .fb_imageblit = drm_fb_helper_cfb_imageblit, >> -}; >> - >> -void >> -nouveau_fbcon_accel_save_disable(struct drm_device *dev) >> -{ >> - struct nouveau_drm *drm = nouveau_drm(dev); >> - if (drm->fbcon && drm->fbcon->helper.info) { >> - drm->fbcon->saved_flags = drm->fbcon->helper.info->flags; >> -...

...-static int nouveau_nofbaccel = 0; +int nouveau_nofbaccel = 0; module_param_named(nofbaccel, nouveau_nofbaccel, int, 0400); static void diff --git a/drm/nouveau_fbcon.h b/drm/nouveau_fbcon.h index 34658cf..1e2e9e2 100644 --- a/drm/nouveau_fbcon.h +++ b/drm/nouveau_fbcon.h @@ -72,5 +72,8 @@ void nouveau_fbcon_accel_save_disable(struct drm_device *dev); void nouveau_fbcon_accel_restore(struct drm_device *dev); void nouveau_fbcon_output_poll_changed(struct drm_device *dev); + +extern int nouveau_nofbaccel; + #endif /* __NV50_FBCON_H__ */ diff --git a/drm/nouveau_sysfs.c b/drm/nouveau_sysfs.c index 3c6962d..8fbbf30 10...

...nouveau_fbcon_sw_ops = { - .owner = THIS_MODULE, - DRM_FB_HELPER_DEFAULT_OPS, - .fb_open = nouveau_fbcon_open, - .fb_release = nouveau_fbcon_release, - .fb_fillrect = drm_fb_helper_cfb_fillrect, - .fb_copyarea = drm_fb_helper_cfb_copyarea, - .fb_imageblit = drm_fb_helper_cfb_imageblit, -}; - -void -nouveau_fbcon_accel_save_disable(struct drm_device *dev) -{ - struct nouveau_drm *drm = nouveau_drm(dev); - if (drm->fbcon && drm->fbcon->helper.info) { - drm->fbcon->saved_flags = drm->fbcon->helper.info->flags; - drm->fbcon->helper.info->flags |= FBINFO_HWACCEL_DISABLED; - } -} - -void...

...+41,9 @@ struct nouveau_fbdev { struct nvif_object gdi; struct nvif_object blit; struct nvif_object twod; + + struct mutex hotplug_lock; + bool hotplug_waiting; }; void nouveau_fbcon_restore(void); @@ -68,6 +71,8 @@ void nouveau_fbcon_set_suspend(struct drm_device *dev, int state); void nouveau_fbcon_accel_save_disable(struct drm_device *dev); void nouveau_fbcon_accel_restore(struct drm_device *dev); +void nouveau_fbcon_output_poll_changed(struct drm_device *dev); +void nouveau_fbcon_hotplug_resume(struct nouveau_fbdev *fbcon); extern int nouveau_nofbaccel; #endif /* __NV50_FBCON_H__ */ -- 2.17.1

...+41,9 @@ struct nouveau_fbdev { struct nvif_object gdi; struct nvif_object blit; struct nvif_object twod; + + struct mutex hotplug_lock; + bool hotplug_waiting; }; void nouveau_fbcon_restore(void); @@ -68,6 +71,10 @@ void nouveau_fbcon_set_suspend(struct drm_device *dev, int state); void nouveau_fbcon_accel_save_disable(struct drm_device *dev); void nouveau_fbcon_accel_restore(struct drm_device *dev); +void nouveau_fbcon_output_poll_changed(struct drm_device *dev); +void nouveau_fbcon_hotplug_resume(struct nouveau_fbdev *fbcon); +bool nouveau_fbcon_hotplugged_in_suspend(struct nouveau_fbdev *fbcon); + extern i...

Hi folks, Currently, when the device is removed (or the driver is unbound) the nouveau_drm structure de-allocated. However, it's still accessible from and used by some DRM layer callbacks. For example, file handles can be closed after the device has been removed (physically or otherwise). This series converts the Nouveau device structure to be allocated and de-allocated with the

Latest version of https://patchwork.freedesktop.org/series/46815/ , with one small change re: ilia Lyude Paul (5): drm/nouveau: Fix bogus drm_kms_helper_poll_enable() placement drm/nouveau: Remove duplicate poll_enable() in pmops_runtime_suspend() drm/nouveau: Fix deadlock with fb_helper with async RPM requests drm/nouveau: Use pm_runtime_get_noresume() in connector_detect()

...ect gdi; > struct nvif_object blit; > struct nvif_object twod; > + > + struct mutex hotplug_lock; > + bool hotplug_waiting; > }; > > void nouveau_fbcon_restore(void); > @@ -68,6 +71,8 @@ void nouveau_fbcon_set_suspend(struct drm_device *dev, int state); > void nouveau_fbcon_accel_save_disable(struct drm_device *dev); > void nouveau_fbcon_accel_restore(struct drm_device *dev); > > +void nouveau_fbcon_output_poll_changed(struct drm_device *dev); > +void nouveau_fbcon_hotplug_resume(struct nouveau_fbdev *fbcon); > extern int nouveau_nofbaccel; > > #endif /* __NV...

Next version of https://patchwork.freedesktop.org/series/46815/ Same as previous version, but some small changes made to commit messages and acks/rbs have been added Lyude Paul (5): drm/nouveau: Fix bogus drm_kms_helper_poll_enable() placement drm/nouveau: Remove duplicate poll_enable() in pmops_runtime_suspend() drm/nouveau: Fix deadlock with fb_helper with async RPM requests