dri-devel May 2016

dri-devel@lists.freedesktop.org

212 participants
1404 discussions

[PATCH 00/12] drm/atomic: Rename async to nonblocking.
by Maarten Lankhorst 02 May '16

02 May '16

Userspace calls the atomic ioctl with DRM_MODE_ATOMIC_NONBLOCK when it doesn't want the ioctl to block for completion. In the kernel it's called async commit, which is confusing. The legacy pageflip and atomic ioctl also have a flag DRM_MODE_PAGE_FLIP_ASYNC which means that the flip should happen as soon as possible, even if it causes tearing. Because of this we should rename async commit to nonblocking commit, which should reduce a lot of the confusion. This is done partially with a cocci script, and inspection by hand because cocci doesn't patch comments. Maarten Lankhorst (12): drm/atomic: Rename async parameter to nonblocking. drm/atomic: Rename drm_atomic_async_commit to nonblocking. drm/arm/hdlcd: Rename async to nonblock. drm/exynos: Rename async to nonblock. drm/msm: Rename async to nonblock. drm/omapdrm: Rename async to nonblock. drm/rcar-du: Rename async to nonblock. drm/sti: Rename async to nonblock. drm/tegra: Rename async to nonblock. drm/rockchip: Rename async to nonblock. drm/vc4: Rename async to nonblock. drm/i915: Rename async to nonblock. drivers/gpu/drm/arm/hdlcd_drv.c | 2 +- drivers/gpu/drm/drm_atomic.c | 12 +++++------ drivers/gpu/drm/drm_atomic_helper.c | 34 +++++++++++++++--------------- drivers/gpu/drm/exynos/exynos_drm_drv.c | 4 ++-- drivers/gpu/drm/exynos/exynos_drm_drv.h | 2 +- drivers/gpu/drm/i915/intel_display.c | 16 +++++++------- drivers/gpu/drm/msm/msm_atomic.c | 9 ++++---- drivers/gpu/drm/msm/msm_drv.h | 2 +- drivers/gpu/drm/omapdrm/omap_drv.c | 4 ++-- drivers/gpu/drm/rcar-du/rcar_du_kms.c | 5 +++-- drivers/gpu/drm/rockchip/rockchip_drm_fb.c | 6 +++--- drivers/gpu/drm/sti/sti_drv.c | 6 +++--- drivers/gpu/drm/tegra/drm.c | 6 +++--- drivers/gpu/drm/vc4/vc4_kms.c | 6 +++--- include/drm/drm_atomic.h | 2 +- include/drm/drm_atomic_helper.h | 2 +- include/drm/drm_crtc.h | 8 +++---- 17 files changed, 63 insertions(+), 63 deletions(-) -- 2.5.5

5 17

[Bug 95017] [drm:radeon_ib_ring_tests] *ERROR* radeon: failed testing IB on GFX ring (-35).
by bugzilla-daemon＠freedesktop.org 02 May '16

02 May '16

https://bugs.freedesktop.org/show_bug.cgi?id=95017 --- Comment #6 from Mathieu Malaterre <mathieu.malaterre(a)gmail.com> --- (In reply to Michel Dänzer from comment #4) > (In reply to Mathieu Malaterre from comment #3) > > If AGP is simply not supported on PowerPC, it would be nice to have a > > clearer message (or at least a warning). > > It is supported, it's just unstable on many PowerMacs. > > A patch disabling AGP by default on PowerMacs or even PPC in general > probably wouldn't be rejected. :) Something like the patch: defaulttopci_powerpc.patch ? Should I submit it first on mesa-dev ? -- You are receiving this mail because: You are the assignee for the bug.

1 0

[Bug 95017] [drm:radeon_ib_ring_tests] *ERROR* radeon: failed testing IB on GFX ring (-35).
by bugzilla-daemon＠freedesktop.org 02 May '16

02 May '16

https://bugs.freedesktop.org/show_bug.cgi?id=95017 --- Comment #5 from Mathieu Malaterre <mathieu.malaterre(a)gmail.com> --- Created attachment 123411 --> https://bugs.freedesktop.org/attachment.cgi?id=123411&action=edit default to PCI for PowerPC -- You are receiving this mail because: You are the assignee for the bug.

1 0

[PATCH v4 0/7] drm: Add fbdev deferred io support to helpers
by Noralf Trønnes 02 May '16

02 May '16

This patchset adds fbdev deferred io support to drm_fb_helper and drm_fb_cma_helper. It channels fbdev mmap and fb_{write,fillrect,copyarea,imageblit} damage through the (struct drm_framebuffer_funcs)->dirty callback on the fb_helper framebuffer which will always run in process context. I have also added patches that converts qxl and udl to use this deferred io support. I have only compile tested it, no functional testing. I know that qxl is purely a software thing so I could actually test it, but I have never used qemu so I'm not keen on spending a lot of time on that. This was originally part of the tinydrm patchset. Changes since v3: - drm/fb-helper: Add fb_deferred_io support - Don't use forward decl, move drm_fb_helper_dirty_work() - Use DIV_ROUND_UP in drm_fb_helper_deferred_io() Changes since v2: - drm/rect: Add some drm_clip_rect utility functions - This patch is dropped - drm/fb-helper: Add fb_deferred_io support - FB_DEFERRED_IO is now always selected by DRM_KMS_FB_HELPER, ifdef removed - The drm_clip_rect utility functions are dropped, so open code it - docs: use & to denote structs - drm/qxl: Use drm_fb_helper deferred_io support - The drm_clip_rect_{width/height} functions are dropped, so open code it Changes since v1: - drm/fb-helper: Add fb_deferred_io support - Use a dedicated worker to run the framebuffer flushing like qxl does - Add parameter descriptions to drm_fb_helper_deferred_io - fbdev: fb_defio: Export fb_deferred_io_mmap - Expand commit message - drm/qxl: Use drm_fb_helper deferred_io support - Add FIXME about special dirty() callback for fbdev - Remove note in commit message about deferred worker, drm_fb_helper is similar to qxl now. - drm/udl: Use drm_fb_helper deferred_io support - No need to enable deferred_io by default since drm_fb_helper uses a dedicated worker for flushing Changes since RFC: - Fix drm_clip_rect use to be exclusive on x2/y2 - Put drm_clip_rect functions in drm_rect.{h,c} - Take into account that (struct fb_ops *)->fb_{write,...}() can be called from atomic context (spin_lock_irqsave) - Export fb_deferred_io_mmap() - Add some more documentation - Add qxl and udl patches Noralf Trønnes (7): drm/udl: Change drm_fb_helper_sys_*() calls to sys_*() drm/qxl: Change drm_fb_helper_sys_*() calls to sys_*() drm/fb-helper: Add fb_deferred_io support fbdev: fb_defio: Export fb_deferred_io_mmap drm/fb-cma-helper: Add fb_deferred_io support drm/qxl: Use drm_fb_helper deferred_io support drm/udl: Use drm_fb_helper deferred_io support drivers/gpu/drm/Kconfig | 1 + drivers/gpu/drm/drm_fb_cma_helper.c | 178 ++++++++++++++++++++++++++-- drivers/gpu/drm/drm_fb_helper.c | 103 ++++++++++++++++- drivers/gpu/drm/qxl/qxl_display.c | 9 +- drivers/gpu/drm/qxl/qxl_drv.h | 7 +- drivers/gpu/drm/qxl/qxl_fb.c | 223 +++++++++--------------------------- drivers/gpu/drm/qxl/qxl_kms.c | 4 - drivers/gpu/drm/udl/udl_drv.h | 2 - drivers/gpu/drm/udl/udl_fb.c | 140 +--------------------- drivers/video/fbdev/core/fb_defio.c | 3 +- include/drm/drm_fb_cma_helper.h | 14 +++ include/drm/drm_fb_helper.h | 15 +++ include/linux/fb.h | 1 + 13 files changed, 372 insertions(+), 328 deletions(-) -- 2.2.2

3 17

[PATCH] of: include errno.h in of_graph.h
by Arnd Bergmann 02 May '16

02 May '16

When CONFIG_OF is disabled, we have to include linux/errno.h before including of_graph.h, or get build errors like in the newly added sun4i drm driver: In file included from ../drivers/gpu/drm/sun4i/sun4i_drv.c:14:0: include/linux/of_graph.h: In function 'of_graph_parse_endpoint': include/linux/of_graph.h:58:10: error: 'ENOSYS' undeclared (first use in this function) A better solution is to ensure that the header can be included by itself, so let's include linux/errno.h here to fix the error we just got, and any similar future error. Signed-off-by: Arnd Bergmann <arnd(a)arndb.de> Fixes: 9026e0d122ac ("drm: Add Allwinner A10 Display Engine support") --- include/linux/of_graph.h | 1 + 1 file changed, 1 insertion(+) diff --git a/include/linux/of_graph.h b/include/linux/of_graph.h index f8bcd0e21a26..bb3a5a2cd570 100644 --- a/include/linux/of_graph.h +++ b/include/linux/of_graph.h @@ -15,6 +15,7 @@ #define __LINUX_OF_GRAPH_H #include <linux/types.h> +#include <linux/errno.h> /** * struct of_endpoint - the OF graph endpoint data structure -- 2.7.0

2 1

[PATCH 1/2] drm/etnaviv: take etnaviv_gem_obj in etnaviv_gem_mmap_obj
by Lucas Stach 02 May '16

02 May '16

This function will be changed to be called indirectly and this prototype change brings it in line with all the other indirect object calls. Signed-off-by: Lucas Stach <l.stach(a)pengutronix.de> --- drivers/gpu/drm/etnaviv/etnaviv_gem.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem.c b/drivers/gpu/drm/etnaviv/etnaviv_gem.c index 281c6eca20a8..4d2277f6de03 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gem.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gem.c @@ -129,10 +129,9 @@ void etnaviv_gem_put_pages(struct etnaviv_gem_object *etnaviv_obj) /* when we start tracking the pin count, then do something here */ } -static int etnaviv_gem_mmap_obj(struct drm_gem_object *obj, +static int etnaviv_gem_mmap_obj(struct etnaviv_gem_object *etnaviv_obj, struct vm_area_struct *vma) { - struct etnaviv_gem_object *etnaviv_obj = to_etnaviv_bo(obj); pgprot_t vm_page_prot; vma->vm_flags &= ~VM_PFNMAP; @@ -151,9 +150,9 @@ static int etnaviv_gem_mmap_obj(struct drm_gem_object *obj, * in particular in the case of mmap'd dmabufs) */ fput(vma->vm_file); - get_file(obj->filp); + get_file(etnaviv_obj->base.filp); vma->vm_pgoff = 0; - vma->vm_file = obj->filp; + vma->vm_file = etnaviv_obj->base.filp; vma->vm_page_prot = vm_page_prot; } @@ -173,7 +172,7 @@ int etnaviv_gem_mmap(struct file *filp, struct vm_area_struct *vma) } obj = to_etnaviv_bo(vma->vm_private_data); - return etnaviv_gem_mmap_obj(vma->vm_private_data, vma); + return etnaviv_gem_mmap_obj(obj, vma); } int etnaviv_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) -- 2.8.0.rc3

2 4

[PULL] drm-intel-fixes
by Jani Nikula 02 May '16

02 May '16

Hi Dave, i915 fixes for 4.6. A bit more than I'd like at this stage, but OTOH they're all stable material. BR, Jani. The following changes since commit 02da2d72174c61988eb4456b53f405e3ebdebce4: Linux 4.6-rc5 (2016-04-24 16:17:05 -0700) are available in the git repository at: git://anongit.freedesktop.org/drm-intel tags/drm-intel-fixes-2016-05-02 for you to fetch changes up to 4ea3959018d09edfa36a9e7b5ccdbd4ec4b99e49: drm/i915: Make RPS EI/thresholds multiple of 25 on SNB-BDW (2016-04-27 10:57:00 +0300) ---------------------------------------------------------------- Chris Wilson (1): drm/i915: Avoid stalling on pending flips for legacy cursor updates Imre Deak (2): drm/i915: Fix system resume if PCI device remained enabled drm/i915/ddi: Fix eDP VDD handling during booting and suspend/resume Mika Kahola (1): drm/i915: Fix eDP low vswing for Broadwell Shashank Sharma (1): drm/i915: Fake HDMI live status Ville Syrjälä (1): drm/i915: Make RPS EI/thresholds multiple of 25 on SNB-BDW drivers/gpu/drm/i915/i915_drv.c | 32 +++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/i915_reg.h | 9 ++++++++- drivers/gpu/drm/i915/intel_ddi.c | 22 +++++++++++++--------- drivers/gpu/drm/i915/intel_display.c | 3 +++ drivers/gpu/drm/i915/intel_dp.c | 4 ++-- drivers/gpu/drm/i915/intel_drv.h | 2 ++ drivers/gpu/drm/i915/intel_hdmi.c | 12 ++++++++++-- 7 files changed, 69 insertions(+), 15 deletions(-) -- Jani Nikula, Intel Open Source Technology Center

1 0

[PATCH 0/3] drm/msm: Remove drm_driver load/unload ops
by Archit Taneja 02 May '16

02 May '16

Registering the drm_device using the drm_driver load/unload ops is deprecated since it is prone to race conditions. The second and third patches removes the usage of these ops. The first patch prevents warnings when we try to remove the drm/msm kernel module. Archit Taneja (3): drm/msm/hdmi: Prevent gpio_free related kernel warnings drm/msm: Centralize connector registration/unregistration drm/msm: Drop load/unload drm_driver ops drivers/gpu/drm/msm/dsi/dsi_manager.c | 27 ++--- drivers/gpu/drm/msm/edp/edp_connector.c | 20 +-- drivers/gpu/drm/msm/hdmi/hdmi_connector.c | 36 +++--- drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c | 16 +-- drivers/gpu/drm/msm/msm_drv.c | 134 +++++++++++++-------- 5 files changed, 113 insertions(+), 120 deletions(-) -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation

4 16

blank screen on boot with i915/DRM_FBDEV_EMULATION
by Florian Zumbiehl 02 May '16

02 May '16

Hi, I just compiled a new kernel version 4.4.5, replacing a previous 4.1.9 for my thinkpad x40, which replaced CONFIG_DRM_I915_FBDEV with CONFIG_DRM_FBDEV_EMULATION--now, when the i915 and fbcon modules get loaded during boot, the screen goes blank (with backlight still on), which didn't happen with the old kernel. The funny thing is that starting an X server makes text reappear on the console!? (And the X server also works just fine, as far as I can tell so far ...) i915 is loaded with modeset=1 (I remember I had some problem without it with some earlier kernels, I don't really remember the details anymore ...). Any hints what to try? Any further information that could be useful? Regards, Florian

5 13

[Bug 95085] Invalid sampling of second texture in fragment shader that have two samplers with different parameters.
by bugzilla-daemon＠freedesktop.org 01 May '16

01 May '16

https://bugs.freedesktop.org/show_bug.cgi?id=95085 --- Comment #18 from Leonid Maksymchuk <leonmaxx(a)gmail.com> --- Sorry for bothering You with our bugs. Next time I'll review the code more precisely, before reporting anything. Thanks everyone for your help. -- You are receiving this mail because: You are the assignee for the bug.

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel May 2016