dri-devel July 2018

dri-devel@lists.freedesktop.org

277 participants
1883 discussions

[Bug 101651] [radeonsi][hawaii] Borderlands 2 rendering issues with recent mesa/llvm
by bugzilla-daemon＠freedesktop.org 02 Aug '18

02 Aug '18

https://bugs.freedesktop.org/show_bug.cgi?id=101651 Bug ID: 101651 Summary: [radeonsi][hawaii] Borderlands 2 rendering issues with recent mesa/llvm Product: Mesa Version: git Hardware: Other OS: All Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel(a)lists.freedesktop.org Reporter: ptpzz(a)yandex.ru QA Contact: dri-devel(a)lists.freedesktop.org Created attachment 132354 --> https://bugs.freedesktop.org/attachment.cgi?id=132354&action=edit screenshot (wrong rendering) AFAICS it's related to the VCCZ bug. There is a workaround for SI/CIK in the AMDGPU LLVM backend (SIInsertWaitcnts.cpp) but it doesn't handle this case. The attached patch fixes it for me by relaxing the workaround condition, but probably it can be improved if there is a detailed info on that bug. OTOH I guess it won't have a noticeable effect on performance anyway, so maybe it's safer to use the patch as is. The issue is reproducible with the reduced OpenCL test in the attachment: $ make // run on cpu - reference result $ ./testcpu 117 // OpenCL/gpu - wrong result $ LD_PRELOAD=/opt/llvm64_master/lib/libLLVM-5.0svn.so ./testgpu 99 // OpenCL/gpu + patch - correct result $ LD_PRELOAD=/opt/llvm64_patched/lib/libLLVM-5.0svn.so ./testgpu 117 -- You are receiving this mail because: You are the assignee for the bug.

1 6

[Bug 99528] Wine game doesn't redraw properly in fullscreen
by bugzilla-daemon＠freedesktop.org 02 Aug '18

02 Aug '18

https://bugs.freedesktop.org/show_bug.cgi?id=99528 Bug ID: 99528 Summary: Wine game doesn't redraw properly in fullscreen Product: Mesa Version: git Hardware: Other OS: Linux (All) Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel(a)lists.freedesktop.org Reporter: dark.shadow4(a)web.de QA Contact: dri-devel(a)lists.freedesktop.org Created attachment 129133 --> https://bugs.freedesktop.org/attachment.cgi?id=129133&action=edit Trace from the game The windows game DXBall doesn't properly redraw in fullscreen mode. I assume it's an radeonsi bug, since others reported it working fine. Also, LIBGL_ALWAYS_SOFTWARE makes it render properly, although fairly slow. Apitrace attached, notice how the upper half is a static image while the lower half is how it should look like. Also noteworthy is that it works sometimes. Seemingly random, the fullscreen game either renders fine or displays just a static image from the first menu. System -Arch 64bit -Radeon R9 285 using amdgpu and opensource driver -Mesa 17.1.0-devel (git-0054dded03) Game in question: http://www.blitwise.com/Scorched_Reality/DXBall19.EXE Just run it with wine2.0 -- You are receiving this mail because: You are the assignee for the bug.

1 10

[Bug 102638] Incorrect rendering in OpenGL 4 (Unigine Superposition + others)
by bugzilla-daemon＠freedesktop.org 02 Aug '18

02 Aug '18

https://bugs.freedesktop.org/show_bug.cgi?id=102638 Bug ID: 102638 Summary: Incorrect rendering in OpenGL 4 (Unigine Superposition + others) Product: Mesa Version: 17.1 Hardware: x86-64 (AMD64) OS: Linux (All) Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel(a)lists.freedesktop.org Reporter: mj.wilson.uk(a)gmail.com QA Contact: dri-devel(a)lists.freedesktop.org Ryzen 1600X, 16GiB RAM, AMD RX570, Fedora 26 (clean install) mesa-dri-drivers-17.1.7-1.fc26 Unigine Superposition Benchmark v1.0 (build 6184) Basic Superposition and other OpenGL 4 (I guess) applications fail to render correctly. It seems that only one of the rendering stages is being shown -- it looks like the specular highlights from a cubemap or similar. Tacoma (running on Unity) does the same. Furthermore, there is a large disparity between the reported frame rate (30+) and the actual frame rate I see (perhaps 1/3 or 1/4 of that). Both worked perfectly in Fedora 25 (same hardware) with mesa-dri-drivers-17.0.5-3.fc2. I'm guessing this is a problem in the radeonsi driver. -- You are receiving this mail because: You are the assignee for the bug.

1 6

[Bug 106531] Age of Wonders 3 rendering failure on 18.1.0-rc
by bugzilla-daemon＠freedesktop.org 02 Aug '18

02 Aug '18

https://bugs.freedesktop.org/show_bug.cgi?id=106531 Bug ID: 106531 Summary: Age of Wonders 3 rendering failure on 18.1.0-rc Product: Mesa Version: git Hardware: Other OS: Linux (All) Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel(a)lists.freedesktop.org Reporter: johnz(a)lunarg.com QA Contact: dri-devel(a)lists.freedesktop.org Created attachment 139582 --> https://bugs.freedesktop.org/attachment.cgi?id=139582&action=edit Picture of Age Of Wonders 3 rendering problem on 18.1.0-rc's NOTE: This is for the 18.1 branch for the current RC's, you can reproduce this bug on the radeonsi driver using the git tag 'mesa-18.1.0-rc4'. This bug has NOT been replicated on the 18.0 branch. Overview: Age of Wonders 3 game displays large white squares on menu screen. No buttons are available in menu screen. Steps To Reproduce: Fire up Age of Wonders 3 in steam using radeonsi driver on 18.1.0-rc* releases. The problem should present itself in the menu of the game. Hardware: R9280, RX480, and R9Fury have this problem. Builds: This bug has been seen on Mesa RC's of 18.1.0 Attached is a screenshot of the problem. -- You are receiving this mail because: You are the assignee for the bug.

1 2

[Bug 87059] Graphical glitches r9290
by bugzilla-daemon＠freedesktop.org 02 Aug '18

02 Aug '18

https://bugs.freedesktop.org/show_bug.cgi?id=87059 Bug ID: 87059 Summary: Graphical glitches r9290 Product: Mesa Version: git Hardware: Other OS: All Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel(a)lists.freedesktop.org Reporter: jarkko_korpi(a)hotmail.com Created attachment 110519 --> https://bugs.freedesktop.org/attachment.cgi?id=110519&action=edit glitches I just bought This war of mine and there is some animation at start and it has weird black boxes. -- You are receiving this mail because: You are the assignee for the bug.

1 14

[PATCH] drm/dp: add missing ')' to I2C nack debug message
by Paulo Zanoni 01 Aug '18

01 Aug '18

"(an unmatched left parenthesis creates an unresolved tension that will stay with you all day." -- Randall Munroe Signed-off-by: Paulo Zanoni <paulo.r.zanoni(a)intel.com> --- drivers/gpu/drm/drm_dp_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 0cccbcb2d03e..8c6b9fd89f8a 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -850,7 +850,8 @@ static int drm_dp_i2c_do_msg(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) return ret; case DP_AUX_I2C_REPLY_NACK: - DRM_DEBUG_KMS("I2C nack (result=%d, size=%zu\n", ret, msg->size); + DRM_DEBUG_KMS("I2C nack (result=%d, size=%zu)\n", + ret, msg->size); aux->i2c_nack_count++; return -EREMOTEIO; -- 2.14.4

3 2

[PATCH] drm/rockchip: Convert drm_atomic_helper_suspend/resume()
by Souptick Joarder 01 Aug '18

01 Aug '18

convert drm_atomic_helper_suspend/resume() to use drm_mode_config_helper_suspend/resume(). With this conversion, rockchip_drm_fb_resume() and rockchip_drm_fb_suspend() will not be used anymore. Both of these functions can be removed. Also, in struct rockchip_drm_private state will not be used anymore. So this can be removed forever. Signed-off-by: Souptick Joarder <jrdr.linux(a)gmail.com> Signed-off-by: Ajit Negi <ajitn.linux(a)gmail.com> --- drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 46 ++--------------------------- drivers/gpu/drm/rockchip/rockchip_drm_drv.h | 1 - 2 files changed, 2 insertions(+), 45 deletions(-) diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c index f814d37..dd12390 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c @@ -243,60 +243,18 @@ static void rockchip_drm_unbind(struct device *dev) }; #ifdef CONFIG_PM_SLEEP -static void rockchip_drm_fb_suspend(struct drm_device *drm) -{ - struct rockchip_drm_private *priv = drm->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(&priv->fbdev_helper, 1); - console_unlock(); -} - -static void rockchip_drm_fb_resume(struct drm_device *drm) -{ - struct rockchip_drm_private *priv = drm->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(&priv->fbdev_helper, 0); - console_unlock(); -} - static int rockchip_drm_sys_suspend(struct device *dev) { struct drm_device *drm = dev_get_drvdata(dev); - struct rockchip_drm_private *priv; - - if (!drm) - return 0; - - drm_kms_helper_poll_disable(drm); - rockchip_drm_fb_suspend(drm); - priv = drm->dev_private; - priv->state = drm_atomic_helper_suspend(drm); - if (IS_ERR(priv->state)) { - rockchip_drm_fb_resume(drm); - drm_kms_helper_poll_enable(drm); - return PTR_ERR(priv->state); - } - - return 0; + return drm_mode_config_helper_suspend(drm); } static int rockchip_drm_sys_resume(struct device *dev) { struct drm_device *drm = dev_get_drvdata(dev); - struct rockchip_drm_private *priv; - if (!drm) - return 0; - - priv = drm->dev_private; - drm_atomic_helper_resume(drm, priv->state); - rockchip_drm_fb_resume(drm); - drm_kms_helper_poll_enable(drm); - - return 0; + return drm_mode_config_helper_resume(drm); } #endif diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.h b/drivers/gpu/drm/rockchip/rockchip_drm_drv.h index 3a6ebfc..d67ad0a 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_drv.h +++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.h @@ -51,7 +51,6 @@ struct rockchip_crtc_state { struct rockchip_drm_private { struct drm_fb_helper fbdev_helper; struct drm_gem_object *fbdev_bo; - struct drm_atomic_state *state; struct iommu_domain *domain; struct mutex mm_lock; struct drm_mm mm; -- 1.9.1

2 5

[PATCH 0/5] R-Car D3 LVDS/HDMI support
by Ulrich Hecht 01 Aug '18

01 Aug '18

Hi! This adds D3 support to the DU and LVDS drivers, not including LVDS PLL support. It also adds LVDS encoders to the D3 device tree, and LVDS decoder, HDMI encoder and HDMI output connector to the Draak device tree. In theory that should be good enough to provide HDMI output on the Draak board, but in practice the lack of LVDS PLL support prevents generation of close-enough dot clock frequencies. This series is based on renesas-drivers-2018-05-02-v4.17-rc3 and requires Jacopo's "drm: bridge: Add thc63lvd1024 LVDS decoder driver" patch. CU Uli Kieran Bingham (1): arm64: dts: renesas: r8a77995: Add LVDS support Koji Matsuoka (1): drm: rcar-du: Add r8a77995 device support Ulrich Hecht (3): drm: rcar-du: lvds: Add R8A77995 support arm64: dts: renesas: r8a77995-draak: add HDMI output arm64: dts: renesas: r8a77995-draak: add X12 input dot clock arch/arm64/boot/dts/renesas/r8a77995-draak.dts | 91 ++++++++++++++++++++++++++ arch/arm64/boot/dts/renesas/r8a77995.dtsi | 56 ++++++++++++++++ drivers/gpu/drm/rcar-du/rcar_du_drv.c | 26 ++++++++ drivers/gpu/drm/rcar-du/rcar_lvds.c | 6 ++ 4 files changed, 179 insertions(+) -- 2.7.4

7 25

[PATCH] drm/nouveau: Fix potential memory access error in nouveau_debugfs_pstate_get()
by Lyude Paul 01 Aug '18

01 Aug '18

nouveau_debugfs(drm) will never be NULL, because we're taking the value of the potentially null device pointer and adding to it so it isn't 0x0. So, check if drm is NULL instead. Signed-off-by: Lyude Paul <lyude(a)redhat.com> Cc: Karol Herbst <karolherbst(a)gmail.com> --- drivers/gpu/drm/nouveau/nouveau_debugfs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c index 963a4dba8213..d83313c02913 100644 --- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c +++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c @@ -50,13 +50,14 @@ static int nouveau_debugfs_pstate_get(struct seq_file *m, void *data) { struct drm_device *drm = m->private; - struct nouveau_debugfs *debugfs = nouveau_debugfs(drm); + struct nouveau_debugfs *debugfs; struct nvif_object *ctrl = &debugfs->ctrl; struct nvif_control_pstate_info_v0 info = {}; int ret, i; - if (!debugfs) + if (!drm) return -ENODEV; + debugfs = nouveau_debugfs(drm); ret = nvif_mthd(ctrl, NVIF_CONTROL_PSTATE_INFO, &info, sizeof(info)); if (ret) -- 2.17.1

2 1

[PATCH 1/2] drm/scheduler: Add job dependency trace.
by Andrey Grodzovsky 01 Aug '18

01 Aug '18

During debug sessions I encountered a need to trace back a job dependecy a few steps back to the first failing job. This trace helpped me a lot. Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky(a)amd.com> --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 8 ++++++-- drivers/gpu/drm/scheduler/gpu_scheduler_trace.h | 24 ++++++++++++++++++++++++ 2 files changed, 30 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index 3f2fc5e..45703b9 100644 --- a/drivers/gpu/drm/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/scheduler/gpu_scheduler.c @@ -506,9 +506,13 @@ drm_sched_entity_pop_job(struct drm_sched_entity *entity) if (!sched_job) return NULL; - while ((entity->dependency = sched->ops->dependency(sched_job, entity))) - if (drm_sched_entity_add_dependency_cb(entity)) + while ((entity->dependency = sched->ops->dependency(sched_job, entity))) { + if (drm_sched_entity_add_dependency_cb(entity)) { + + trace_drm_sched_job_wait_dep(sched_job, entity->dependency); return NULL; + } + } /* skip jobs from entity that marked guilty */ if (entity->guilty && atomic_read(entity->guilty)) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler_trace.h b/drivers/gpu/drm/scheduler/gpu_scheduler_trace.h index 4998ad9..1626f39 100644 --- a/drivers/gpu/drm/scheduler/gpu_scheduler_trace.h +++ b/drivers/gpu/drm/scheduler/gpu_scheduler_trace.h @@ -74,6 +74,30 @@ TRACE_EVENT(drm_sched_process_job, TP_printk("fence=%p signaled", __entry->fence) ); +TRACE_EVENT(drm_sched_job_wait_dep, + TP_PROTO(struct drm_sched_job *sched_job, struct dma_fence *fence), + TP_ARGS(sched_job, fence), + TP_STRUCT__entry( + __field(const char *,name) + __field(uint64_t, id) + __field(struct dma_fence *, fence) + __field(uint64_t, ctx) + __field(unsigned, seqno) + ), + + TP_fast_assign( + __entry->name = sched_job->sched->name; + __entry->id = sched_job->id; + __entry->fence = fence; + __entry->ctx = fence->context; + __entry->seqno = fence->seqno; + ), + TP_printk("job ring=%s, id=%llu, depends fence=%p, context=%llu, seq=%u", + __entry->name, __entry->id, + __entry->fence, __entry->ctx, + __entry->seqno) +); + #endif /* This part must be outside protection */ -- 2.7.4

3 3

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel July 2018