dri-devel September 2017

dri-devel@lists.freedesktop.org

210 participants
1526 discussions

[Bug 101160] Venus PRO R9 M265X amdgpu: ring 0 test failed
by bugzilla-daemon＠freedesktop.org 05 Sep '17

05 Sep '17

https://bugs.freedesktop.org/show_bug.cgi?id=101160 Bug ID: 101160 Summary: Venus PRO R9 M265X amdgpu: ring 0 test failed Product: DRI Version: unspecified Hardware: Other OS: All Status: NEW Severity: normal Priority: medium Component: DRM/AMDgpu Assignee: dri-devel(a)lists.freedesktop.org Reporter: mercuriete(a)yahoo.es Similar bug to https://bugs.freedesktop.org/show_bug.cgi?id=92270 System freeze when booting. tty1 output (something like): amdgpu: ring 0 test failed lspci -nn 01:00.0 Display controller [0380]: Advanced Micro Devices, Inc. [AMD/ATI] Venus PRO [Radeon HD 8850M / R9 M265X] [1002:6823] uname -a Linux localhost.localdomain 4.11.1-gentoo #7 SMP PREEMPT Wed May 24 00:50:03 WEST 2017 x86_64 Intel(R) Core(TM) i7-4720HQ CPU @ 2.60GHz GenuineIntel GNU/Linux The graphic card works properly when using radeonsi kernel module. I am trying to use amdgpu in this SI card I canot provide any more logs because my system frezees completely when the module is loaded. I tried to update to latest linux firmware but I am already in the latest ones. I think do a bisect doesn't makes sense because i feel this card have never been working with amdgpu. I need help to have useful logs in order to help you to fix the problem. -- You are receiving this mail because: You are the assignee for the bug.

1 6

[PATCH 1/2] drm/atomic: Consitfy mode parameter to drm_atomic_set_mode_for_crtc()
by ville.syrjala＠linux.intel.com 05 Sep '17

05 Sep '17

From: Ville Syrjälä <ville.syrjala(a)linux.intel.com> drm_atomic_set_mode_for_crtc() doesn't modify the passed mode, so let's make it const. Signed-off-by: Ville Syrjälä <ville.syrjala(a)linux.intel.com> --- drivers/gpu/drm/drm_atomic.c | 2 +- include/drm/drm_atomic.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index cdec19a86af3..21cb78202c90 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -328,7 +328,7 @@ static s32 __user *get_out_fence_for_crtc(struct drm_atomic_state *state, * Zero on success, error code on failure. Cannot return -EDEADLK. */ int drm_atomic_set_mode_for_crtc(struct drm_crtc_state *state, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct drm_mode_modeinfo umode; diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h index 8645dcdef031..0196f264a418 100644 --- a/include/drm/drm_atomic.h +++ b/include/drm/drm_atomic.h @@ -520,7 +520,7 @@ __drm_atomic_get_current_plane_state(struct drm_atomic_state *state, int __must_check drm_atomic_set_mode_for_crtc(struct drm_crtc_state *state, - struct drm_display_mode *mode); + const struct drm_display_mode *mode); int __must_check drm_atomic_set_mode_prop_for_crtc(struct drm_crtc_state *state, struct drm_property_blob *blob); -- 2.10.2

3 4

[Bug 102338] QXL driver causes oops on boot, kernel 4.12
by bugzilla-daemon＠freedesktop.org 05 Sep '17

05 Sep '17

https://bugs.freedesktop.org/show_bug.cgi?id=102338 Bug ID: 102338 Summary: QXL driver causes oops on boot, kernel 4.12 Product: DRI Version: unspecified Hardware: x86-64 (AMD64) OS: All Status: NEW Severity: major Priority: medium Component: DRM/other Assignee: dri-devel(a)lists.freedesktop.org Reporter: chuckebbert.lk(a)gmail.com This is from 4.12-rc6, still happening on 4.12.5. Taken from https://bugzilla.redhat.com/show_bug.cgi?id=1465148 int ttm_bo_kmap(struct ttm_buffer_object *bo, unsigned long start_page, unsigned long num_pages, struct ttm_bo_kmap_obj *map) { struct ttm_mem_type_manager *man = &bo->bdev->man[bo->mem.mem_type]; unsigned long offset, size; int ret; 589-> BUG_ON(!list_empty(&bo->swap)); Jun 26 10:51:00 bld-fedora27 kernel: ------------[ cut here ]------------ Jun 26 10:51:00 bld-fedora27 kernel: kernel BUG at drivers/gpu/drm/ttm/ttm_bo_util.c:589! Jun 26 10:51:00 bld-fedora27 kernel: invalid opcode: 0000 [#1] SMP Jun 26 10:51:00 bld-fedora27 kernel: Modules linked in: sunrpc crct10dif_pclmul crc32_pclmul ghash_clmulni_intel ppdev snd_hda_codec_generic snd_hda_intel snd_hda_codec snd_hda_core snd_hwdep snd_seq Jun 26 10:51:00 bld-fedora27 kernel: CPU: 1 PID: 339 Comm: plymouthd Not tainted 4.12.0-0.rc6.git3.1.fc27.x86_64 #1 Jun 26 10:51:00 bld-fedora27 kernel: Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011 Jun 26 10:51:00 bld-fedora27 kernel: task: ffff91dd77990000 task.stack: ffffb69e00720000 Jun 26 10:51:00 bld-fedora27 kernel: RIP: 0010:ttm_bo_kmap+0x1af/0x250 [ttm] Jun 26 10:51:00 bld-fedora27 kernel: RSP: 0018:ffffb69e00723ba8 EFLAGS: 00010206 Jun 26 10:51:00 bld-fedora27 kernel: RAX: ffff91dd77987b50 RBX: ffff91dd779dc858 RCX: ffff91dd779dcc08 Jun 26 10:51:00 bld-fedora27 kernel: RDX: 0000000000000300 RSI: 0000000000000000 RDI: ffff91dd77430d90 Jun 26 10:51:00 bld-fedora27 kernel: RBP: ffffb69e00723be0 R08: 0000000000010004 R09: ffff91dd779dc928 Jun 26 10:51:00 bld-fedora27 kernel: R10: 0000000000000008 R11: 0000000000000000 R12: ffffb69e00723c48 Jun 26 10:51:00 bld-fedora27 kernel: R13: 0000000000000000 R14: ffff91dd77430d70 R15: 0000000000000000 Jun 26 10:51:00 bld-fedora27 kernel: FS: 00007fa4e051fd00(0000) GS:ffff91dd7d400000(0000) knlGS:0000000000000000 Jun 26 10:51:00 bld-fedora27 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Jun 26 10:51:00 bld-fedora27 kernel: CR2: 00005616290356f8 CR3: 0000000035b8b000 CR4: 00000000001406e0 Jun 26 10:51:00 bld-fedora27 kernel: Call Trace: Jun 26 10:51:00 bld-fedora27 kernel: qxl_bo_kmap+0x42/0x70 [qxl] Jun 26 10:51:00 bld-fedora27 kernel: qxl_draw_dirty_fb+0x1f5/0x420 [qxl] Jun 26 10:51:00 bld-fedora27 kernel: qxl_framebuffer_surface_dirty+0xa0/0xf0 [qxl] Jun 26 10:51:00 bld-fedora27 kernel: drm_mode_dirtyfb_ioctl+0x190/0x1d0 [drm] Jun 26 10:51:00 bld-fedora27 kernel: drm_ioctl+0x232/0x500 [drm] Jun 26 10:51:00 bld-fedora27 kernel: ? drm_mode_getfb+0x110/0x110 [drm] Jun 26 10:51:00 bld-fedora27 kernel: ? __might_sleep+0x4a/0x80 Jun 26 10:51:00 bld-fedora27 kernel: ? selinux_file_ioctl+0x124/0x1e0 Jun 26 10:51:00 bld-fedora27 kernel: do_vfs_ioctl+0xa6/0x720 Jun 26 10:51:00 bld-fedora27 kernel: ? getnstimeofday64+0xe/0x20 Jun 26 10:51:00 bld-fedora27 kernel: ? security_file_ioctl+0x43/0x60 Jun 26 10:51:00 bld-fedora27 kernel: SyS_ioctl+0x79/0x90 Jun 26 10:51:00 bld-fedora27 kernel: do_syscall_64+0x6c/0x1c0 Jun 26 10:51:00 bld-fedora27 kernel: entry_SYSCALL64_slow_path+0x25/0x25 Jun 26 10:51:00 bld-fedora27 kernel: RIP: 0033:0x7fa4df504eb7 Jun 26 10:51:00 bld-fedora27 kernel: RSP: 002b:00007ffca83efbc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 Jun 26 10:51:00 bld-fedora27 kernel: RAX: ffffffffffffffda RBX: 0000000000001000 RCX: 00007fa4df504eb7 Jun 26 10:51:00 bld-fedora27 kernel: RDX: 00007ffca83efc00 RSI: 00000000c01864b1 RDI: 0000000000000009 Jun 26 10:51:00 bld-fedora27 kernel: RBP: 00007ffca83efc00 R08: 00007fa4dd42577c R09: 0000000000000010 Jun 26 10:51:00 bld-fedora27 kernel: R10: 00007fa4dd424884 R11: 0000000000000246 R12: 00000000c01864b1 Jun 26 10:51:00 bld-fedora27 kernel: R13: 0000000000000009 R14: 0000555e394ef330 R15: 00007fa4ddf8278c Jun 26 10:51:00 bld-fedora27 kernel: Code: d0 48 8b bb 80 00 00 00 48 c1 e6 0c f6 43 62 04 74 44 48 03 7b 70 4c 01 ef e8 8e 21 bd ef 49 89 04 24 44 8b 45 d0 e9 18 ff ff ff <0f> 0b 4b 8d 7c 3e 58 89 4 Jun 26 10:51:00 bld-fedora27 kernel: RIP: ttm_bo_kmap+0x1af/0x250 [ttm] RSP: ffffb69e00723ba8 Jun 26 10:51:00 bld-fedora27 kernel: ---[ end trace 13b661e538049e06 ]--- -- You are receiving this mail because: You are the assignee for the bug.

1 4

[PATCH 00/14] atomic modesetting for cirrus
by Varad Gautam 05 Sep '17

05 Sep '17

Hello, This series converts drm/cirrus to support atomic modesetting for the emulated CL GD-54xx hw in qemu. Patches 1-7 are general fixes and improvements, in preparation for atomic. Patch 8 provides dmabuf export for cirrus. Patch 9-11 add universal primary plane support and rework the driver to use atomic transition helpers for legacy modesetting. Patch 12 moves away from legacy modesetting entirely to use atomic handlers. Patch 13 adds atomic cursor plane. Patch 14 flips on DRIVER_ATOMIC. I have tested this with igt and chromeos/drm-tests (details in patch 14). Fun trivia from the GD-54xx TRM, 1996! It boasts 'optimized PCI burst write, which supports PCI writes to the frame buffer at greater than 55 Mbytes per second.' Dominik Behr (1): drm/cirrus: initialize start and size fields Stéphane Marchesin (1): drm/cirrus: Use 32bpp by default Varad Gautam (10): drm/cirrus: split out bo unpinning from cirrus_bo_push_sysram drm/cirrus: unregister connector on destroy drm/cirrus: add drm_read to cirrus_driver_fops drm/cirrus: do not disable outputs on fbdev init for atomic. drm/cirrus: use universal plane interfaces for primary plane drm/cirrus: use atomic transition helpers for plane and crtc drm/cirrus: send vblank on crtc atomic_flush drm/cirrus: use atomic handlers for plane and crtc drm/cirrus: implement atomic hardware cursor support drm/cirrus: advertise DRIVER_ATOMIC Zach Reizner (2): drm/cirrus: hardcode vram size drm/cirrus: implement PRIME export for cirrus drivers/gpu/drm/cirrus/Makefile | 2 +- drivers/gpu/drm/cirrus/cirrus_drv.c | 17 +- drivers/gpu/drm/cirrus/cirrus_drv.h | 25 ++ drivers/gpu/drm/cirrus/cirrus_fbdev.c | 6 +- drivers/gpu/drm/cirrus/cirrus_main.c | 23 +- drivers/gpu/drm/cirrus/cirrus_mode.c | 681 ++++++++++++++++++++++++++-------- drivers/gpu/drm/cirrus/cirrus_prime.c | 63 ++++ drivers/gpu/drm/cirrus/cirrus_ttm.c | 55 ++- 8 files changed, 705 insertions(+), 167 deletions(-) create mode 100644 drivers/gpu/drm/cirrus/cirrus_prime.c Thanks, Varad -- 2.13.1

3 21

[PATCH] drm/exynos/hdmi: Fix unsafe list iteration
by Maciej Purski 05 Sep '17

05 Sep '17

Function hdmi_mode_fixup() used bare list_for_each entry, which was unsafe and caused memory corruption detected by kasan. It now uses drm_for_each_connector_iter macro, which is now recommended by the documentation and safe. Signed-off-by: Maciej Purski <m.purski(a)samsung.com> --- drivers/gpu/drm/exynos/exynos_hdmi.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c index 214fa5e..0109ff4 100644 --- a/drivers/gpu/drm/exynos/exynos_hdmi.c +++ b/drivers/gpu/drm/exynos/exynos_hdmi.c @@ -944,22 +944,27 @@ static bool hdmi_mode_fixup(struct drm_encoder *encoder, struct drm_device *dev = encoder->dev; struct drm_connector *connector; struct drm_display_mode *m; + struct drm_connector_list_iter conn_iter; int mode_ok; drm_mode_set_crtcinfo(adjusted_mode, 0); - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + drm_connector_list_iter_begin(dev, &conn_iter); + drm_for_each_connector_iter(connector, &conn_iter) { if (connector->encoder == encoder) break; } + if (connector) + drm_connector_get(connector); + drm_connector_list_iter_end(&conn_iter); - if (connector->encoder != encoder) + if (!connector) return true; mode_ok = hdmi_mode_valid(connector, adjusted_mode); if (mode_ok == MODE_OK) - return true; + goto cleanup; /* * Find the most suitable mode and copy it to adjusted_mode. @@ -979,6 +984,9 @@ static bool hdmi_mode_fixup(struct drm_encoder *encoder, } } +cleanup: + drm_connector_put(connector); + return true; } -- 2.7.4

1 0

[PATCH] drm/ttm: Fix configuration error around populate_and_map() functions
by Tom St Denis 05 Sep '17

05 Sep '17

Fixed kbuild errors when IOMMU/SWIOTLB are disabled. Signed-off-by: Tom St Denis <tom.stdenis(a)amd.com> --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index 6a660d196d87..052e1f102113 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -920,6 +920,7 @@ void ttm_pool_unpopulate(struct ttm_tt *ttm) } EXPORT_SYMBOL(ttm_pool_unpopulate); +#if defined(CONFIG_SWIOTLB) || defined(CONFIG_INTEL_IOMMU) int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt) { unsigned i; @@ -960,6 +961,7 @@ void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_dma_tt *tt) ttm_pool_unpopulate(&tt->ttm); } EXPORT_SYMBOL(ttm_unmap_and_unpopulate_pages); +#endif int ttm_page_alloc_debugfs(struct seq_file *m, void *data) { -- 2.12.0

2 1

[PATCH] drm: gma500: fix logic error
by Arnd Bergmann 05 Sep '17

05 Sep '17

gcc-8 points out a condition that almost certainly doesn't do what the author had in mind: drivers/gpu/drm/gma500/mdfld_intel_display.c: In function 'mdfldWaitForPipeEnable': drivers/gpu/drm/gma500/mdfld_intel_display.c:102:37: error: bitwise comparison always evaluates to false [-Werror=tautological-compare] This changes it to a simple bit mask operation to check whether the bit is set. Fixes: 026abc333205 ("gma500: initial medfield merge") Signed-off-by: Arnd Bergmann <arnd(a)arndb.de> --- drivers/gpu/drm/gma500/mdfld_intel_display.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/gma500/mdfld_intel_display.c b/drivers/gpu/drm/gma500/mdfld_intel_display.c index 531e4450c000..5c066448be5b 100644 --- a/drivers/gpu/drm/gma500/mdfld_intel_display.c +++ b/drivers/gpu/drm/gma500/mdfld_intel_display.c @@ -99,7 +99,7 @@ void mdfldWaitForPipeEnable(struct drm_device *dev, int pipe) /* Wait for for the pipe enable to take effect. */ for (count = 0; count < COUNT_MAX; count++) { temp = REG_READ(map->conf); - if ((temp & PIPEACONF_PIPE_STATE) == 1) + if (temp & PIPEACONF_PIPE_STATE) break; } } -- 2.9.0

2 1

[Bug 102543] [BAT][HSW] igt@tools_test@tools_test - Unclaimed read from register 0x[4c]400c
by bugzilla-daemon＠freedesktop.org 05 Sep '17

05 Sep '17

https://bugs.freedesktop.org/show_bug.cgi?id=102543 Chris Wilson <chris(a)chris-wilson.co.uk> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|DRM/Intel |IGT Assignee|intel-gfx-bugs(a)lists.freede |dri-devel(a)lists.freedesktop |sktop.org |.org QA Contact|intel-gfx-bugs(a)lists.freede | |sktop.org | --- Comment #1 from Chris Wilson <chris(a)chris-wilson.co.uk> --- That looks entirely to be a test bug, intel_reg_read reading unknown registers which is then caught by the kernel on its next mmio. Since the kernel is not aware of a third party messing around with registers, it takes the blame upon itself. -- You are receiving this mail because: You are the assignee for the bug.

1 0

[PATCH] drm: exynos: include linux/irq.h
by Arnd Bergmann 05 Sep '17

05 Sep '17

I ran into a build error on x86: drivers/gpu/drm/exynos/exynos5433_drm_decon.c: In function 'decon_conf_irq': drivers/gpu/drm/exynos/exynos5433_drm_decon.c:706:2: error: implicit declaration of function 'irq_set_status_flags'; did you mean 'dquot_state_flag'? [-Werror=implicit-function-declaration] irq_set_status_flags(irq, IRQ_NOAUTOEN); Adding the missing include fixes the error. Fixes: b37d53a0382c ("drm/exynos/decon5433: move TE handling to DECON") Signed-off-by: Arnd Bergmann <arnd(a)arndb.de> --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 730b8d9db187..6be5b53c3b27 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -14,6 +14,7 @@ #include <linux/clk.h> #include <linux/component.h> #include <linux/iopoll.h> +#include <linux/irq.h> #include <linux/mfd/syscon.h> #include <linux/of_device.h> #include <linux/of_gpio.h> -- 2.9.0

1 0

[PATCH] drm/vblank: Fix delta_ns to an absolute value
by Hoegeun Kwon 05 Sep '17

05 Sep '17

If scanout started, we should reduce etime by delta_ns. But delta_ns is negative if scanout has not started. If delta_ns is negative, subtraction of delta_ns from etime increases etime. This is wrong, the etime should not be increased, so you have to make delta_ns an absolute value. Signed-off-by: Hoegeun Kwon <hoegeun.kwon(a)samsung.com> --- Hello all, I think that the etime should not be increased. In cases where delta_ns is negative, if you get time again after an interrupt call, there is a problem that the time obtained from the interrupt becomes the future time instead of the past time. Please let me know if this patch is wrong. Best regards, Hoegeun drivers/gpu/drm/drm_vblank.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c index 70f2b95..a3e0176 100644 --- a/drivers/gpu/drm/drm_vblank.c +++ b/drivers/gpu/drm/drm_vblank.c @@ -684,7 +684,7 @@ bool drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, /* Subtract time delta from raw timestamp to get final * vblank_time timestamp for end of vblank. */ - etime = ktime_sub_ns(etime, delta_ns); + etime = ktime_sub_ns(etime, abs(delta_ns)); *vblank_time = ktime_to_timeval(etime); DRM_DEBUG_VBL("crtc %u : v p(%d,%d)@ %ld.%ld -> %ld.%ld [e %d us, %d rep]\n", -- 1.9.1

3 3

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel September 2017