dri-devel January 2018

dri-devel@lists.freedesktop.org

226 participants
1677 discussions

[-next PATCH 0/4] sysfs and DEVICE_ATTR_<foo>
by Joe Perches 10 Jan '18

10 Jan '18

17 25

[PATCH] drm/vmwgfx: Potential off by one in vmw_view_add()
by Dan Carpenter 10 Jan '18

10 Jan '18

The vmw_view_cmd_to_type() function returns vmw_view_max (3) on error. It's one element beyond the end of the vmw_view_cotables[] table. My read on this is that it's possible to hit this failure. header->id comes from vmw_cmd_check() and it's a user controlled number between 1040 and 1225 so we can hit that error. But I don't have the hardware to test this code. Fixes: d80efd5cb3de ("drm/vmwgfx: Initial DX support") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index b700667f6f0b..c9d5cc237124 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -2731,6 +2731,8 @@ static int vmw_cmd_dx_view_define(struct vmw_private *dev_priv, } view_type = vmw_view_cmd_to_type(header->id); + if (view_type == vmw_view_max) + return -EINVAL; cmd = container_of(header, typeof(*cmd), header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter,

2 1

[PATCH] dma-buf: make returning the exclusive fence optional
by Christian König 10 Jan '18

10 Jan '18

Change reservation_object_get_fences_rcu to make the exclusive fence pointer optional. If not specified the exclusive fence is put into the fence array as well. This is helpful for a couple of cases where we need all fences in a single array. Signed-off-by: Christian König <christian.koenig(a)amd.com> --- drivers/dma-buf/reservation.c | 31 ++++++++++++++++++++++--------- 1 file changed, 22 insertions(+), 9 deletions(-) diff --git a/drivers/dma-buf/reservation.c b/drivers/dma-buf/reservation.c index b759a569b7b8..461afa9febd4 100644 --- a/drivers/dma-buf/reservation.c +++ b/drivers/dma-buf/reservation.c @@ -374,8 +374,9 @@ EXPORT_SYMBOL(reservation_object_copy_fences); * @pshared: the array of shared fence ptrs returned (array is krealloc'd to * the required size, and must be freed by caller) * - * RETURNS - * Zero or -errno + * Retrieve all fences from the reservation object. If the pointer for the + * exclusive fence is not specified the fence is put into the array of the + * shared fences as well. Returns either zero or -ENOMEM. */ int reservation_object_get_fences_rcu(struct reservation_object *obj, struct dma_fence **pfence_excl, @@ -389,8 +390,8 @@ int reservation_object_get_fences_rcu(struct reservation_object *obj, do { struct reservation_object_list *fobj; - unsigned seq; - unsigned int i; + unsigned int i, seq; + size_t sz = 0; shared_count = i = 0; @@ -402,9 +403,14 @@ int reservation_object_get_fences_rcu(struct reservation_object *obj, goto unlock; fobj = rcu_dereference(obj->fence); - if (fobj) { + if (fobj) + sz += sizeof(*shared) * fobj->shared_max; + + if (!pfence_excl && fence_excl) + sz += sizeof(*shared); + + if (sz) { struct dma_fence **nshared; - size_t sz = sizeof(*shared) * fobj->shared_max; nshared = krealloc(shared, sz, GFP_NOWAIT | __GFP_NOWARN); @@ -420,13 +426,19 @@ int reservation_object_get_fences_rcu(struct reservation_object *obj, break; } shared = nshared; - shared_count = fobj->shared_count; - + shared_count = fobj ? fobj->shared_count : 0; for (i = 0; i < shared_count; ++i) { shared[i] = rcu_dereference(fobj->shared[i]); if (!dma_fence_get_rcu(shared[i])) break; } + + if (!pfence_excl && fence_excl) { + shared[i] = fence_excl; + fence_excl = NULL; + ++i; + ++shared_count; + } } if (i != shared_count || read_seqcount_retry(&obj->seq, seq)) { @@ -448,7 +460,8 @@ int reservation_object_get_fences_rcu(struct reservation_object *obj, *pshared_count = shared_count; *pshared = shared; - *pfence_excl = fence_excl; + if (pfence_excl) + *pfence_excl = fence_excl; return ret; } -- 2.14.1

2 3

AMD DC test results
by Daniel Drake 10 Jan '18

10 Jan '18

Hi, Thanks for the hard work on AMD DC development! Here are some new test results - hope they are interesting/useful. CONTEXT We have been tracking DC for a while as we work with multiple products where amdgpu previously was not able to support the HDMI audio output. We are hoping to ship the new driver as soon as possible to fix this Linux compatibility issue. TEST SETUP With an eye to shipping this ASAP, we backported DC to the current stable release, Linux 4.14. To do this we took the amdgpu/DC commits from: https://cgit.freedesktop.org/~agd5f/linux/log/?h=upstream-4.14-drm-next-amd… And added these two pull requests: https://www.spinics.net/lists/dri-devel/msg156828.html https://lists.freedesktop.org/archives/dri-devel/2017-November/157687.html At this point we diff the resultant amdgpu directories with Linus tree as of commit f6705bf959efac87bca76d40050d342f1d212587 (when he accepted the DC driver) and the diff is zero. This is published on the master branch of https://github.com/endlessm/linux There have been more AMD DC changes merged into Linus tree since then, however we tried and the backporting seems like loads of work, so we have not gone beyond the initial merge into Linus tree for now. We then asked our QA team to run our standard tests with this kernel. They tested 3 platforms: Acer Aspire E5-553G (AMD FX-9800P RADEON R7) Acer Aspire E5-523G (AMD E2-9010 RADEON R2) Acer Aspire A315-21 (AMD A4-9120 RADEON R3) CONFIG_DRM_AMD_DC_PRE_VEGA was set in order to have AMD DC support this hardware. TEST RESULTS HDMI audio works now, however we have encountered some regressions. In all cases the 3 platforms behaved the same (either all 3 pass the individual tests, or all 3 show the same problem). The regressions are: 1. System hangs on S3 resume. Bug can't be reproduced when booting with amdgpu.dc=0 Also reproduced with DC on Linus master (4.15-rc), but appears fixed on agd5f/amd-staging-drm-next https://bugs.freedesktop.org/show_bug.cgi?id=104281 2. Display corruption when using multiple displays Bug can't be reproduced when booting with amdgpu.dc=0 Also reproduced with DC on Linus master (4.15-rc), but appears fixed on agd5f/amd-staging-drm-next https://bugs.freedesktop.org/show_bug.cgi?id=104319 3. HDMI audio device still shown as present and active even after disconnecting HDMI cable Bug can't be reproduced when booting with amdgpu.dc=0 Appears fixed with DC on Linus master (4.15-rc). Full test results spreadsheet (including tests passed): http://goo.gl/oqnMDx Thanks to my colleagues: Carlo Caione for doing the backport and Vanessa Chang for managing the testing. NEXT STEPS Unfortunately we can't ship the driver in current state with these regressions, but on the positive side it looks like it will be in better shape for Linux 4.16. We will now focus on making 4.15 regression-free before revisiting the idea of backporting DC to earlier kernels. After the holiday break we will try to identify the relevant fixes in amd-staging-drm-next and see if they can be included in Linux 4.15. We'll send these backports upstream as soon as we find them. Daniel

3 2

[RFC PATCH v2 0/1] Fix drm driver hang on tegra124-nyan-big
by Guillaume Tucker 10 Jan '18

10 Jan '18

This patch was initially sent along with another one to fix a first hang in the nouveau drm driver[1]. I'm now sending it again as a separate patch as it's to fix a second hang which is not strictly related. It is hidden by the first hang though, as this happens later on during the driver initialisation. The issue this is addressing is due to a rework[2] of the driver which apparently wasn't fully tested on tegra124 and possibly also on tegra210. The sor->clk_out is not used on these platforms as HDMI and DP are both not supported (but eDP is), so trying to assign the parent clock to clk_out does nothing. The driver then goes on to configure the clock and PLL registers in tegra_sor_edp_enable and this is where it hangs. As this is hard to debug further without full documentation or a hardware debugger, this patch essentially makes the driver do what it used to do on these platforms before the rework (i.e. assign the parent clock to sor->clk instead). There might be a more correct way of fixing this, hence the RFC. [1] https://patchwork.kernel.org/patch/10124637/ [2] e1335e2f0cfc ("drm/tegra: sor: Reimplement pad clock") Guillaume Tucker (1): drm/tegra: sor: Fix hang on tegra124 due to NULL clk_out drivers/gpu/drm/tegra/sor.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) -- 2.11.0

3 4

[PATCH] drm/tegra: sor: Fix hang on Tegra124 eDP
by Thierry Reding 10 Jan '18

10 Jan '18

From: Thierry Reding <treding(a)nvidia.com> The SOR0 found on Tegra124 and Tegra210 only supports eDP and LVDS and therefore has a slightly different clock tree than the SOR1 which does not support eDP, but HDMI and DP instead. Commit e1335e2f0cfc ("drm/tegra: sor: Reimplement pad clock") breaks setups with eDP because the sor->clk_out clock is uninitialized and therefore setting the parent clock (either the safe clock or either of the display PLLs) fails, which can cause hangs later on since there is no clock driving the module. Fix this by falling back to the module clock for sor->clk_out on those setups. This guarantees that the module will always be clocked by an enabled clock and hence prevents those hangs. Fixes: e1335e2f0cfc ("drm/tegra: sor: Reimplement pad clock") Reported-by: Guillaume Tucker <guillaume.tucker(a)collabora.com> Tested-by: Jon Hunter <jonathanh(a)nvidia.com> Signed-off-by: Thierry Reding <treding(a)nvidia.com> --- drivers/gpu/drm/tegra/sor.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c index b0a1dedac802..476079f1255f 100644 --- a/drivers/gpu/drm/tegra/sor.c +++ b/drivers/gpu/drm/tegra/sor.c @@ -2656,6 +2656,9 @@ static int tegra_sor_probe(struct platform_device *pdev) name, err); goto remove; } + } else { + /* fall back to the module clock on SOR0 (eDP/LVDS only) */ + sor->clk_out = sor->clk; } sor->clk_parent = devm_clk_get(&pdev->dev, "parent"); -- 2.15.1

1 0

[PATCH] radeon: completely remove lut leftovers
by Daniel Vetter 10 Jan '18

10 Jan '18

This is an oversight from commit 42585395ebc1034a98937702849669f17eadb35f Author: Peter Rosin <peda(a)axentia.se> Date: Thu Jul 13 18:25:36 2017 +0200 drm: radeon: remove dead code and pointless local lut storage Cc: Peter Rosin <peda(a)axentia.se> Cc: Alex Deucher <alexander.deucher(a)amd.com> Cc: Michel Dänzer <michel.daenzer(a)amd.com> Signed-off-by: Daniel Vetter <daniel.vetter(a)intel.com> --- drivers/gpu/drm/radeon/radeon_display.c | 6 ------ drivers/gpu/drm/radeon/radeon_mode.h | 1 - 2 files changed, 7 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index dfda5e0ed166..d22f4b6a8828 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -698,12 +698,6 @@ static void radeon_crtc_init(struct drm_device *dev, int index) radeon_crtc->mode_set.num_connectors = 0; #endif - for (i = 0; i < 256; i++) { - radeon_crtc->lut_r[i] = i << 2; - radeon_crtc->lut_g[i] = i << 2; - radeon_crtc->lut_b[i] = i << 2; - } - if (rdev->is_atom_bios && (ASIC_IS_AVIVO(rdev) || radeon_r4xx_atom)) radeon_atombios_init_crtc(dev, radeon_crtc); else diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index 3243e5e01432..5f61facafe36 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -328,7 +328,6 @@ enum radeon_flip_status { struct radeon_crtc { struct drm_crtc base; int crtc_id; - u16 lut_r[256], lut_g[256], lut_b[256]; bool enabled; bool can_tile; bool cursor_out_of_bounds; -- 2.15.1

2 1

[PATCH v7 0/3] Update ROCKCHIP DSI driver that uses dw-mipi-dsi bridge
by Nickey Yang 10 Jan '18

10 Jan '18

We now have a generic dw-mipi-dsi bridge driver.So we send this patchs to moving rockchip dw-mipi-dsi driver to that in order to add new features(dual mipi support). Update ROCKCHIP DSI controller driver that uses the Synopsys DesignWare MIPI DSI host controller bridge. ChangeLog: v2: add err_pllref、remove unnecessary encoder.enable & disable correct spelling mistakes v3: add Brian's patch "drm/bridge/synopsys: stop clobbering drvdata" (Link:https://patchwork.kernel.org/patch/10078493/) adjust drm/stm/dsi code for above drm/rockchip: Add ROCKCHIP DW MIPI DSI controller driver: call dw_mipi_dsi_unbind() in dw_mipi_dsi_rockchip_unbind() fix typo, use of_device_get_match_data(), change some ‘bind()’ logic into 'probe()' add 'dev_set_drvdata()' v4: keep "drm/stm: dsi: Adjust dw_mipi_dsi_probe and remove" into Brians "drm/bridge/synopsys: stop clobbering drvdata" add review tag and add some comments v5: 1/3: fix stm part remove() directly referencing the static dw_mipi_dsi_stm_plat_data struct. 3/3: keep our power domain enabled while touching GRF v6: 3/3: just change func dw_mipi_encoder_disable name to dw_mipi_dsi_encoder_disable v7: 1/3: add missing platform_set_drvdata in stm part. Nickey Yang (3): drm/bridge/synopsys: dsi: stop clobbering drvdata dt-bindings: display: rockchip: update DSI controller drm/rockchip: Add ROCKCHIP DW MIPI DSI controller driver .../display/rockchip/dw_mipi_dsi_rockchip.txt | 23 +- drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c | 36 +- drivers/gpu/drm/rockchip/Kconfig | 2 +- drivers/gpu/drm/rockchip/Makefile | 2 +- drivers/gpu/drm/rockchip/dw-mipi-dsi.c | 1349 -------------------- drivers/gpu/drm/rockchip/dw-mipi-dsi_rockchip.c | 785 ++++++++++++ drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 2 +- drivers/gpu/drm/rockchip/rockchip_drm_drv.h | 2 +- drivers/gpu/drm/stm/dw_mipi_dsi-stm.c | 12 +- include/drm/bridge/dw_mipi_dsi.h | 17 +- 10 files changed, 842 insertions(+), 1388 deletions(-) delete mode 100644 drivers/gpu/drm/rockchip/dw-mipi-dsi.c create mode 100644 drivers/gpu/drm/rockchip/dw-mipi-dsi_rockchip.c -- 1.9.1

3 5

[PATCH] drm/atomic: Fix memleak on ERESTARTSYS during non-blocking commits
by sunpeng.li＠amd.com 10 Jan '18

10 Jan '18

From: "Leo (Sunpeng) Li" <sunpeng.li(a)amd.com> During a non-blocking commit, it is possible to return before the commit_tail work is queued (-ERESTARTSYS, for example). Since a reference on the crtc commit object is obtained for the pending vblank event when preparing the commit, the above situation will leave us with an extra reference. Therefore, if the commit_tail worker has not consumed the event at the end of a commit, release it's reference. Signed-off-by: Leo (Sunpeng) Li <sunpeng.li(a)amd.com> --- drivers/gpu/drm/drm_atomic_helper.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index ab40321..4253f57 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -3421,6 +3421,15 @@ EXPORT_SYMBOL(drm_atomic_helper_crtc_duplicate_state); void __drm_atomic_helper_crtc_destroy_state(struct drm_crtc_state *state) { if (state->commit) { + /* + * In the event that a non-blocking commit returns + * -ERESTARTSYS before the commit_tail work is queued, we will + * have an extra reference to the commit object. Release it, if + * the event has not been consumed by the worker. + */ + if (state->event) + drm_crtc_commit_put(state->commit); + kfree(state->commit->event); state->commit->event = NULL; drm_crtc_commit_put(state->commit); -- 2.7.4

5 4

[PATCH v2 0/6] drm/tinydrm: Cleanup
by Noralf Trønnes 10 Jan '18

10 Jan '18

Fix a few things that came up during the tinydrm review but wasn't addressed at the time. Noralf. Changes since version 1: - Add mipi_dbi_enable_flush() instead of the pipe enable helper mipi_dbi_pipe_enable() that doesn't really fit anymore. - Add common poweron-reset functionality. - Use drm_mode_copy() instead of direct assignment Noralf Trønnes (6): drm/tinydrm/mi0283qt: Use common include order drm/tinydrm/mi0283qt: Remove ili9341.h drm/tinydrm/mipi-dbi: Add mipi_dbi_enable_flush() drm/tinydrm/mipi-dbi: Add poweron-reset functions drm/tinydrm/mi0283qt: Let the display pipe handle power drm/tinydrm: Embed the mode in tinydrm_connector drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c | 34 ++++----- drivers/gpu/drm/tinydrm/ili9225.c | 6 +- drivers/gpu/drm/tinydrm/mi0283qt.c | 104 ++++++++++++---------------- drivers/gpu/drm/tinydrm/mipi-dbi.c | 97 +++++++++++++++++++++----- drivers/gpu/drm/tinydrm/st7586.c | 15 ++-- drivers/gpu/drm/tinydrm/st7735r.c | 10 +-- include/drm/tinydrm/ili9341.h | 54 --------------- include/drm/tinydrm/mipi-dbi.h | 5 +- 8 files changed, 147 insertions(+), 178 deletions(-) delete mode 100644 include/drm/tinydrm/ili9341.h -- 2.14.2

3 18

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel January 2018