dri-devel May 2022

dri-devel@lists.freedesktop.org

344 participants
736 discussions

[PATCH] fbdev: vesafb: Allow to be built if COMPILE_TEST is enabled
by Javier Martinez Canillas 13 May '22

13 May '22

The driver has runtime but no build time dependency with X86, so it can be built for testing purposes if the COMPILE_TEST option is enabled. This is useful to have more build coverage and make sure that the driver is not affected by changes that could cause build regressions. Signed-off-by: Javier Martinez Canillas <javierm(a)redhat.com> --- drivers/video/fbdev/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/video/fbdev/Kconfig b/drivers/video/fbdev/Kconfig index f2a6b81e45c4..bd849013f16f 100644 --- a/drivers/video/fbdev/Kconfig +++ b/drivers/video/fbdev/Kconfig @@ -616,7 +616,7 @@ config FB_UVESA config FB_VESA bool "VESA VGA graphics support" - depends on (FB = y) && X86 + depends on (FB = y) && (X86 || COMPILE_TEST) select FB_CFB_FILLRECT select FB_CFB_COPYAREA select FB_CFB_IMAGEBLIT -- 2.35.1

1 1

[PATCH 1/2] drm/msm: don't free the IRQ if it was not requested
by Dmitry Baryshkov 13 May '22

13 May '22

As msm_drm_uninit() is called from the msm_drm_init() error path, additional care should be necessary as not to call the free_irq() for the IRQ that was not requested before (because an error occured earlier than the request_irq() call). This fixed the issue reported with the following backtrace: [ 8.571329] Trying to free already-free IRQ 187 [ 8.571339] WARNING: CPU: 0 PID: 76 at kernel/irq/manage.c:1895 free_irq+0x1e0/0x35c [ 8.588746] Modules linked in: pmic_glink pdr_interface fastrpc qrtr_smd snd_soc_hdmi_codec msm fsa4480 gpu_sched drm_dp_aux_bus qrtr i2c_qcom_geni crct10dif_ce qcom_stats qcom_q6v5_pas drm_display_helper gpi qcom_pil_info drm_kms_helper qcom_q6v5 qcom_sysmon qcom_common qcom_glink_smem qcom_rng mdt_loader qmi_helpers phy_qcom_qmp ufs_qcom typec qnoc_sm8350 socinfo rmtfs_mem fuse drm ipv6 [ 8.624154] CPU: 0 PID: 76 Comm: kworker/u16:2 Not tainted 5.18.0-rc5-next-20220506-00033-g6cee8cab6089-dirty #419 [ 8.624161] Hardware name: Qualcomm Technologies, Inc. SM8350 HDK (DT) [ 8.641496] Workqueue: events_unbound deferred_probe_work_func [ 8.647510] pstate: 604000c5 (nZCv daIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 8.654681] pc : free_irq+0x1e0/0x35c [ 8.658454] lr : free_irq+0x1e0/0x35c [ 8.662228] sp : ffff800008ab3950 [ 8.665642] x29: ffff800008ab3950 x28: 0000000000000000 x27: ffff16350f56a700 [ 8.672994] x26: ffff1635025df080 x25: ffff16350251badc x24: ffff16350251bb90 [ 8.680343] x23: 0000000000000000 x22: 00000000000000bb x21: ffff16350e8f9800 [ 8.687690] x20: ffff16350251ba00 x19: ffff16350cbd5880 x18: ffffffffffffffff [ 8.695039] x17: 0000000000000000 x16: ffffa2dd12179434 x15: ffffa2dd1431d02d [ 8.702391] x14: 0000000000000000 x13: ffffa2dd1431d028 x12: 662d79646165726c [ 8.709740] x11: ffffa2dd13fd2438 x10: 000000000000000a x9 : 00000000000000bb [ 8.717111] x8 : ffffa2dd13fd23f0 x7 : ffff800008ab3750 x6 : 00000000fffff202 [ 8.724487] x5 : ffff16377e870a18 x4 : 00000000fffff202 x3 : ffff735a6ae1b000 [ 8.731851] x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffff1635015f8000 [ 8.739217] Call trace: [ 8.741755] free_irq+0x1e0/0x35c [ 8.745198] msm_drm_uninit.isra.0+0x14c/0x294 [msm] [ 8.750548] msm_drm_bind+0x28c/0x5d0 [msm] [ 8.755081] try_to_bring_up_aggregate_device+0x164/0x1d0 [ 8.760657] __component_add+0xa0/0x170 [ 8.764626] component_add+0x14/0x20 [ 8.768337] dp_display_probe+0x2a4/0x464 [msm] [ 8.773242] platform_probe+0x68/0xe0 [ 8.777043] really_probe.part.0+0x9c/0x28c [ 8.781368] __driver_probe_device+0x98/0x144 [ 8.785871] driver_probe_device+0x40/0x140 [ 8.790191] __device_attach_driver+0xb4/0x120 [ 8.794788] bus_for_each_drv+0x78/0xd0 [ 8.798751] __device_attach+0xdc/0x184 [ 8.802713] device_initial_probe+0x14/0x20 [ 8.807031] bus_probe_device+0x9c/0xa4 [ 8.810991] deferred_probe_work_func+0x88/0xc0 [ 8.815667] process_one_work+0x1d0/0x320 [ 8.819809] worker_thread+0x14c/0x444 [ 8.823688] kthread+0x10c/0x110 [ 8.827036] ret_from_fork+0x10/0x20 Reported-by: Bjorn Andersson <bjorn.andersson(a)linaro.org> Fixes: f026e431cf86 ("drm/msm: Convert to Linux IRQ interfaces") Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov(a)linaro.org> --- drivers/gpu/drm/msm/msm_drv.c | 7 ++++++- drivers/gpu/drm/msm/msm_kms.h | 1 + 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c index 4a3dda23e3e0..44485363f37a 100644 --- a/drivers/gpu/drm/msm/msm_drv.c +++ b/drivers/gpu/drm/msm/msm_drv.c @@ -113,6 +113,8 @@ static int msm_irq_postinstall(struct drm_device *dev) static int msm_irq_install(struct drm_device *dev, unsigned int irq) { + struct msm_drm_private *priv = dev->dev_private; + struct msm_kms *kms = priv->kms; int ret; if (irq == IRQ_NOTCONNECTED) @@ -124,6 +126,8 @@ static int msm_irq_install(struct drm_device *dev, unsigned int irq) if (ret) return ret; + kms->irq_requested = true; + ret = msm_irq_postinstall(dev); if (ret) { free_irq(irq, dev); @@ -139,7 +143,8 @@ static void msm_irq_uninstall(struct drm_device *dev) struct msm_kms *kms = priv->kms; kms->funcs->irq_uninstall(kms); - free_irq(kms->irq, dev); + if (kms->irq_requested) + free_irq(kms->irq, dev); } struct msm_vblank_work { diff --git a/drivers/gpu/drm/msm/msm_kms.h b/drivers/gpu/drm/msm/msm_kms.h index ab25fff271f9..f8ed7588928c 100644 --- a/drivers/gpu/drm/msm/msm_kms.h +++ b/drivers/gpu/drm/msm/msm_kms.h @@ -148,6 +148,7 @@ struct msm_kms { /* irq number to be passed on to msm_irq_install */ int irq; + bool irq_requested; /* mapper-id used to request GEM buffer mapped for scanout: */ struct msm_gem_address_space *aspace; -- 2.35.1

3 10

[PATCH V3 RESEND] dma-buf: ensure unique directory name for dmabuf stats
by Charan Teja Kalla 13 May '22

13 May '22

The dmabuf file uses get_next_ino()(through dma_buf_getfile() -> alloc_anon_inode()) to get an inode number and uses the same as a directory name under /sys/kernel/dmabuf/buffers/<ino>. This directory is used to collect the dmabuf stats and it is created through dma_buf_stats_setup(). At current, failure to create this directory entry can make the dma_buf_export() to fail. Now, as the get_next_ino() can definitely give a repetitive inode no causing the directory entry creation to fail with -EEXIST. This is a problem on the systems where dmabuf stats functionality is enabled on the production builds can make the dma_buf_export(), though the dmabuf memory is allocated successfully, to fail just because it couldn't create stats entry. This issue we are able to see on the snapdragon system within 13 days where there already exists a directory with inode no "122602" so dma_buf_stats_setup() failed with -EEXIST as it is trying to create the same directory entry. To make the dentry name as unique, use the dmabuf fs specific inode which is based on the simple atomic variable increment. There is tmpfs subsystem too which relies on its own inode generation rather than relying on the get_next_ino() for the same reason of avoiding the duplicate inodes[1]. [1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/patch/?id=… Signed-off-by: Charan Teja Kalla <quic_charante(a)quicinc.com> Cc: <stable(a)vger.kernel.org> # 5.15.x+ Reviewed-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- Changes in V3-resend: -- Collect all the tags and apply stable tag. Changes in V3: -- Used the atomic64 variable to have dmabuf files its own inodes. -- Ensured no UAPI breakage as suggested by Christian. Changes in V2: -- Used the atomic64_t variable to generate a unique_id to be appended to inode to have an unique directory with name <inode_number-unique_id> -- Suggested by christian -- Updated the ABI documentation -- Identified by Greg. -- Massaged the commit log. -- https://lore.kernel.org/all/1652191562-18700-1-git-send-email-quic_charante… Changes in V1: -- Used the inode->i_ctime->tv_secs as an id appended to inode to create the unique directory with name <inode_number-time_in_secs>. -- https://lore.kernel.org/all/1652178212-22383-1-git-send-email-quic_charante… drivers/dma-buf/dma-buf.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index a6fc96e..0ad5039 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -407,6 +407,7 @@ static inline int is_dma_buf_file(struct file *file) static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags) { + static atomic64_t dmabuf_inode = ATOMIC64_INIT(0); struct file *file; struct inode *inode = alloc_anon_inode(dma_buf_mnt->mnt_sb); @@ -416,6 +417,13 @@ static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags) inode->i_size = dmabuf->size; inode_set_bytes(inode, dmabuf->size); + /* + * The ->i_ino acquired from get_next_ino() is not unique thus + * not suitable for using it as dentry name by dmabuf stats. + * Override ->i_ino with the unique and dmabuffs specific + * value. + */ + inode->i_ino = atomic64_add_return(1, &dmabuf_inode); file = alloc_file_pseudo(inode, dma_buf_mnt, "dmabuf", flags, &dma_buf_fops); if (IS_ERR(file)) -- 2.7.4

1 0

[PATCH v2 11/11] dt-bindings: display: convert Arm Komeda to DT schema
by Andre Przywara 13 May '22

13 May '22

The Arm Komeda (aka Mali-D71) is a display controller that scans out a framebuffer and hands a signal to a digital encoder to generate a DVI or HDMI signal. It supports up to two pipelines, each frame can be composed of up to four layers. Convert the existing DT binding to DT schema. Signed-off-by: Andre Przywara <andre.przywara(a)arm.com> --- .../bindings/display/arm,komeda.txt | 78 ----------- .../bindings/display/arm,komeda.yaml | 130 ++++++++++++++++++ 2 files changed, 130 insertions(+), 78 deletions(-) delete mode 100644 Documentation/devicetree/bindings/display/arm,komeda.txt create mode 100644 Documentation/devicetree/bindings/display/arm,komeda.yaml diff --git a/Documentation/devicetree/bindings/display/arm,komeda.txt b/Documentation/devicetree/bindings/display/arm,komeda.txt deleted file mode 100644 index 8513695ee47fe..0000000000000 --- a/Documentation/devicetree/bindings/display/arm,komeda.txt +++ /dev/null @@ -1,78 +0,0 @@ -Device Tree bindings for Arm Komeda display driver - -Required properties: -- compatible: Should be "arm,mali-d71" -- reg: Physical base address and length of the registers in the system -- interrupts: the interrupt line number of the device in the system -- clocks: A list of phandle + clock-specifier pairs, one for each entry - in 'clock-names' -- clock-names: A list of clock names. It should contain: - - "aclk": for the main processor clock -- #address-cells: Must be 1 -- #size-cells: Must be 0 -- iommus: configure the stream id to IOMMU, Must be configured if want to - enable iommu in display. for how to configure this node please reference - devicetree/bindings/iommu/arm,smmu-v3.txt, - devicetree/bindings/iommu/iommu.txt - -Required properties for sub-node: pipeline@nq -Each device contains one or two pipeline sub-nodes (at least one), each -pipeline node should provide properties: -- reg: Zero-indexed identifier for the pipeline -- clocks: A list of phandle + clock-specifier pairs, one for each entry - in 'clock-names' -- clock-names: should contain: - - "pxclk": pixel clock - -- port: each pipeline connect to an encoder input port. The connection is - modeled using the OF graph bindings specified in - Documentation/devicetree/bindings/graph.txt - -Optional properties: - - memory-region: phandle to a node describing memory (see - Documentation/devicetree/bindings/reserved-memory/reserved-memory.txt) - to be used for the framebuffer; if not present, the framebuffer may - be located anywhere in memory. - -Example: -/ { - ... - - dp0: display@c00000 { - #address-cells = <1>; - #size-cells = <0>; - compatible = "arm,mali-d71"; - reg = <0xc00000 0x20000>; - interrupts = <0 168 4>; - clocks = <&dpu_aclk>; - clock-names = "aclk"; - iommus = <&smmu 0>, <&smmu 1>, <&smmu 2>, <&smmu 3>, - <&smmu 4>, <&smmu 5>, <&smmu 6>, <&smmu 7>, - <&smmu 8>, <&smmu 9>; - - dp0_pipe0: pipeline@0 { - clocks = <&fpgaosc2>; - clock-names = "pxclk"; - reg = <0>; - - port { - dp0_pipe0_out: endpoint { - remote-endpoint = <&db_dvi0_in>; - }; - }; - }; - - dp0_pipe1: pipeline@1 { - clocks = <&fpgaosc2>; - clock-names = "pxclk"; - reg = <1>; - - port { - dp0_pipe1_out: endpoint { - remote-endpoint = <&db_dvi1_in>; - }; - }; - }; - }; - ... -}; diff --git a/Documentation/devicetree/bindings/display/arm,komeda.yaml b/Documentation/devicetree/bindings/display/arm,komeda.yaml new file mode 100644 index 0000000000000..9f4aade97f10a --- /dev/null +++ b/Documentation/devicetree/bindings/display/arm,komeda.yaml @@ -0,0 +1,130 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/display/arm,komeda.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: Arm Komeda display processor + +maintainers: + - Liviu Dudau <Liviu.Dudau(a)arm.com> + - Andre Przywara <andre.przywara(a)arm.com> + +description: + The Arm Mali D71 display processor supports up to two displays with up + to a 4K resolution each. Each pipeline can be composed of up to four + layers. It is typically connected to a digital display connector like HDMI. + +properties: + compatible: + oneOf: + - items: + - const: arm,mali-d32 + - const: arm,mali-d71 + - const: arm,mali-d71 + + reg: + maxItems: 1 + + interrupts: + maxItems: 1 + + clock-names: + const: aclk + + clocks: + maxItems: 1 + description: The main DPU processor clock + + "#address-cells": + const: 1 + + "#size-cells": + const: 0 + + memory-region: + maxItems: 1 + description: + Phandle to a node describing memory to be used for the framebuffer. + If not present, the framebuffer may be located anywhere in memory. + + iommus: + description: + The stream IDs for each of the used pipelines, each four IDs for the + four layers, plus one for the write-back stream. + minItems: 5 + maxItems: 10 + +patternProperties: + '^pipeline@[01]$': + type: object + description: + clocks + + properties: + reg: + enum: [ 0, 1 ] + + clock-names: + const: pxclk + + clocks: + maxItems: 1 + description: The input reference for the pixel clock. + + port: + $ref: /schemas/graph.yaml#/$defs/port-base + unevaluatedProperties: false + +additionalProperties: false + +required: + - "#address-cells" + - "#size-cells" + - compatible + - reg + - interrupts + - clock-names + - clocks + - pipeline@0 + +examples: + - | + display@c00000 { + #address-cells = <1>; + #size-cells = <0>; + compatible = "arm,mali-d71"; + reg = <0xc00000 0x20000>; + interrupts = <168>; + clocks = <&dpu_aclk>; + clock-names = "aclk"; + iommus = <&smmu 0>, <&smmu 1>, <&smmu 2>, <&smmu 3>, + <&smmu 8>, + <&smmu 4>, <&smmu 5>, <&smmu 6>, <&smmu 7>, + <&smmu 9>; + + dp0_pipe0: pipeline@0 { + clocks = <&fpgaosc2>; + clock-names = "pxclk"; + reg = <0>; + + port { + dp0_pipe0_out: endpoint { + remote-endpoint = <&db_dvi0_in>; + }; + }; + }; + + dp0_pipe1: pipeline@1 { + clocks = <&fpgaosc2>; + clock-names = "pxclk"; + reg = <1>; + + port { + dp0_pipe1_out: endpoint { + remote-endpoint = <&db_dvi1_in>; + }; + }; + }; + }; +... -- 2.25.1

2 1

[PATCH V3] dma-buf: ensure unique directory name for dmabuf stats
by Charan Teja Kalla 13 May '22

13 May '22

The dmabuf file uses get_next_ino()(through dma_buf_getfile() -> alloc_anon_inode()) to get an inode number and uses the same as a directory name under /sys/kernel/dmabuf/buffers/<ino>. This directory is used to collect the dmabuf stats and it is created through dma_buf_stats_setup(). At current, failure to create this directory entry can make the dma_buf_export() to fail. Now, as the get_next_ino() can definitely give a repetitive inode no causing the directory entry creation to fail with -EEXIST. This is a problem on the systems where dmabuf stats functionality is enabled on the production builds can make the dma_buf_export(), though the dmabuf memory is allocated successfully, to fail just because it couldn't create stats entry. This issue we are able to see on the snapdragon system within 13 days where there already exists a directory with inode no "122602" so dma_buf_stats_setup() failed with -EEXIST as it is trying to create the same directory entry. To make the dentry name as unique, use the dmabuf fs specific inode which is based on the simple atomic variable increment. There is tmpfs subsystem too which relies on its own inode generation rather than relying on the get_next_ino() for the same reason of avoiding the duplicate inodes[1]. [1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/patch/?id=… Reported-by: kernel test robot <lkp(a)intel.com> Signed-off-by: Charan Teja Kalla <quic_charante(a)quicinc.com> --- Changes in V3: -- Used the atomic64 variable to have dmabuf files its own inodes. -- Ensured no UAPI breakage as suggested by Christian. Changes in V2: -- Used the atomic64_t variable to generate a unique_id to be appended to inode to have an unique directory with name <inode_number-unique_id> -- Suggested by christian -- Updated the ABI documentation -- Identified by Greg. -- Massaged the commit log. -- https://lore.kernel.org/all/1652191562-18700-1-git-send-email-quic_charante… Changes in V1: -- Used the inode->i_ctime->tv_secs as an id appended to inode to create the unique directory with name <inode_number-time_in_secs>. -- https://lore.kernel.org/all/1652178212-22383-1-git-send-email-quic_charante… drivers/dma-buf/dma-buf.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index a6fc96e..0ad5039 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -407,6 +407,7 @@ static inline int is_dma_buf_file(struct file *file) static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags) { + static atomic64_t dmabuf_inode = ATOMIC64_INIT(0); struct file *file; struct inode *inode = alloc_anon_inode(dma_buf_mnt->mnt_sb); @@ -416,6 +417,13 @@ static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags) inode->i_size = dmabuf->size; inode_set_bytes(inode, dmabuf->size); + /* + * The ->i_ino acquired from get_next_ino() is not unique thus + * not suitable for using it as dentry name by dmabuf stats. + * Override ->i_ino with the unique and dmabuffs specific + * value. + */ + inode->i_ino = atomic64_add_return(1, &dmabuf_inode); file = alloc_file_pseudo(inode, dma_buf_mnt, "dmabuf", flags, &dma_buf_fops); if (IS_ERR(file)) -- 2.7.4

3 6

DSI Bridge switching
by Jagan Teki 13 May '22

13 May '22

Hi, I think this seems to be a known use case for industrial these days with i.mx8m. The host DSI would configure with two bridges one for DSI to LVDS (SN65DSI83) and another for DSI to HDMI Out (ADV7535). Technically we can use only one bridge at a time as host DSI support single out port. So we can have two separate device tree files for LVDS and HDMI and load them static. But, one of the use cases is to support both of them in single dts, and - Turn On LVDS (default) - Turn Off LVDS then Turn On HDMI when cable plug-in. The HDMI event can be detected via some HDMI-INT GPIO on-board design. The possible solution, I'm thinking of adding LVDS on port 1, HDMI on port 2 in the DSI host node, and trying to attach the respective bridge based on HDMI-INT like repeating the bridge attachment cycle based on the HDMI-INT. Can it be possible to do bridge attachment at runtime? something like a bridge hotplug event? or any other possible solutions? Any suggestions? Thanks, Jagan.

6 15

[PATCH] drm/i915: Fix CFI violation with show_dynamic_id()
by Nathan Chancellor 13 May '22

13 May '22

When an attribute group is created with sysfs_create_group(), the ->sysfs_ops() callback is set to kobj_sysfs_ops, which sets the ->show() callback to kobj_attr_show(). kobj_attr_show() uses container_of() to get the ->show() callback from the attribute it was passed, meaning the ->show() callback needs to be the same type as the ->show() callback in 'struct kobj_attribute'. However, show_dynamic_id() has the type of the ->show() callback in 'struct device_attribute', which causes a CFI violation when opening the 'id' sysfs node under drm/card0/metrics. This happens to work because the layout of 'struct kobj_attribute' and 'struct device_attribute' are the same, so the container_of() cast happens to allow the ->show() callback to still work. Change the type of show_dynamic_id() to match the ->show() callback in 'struct kobj_attributes' and update the type of sysfs_metric_id to match, which resolves the CFI violation. Fixes: f89823c21224 ("drm/i915/perf: Implement I915_PERF_ADD/REMOVE_CONFIG interface") Signed-off-by: Nathan Chancellor <nathan(a)kernel.org> --- drivers/gpu/drm/i915/i915_perf.c | 4 ++-- drivers/gpu/drm/i915/i915_perf_types.h | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 0a9c3fcc09b1..1577ab6754db 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -4050,8 +4050,8 @@ static struct i915_oa_reg *alloc_oa_regs(struct i915_perf *perf, return ERR_PTR(err); } -static ssize_t show_dynamic_id(struct device *dev, - struct device_attribute *attr, +static ssize_t show_dynamic_id(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) { struct i915_oa_config *oa_config = diff --git a/drivers/gpu/drm/i915/i915_perf_types.h b/drivers/gpu/drm/i915/i915_perf_types.h index 473a3c0544bb..05cb9a335a97 100644 --- a/drivers/gpu/drm/i915/i915_perf_types.h +++ b/drivers/gpu/drm/i915/i915_perf_types.h @@ -55,7 +55,7 @@ struct i915_oa_config { struct attribute_group sysfs_metric; struct attribute *attrs[2]; - struct device_attribute sysfs_metric_id; + struct kobj_attribute sysfs_metric_id; struct kref ref; struct rcu_head rcu; base-commit: 7ecc3cc8a7b39f08eee9aea7b718187583342a70 -- 2.36.1

4 3

[PATCH v3] media: mediatek: vcodec: add vp9 decoder driver for mt8186
by Yunfei Dong 13 May '22

13 May '22

Add vp9 single core decode driver to support mt8186. Signed-off-by: Mingjia Zhang <mingjia.zhang(a)mediatek.com> Signed-off-by: Xiaoyong Lu <xiaoyong.lu(a)mediatek.com> Signed-off-by: Yunfei Dong <yunfei.dong(a)mediatek.com> --- This patch depends on "add h264 decoder driver for mt8186"[1] [1] https://patchwork.kernel.org/project/linux-mediatek/cover/20220512034620.30… --- changed with v2: - fix __iomem not reasonable, align share memory to dram. changed with v1: - fix sparse and smatch check fail. --- .../vcodec/vdec/vdec_vp9_req_lat_if.c | 197 +++++++++++++++++- .../platform/mediatek/vcodec/vdec_drv_if.c | 2 +- 2 files changed, 195 insertions(+), 4 deletions(-) diff --git a/drivers/media/platform/mediatek/vcodec/vdec/vdec_vp9_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/vdec/vdec_vp9_req_lat_if.c index 023aba4ec2c4..fba06f321baa 100644 --- a/drivers/media/platform/mediatek/vcodec/vdec/vdec_vp9_req_lat_if.c +++ b/drivers/media/platform/mediatek/vcodec/vdec/vdec_vp9_req_lat_if.c @@ -439,6 +439,8 @@ struct vdec_vp9_slice_ref { * @init_vsi: vsi used for initialized VP9 instance * @vsi: vsi used for decoding/flush ... * @core_vsi: vsi used for Core stage + * + * @sc_pfc: per frame context single core * @counts_map: used map to counts_helper * @counts_helper: counts table according to newest kernel spec */ @@ -487,6 +489,7 @@ struct vdec_vp9_slice_instance { }; struct vdec_vp9_slice_vsi *core_vsi; + struct vdec_vp9_slice_pfc sc_pfc; struct vdec_vp9_slice_counts_map counts_map; struct v4l2_vp9_frame_symbol_counts counts_helper; }; @@ -692,6 +695,25 @@ static int vdec_vp9_slice_tile_offset(int idx, int mi_num, int tile_log2) return offset < mi_num ? offset : mi_num; } +static +int vdec_vp9_slice_setup_single_from_src_to_dst(struct vdec_vp9_slice_instance *instance) +{ + struct vb2_v4l2_buffer *src; + struct vb2_v4l2_buffer *dst; + + src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx); + if (!src) + return -EINVAL; + + dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx); + if (!dst) + return -EINVAL; + + v4l2_m2m_buf_copy_metadata(src, dst, true); + + return 0; +} + static int vdec_vp9_slice_setup_lat_from_src_buf(struct vdec_vp9_slice_instance *instance, struct vdec_lat_buf *lat_buf) { @@ -1567,6 +1589,33 @@ static int vdec_vp9_slice_update_prob(struct vdec_vp9_slice_instance *instance, return 0; } +static int vdec_vp9_slice_update_single(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi; + + vsi = &pfc->vsi; + memcpy(&pfc->state[0], &vsi->state, sizeof(vsi->state)); + + mtk_vcodec_debug(instance, "Frame %u Y_CRC %08x %08x %08x %08x\n", + pfc->seq, + vsi->state.crc[0], vsi->state.crc[1], + vsi->state.crc[2], vsi->state.crc[3]); + mtk_vcodec_debug(instance, "Frame %u C_CRC %08x %08x %08x %08x\n", + pfc->seq, + vsi->state.crc[4], vsi->state.crc[5], + vsi->state.crc[6], vsi->state.crc[7]); + + vdec_vp9_slice_update_prob(instance, vsi); + + instance->width = vsi->frame.uh.frame_width; + instance->height = vsi->frame.uh.frame_height; + instance->frame_type = vsi->frame.uh.frame_type; + instance->show_frame = vsi->frame.uh.show_frame; + + return 0; +} + static int vdec_vp9_slice_update_lat(struct vdec_vp9_slice_instance *instance, struct vdec_lat_buf *lat_buf, struct vdec_vp9_slice_pfc *pfc) @@ -1690,6 +1739,40 @@ static int vdec_vp9_slice_setup_core_buffer(struct vdec_vp9_slice_instance *inst return 0; } +static void vdec_vp9_slice_setup_single_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc, + struct vdec_vp9_slice_vsi *vsi, + struct mtk_vcodec_mem *bs, + struct vdec_fb *fb) +{ + int i; + + vsi->bs.buf.dma_addr = bs->dma_addr; + vsi->bs.buf.size = bs->size; + vsi->bs.frame.dma_addr = bs->dma_addr; + vsi->bs.frame.size = bs->size; + + for (i = 0; i < 2; i++) { + vsi->mv[i].dma_addr = instance->mv[i].dma_addr; + vsi->mv[i].size = instance->mv[i].size; + } + for (i = 0; i < 2; i++) { + vsi->seg[i].dma_addr = instance->seg[i].dma_addr; + vsi->seg[i].size = instance->seg[i].size; + } + vsi->tile.dma_addr = instance->tile.dma_addr; + vsi->tile.size = instance->tile.size; + vsi->prob.dma_addr = instance->prob.dma_addr; + vsi->prob.size = instance->prob.size; + vsi->counts.dma_addr = instance->counts.dma_addr; + vsi->counts.size = instance->counts.size; + + vsi->row_info.buf = 0; + vsi->row_info.size = 0; + + vdec_vp9_slice_setup_core_buffer(instance, pfc, vsi, fb, NULL); +} + static int vdec_vp9_slice_setup_core(struct vdec_vp9_slice_instance *instance, struct vdec_fb *fb, struct vdec_lat_buf *lat_buf, @@ -1716,6 +1799,43 @@ static int vdec_vp9_slice_setup_core(struct vdec_vp9_slice_instance *instance, return ret; } +static int vdec_vp9_slice_setup_single(struct vdec_vp9_slice_instance *instance, + struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi = &pfc->vsi; + int ret; + + ret = vdec_vp9_slice_setup_single_from_src_to_dst(instance); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_pfc(instance, pfc); + if (ret) + goto err; + + ret = vdec_vp9_slice_alloc_working_buffer(instance, vsi); + if (ret) + goto err; + + vdec_vp9_slice_setup_single_buffer(instance, pfc, vsi, bs, fb); + vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[0]); + + ret = vdec_vp9_slice_setup_prob_buffer(instance, vsi); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_tile_buffer(instance, vsi, bs); + if (ret) + goto err; + + return 0; + +err: + return ret; +} + static int vdec_vp9_slice_update_core(struct vdec_vp9_slice_instance *instance, struct vdec_lat_buf *lat_buf, struct vdec_vp9_slice_pfc *pfc) @@ -1813,8 +1933,8 @@ static int vdec_vp9_slice_flush(void *h_vdec, struct mtk_vcodec_mem *bs, struct vdec_vp9_slice_instance *instance = h_vdec; mtk_vcodec_debug(instance, "flush ...\n"); - - vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue); + if (instance->ctx->dev->vdec_pdata->hw_arch != MTK_VDEC_PURE_SINGLE_CORE) + vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue); return vpu_dec_reset(&instance->vpu); } @@ -1867,6 +1987,63 @@ static int vdec_vp9_slice_get_param(void *h_vdec, enum vdec_get_param_type type, return 0; } +static int vdec_vp9_slice_single_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + struct vdec_vp9_slice_pfc *pfc = &(instance->sc_pfc); + struct vdec_vp9_slice_vsi *vsi; + struct mtk_vcodec_ctx *ctx; + int ret; + + if (!instance || !instance->ctx) + return -EINVAL; + ctx = instance->ctx; + + /* bs NULL means flush decoder */ + if (!bs) + return vdec_vp9_slice_flush(h_vdec, bs, fb, res_chg); + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) + return -EBUSY; + + vsi = &pfc->vsi; + + ret = vdec_vp9_slice_setup_single(instance, bs, fb, pfc); + if (ret) { + mtk_vcodec_err(instance, "Failed to setup VP9 single ret %d\n", ret); + return ret; + } + vdec_vp9_slice_vsi_to_remote(vsi, instance->vsi); + + ret = vpu_dec_start(&instance->vpu, NULL, 0); + if (ret) { + mtk_vcodec_err(instance, "Failed to dec VP9 ret %d\n", ret); + return ret; + } + + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vcodec_err(instance, "VP9 decode timeout %d\n", ret); + WRITE_ONCE(instance->vsi->state.timeout, 1); + } + + vpu_dec_end(&instance->vpu); + + vdec_vp9_slice_vsi_from_remote(vsi, instance->vsi, 0); + ret = vdec_vp9_slice_update_single(instance, pfc); + if (ret) { + mtk_vcodec_err(instance, "VP9 decode error: %d\n", ret); + return ret; + } + + instance->ctx->decoded_frame_cnt++; + return 0; +} + static int vdec_vp9_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, struct vdec_fb *fb, bool *res_chg) { @@ -1946,6 +2123,20 @@ static int vdec_vp9_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, return 0; } +static int vdec_vp9_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + int ret; + + if (instance->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE) + ret = vdec_vp9_slice_single_decode(h_vdec, bs, fb, res_chg); + else + ret = vdec_vp9_slice_lat_decode(h_vdec, bs, fb, res_chg); + + return ret; +} + static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf) { struct vdec_vp9_slice_instance *instance; @@ -2024,7 +2215,7 @@ static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf) const struct vdec_common_if vdec_vp9_slice_lat_if = { .init = vdec_vp9_slice_init, - .decode = vdec_vp9_slice_lat_decode, + .decode = vdec_vp9_slice_decode, .get_param = vdec_vp9_slice_get_param, .deinit = vdec_vp9_slice_deinit, }; diff --git a/drivers/media/platform/mediatek/vcodec/vdec_drv_if.c b/drivers/media/platform/mediatek/vcodec/vdec_drv_if.c index 27b4b35039cf..f3807f03d880 100644 --- a/drivers/media/platform/mediatek/vcodec/vdec_drv_if.c +++ b/drivers/media/platform/mediatek/vcodec/vdec_drv_if.c @@ -47,7 +47,7 @@ int vdec_if_init(struct mtk_vcodec_ctx *ctx, unsigned int fourcc) break; case V4L2_PIX_FMT_VP9_FRAME: ctx->dec_if = &vdec_vp9_slice_lat_if; - ctx->hw_id = MTK_VDEC_LAT0; + ctx->hw_id = IS_VDEC_LAT_ARCH(hw_arch) ? MTK_VDEC_LAT0 : MTK_VDEC_CORE; break; default: return -EINVAL; -- 2.18.0

1 0

[5.18 regression] drm/i915 BYT rendering broken due to "Remove short-term pins from execbuf, v6"
by Hans de Goede 13 May '22

13 May '22

Hi All, When running a 5.18-rc4 (and -rc5) kernel on a Chuwi Hi 8, which is a Bay Trail based tablet with 2G RAM and a 1200x1920 DSI panel. I noticed that gnome-shell was misrendering. Many UI elements were missing (they were all black) and at the gdm login screen (which is a special gnome-shell session) the screen often was entirely black until I move the cursor around and then various things got highlighted after which they sometimes stuck around and sometimes they disappeared again after the highlight. Since this problem does not happen with various 5.17.y kernels I believe that this is a kernel regression in 5.18. I've bisected this and the bisect points to: commit b5cfe6f7a6e1 ("drm/i915: Remove short-term pins from execbuf, v6.") from Maarten. This commit cleanly reverts on top of 5.18-rc5 and I can confirm that 5.18-rc5 with b5cfe6f7a6e1 reverted fixes things. I would be more then happy to test any possible fixes for this. Regards, Hans

4 3

[PATCH] drm/amd/display: Remove macro DC_DEFAULT_LOG_MASK
by pengfuyuan 13 May '22

13 May '22

[Why & How] The DC_DEFAULT_LOG_MASK macro has not been used for a long time, so remove it. Signed-off-by: pengfuyuan <pengfuyuan(a)kylinos.cn> --- .../drm/amd/display/include/logger_types.h | 34 ------------------- 1 file changed, 34 deletions(-) diff --git a/drivers/gpu/drm/amd/display/include/logger_types.h b/drivers/gpu/drm/amd/display/include/logger_types.h index f093b49c5e6e..a31d7c959f2c 100644 --- a/drivers/gpu/drm/amd/display/include/logger_types.h +++ b/drivers/gpu/drm/amd/display/include/logger_types.h @@ -131,37 +131,3 @@ enum dc_log_type { #define DC_MIN_LOG_MASK ((1 << LOG_ERROR) | \ (1 << LOG_DETECTION_EDID_PARSER)) -#define DC_DEFAULT_LOG_MASK ((1ULL << LOG_ERROR) | \ - (1ULL << LOG_WARNING) | \ - (1ULL << LOG_EVENT_MODE_SET) | \ - (1ULL << LOG_EVENT_DETECTION) | \ - (1ULL << LOG_EVENT_LINK_TRAINING) | \ - (1ULL << LOG_EVENT_LINK_LOSS) | \ - (1ULL << LOG_EVENT_UNDERFLOW) | \ - (1ULL << LOG_RESOURCE) | \ - (1ULL << LOG_FEATURE_OVERRIDE) | \ - (1ULL << LOG_DETECTION_EDID_PARSER) | \ - (1ULL << LOG_DC) | \ - (1ULL << LOG_HW_HOTPLUG) | \ - (1ULL << LOG_HW_SET_MODE) | \ - (1ULL << LOG_HW_RESUME_S3) | \ - (1ULL << LOG_HW_HPD_IRQ) | \ - (1ULL << LOG_SYNC) | \ - (1ULL << LOG_BANDWIDTH_VALIDATION) | \ - (1ULL << LOG_MST) | \ - (1ULL << LOG_DETECTION_DP_CAPS) | \ - (1ULL << LOG_BACKLIGHT)) | \ - (1ULL << LOG_I2C_AUX) | \ - (1ULL << LOG_IF_TRACE) | \ - (1ULL << LOG_HDMI_FRL) | \ - (1ULL << LOG_SCALER) | \ - (1ULL << LOG_DTN) /* | \ - (1ULL << LOG_DEBUG) | \ - (1ULL << LOG_BIOS) | \ - (1ULL << LOG_SURFACE) | \ - (1ULL << LOG_DML) | \ - (1ULL << LOG_HW_LINK_TRAINING) | \ - (1ULL << LOG_HW_AUDIO)| \ - (1ULL << LOG_BANDWIDTH_CALCS)*/ - -#endif /* __DAL_LOGGER_TYPES_H__ */ -- 2.25.1

2 2

← Newer
1
...
44
45
46
47
48
49
50
...
74
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel May 2022