dri-devel January 2021

dri-devel@lists.freedesktop.org

322 participants
737 discussions

[PATCH v4 1/2] drm/msm: Add speed-bin support to a618 gpu
by Akhil P Oommen 03 Feb '21

03 Feb '21

Some GPUs support different max frequencies depending on the platform. To identify the correct variant, we should check the gpu speedbin fuse value. Add support for this speedbin detection to a6xx family along with the required fuse details for a618 gpu. Signed-off-by: Akhil P Oommen <akhilpo(a)codeaurora.org> --- Changes from v2: 1. Made the changes a6xx specific to save space. Changes from v1: 1. Added the changes to support a618 sku to the series. 2. Avoid failing probe in case of an unsupported sku. (Rob) Changes from v3: 1. Replace a618_speedbins[] with a function. (Jordan) drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 83 +++++++++++++++++++++++++++++++++++ drivers/gpu/drm/msm/adreno/a6xx_gpu.h | 2 + 2 files changed, 85 insertions(+) diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index 1306618..499d134 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -10,6 +10,7 @@ #include <linux/bitfield.h> #include <linux/devfreq.h> +#include <linux/nvmem-consumer.h> #include <linux/soc/qcom/llcc-qcom.h> #define GPU_PAS_ID 13 @@ -1208,6 +1209,10 @@ static void a6xx_destroy(struct msm_gpu *gpu) a6xx_gmu_remove(a6xx_gpu); adreno_gpu_cleanup(adreno_gpu); + + if (a6xx_gpu->opp_table) + dev_pm_opp_put_supported_hw(a6xx_gpu->opp_table); + kfree(a6xx_gpu); } @@ -1264,6 +1269,78 @@ static uint32_t a6xx_get_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring) return ring->memptrs->rptr = gpu_read(gpu, REG_A6XX_CP_RB_RPTR); } +static u32 a618_get_speed_bin(u32 fuse) +{ + if (fuse == 0) + return 0; + else if (fuse == 169) + return 1; + else if (fuse == 174) + return 2; + + return UINT_MAX; +} + +static u32 fuse_to_supp_hw(struct device *dev, u32 revn, u32 fuse) +{ + u32 val = UINT_MAX; + + if (revn == 618) + val = a618_get_speed_bin(fuse); + + if (val == UINT_MAX) { + DRM_DEV_ERROR(dev, + "missing support for speed-bin: %u. Some OPPs may not be supported by hardware", + fuse); + return UINT_MAX; + } + + return (1 << val); +} + +static int a6xx_set_supported_hw(struct device *dev, struct a6xx_gpu *a6xx_gpu, + u32 revn) +{ + struct opp_table *opp_table; + struct nvmem_cell *cell; + u32 supp_hw = UINT_MAX; + void *buf; + + cell = nvmem_cell_get(dev, "speed_bin"); + /* + * -ENOENT means that the platform doesn't support speedbin which is + * fine + */ + if (PTR_ERR(cell) == -ENOENT) + return 0; + else if (IS_ERR(cell)) { + DRM_DEV_ERROR(dev, + "failed to read speed-bin. Some OPPs may not be supported by hardware"); + goto done; + } + + buf = nvmem_cell_read(cell, NULL); + if (IS_ERR(buf)) { + nvmem_cell_put(cell); + DRM_DEV_ERROR(dev, + "failed to read speed-bin. Some OPPs may not be supported by hardware"); + goto done; + } + + supp_hw = fuse_to_supp_hw(dev, revn, *((u32 *) buf)); + + kfree(buf); + nvmem_cell_put(cell); + +done: + opp_table = dev_pm_opp_set_supported_hw(dev, &supp_hw, 1); + if (IS_ERR(opp_table)) + return PTR_ERR(opp_table); + + a6xx_gpu->opp_table = opp_table; + return 0; +} + static const struct adreno_gpu_funcs funcs = { .base = { .get_param = adreno_get_param, @@ -1325,6 +1402,12 @@ struct msm_gpu *a6xx_gpu_init(struct drm_device *dev) a6xx_llc_slices_init(pdev, a6xx_gpu); + ret = a6xx_set_supported_hw(&pdev->dev, a6xx_gpu, info->revn); + if (ret) { + a6xx_destroy(&(a6xx_gpu->base.base)); + return ERR_PTR(ret); + } + ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, 1); if (ret) { a6xx_destroy(&(a6xx_gpu->base.base)); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h index e793d32..ce0610c 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h @@ -33,6 +33,8 @@ struct a6xx_gpu { void *llc_slice; void *htw_llc_slice; bool have_mmu500; + + struct opp_table *opp_table; }; #define to_a6xx_gpu(x) container_of(x, struct a6xx_gpu, base) -- 2.7.4

3 4

[PATCH] drm/v3d: Fix incorrect return type to timedout_job
by Patrik Jakobsson 03 Feb '21

03 Feb '21

The commit a6a1f036c74e ("drm/scheduler: Job timeout handler returns status (v3)") incorrectly uses "enum drm_task_status" for v3d and causes a build failure. "enum drm_task_status" got changed into "enum drm_gpu_sched_status" in v3 of the patch but the change for v3d got lost. Fixes: ("drm/scheduler: Job timeout handler returns status (v3)") Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson(a)gmail.com> --- drivers/gpu/drm/v3d/v3d_sched.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index ef2338a294ca..ceb33f8e4379 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -259,7 +259,7 @@ v3d_cache_clean_job_run(struct drm_sched_job *sched_job) return NULL; } -static enum drm_gpu_sched_status +static enum drm_gpu_sched_stat v3d_gpu_reset_for_timeout(struct v3d_dev *v3d, struct drm_sched_job *sched_job) { enum v3d_queue q; @@ -294,7 +294,7 @@ v3d_gpu_reset_for_timeout(struct v3d_dev *v3d, struct drm_sched_job *sched_job) * could fail if the GPU got in an infinite loop in the CL, but that * is pretty unlikely outside of an i-g-t testcase. */ -static enum drm_task_status +static enum drm_gpu_sched_stat v3d_cl_job_timedout(struct drm_sched_job *sched_job, enum v3d_queue q, u32 *timedout_ctca, u32 *timedout_ctra) { @@ -312,7 +312,7 @@ v3d_cl_job_timedout(struct drm_sched_job *sched_job, enum v3d_queue q, return v3d_gpu_reset_for_timeout(v3d, sched_job); } -static enum drm_task_status +static enum drm_gpu_sched_stat v3d_bin_job_timedout(struct drm_sched_job *sched_job) { struct v3d_bin_job *job = to_bin_job(sched_job); @@ -321,7 +321,7 @@ v3d_bin_job_timedout(struct drm_sched_job *sched_job) &job->timedout_ctca, &job->timedout_ctra); } -static enum drm_task_status +static enum drm_gpu_sched_stat v3d_render_job_timedout(struct drm_sched_job *sched_job) { struct v3d_render_job *job = to_render_job(sched_job); @@ -330,7 +330,7 @@ v3d_render_job_timedout(struct drm_sched_job *sched_job) &job->timedout_ctca, &job->timedout_ctra); } -static enum drm_task_status +static enum drm_gpu_sched_stat v3d_generic_job_timedout(struct drm_sched_job *sched_job) { struct v3d_job *job = to_v3d_job(sched_job); @@ -338,7 +338,7 @@ v3d_generic_job_timedout(struct drm_sched_job *sched_job) return v3d_gpu_reset_for_timeout(job->v3d, sched_job); } -static enum drm_task_status +static enum drm_gpu_sched_stat v3d_csd_job_timedout(struct drm_sched_job *sched_job) { struct v3d_csd_job *job = to_csd_job(sched_job); -- 2.30.0

4 4

[PATCH] drm/todo: Add entry for moving to dma_resv_lock
by Daniel Vetter 03 Feb '21

03 Feb '21

Requested by Thomas. I think it justifies a new level, since I tried to make some forward progress on this last summer, and gave up (for now). This is very tricky. Signed-off-by: Daniel Vetter <daniel.vetter(a)intel.com> Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com> Cc: Maxime Ripard <mripard(a)kernel.org> Cc: Thomas Zimmermann <tzimmermann(a)suse.de> Cc: David Airlie <airlied(a)linux.ie> Cc: Daniel Vetter <daniel(a)ffwll.ch> Cc: Sumit Semwal <sumit.semwal(a)linaro.org> Cc: "Christian König" <christian.koenig(a)amd.com> Cc: linux-media(a)vger.kernel.org Cc: linaro-mm-sig(a)lists.linaro.org --- Documentation/gpu/todo.rst | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst index dea9082c0e5f..f872d3d33218 100644 --- a/Documentation/gpu/todo.rst +++ b/Documentation/gpu/todo.rst @@ -23,6 +23,9 @@ Advanced: Tricky tasks that need fairly good understanding of the DRM subsystem and graphics topics. Generally need the relevant hardware for development and testing. +Expert: Only attempt these if you've successfully completed some tricky +refactorings already and are an expert in the specific area + Subsystem-wide refactorings =========================== @@ -168,6 +171,22 @@ Contact: Daniel Vetter, respective driver maintainers Level: Advanced +Move Buffer Object Locking to dma_resv_lock() +--------------------------------------------- + +Many drivers have their own per-object locking scheme, usually using +mutex_lock(). This causes all kinds of trouble for buffer sharing, since +depending which driver is the exporter and importer, the locking hierarchy is +reversed. + +To solve this we need one standard per-object locking mechanism, which is +dma_resv_lock(). This lock needs to be called as the outermost lock, with all +other driver specific per-object locks removed. The problem is tha rolling out +the actual change to the locking contract is a flag day, due to struct dma_buf +buffer sharing. + +Level: Expert + Convert logging to drm_* functions with drm_device paramater ------------------------------------------------------------ -- 2.30.0

4 4

[PATCH v4 0/5] drm/qxl: fix driver shutdown issues.
by Gerd Hoffmann 03 Feb '21

03 Feb '21

Almost there. Still getting this on driver unbind: kobject: '(null)' ((____ptrval____)): is not initialized, yet kobject_put(= ) is being called [ ... ] Call Trace: ttm_device_fini+0x133/0x1b0 [ttm] qxl_ttm_fini+0x2f/0x40 [qxl] qxl_device_fini+0x88/0x120 [qxl] drm_minor_release+0x3d/0x60 but I don't think this is the qxl driver's fault. Gerd Hoffmann (5): drm/qxl: use drmm_mode_config_init drm/qxl: unpin release objects drm/qxl: release shadow on shutdown drm/qxl: handle shadow in primary destroy drm/qxl: properly free qxl releases drivers/gpu/drm/qxl/qxl_drv.h | 1 + drivers/gpu/drm/qxl/qxl_display.c | 11 +++++++++-- drivers/gpu/drm/qxl/qxl_kms.c | 22 ++++++++++++++++++++-- drivers/gpu/drm/qxl/qxl_release.c | 3 +++ 4 files changed, 33 insertions(+), 4 deletions(-) --=20 2.29.2

2 7

[PATCH] procfs/dmabuf: Add /proc/<pid>/task/<tid>/dmabuf_fds
by Kalesh Singh 03 Feb '21

03 Feb '21

In order to measure how much memory a process actually consumes, it is necessary to include the DMA buffer sizes for that process in the memory accounting. Since the handle to DMA buffers are raw FDs, it is important to be able to identify which processes have FD references to a DMA buffer. Currently, DMA buffer FDs can be accounted using /proc/<pid>/fd/* and /proc/<pid>/fdinfo -- both of which are only root readable, as follows: 1. Do a readlink on each FD. 2. If the target path begins with "/dmabuf", then the FD is a dmabuf FD. 3. stat the file to get the dmabuf inode number. 4. Read/ proc/<pid>/fdinfo/<fd>, to get the DMA buffer size. Android captures per-process system memory state when certain low memory events (e.g a foreground app kill) occur, to identify potential memory hoggers. To include a process’s dmabuf usage as part of its memory state, the data collection needs to be fast enough to reflect the memory state at the time of such events. Since reading /proc/<pid>/fd/ and /proc/<pid>/fdinfo/ requires root privileges, this approach is not suitable for production builds. Granting root privileges even to a system process increases the attack surface and is highly undesirable. Additionally this is slow as it requires many context switches for searching and getting the dma-buf info. With the addition of per-buffer dmabuf stats in sysfs [1], the DMA buffer details can be queried using their unique inode numbers. This patch proposes adding a /proc/<pid>/task/<tid>/dmabuf_fds interface. /proc/<pid>/task/<tid>/dmabuf_fds contains a list of inode numbers for every DMA buffer FD that the task has. Entries with the same inode number can appear more than once, indicating the total FD references for the associated DMA buffer. If a thread shares the same files as the group leader then its dmabuf_fds file will be empty, as these dmabufs are reported by the group leader. The interface requires PTRACE_MODE_READ_FSCRED (same as /proc/<pid>/maps) and allows the efficient accounting of per-process DMA buffer usage without requiring root privileges. (See data below) Performance Comparison: ----------------------- The following data compares the time to capture the sizes of all DMA buffers referenced by FDs for all processes on an arm64 android device. ------------------------------------------------------- | Core 0 (Little) | Core 7 (Big) | ------------------------------------------------------- >From <pid>/fdinfo | 318 ms | 145 ms | ------------------------------------------------------- Inodes from | 114 ms | 27 ms | dmabuf_fds; | (2.8x ^) | (5.4x ^) | data from sysfs | | | ------------------------------------------------------- It can be inferred that in the worst case there is a 2.8x speedup for determining per-process DMA buffer FD sizes, when using the proposed interfaces. [1] https://lore.kernel.org/dri-devel/20210119225723.388883-1-hridya@google.com/ Signed-off-by: Kalesh Singh <kaleshsingh(a)google.com> --- Documentation/filesystems/proc.rst | 30 ++++++ drivers/dma-buf/dma-buf.c | 7 +- fs/proc/Makefile | 1 + fs/proc/base.c | 1 + fs/proc/dma_bufs.c | 159 +++++++++++++++++++++++++++++ fs/proc/internal.h | 1 + include/linux/dma-buf.h | 5 + 7 files changed, 198 insertions(+), 6 deletions(-) create mode 100644 fs/proc/dma_bufs.c diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst index 2fa69f710e2a..757dd47ab679 100644 --- a/Documentation/filesystems/proc.rst +++ b/Documentation/filesystems/proc.rst @@ -47,6 +47,7 @@ fixes/update part 1.1 Stefani Seibold <stefani(a)seibold.net> June 9 2009 3.10 /proc/<pid>/timerslack_ns - Task timerslack value 3.11 /proc/<pid>/patch_state - Livepatch patch operation state 3.12 /proc/<pid>/arch_status - Task architecture specific information + 3.13 /proc/<pid>/task/<tid>/dmabuf_fds - DMA buffers referenced by an FD 4 Configuring procfs 4.1 Mount options @@ -2131,6 +2132,35 @@ AVX512_elapsed_ms the task is unlikely an AVX512 user, but depends on the workload and the scheduling scenario, it also could be a false negative mentioned above. +3.13 /proc/<pid>/task/<tid>/dmabuf_fds - DMA buffers referenced by an FD +------------------------------------------------------------------------- +This file exposes a list of the inode numbers for every DMA buffer +FD that the task has. + +The same inode number can appear more than once, indicating the total +FD references for the associated DMA buffer. + +The inode number can be used to lookup the DMA buffer information in +the sysfs interface /sys/kernel/dmabuf/buffers/<inode-no>/. + +Example Output +~~~~~~~~~~~~~~ +$ cat /proc/612/task/612/dmabuf_fds +30972 30973 45678 49326 + +Permission to access this file is governed by a ptrace access mode +PTRACE_MODE_READ_FSCREDS. + +Threads can have different files when created without specifying +the CLONE_FILES flag. For this reason the interface is presented as +/proc/<pid>/task/<tid>/dmabuf_fds and not /proc/<pid>/dmabuf_fds. +This simplifies kernel code and aggregation can be handled in +userspace. + +If a thread has the same files as its group leader, then its dmabuf_fds +file will be empty as these dmabufs are already reported by the +group leader. + Chapter 4: Configuring procfs ============================= diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index 9ad6397aaa97..0660c06be4c6 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -29,8 +29,6 @@ #include <uapi/linux/dma-buf.h> #include <uapi/linux/magic.h> -static inline int is_dma_buf_file(struct file *); - struct dma_buf_list { struct list_head head; struct mutex lock; @@ -434,10 +432,7 @@ static const struct file_operations dma_buf_fops = { .show_fdinfo = dma_buf_show_fdinfo, }; -/* - * is_dma_buf_file - Check if struct file* is associated with dma_buf - */ -static inline int is_dma_buf_file(struct file *file) +int is_dma_buf_file(struct file *file) { return file->f_op == &dma_buf_fops; } diff --git a/fs/proc/Makefile b/fs/proc/Makefile index bd08616ed8ba..91a67f43ddf4 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -16,6 +16,7 @@ proc-y += cmdline.o proc-y += consoles.o proc-y += cpuinfo.o proc-y += devices.o +proc-y += dma_bufs.o proc-y += interrupts.o proc-y += loadavg.o proc-y += meminfo.o diff --git a/fs/proc/base.c b/fs/proc/base.c index b3422cda2a91..af15a60b9831 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3598,6 +3598,7 @@ static const struct pid_entry tid_base_stuff[] = { #ifdef CONFIG_SECCOMP_CACHE_DEBUG ONE("seccomp_cache", S_IRUSR, proc_pid_seccomp_cache), #endif + REG("dmabuf_fds", 0444, proc_tid_dmabuf_fds_operations), }; static int proc_tid_base_readdir(struct file *file, struct dir_context *ctx) diff --git a/fs/proc/dma_bufs.c b/fs/proc/dma_bufs.c new file mode 100644 index 000000000000..46ea9cf968ed --- /dev/null +++ b/fs/proc/dma_bufs.c @@ -0,0 +1,159 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Per-process DMA-BUF Stats + * + * Copyright (C) 2021 Google LLC. + */ + +#include <linux/dma-buf.h> +#include <linux/fdtable.h> +#include <linux/ptrace.h> +#include <linux/seq_file.h> + +#include "internal.h" + +struct dmabuf_fds_private { + struct inode *inode; + struct task_struct *task; + struct file *dmabuf_file; +}; + +static loff_t *next_dmabuf(struct dmabuf_fds_private *priv, + loff_t *pos) +{ + struct fdtable *fdt; + struct file *file; + + rcu_read_lock(); + fdt = files_fdtable(priv->task->files); + for (; *pos < fdt->max_fds; ++*pos) { + file = files_lookup_fd_rcu(priv->task->files, (unsigned int) *pos); + if (file && is_dma_buf_file(file) && get_file_rcu(file)) { + priv->dmabuf_file = file; + break; + } + } + if (*pos >= fdt->max_fds) + pos = NULL; + rcu_read_unlock(); + + return pos; +} + +static void *dmabuf_fds_seq_start(struct seq_file *s, loff_t *pos) +{ + struct dmabuf_fds_private *priv = s->private; + struct files_struct *group_leader_files; + + priv->task = get_proc_task(priv->inode); + + if (!priv->task) + return ERR_PTR(-ESRCH); + + /* Hold task lock for duration that files need to be stable */ + task_lock(priv->task); + + /* + * If this task is not the group leader but shares the same files, leave file empty. + * These dmabufs are already reported in the group leader's dmabuf_fds. + */ + group_leader_files = priv->task->group_leader->files; + if (priv->task != priv->task->group_leader && priv->task->files == group_leader_files) { + task_unlock(priv->task); + put_task_struct(priv->task); + priv->task = NULL; + return NULL; + } + + return next_dmabuf(priv, pos); +} + +static void *dmabuf_fds_seq_next(struct seq_file *s, void *v, loff_t *pos) +{ + ++*pos; + return next_dmabuf(s->private, pos); +} + +static void dmabuf_fds_seq_stop(struct seq_file *s, void *v) +{ + struct dmabuf_fds_private *priv = s->private; + + if (priv->task) { + task_unlock(priv->task); + put_task_struct(priv->task); + + } + if (priv->dmabuf_file) + fput(priv->dmabuf_file); +} + +static int dmabuf_fds_seq_show(struct seq_file *s, void *v) +{ + struct dmabuf_fds_private *priv = s->private; + struct file *file = priv->dmabuf_file; + struct dma_buf *dmabuf = file->private_data; + + if (!dmabuf) + return -ESRCH; + + seq_printf(s, "%8lu ", file_inode(file)->i_ino); + + fput(priv->dmabuf_file); + priv->dmabuf_file = NULL; + + return 0; +} + +static const struct seq_operations proc_tid_dmabuf_fds_seq_ops = { + .start = dmabuf_fds_seq_start, + .next = dmabuf_fds_seq_next, + .stop = dmabuf_fds_seq_stop, + .show = dmabuf_fds_seq_show +}; + +static int proc_dmabuf_fds_open(struct inode *inode, struct file *file, + const struct seq_operations *ops) +{ + struct dmabuf_fds_private *priv; + struct task_struct *task; + bool allowed = false; + + task = get_proc_task(inode); + if (!task) + return -ESRCH; + + allowed = ptrace_may_access(task, PTRACE_MODE_READ_FSCREDS); + put_task_struct(task); + + if (!allowed) + return -EACCES; + + priv = __seq_open_private(file, ops, sizeof(*priv)); + if (!priv) + return -ENOMEM; + + priv->inode = inode; + priv->task = NULL; + priv->dmabuf_file = NULL; + + return 0; +} + +static int proc_dmabuf_fds_release(struct inode *inode, struct file *file) +{ + return seq_release_private(inode, file); +} + +static int tid_dmabuf_fds_open(struct inode *inode, struct file *file) +{ + return proc_dmabuf_fds_open(inode, file, + &proc_tid_dmabuf_fds_seq_ops); +} + +const struct file_operations proc_tid_dmabuf_fds_operations = { + .open = tid_dmabuf_fds_open, + .read = seq_read, + .llseek = seq_lseek, + .release = proc_dmabuf_fds_release, +}; + diff --git a/fs/proc/internal.h b/fs/proc/internal.h index f60b379dcdc7..4ca74220db9c 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -303,6 +303,7 @@ extern const struct file_operations proc_pid_smaps_operations; extern const struct file_operations proc_pid_smaps_rollup_operations; extern const struct file_operations proc_clear_refs_operations; extern const struct file_operations proc_pagemap_operations; +extern const struct file_operations proc_tid_dmabuf_fds_operations; extern unsigned long task_vsize(struct mm_struct *); extern unsigned long task_statm(struct mm_struct *, diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h index cf72699cb2bc..087e11f7f193 100644 --- a/include/linux/dma-buf.h +++ b/include/linux/dma-buf.h @@ -27,6 +27,11 @@ struct device; struct dma_buf; struct dma_buf_attachment; +/** + * Check if struct file* is associated with dma_buf. + */ +int is_dma_buf_file(struct file *file); + /** * struct dma_buf_ops - operations possible on struct dma_buf * @vmap: [optional] creates a virtual mapping for the buffer into kernel -- 2.30.0.280.ga3ce27912f-goog

7 16

[RFC][PATCH 1/3] dma-buf: heaps: Add deferred-free-helper library code
by John Stultz 03 Feb '21

03 Feb '21

This patch provides infrastructure for deferring buffer frees. This is a feature ION provided which when used with some form of a page pool, provides a nice performance boost in an allocation microbenchmark. The reason it helps is it allows the page-zeroing to be done out of the normal allocation/free path, and pushed off to a kthread. As not all heaps will find this useful, its implemented as a optional helper library that heaps can utilize. Cc: Sumit Semwal <sumit.semwal(a)linaro.org> Cc: Liam Mark <lmark(a)codeaurora.org> Cc: Chris Goldsworthy <cgoldswo(a)codeaurora.org> Cc: Laura Abbott <labbott(a)kernel.org> Cc: Brian Starkey <Brian.Starkey(a)arm.com> Cc: Hridya Valsaraju <hridya(a)google.com> Cc: Suren Baghdasaryan <surenb(a)google.com> Cc: Sandeep Patil <sspatil(a)google.com> Cc: Daniel Mentz <danielmentz(a)google.com> Cc: Ørjan Eide <orjan.eide(a)arm.com> Cc: Robin Murphy <robin.murphy(a)arm.com> Cc: Ezequiel Garcia <ezequiel(a)collabora.com> Cc: Simon Ser <contact(a)emersion.fr> Cc: James Jones <jajones(a)nvidia.com> Cc: linux-media(a)vger.kernel.org Cc: dri-devel(a)lists.freedesktop.org Signed-off-by: John Stultz <john.stultz(a)linaro.org> --- drivers/dma-buf/heaps/Kconfig | 3 + drivers/dma-buf/heaps/Makefile | 1 + drivers/dma-buf/heaps/deferred-free-helper.c | 136 +++++++++++++++++++ drivers/dma-buf/heaps/deferred-free-helper.h | 15 ++ 4 files changed, 155 insertions(+) create mode 100644 drivers/dma-buf/heaps/deferred-free-helper.c create mode 100644 drivers/dma-buf/heaps/deferred-free-helper.h diff --git a/drivers/dma-buf/heaps/Kconfig b/drivers/dma-buf/heaps/Kconfig index a5eef06c4226..ecf65204f714 100644 --- a/drivers/dma-buf/heaps/Kconfig +++ b/drivers/dma-buf/heaps/Kconfig @@ -1,3 +1,6 @@ +config DMABUF_HEAPS_DEFERRED_FREE + bool + config DMABUF_HEAPS_SYSTEM bool "DMA-BUF System Heap" depends on DMABUF_HEAPS diff --git a/drivers/dma-buf/heaps/Makefile b/drivers/dma-buf/heaps/Makefile index 974467791032..4e7839875615 100644 --- a/drivers/dma-buf/heaps/Makefile +++ b/drivers/dma-buf/heaps/Makefile @@ -1,3 +1,4 @@ # SPDX-License-Identifier: GPL-2.0 +obj-$(CONFIG_DMABUF_HEAPS_DEFERRED_FREE) += deferred-free-helper.o obj-$(CONFIG_DMABUF_HEAPS_SYSTEM) += system_heap.o obj-$(CONFIG_DMABUF_HEAPS_CMA) += cma_heap.o diff --git a/drivers/dma-buf/heaps/deferred-free-helper.c b/drivers/dma-buf/heaps/deferred-free-helper.c new file mode 100644 index 000000000000..b8f54860454f --- /dev/null +++ b/drivers/dma-buf/heaps/deferred-free-helper.c @@ -0,0 +1,136 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Deferred dmabuf freeing helper + * + * Copyright (C) 2020 Linaro, Ltd. + * + * Based on the ION page pool code + * Copyright (C) 2011 Google, Inc. + */ + +#include <linux/freezer.h> +#include <linux/list.h> +#include <linux/slab.h> +#include <linux/swap.h> +#include <linux/sched/signal.h> + +#include "deferred-free-helper.h" + +static LIST_HEAD(free_list); +static size_t list_size; +wait_queue_head_t freelist_waitqueue; +struct task_struct *freelist_task; +static DEFINE_MUTEX(free_list_lock); + +enum { + USE_POOL = 0, + SKIP_POOL = 1, +}; + +void deferred_free(struct deferred_freelist_item *item, + void (*free)(struct deferred_freelist_item*, bool), + size_t size) +{ + INIT_LIST_HEAD(&item->list); + item->size = size; + item->free = free; + + mutex_lock(&free_list_lock); + list_add(&item->list, &free_list); + list_size += size; + mutex_unlock(&free_list_lock); + wake_up(&freelist_waitqueue); +} + +static size_t free_one_item(bool nopool) +{ + size_t size = 0; + struct deferred_freelist_item *item; + + mutex_lock(&free_list_lock); + if (list_empty(&free_list)) { + mutex_unlock(&free_list_lock); + return 0; + } + item = list_first_entry(&free_list, struct deferred_freelist_item, list); + list_del(&item->list); + size = item->size; + list_size -= size; + mutex_unlock(&free_list_lock); + + item->free(item, nopool); + return size; +} + +static unsigned long get_freelist_size(void) +{ + unsigned long size; + + mutex_lock(&free_list_lock); + size = list_size; + mutex_unlock(&free_list_lock); + return size; +} + +static unsigned long freelist_shrink_count(struct shrinker *shrinker, + struct shrink_control *sc) +{ + return get_freelist_size(); +} + +static unsigned long freelist_shrink_scan(struct shrinker *shrinker, + struct shrink_control *sc) +{ + int total_freed = 0; + + if (sc->nr_to_scan == 0) + return 0; + + while (total_freed < sc->nr_to_scan) { + int freed = free_one_item(SKIP_POOL); + + if (!freed) + break; + + total_freed += freed; + } + + return total_freed; +} + +static struct shrinker freelist_shrinker = { + .count_objects = freelist_shrink_count, + .scan_objects = freelist_shrink_scan, + .seeks = DEFAULT_SEEKS, + .batch = 0, +}; + +static int deferred_free_thread(void *data) +{ + while (true) { + wait_event_freezable(freelist_waitqueue, + get_freelist_size() > 0); + + free_one_item(USE_POOL); + } + + return 0; +} + +static int deferred_freelist_init(void) +{ + list_size = 0; + + init_waitqueue_head(&freelist_waitqueue); + freelist_task = kthread_run(deferred_free_thread, NULL, + "%s", "dmabuf-deferred-free-worker"); + if (IS_ERR(freelist_task)) { + pr_err("%s: creating thread for deferred free failed\n", + __func__); + return -1; + } + sched_set_normal(freelist_task, 19); + + return register_shrinker(&freelist_shrinker); +} +device_initcall(deferred_freelist_init); diff --git a/drivers/dma-buf/heaps/deferred-free-helper.h b/drivers/dma-buf/heaps/deferred-free-helper.h new file mode 100644 index 000000000000..09a2274a897c --- /dev/null +++ b/drivers/dma-buf/heaps/deferred-free-helper.h @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifndef DEFERRED_FREE_HELPER_H +#define DEFERRED_FREE_HELPER_H + +struct deferred_freelist_item { + size_t size; + void (*free)(struct deferred_freelist_item *i, bool no_pool); + struct list_head list; +}; + +void deferred_free(struct deferred_freelist_item *item, + void (*free)(struct deferred_freelist_item *i, bool no_pool), + size_t size); +#endif -- 2.17.1

4 11

[PATCH] drm/nouveau: remove set but not used variable ‘pdev’ in nouveau_bios_init
by Ye Bin 03 Feb '21

03 Feb '21

Fix follow warning: drivers/gpu/drm/nouveau/nouveau_bios.c:2086:18: warning: variable ‘pdev’ set but not used [-Wunused-but-set-variable] struct pci_dev *pdev; ^~~~ Reported-by: Hulk Robot <hulkci(a)huawei.com> Signed-off-by: Ye Bin <yebin10(a)huawei.com> --- drivers/gpu/drm/nouveau/nouveau_bios.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c index 7cc683b8dc7a..e8c445eb1100 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.c +++ b/drivers/gpu/drm/nouveau/nouveau_bios.c @@ -2083,13 +2083,11 @@ nouveau_bios_init(struct drm_device *dev) { struct nouveau_drm *drm = nouveau_drm(dev); struct nvbios *bios = &drm->vbios; - struct pci_dev *pdev; int ret; /* only relevant for PCI devices */ if (!dev_is_pci(dev->dev)) return 0; - pdev = to_pci_dev(dev->dev); if (!NVInitVBIOS(dev)) return -ENODEV; -- 2.25.4

2 1

[PATCH 0/5] Clock fixes for DSI 10nm PLL
by AngeloGioacchino Del Regno 02 Feb '21

02 Feb '21

The DSI 10nm PLL driver was apparently ported from downstream, but some of its "features" were not ported over, for a good reason. Pity is that the removal of the downstream dependencies broke the clock calculation logic for this driver and that made it impossible to use any DSI display on at least MSM8998. This patch series fixes the calculation issues and also solves some TODOs that I've found in this driver. Tested on: - Sony Xperia XZ Premium (MSM8998) dual-dsi command-mode LCD display - F(x)Tec Pro1 (MSM8998) single dsi, video-mode OLED display AngeloGioacchino Del Regno (5): drm/msm/dsi_pll_10nm: Fix dividing the same numbers twice drm/msm/dsi_pll_10nm: Solve TODO for multiplier frac_bits assignment drm/msm/dsi_pll_10nm: Fix bad VCO rate calculation and prescaler drm/msm/dsi_pll_10nm: Fix variable usage for pll_lockdet_rate drm/msm/dsi_pll_10nm: Convert pr_err prints to DRM_DEV_ERROR drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c | 43 ++++++++++------------ 1 file changed, 20 insertions(+), 23 deletions(-) -- 2.29.2

2 16

[PATCH 00/29] treewide: Convert comma separated statements
by Joe Perches 02 Feb '21

02 Feb '21

There are many comma separated statements in the kernel. See:https://lore.kernel.org/lkml/alpine.DEB.2.22.394.2008201856110.2524@had… Convert the comma separated statements that are in if/do/while blocks to use braces and semicolons. Many comma separated statements still exist but those are changes for another day. Joe Perches (29): coding-style.rst: Avoid comma statements alpha: Avoid comma separated statements ia64: Avoid comma separated statements sparc: Avoid comma separated statements ata: Avoid comma separated statements drbd: Avoid comma separated statements lp: Avoid comma separated statements dma-buf: Avoid comma separated statements drm/gma500: Avoid comma separated statements drm/i915: Avoid comma separated statements hwmon: (scmi-hwmon): Avoid comma separated statements Input: MT - Avoid comma separated statements bcache: Avoid comma separated statements media: Avoid comma separated statements mtd: Avoid comma separated statements 8390: Avoid comma separated statements fs_enet: Avoid comma separated statements wan: sbni: Avoid comma separated statements s390/tty3270: Avoid comma separated statements scai/arm: Avoid comma separated statements media: atomisp: Avoid comma separated statements video: fbdev: Avoid comma separated statements fuse: Avoid comma separated statements reiserfs: Avoid comma separated statements lib/zlib: Avoid comma separated statements lib: zstd: Avoid comma separated statements ipv6: fib6: Avoid comma separated statements sunrpc: Avoid comma separated statements tools: Avoid comma separated statements Documentation/process/coding-style.rst | 17 + arch/alpha/kernel/pci_iommu.c | 8 +- arch/alpha/oprofile/op_model_ev4.c | 22 +- arch/alpha/oprofile/op_model_ev5.c | 8 +- arch/ia64/kernel/smpboot.c | 7 +- arch/sparc/kernel/smp_64.c | 7 +- drivers/ata/pata_icside.c | 21 +- drivers/block/drbd/drbd_receiver.c | 6 +- drivers/char/lp.c | 6 +- drivers/dma-buf/st-dma-fence.c | 7 +- drivers/gpu/drm/gma500/mdfld_intel_display.c | 44 ++- drivers/gpu/drm/i915/gt/gen8_ppgtt.c | 8 +- drivers/gpu/drm/i915/gt/intel_gt_requests.c | 6 +- .../gpu/drm/i915/gt/selftest_workarounds.c | 6 +- drivers/gpu/drm/i915/intel_runtime_pm.c | 6 +- drivers/hwmon/scmi-hwmon.c | 6 +- drivers/input/input-mt.c | 11 +- drivers/md/bcache/bset.c | 12 +- drivers/md/bcache/sysfs.c | 6 +- drivers/media/i2c/msp3400-kthreads.c | 12 +- drivers/media/pci/bt8xx/bttv-cards.c | 6 +- drivers/media/pci/saa7134/saa7134-video.c | 7 +- drivers/mtd/devices/lart.c | 10 +- drivers/net/ethernet/8390/axnet_cs.c | 19 +- drivers/net/ethernet/8390/lib8390.c | 14 +- drivers/net/ethernet/8390/pcnet_cs.c | 6 +- .../ethernet/freescale/fs_enet/fs_enet-main.c | 11 +- drivers/net/wan/sbni.c | 101 +++--- drivers/s390/char/tty3270.c | 6 +- drivers/scsi/arm/cumana_2.c | 19 +- drivers/scsi/arm/eesox.c | 9 +- drivers/scsi/arm/powertec.c | 9 +- .../media/atomisp/pci/atomisp_subdev.c | 6 +- drivers/video/fbdev/tgafb.c | 12 +- fs/fuse/dir.c | 24 +- fs/reiserfs/fix_node.c | 36 ++- lib/zlib_deflate/deftree.c | 49 ++- lib/zstd/compress.c | 120 ++++--- lib/zstd/fse_compress.c | 24 +- lib/zstd/huf_compress.c | 6 +- net/ipv6/ip6_fib.c | 12 +- net/sunrpc/sysctl.c | 6 +- tools/lib/subcmd/help.c | 10 +- tools/power/cpupower/utils/cpufreq-set.c | 14 +- tools/testing/selftests/vm/gup_benchmark.c | 18 +- tools/testing/selftests/vm/userfaultfd.c | 296 +++++++++++------- 46 files changed, 694 insertions(+), 382 deletions(-) -- 2.26.0

10 21

[PATCH] drm/vc4: hvs: Fix buffer overflow with the dlist handling
by Maxime Ripard 02 Feb '21

02 Feb '21

Commit 0a038c1c29a7 ("drm/vc4: Move LBM creation out of vc4_plane_mode_set()") changed the LBM allocation logic from first allocating the LBM memory for the plane to running mode_set, adding a gap in the LBM, and then running the dlist allocation filling that gap. The gap was introduced by incrementing the dlist array index, but was never checking whether or not we were over the array length, leading eventually to memory corruptions if we ever crossed this limit. vc4_dlist_write had that logic though, and was reallocating a larger dlist array when reaching the end of the buffer. Let's share the logic between both functions. Cc: Boris Brezillon <boris.brezillon(a)collabora.com> Cc: Eric Anholt <eric(a)anholt.net> Fixes: 0a038c1c29a7 ("drm/vc4: Move LBM creation out of vc4_plane_mode_set()") Signed-off-by: Maxime Ripard <maxime(a)cerno.tech> --- drivers/gpu/drm/vc4/vc4_plane.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c index b5586c92bfe5..3d33fe3dacea 100644 --- a/drivers/gpu/drm/vc4/vc4_plane.c +++ b/drivers/gpu/drm/vc4/vc4_plane.c @@ -227,7 +227,7 @@ static void vc4_plane_reset(struct drm_plane *plane) __drm_atomic_helper_plane_reset(plane, &vc4_state->base); } -static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val) +static void vc4_dlist_counter_increment(struct vc4_plane_state *vc4_state) { if (vc4_state->dlist_count == vc4_state->dlist_size) { u32 new_size = max(4u, vc4_state->dlist_count * 2); @@ -242,7 +242,15 @@ static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val) vc4_state->dlist_size = new_size; } - vc4_state->dlist[vc4_state->dlist_count++] = val; + vc4_state->dlist_count++; +} + +static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val) +{ + unsigned int idx = vc4_state->dlist_count; + + vc4_dlist_counter_increment(vc4_state); + vc4_state->dlist[idx] = val; } /* Returns the scl0/scl1 field based on whether the dimensions need to @@ -1057,8 +1065,10 @@ static int vc4_plane_mode_set(struct drm_plane *plane, * be set when calling vc4_plane_allocate_lbm(). */ if (vc4_state->y_scaling[0] != VC4_SCALING_NONE || - vc4_state->y_scaling[1] != VC4_SCALING_NONE) - vc4_state->lbm_offset = vc4_state->dlist_count++; + vc4_state->y_scaling[1] != VC4_SCALING_NONE) { + vc4_state->lbm_offset = vc4_state->dlist_count; + vc4_dlist_counter_increment(vc4_state); + } if (num_planes > 1) { /* Emit Cb/Cr as channel 0 and Y as channel -- 2.29.2

3 3

← Newer
1
...
16
17
18
19
20
21
22
...
74
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel January 2021