kernel/drivers/gpu/drm/exynos/exynos_drm_g2d.c, branch linux-4.17.y

drm/exynos: g2d: Delete an error message for a failed memory allocation in two functions

2018-02-20T00:16:18Z

Omit an extra message for a memory allocation failure in these functions. This issue was detected by using the Coccinelle software. Signed-off-by: Markus Elfring Signed-off-by: Inki Dae

drm/exynos: g2d: use monotonic timestamps

2018-02-20T00:16:18Z

The exynos DRM driver uses real-time 'struct timeval' values for exporting its timestamps to user space. This has multiple problems: 1. signed seconds overflow in y2038 2. the 'struct timeval' definition is deprecated in the kernel 3. time may jump or go backwards after a 'settimeofday()' syscall 4. other DRM timestamps are in CLOCK_MONOTONIC domain, so they can't be compared 5. exporting microseconds requires a division by 1000, which may be slow on some architectures. The code existed in two places before, but the IPP portion was removed in 8ded59413ccc ("drm/exynos: ipp: Remove Exynos DRM IPP subsystem"), so we no longer need to worry about it. Ideally timestamps should just use 64-bit nanoseconds instead, but of course we can't change that now. Instead, this tries to address the first four points above by using monotonic 'timespec' values. According to Tobias Jakobi, user space doesn't care about the timestamp at the moment, so we can change the format. Even if there is something looking at them, it will work just fine with monotonic times as long as the application only looks at the relative values between two events. Link: https://patchwork.kernel.org/patch/10038593/ Cc: Tobias Jakobi Signed-off-by: Arnd Bergmann Reviewed-by: Tobias Jakobi Signed-off-by: Inki Dae

drm/exynos: g2d: change platform driver name to 'exynos-drm-g2d'

2017-02-07T04:52:49Z

The current name is 's5p-g2d', which is identical with the driver name of the old V4L2 driver in media/platform. This is probably due to the DRM driver being based on the V4L2 driver when it was initially created. Still the clashing of driver names is confusing, so rename it to something in line with the other DRM subdrivers. Signed-off-by: Tobias Jakobi Signed-off-by: Inki Dae

drm/exynos: g2d: prevent integer overflow in

2017-01-30T23:50:35Z

The size computations done in the ioctl function use an integer. If userspace submits a request with req->cmd_nr or req->cmd_buf_nr set to INT_MAX, the integer computations overflow later, leading to potential (kernel) memory corruption. Prevent this issue by enforcing a limit on the number of submitted commands, so that we have enough headroom later for the size computations. Note that this change has no impact on the currently available users in userspace, like e.g. libdrm/exynos. While at it, also make a comment about the size computation more detailed. Signed-off-by: Joonyoung Shim Signed-off-by: Tobias Jakobi Signed-off-by: Inki Dae

mm: replace get_vaddr_frames() write/force parameters with gup_flags

2016-10-19T15:11:24Z

This removes the 'write' and 'force' from get_vaddr_frames() and replaces them with 'gup_flags' to make the use of FOLL_FORCE explicit in callers as use of this flag can result in surprising behaviour (and hence bugs) within the mm subsystem. Signed-off-by: Lorenzo Stoakes Acked-by: Michal Hocko Reviewed-by: Jan Kara Signed-off-by: Linus Torvalds

drm/exynos: g2d: simplify g2d_free_runqueue_node()

2016-09-30T15:39:42Z

The function is never called with zero 'runqueue_node'. Signed-off-by: Tobias Jakobi Acked-by: Marek Szyprowski Signed-off-by: Inki Dae

drm/exynos: g2d: use autosuspend mode for PM runtime

2016-09-30T15:39:41Z

The runqueue worker currently issues a get() when a new node is processed, and a put() once a node is completed. The corresponding suspend and resume calls currently only do clock gating, but with the upcoming introduction of IOMMU runpm also the corresponding IOMMU domain gets enabled (for get()) and disabled (for put()). This introduces performance regressions with we mitigate here. Switch PM runtime to autosuspend, such that clock gating and IOMMU control only happens when the engine is idle for a 'long' time. Signed-off-by: Tobias Jakobi Acked-by: Marek Szyprowski Signed-off-by: Inki Dae

drm/exynos: g2d: wait for engine to finish

2016-09-30T15:39:40Z

While the engine works on a runqueue node it does memory access to the buffers associated with that node. Make sure that the engine is idle when g2d_close() and/or g2d_remove() are called, i.e. buffer associated with the process (for g2d_close()), or all buffers (for g2d_remove()) can be safely be unmapped. We have to take into account that the engine might be in an undefined state, i.e. it hangs and doesn't become idle. In this case, we issue a hardware reset to return the hardware and the driver context into a proper state. Signed-off-by: Tobias Jakobi Acked-by: Marek Szyprowski Signed-off-by: Inki Dae

drm/exynos: g2d: remove runqueue nodes in g2d_{close,remove}()

2016-09-30T15:39:39Z

The driver might be closed (and/or removed) while there are still nodes queued for processing. Make sure to remove these nodes, which means all of them in the case of g2d_remove() and only those belonging to the corresponding process in g2d_close(). Signed-off-by: Tobias Jakobi Acked-by: Marek Szyprowski Signed-off-by: Inki Dae

drm/exynos: g2d: move PM management to runqueue worker

2016-09-30T15:39:39Z

Do all pm_runtime_{get,put}() calls in the runqueue worker. Also keep track of the engine's idle/busy state. Signed-off-by: Tobias Jakobi Acked-by: Marek Szyprowski Signed-off-by: Inki Dae