[PATCH v7 13/15] drm/tegra: Implement job submission part of new UAPI

[Jon Hunter]

Hi Mikko,

On 10/06/2021 12:04, Mikko Perttunen wrote:
> Implement the job submission IOCTL with a minimum feature set.
> 
> Signed-off-by: Mikko Perttunen <mperttunen at nvidia.com>
> ---
> v7:
> * Allocate gather BO with DMA API to get page-aligned
>   memory
> * Add error prints to a few places where they were missing
> v6:
> * Remove sgt bypass path in gather_bo - this would cause
>   cache maintenance to be skipped and is unnecessary in
>   general.
> * Changes related to moving to using syncpoint IDs
> * Add syncobj related code
> * Print warning on submit failure describing the issue
> * Use ARCH_DMA_ADDR_T_64BIT to check if that is indeed
>   the case
> * Add support for relative syncpoint wait
> * Use pm_runtime_resume_and_get
> * Only try to resume engines that support runtime PM
> * Removed uapi subdirectory
> * Don't use "copy_err" variables for copy_from_user
>   return value
> * Fix setting of blocklinear flag
> v5:
> * Add 16K size limit to copies from userspace.
> * Guard RELOC_BLOCKLINEAR flag handling to only exist in ARM64
>   to prevent oversized shift on 32-bit platforms.
> v4:
> * Remove all features that are not strictly necessary.
> * Split into two patches.
> v3:
> * Remove WRITE_RELOC. Relocations are now patched implicitly
>   when patching is needed.
> * Directly call PM runtime APIs on devices instead of using
>   power_on/power_off callbacks.
> * Remove incorrect mutex unlock in tegra_drm_ioctl_channel_open
> * Use XA_FLAGS_ALLOC1 instead of XA_FLAGS_ALLOC
> * Accommodate for removal of timeout field and inlining of
>   syncpt_incrs array.
> * Copy entire user arrays at a time instead of going through
>   elements one-by-one.
> * Implement waiting of DMA reservations.
> * Split out gather_bo implementation into a separate file.
> * Fix length parameter passed to sg_init_one in gather_bo
> * Cosmetic cleanup.
> ---
>  drivers/gpu/drm/tegra/Makefile    |   2 +
>  drivers/gpu/drm/tegra/drm.c       |   4 +-
>  drivers/gpu/drm/tegra/gather_bo.c |  82 +++++
>  drivers/gpu/drm/tegra/gather_bo.h |  24 ++
>  drivers/gpu/drm/tegra/submit.c    | 549 ++++++++++++++++++++++++++++++
>  drivers/gpu/drm/tegra/submit.h    |  17 +
>  6 files changed, 677 insertions(+), 1 deletion(-)
>  create mode 100644 drivers/gpu/drm/tegra/gather_bo.c
>  create mode 100644 drivers/gpu/drm/tegra/gather_bo.h
>  create mode 100644 drivers/gpu/drm/tegra/submit.c
>  create mode 100644 drivers/gpu/drm/tegra/submit.h

...

> diff --git a/drivers/gpu/drm/tegra/submit.c b/drivers/gpu/drm/tegra/submit.c
> new file mode 100644
> index 000000000000..e3200c10ca9e
> --- /dev/null
> +++ b/drivers/gpu/drm/tegra/submit.c
> @@ -0,0 +1,549 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/* Copyright (c) 2020 NVIDIA Corporation */
> +
> +#include <linux/dma-fence-array.h>
> +#include <linux/file.h>
> +#include <linux/host1x.h>
> +#include <linux/iommu.h>
> +#include <linux/kref.h>
> +#include <linux/list.h>
> +#include <linux/nospec.h>
> +#include <linux/pm_runtime.h>
> +#include <linux/sync_file.h>
> +
> +#include <drm/drm_drv.h>
> +#include <drm/drm_file.h>
> +#include <drm/drm_syncobj.h>
> +
> +#include "drm.h"
> +#include "gather_bo.h"
> +#include "gem.h"
> +#include "submit.h"
> +#include "uapi.h"
> +
> +#define SUBMIT_ERR(ctx, fmt, ...) \
> +	dev_err_ratelimited(ctx->client->base.dev, \
> +		"%s: job submission failed: " fmt "\n", \
> +		current->comm __VA_OPT__(,) __VA_ARGS__)


For older compilers that don't support __VA_OPT__ this generates a
compilation error ...

drivers/gpu/drm/tegra/submit.c: In function ‘submit_copy_gather_data’:
drivers/gpu/drm/tegra/submit.c:27:17: error: expected ‘)’ before
‘__VA_OPT__’
   current->comm __VA_OPT__(,) __VA_ARGS__)
                 ^
I think that we may just have to use ##__VA_ARGS__ here.

Cheers
Jon

-- 
nvpublic