From: Daniel Vetter <daniel@ffwll.ch>
To: Matthew Auld <matthew.auld@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH v2 29/37] drm/i915: Introduce DRM_I915_GEM_MMAP_OFFSET
Date: Tue, 30 Jul 2019 11:49:01 +0200 [thread overview]
Message-ID: <20190730094901.GQ15868@phenom.ffwll.local> (raw)
In-Reply-To: <20190627205633.1143-30-matthew.auld@intel.com>
On Thu, Jun 27, 2019 at 09:56:25PM +0100, Matthew Auld wrote:
> From: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>
>
> Add a new CPU mmap implementation that allows multiple fault handlers
> that depends on the object's backing pages.
>
> Note that we multiplex mmap_gtt and mmap_offset through the same ioctl,
> and use the zero extending behaviour of drm to differentiate between
> them, when we inspect the flags.
>
> Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>
> Signed-off-by: Matthew Auld <matthew.auld@intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
So I thought that the plan is to reject invalid mmaps, i.e. mmap modes
which are not compatibale with all placement options. Given that, why do
we need this?
- cpu mmap with all the flags still keep working, as long as the only
placement you select is smem.
- for lmem/stolen the only option we have is a wc mapping, either through
the pci bar or through the gtt. So for objects only sitting in there
also no problem, we can just keep using the current gtt mmap stuff (but
redirect it internally).
- that leaves us with objects which can move around. Only option allows is
WC, and the gtt mmap ioctl does that already. When the object is in smem
we'll need to redirect it to a cpu wc mmap, but I think we need to do
that anyway.
So not really seeing what the uapi problem is you're trying to solve here?
Can you pls explain why we need this?
Thanks, Daniel
> ---
> drivers/gpu/drm/i915/gem/i915_gem_ioctls.h | 2 ++
> drivers/gpu/drm/i915/gem/i915_gem_mman.c | 30 ++++++++++++++++++
> .../gpu/drm/i915/gem/i915_gem_object_types.h | 3 ++
> drivers/gpu/drm/i915/i915_drv.c | 3 +-
> include/uapi/drm/i915_drm.h | 31 +++++++++++++++++++
> 5 files changed, 68 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h
> index ddc7f2a52b3e..5abd5b2172f2 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h
> @@ -30,6 +30,8 @@ int i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
> struct drm_file *file);
> int i915_gem_mmap_gtt_ioctl(struct drm_device *dev, void *data,
> struct drm_file *file);
> +int i915_gem_mmap_offset_ioctl(struct drm_device *dev, void *data,
> + struct drm_file *file_priv);
> int i915_gem_pread_ioctl(struct drm_device *dev, void *data,
> struct drm_file *file);
> int i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
> index 7b46f44d9c20..cbf89e80a97b 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
> @@ -536,12 +536,42 @@ i915_gem_mmap_gtt_ioctl(struct drm_device *dev, void *data,
> struct drm_file *file)
> {
> struct drm_i915_gem_mmap_offset *args = data;
> + struct drm_i915_private *i915 = to_i915(dev);
> +
> + if (args->flags & I915_MMAP_OFFSET_FLAGS)
> + return i915_gem_mmap_offset_ioctl(dev, data, file);
> +
> + if (!HAS_MAPPABLE_APERTURE(i915)) {
> + DRM_ERROR("No aperture, cannot mmap via legacy GTT\n");
> + return -ENODEV;
> + }
>
> return __assign_gem_object_mmap_data(file, args->handle,
> I915_MMAP_TYPE_GTT,
> &args->offset);
> }
>
> +int i915_gem_mmap_offset_ioctl(struct drm_device *dev, void *data,
> + struct drm_file *file)
> +{
> + struct drm_i915_gem_mmap_offset *args = data;
> + enum i915_mmap_type type;
> +
> + if ((args->flags & (I915_MMAP_OFFSET_WC | I915_MMAP_OFFSET_WB)) &&
> + !boot_cpu_has(X86_FEATURE_PAT))
> + return -ENODEV;
> +
> + if (args->flags & I915_MMAP_OFFSET_WC)
> + type = I915_MMAP_TYPE_OFFSET_WC;
> + else if (args->flags & I915_MMAP_OFFSET_WB)
> + type = I915_MMAP_TYPE_OFFSET_WB;
> + else if (args->flags & I915_MMAP_OFFSET_UC)
> + type = I915_MMAP_TYPE_OFFSET_UC;
> +
> + return __assign_gem_object_mmap_data(file, args->handle, type,
> + &args->offset);
> +}
> +
> void i915_mmap_offset_object_release(struct kref *ref)
> {
> struct i915_mmap_offset *mmo = container_of(ref,
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
> index 86f358da8085..f95e54a25426 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
> @@ -63,6 +63,9 @@ struct drm_i915_gem_object_ops {
>
> enum i915_mmap_type {
> I915_MMAP_TYPE_GTT = 0,
> + I915_MMAP_TYPE_OFFSET_WC,
> + I915_MMAP_TYPE_OFFSET_WB,
> + I915_MMAP_TYPE_OFFSET_UC,
> };
>
> struct i915_mmap_offset {
> diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> index 0f1f3b7f3029..8dadd6b9a0a9 100644
> --- a/drivers/gpu/drm/i915/i915_drv.c
> +++ b/drivers/gpu/drm/i915/i915_drv.c
> @@ -459,6 +459,7 @@ static int i915_getparam_ioctl(struct drm_device *dev, void *data,
> case I915_PARAM_HAS_EXEC_BATCH_FIRST:
> case I915_PARAM_HAS_EXEC_FENCE_ARRAY:
> case I915_PARAM_HAS_EXEC_SUBMIT_FENCE:
> + case I915_PARAM_MMAP_OFFSET_VERSION:
> /* For the time being all of these are always true;
> * if some supported hardware does not have one of these
> * features this value needs to be provided from
> @@ -3176,7 +3177,7 @@ static const struct drm_ioctl_desc i915_ioctls[] = {
> DRM_IOCTL_DEF_DRV(I915_GEM_PREAD, i915_gem_pread_ioctl, DRM_RENDER_ALLOW),
> DRM_IOCTL_DEF_DRV(I915_GEM_PWRITE, i915_gem_pwrite_ioctl, DRM_RENDER_ALLOW),
> DRM_IOCTL_DEF_DRV(I915_GEM_MMAP, i915_gem_mmap_ioctl, DRM_RENDER_ALLOW),
> - DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_GTT, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW),
> + DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_OFFSET, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW),
> DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW),
> DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW),
> DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling_ioctl, DRM_RENDER_ALLOW),
> diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
> index 328d05e77d9f..729e729e2282 100644
> --- a/include/uapi/drm/i915_drm.h
> +++ b/include/uapi/drm/i915_drm.h
> @@ -359,6 +359,7 @@ typedef struct _drm_i915_sarea {
> #define DRM_I915_QUERY 0x39
> #define DRM_I915_GEM_VM_CREATE 0x3a
> #define DRM_I915_GEM_VM_DESTROY 0x3b
> +#define DRM_I915_GEM_MMAP_OFFSET DRM_I915_GEM_MMAP_GTT
> /* Must be kept compact -- no holes */
>
> #define DRM_IOCTL_I915_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_I915_INIT, drm_i915_init_t)
> @@ -421,6 +422,7 @@ typedef struct _drm_i915_sarea {
> #define DRM_IOCTL_I915_QUERY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_QUERY, struct drm_i915_query)
> #define DRM_IOCTL_I915_GEM_VM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_VM_CREATE, struct drm_i915_gem_vm_control)
> #define DRM_IOCTL_I915_GEM_VM_DESTROY DRM_IOW (DRM_COMMAND_BASE + DRM_I915_GEM_VM_DESTROY, struct drm_i915_gem_vm_control)
> +#define DRM_IOCTL_I915_GEM_MMAP_OFFSET DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_MMAP_OFFSET, struct drm_i915_gem_mmap_offset)
>
> /* Allow drivers to submit batchbuffers directly to hardware, relying
> * on the security mechanisms provided by hardware.
> @@ -610,6 +612,10 @@ typedef struct drm_i915_irq_wait {
> * See I915_EXEC_FENCE_OUT and I915_EXEC_FENCE_SUBMIT.
> */
> #define I915_PARAM_HAS_EXEC_SUBMIT_FENCE 53
> +
> +/* Mmap offset ioctl */
> +#define I915_PARAM_MMAP_OFFSET_VERSION 54
> +
> /* Must be kept compact -- no holes and well documented */
>
> typedef struct drm_i915_getparam {
> @@ -785,6 +791,31 @@ struct drm_i915_gem_mmap_gtt {
> __u64 offset;
> };
>
> +struct drm_i915_gem_mmap_offset {
> + /** Handle for the object being mapped. */
> + __u32 handle;
> + __u32 pad;
> + /**
> + * Fake offset to use for subsequent mmap call
> + *
> + * This is a fixed-size type for 32/64 compatibility.
> + */
> + __u64 offset;
> +
> + /**
> + * Flags for extended behaviour.
> + *
> + * It is mandatory that either one of the _WC/_WB flags
> + * should be passed here.
> + */
> + __u64 flags;
> +#define I915_MMAP_OFFSET_WC (1 << 0)
> +#define I915_MMAP_OFFSET_WB (1 << 1)
> +#define I915_MMAP_OFFSET_UC (1 << 2)
> +#define I915_MMAP_OFFSET_FLAGS \
> + (I915_MMAP_OFFSET_WC | I915_MMAP_OFFSET_WB | I915_MMAP_OFFSET_UC)
> +};
> +
> struct drm_i915_gem_set_domain {
> /** Handle for the object */
> __u32 handle;
> --
> 2.20.1
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2019-07-30 9:49 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-27 20:55 [PATCH v2 00/37] Introduce memory region concept (including device local memory) Matthew Auld
2019-06-27 20:55 ` [PATCH v2 01/37] drm/i915: buddy allocator Matthew Auld
2019-06-27 22:28 ` Chris Wilson
2019-06-28 9:35 ` Chris Wilson
2019-06-27 20:55 ` [PATCH v2 02/37] drm/i915: introduce intel_memory_region Matthew Auld
2019-06-27 22:47 ` Chris Wilson
2019-06-28 8:09 ` Chris Wilson
2019-06-27 20:55 ` [PATCH v2 03/37] drm/i915/region: support basic eviction Matthew Auld
2019-06-27 22:59 ` Chris Wilson
2019-07-30 16:26 ` Daniel Vetter
2019-08-15 10:48 ` Matthew Auld
2019-08-15 14:26 ` Daniel Vetter
2019-08-15 14:34 ` Daniel Vetter
2019-08-15 14:57 ` Tang, CQ
2019-08-15 16:20 ` Daniel Vetter
2019-08-15 16:35 ` Tang, CQ
2019-08-15 15:26 ` Chris Wilson
2019-08-15 16:23 ` Daniel Vetter
2019-06-27 20:56 ` [PATCH v2 04/37] drm/i915/region: support continuous allocations Matthew Auld
2019-06-27 23:01 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 05/37] drm/i915/region: support volatile objects Matthew Auld
2019-06-27 23:03 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 06/37] drm/i915: Add memory region information to device_info Matthew Auld
2019-06-27 23:05 ` Chris Wilson
2019-06-27 23:08 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 07/37] drm/i915: support creating LMEM objects Matthew Auld
2019-06-27 23:11 ` Chris Wilson
2019-06-27 23:16 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 08/37] drm/i915: setup io-mapping for LMEM Matthew Auld
2019-06-27 20:56 ` [PATCH v2 09/37] drm/i915/lmem: support kernel mapping Matthew Auld
2019-06-27 23:27 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 10/37] drm/i915/blt: support copying objects Matthew Auld
2019-06-27 23:35 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 11/37] drm/i915/selftests: move gpu-write-dw into utils Matthew Auld
2019-06-27 20:56 ` [PATCH v2 12/37] drm/i915/selftests: add write-dword test for LMEM Matthew Auld
2019-06-27 20:56 ` [PATCH v2 13/37] drm/i915/selftests: don't just test CACHE_NONE for huge-pages Matthew Auld
2019-06-27 23:40 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 14/37] drm/i915/selftest: extend coverage to include LMEM huge-pages Matthew Auld
2019-06-27 23:42 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 15/37] drm/i915/lmem: support CPU relocations Matthew Auld
2019-06-27 23:46 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 16/37] drm/i915/lmem: support pread Matthew Auld
2019-06-27 23:50 ` Chris Wilson
2019-07-30 8:58 ` Daniel Vetter
2019-07-30 9:25 ` Matthew Auld
2019-07-30 9:50 ` Daniel Vetter
2019-07-30 12:05 ` Chris Wilson
2019-07-30 12:42 ` Daniel Vetter
2019-06-27 20:56 ` [PATCH v2 17/37] drm/i915/lmem: support pwrite Matthew Auld
2019-06-27 20:56 ` [PATCH v2 18/37] drm/i915: enumerate and init each supported region Matthew Auld
2019-06-27 20:56 ` [PATCH v2 19/37] drm/i915: treat shmem as a region Matthew Auld
2019-06-27 23:55 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 20/37] drm/i915: treat stolen " Matthew Auld
2019-06-27 20:56 ` [PATCH v2 21/37] drm/i915: define HAS_MAPPABLE_APERTURE Matthew Auld
2019-06-27 20:56 ` [PATCH v2 22/37] drm/i915: do not map aperture if it is not available Matthew Auld
2019-06-27 20:56 ` [PATCH v2 23/37] drm/i915: expose missing map_gtt support to users Matthew Auld
2019-06-27 23:59 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 24/37] drm/i915: set num_fence_regs to 0 if there is no aperture Matthew Auld
2019-06-28 0:00 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 25/37] drm/i915/selftests: check for missing aperture Matthew Auld
2019-06-27 20:56 ` [PATCH v2 26/37] drm/i915: error capture with no ggtt slot Matthew Auld
2019-06-27 20:56 ` [PATCH v2 27/37] drm/i915: Don't try to place HWS in non-existing mappable region Matthew Auld
2019-06-27 20:56 ` [PATCH v2 28/37] drm/i915: Allow i915 to manage the vma offset nodes instead of drm core Matthew Auld
2019-06-28 0:05 ` Chris Wilson
2019-06-28 0:08 ` Chris Wilson
2019-06-28 0:09 ` Chris Wilson
2019-06-28 0:10 ` Chris Wilson
2019-06-27 20:56 ` [PATCH v2 29/37] drm/i915: Introduce DRM_I915_GEM_MMAP_OFFSET Matthew Auld
2019-06-28 0:12 ` Chris Wilson
2019-07-30 9:49 ` Daniel Vetter [this message]
2019-07-30 14:28 ` Matthew Auld
2019-07-30 16:22 ` Daniel Vetter
2019-08-12 16:18 ` Daniel Vetter
2019-06-27 20:56 ` [PATCH v2 30/37] drm/i915/lmem: add helper to get CPU accessible offset Matthew Auld
2019-06-27 20:56 ` [PATCH v2 31/37] drm/i915: Add cpu and lmem fault handlers Matthew Auld
2019-06-27 20:56 ` [PATCH v2 32/37] drm/i915: cpu-map based dumb buffers Matthew Auld
2019-06-27 20:56 ` [PATCH v2 33/37] drm/i915: support basic object migration Matthew Auld
2019-06-27 20:56 ` [PATCH v2 34/37] drm/i915: Introduce GEM_OBJECT_SETPARAM with I915_PARAM_MEMORY_REGION Matthew Auld
2019-06-28 0:22 ` Chris Wilson
2019-06-28 5:53 ` Tvrtko Ursulin
2019-07-30 16:17 ` Daniel Vetter
2019-06-27 20:56 ` [PATCH v2 35/37] drm/i915/query: Expose memory regions through the query uAPI Matthew Auld
2019-06-28 5:59 ` Tvrtko Ursulin
2019-06-27 20:56 ` [PATCH v2 36/37] HAX drm/i915: add the fake lmem region Matthew Auld
2019-06-27 20:56 ` [PATCH v2 37/37] HAX drm/i915/lmem: default userspace allocations to LMEM Matthew Auld
2019-06-27 21:36 ` ✗ Fi.CI.CHECKPATCH: warning for Introduce memory region concept (including device local memory) (rev2) Patchwork
2019-06-27 21:50 ` ✗ Fi.CI.SPARSE: " Patchwork
2019-06-28 9:59 ` ✗ Fi.CI.BAT: failure " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190730094901.GQ15868@phenom.ffwll.local \
--to=daniel@ffwll.ch \
--cc=intel-gfx@lists.freedesktop.org \
--cc=matthew.auld@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox