Re: [PATCH v10 2/2] drm/i915: Allow user to set cache at BO creation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2023-05-18 22:11:03,  wrote:
> From: Fei Yang <fei.yang@xxxxxxxxx>
> 
> To comply with the design that buffer objects shall have immutable
> cache setting through out their life cycle, {set, get}_caching ioctl's
> are no longer supported from MTL onward. With that change caching
> policy can only be set at object creation time. The current code
> applies a default (platform dependent) cache setting for all objects.
> However this is not optimal for performance tuning. The patch extends
> the existing gem_create uAPI to let user set PAT index for the object
> at creation time.
> The new extension is platform independent, so UMD's can switch to using
> this extension for older platforms as well, while {set, get}_caching are
> still supported on these legacy paltforms for compatibility reason.
> 
> Test igt@gem_create@create_ext_set_pat posted at
> https://patchwork.freedesktop.org/series/117695/
> 
> Tested with https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22878
> 
> Signed-off-by: Fei Yang <fei.yang@xxxxxxxxx>
> Cc: Chris Wilson <chris.p.wilson@xxxxxxxxxxxxxxx>
> Cc: Matt Roper <matthew.d.roper@xxxxxxxxx>
> Cc: Andi Shyti <andi.shyti@xxxxxxxxxxxxxxx>
> Reviewed-by: Andi Shyti <andi.shyti@xxxxxxxxxxxxxxx>
> Acked-by: Jordan Justen <jordan.l.justen@xxxxxxxxx>

Nevertheless, I'm still disappointed my suggestion was so quickly shot
down.

I tried to look over our usage Mesa of i915 extensions, and found
this:

I915_GEM_CREATE_EXT_MEMORY_REGIONS:

 * If DRM_I915_QUERY_MEMORY_REGIONS is found

I915_GEM_CREATE_EXT_PROTECTED_CONTENT:

 * Probed via the current "robust" method. Resulted in 8s driver
   startup delay in some bad scenarios.

 * Will be guarded by I915_PARAM_PXP_STATUS when available in future

I915_CONTEXT_CREATE_EXT_SETPARAM (I915_CONTEXT_PARAM_ENGINES):

 * If DRM_I915_QUERY_ENGINE_INFO is found

I915_GEM_CREATE_EXT_SET_PAT:

 * When platform is mtl or newer

I think we will continue to try to find workarounds that imply the
extension's existence, but it could be nice to have a generic way to
find out what extensions the kernel knows about.

-Jordan




[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux