[Intel-xe] [PATCH 00/17] uAPI Alignment - take 1
Zhang, Carl
carl.zhang at intel.com
Thu Sep 21 15:06:24 UTC 2023
Hi Rodrigo,
Have a quick review, several questions:
1. we need add pad in drm_xe_engine_class_instance, in the future, media may need to distinguish slim vdbox and full vdbox , or other special configurations.
So, reserve one word should be suitable for future use. and then the structure length become 8 bytes.
2. no clear description about DRM_XE_DEVICE_QUERY_HWCONFIG, even we know it is in intel_hwconfig_types.h
3. about query GT, if the there are multiple tile , and it is symmetrical, which one will be MAIN, which is REMOTE?
4. no method to query HuC readiness .
5. XE_QUERY_CONFIG_NUM_PARAM is only used to represent the number of query type? And it is actually useless, no one should use it.
6.
+/** struct drm_xe_ext_set_property - XE set property extension */
+struct drm_xe_ext_set_property {
It should be VM set property extension ( XE_VM_EXTENSION_SET_PROPERTY)?
ps. I still not clear why we need extension for single bind operation. :)
may raise more question in following days.
thanks
Carl
> -----Original Message-----
> From: Vivi, Rodrigo <rodrigo.vivi at intel.com>
> Sent: Tuesday, September 19, 2023 10:25 PM
> To: intel-xe at lists.freedesktop.org
> Cc: Hazubski, Filip <filip.hazubski at intel.com>; Kopryk, Kamil
> <kamil.kopryk at intel.com>; Chaberek, Jakub <jakub.chaberek at intel.com>;
> Dunajski, Bartosz <bartosz.dunajski at intel.com>; Souza, Jose
> <jose.souza at intel.com>; Yu, Effie <effie.yu at intel.com>; Zhang, Carl
> <carl.zhang at intel.com>; Vivi, Rodrigo <rodrigo.vivi at intel.com>
> Subject: [PATCH 00/17] uAPI Alignment - take 1
>
> As a result of the uAPI review efforts started by Thomas[1], we have identified
> many updates on our uAPI that would lead to breakage in the compatibility.
> What it is not acceptable after we are merged upstream. So, let's break it
> before it is too late, and start upstreaming a good, reliable and clean uapi.
>
> Most of this work on putting these patches together for a single shot was led
> by Francois.
>
> The IGT counter part of this series is available as well[2].
>
> [1] -
> https://lore.kernel.org/all/863bebd0c624d6fc2b38c0a06b63e468b4185128.c
> amel at l\
> inux.intel.com/
> [2] - https://lore.kernel.org/all/20230919142000.91363-1-
> rodrigo.vivi at intel.com
>
> Ashutosh Dixit (1):
> drm/xe/uapi: Use common drm_xe_ext_set_property extension
>
> Francois Dugast (4):
> drm/xe/uapi: Separate VM_BIND's operation and flag
> drm/xe/vm: Remove VM_BIND_OP macro
> drm/xe/uapi: Remove MMIO ioctl
> drm/xe/uapi: Fix naming of
> XE_QUERY_CONFIG_MAX_EXEC_QUEUE_PRIORITY
>
> Matthew Brost (4):
> drm/xe: Fix xe_exec_queue_is_idle for parallel exec queues
> drm/xe: Deprecate XE_EXEC_QUEUE_SET_PROPERTY_COMPUTE_MODE
> implementation
> drm/xe: Rename exec_queue_kill_compute to
> xe_vm_remove_compute_exec_queue
> drm/xe: Remove XE_EXEC_QUEUE_SET_PROPERTY_COMPUTE_MODE from
> uAPI
>
> Rodrigo Vivi (5):
> drm/xe: Kill XE_VM_PROPERTY_BIND_OP_ERROR_CAPTURE_ADDRESS
> extension
> drm/xe/uapi: Document drm_xe_query_gt
> drm/xe/uapi: Replace useless 'instance' per unique gt_id
> drm/xe/uapi: Remove unused field of drm_xe_query_gt
> drm/xe/uapi: Rename gts to gt_list
>
> Umesh Nerlige Ramappa (3):
> drm/xe: Fix array bounds check for queries
> drm/xe: Set the correct type for xe_to_user_engine_class
> drm/xe: Correlate engine and cpu timestamps with better accuracy
>
> drivers/gpu/drm/xe/xe_device.c | 1 -
> drivers/gpu/drm/xe/xe_exec_queue.c | 99 +++-----
> drivers/gpu/drm/xe/xe_exec_queue_types.h | 6 +-
> drivers/gpu/drm/xe/xe_gt_types.h | 2 +-
> drivers/gpu/drm/xe/xe_mmio.c | 102 ---------
> drivers/gpu/drm/xe/xe_pci.c | 4 -
> drivers/gpu/drm/xe/xe_query.c | 187 +++++++++++++--
> drivers/gpu/drm/xe/xe_vm.c | 221 +++++-------------
> drivers/gpu/drm/xe/xe_vm.h | 1 +
> include/uapi/drm/xe_drm.h | 280 ++++++++++++-----------
> 10 files changed, 402 insertions(+), 501 deletions(-)
>
> --
> 2.41.0
More information about the Intel-xe
mailing list