From: John Harrison <john.c.harrison@intel.com>
To: Michal Wajdeczko <michal.wajdeczko@intel.com>,
<intel-gfx@lists.freedesktop.org>,
<intel-xe@lists.freedesktop.org>
Cc: Jani Nikula <jani.nikula@intel.com>,
Rodrigo Vivi <rodrigo.vivi@intel.com>,
Lucas De Marchi <lucas.demarchi@intel.com>,
Matthew Brost <matthew.brost@intel.com>,
Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Subject: Re: [RFC 0/7] Promote GuC ABI headers to shared location
Date: Wed, 12 Jun 2024 14:18:04 -0700 [thread overview]
Message-ID: <3a3a5560-ee3a-47e2-a609-faa64ae78475@intel.com> (raw)
In-Reply-To: <a0d7968d-81c5-4328-8654-9feab682655d@intel.com>
On 6/11/2024 14:45, Michal Wajdeczko wrote:
> On 11.06.2024 22:32, John Harrison wrote:
>> On 6/11/2024 07:30, Michal Wajdeczko wrote:
>>> There are many GuC ABI definitions named in the same way by the i915
>>> and Xe drivers, preventing proper generation of the documentation.
>>>
>>> Promote GuC ABI definitions to shared location that can be used by
>>> both drivers and can be included in documentation.
>> I still very strongly feel that this is the wrong place for such
>> documentation. We do not document any of the hardware registers in the
>> driver, nor the MI_ instructions, etc. Why is this any different? The
>> GuC ABI is not under the control of the Linux kernel driver, either i915
>> or Xe. It is effectively a hardware interface no different to any other
>> hardware interface. It is already fully documented by the owners of that
>> interface. Rather than just copying random chunks of that documentation
>> into the kernel driver, we should just be publishing the official
>> document itself. Same as we do for the rest of the hardware.
> so go publish this official document
My point is that we should be putting effort into making this happen. So
far as I know, I am the only person that ever raises this topic. And I
do keep raising it only to be told it is too low a priority. If more
people actually pushed for it then maybe it would happen.
>
> in the meantime IMO it is useful to show, with really little effort, on
> what grounds the communication between i915/Xe and GuC works, so
> everyone, not just selected engineers, can understand and review our
> implementation and check its correctness
Simply providing a half sentence description of a bitfield definition
does not tell you anything about the correctness or usage of an
interface. Sure, you can tell that it is syntactically correct but the
compiler can do that for you with the appropriate FIELD definitions. It
does not tell you anything about whether you are using the interface
properly or even if this is the proper interface to use. You need the
full descriptive documentation for that. Are you going to copy all of
that into the driver? Then you might as well just publish the document
you are copying from and be done.
>
> furthermore, if you don't like any hw documentation then we should
> revisit what is already in gpu/i915 section and also reconsider all our
> efforts to document all non-static driver functions, as those functions
> are still internal to the driver, not to be used outside
I do not know what you are arguing for here.
Xe in general is being pushed to have much better documentation of the
code than i915 did. But documentation of the driver is a very different
topic to documentation of the hardware.
John.
>
>> John.
>>
>>
>>> Cc: Jani Nikula <jani.nikula@intel.com>
>>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
>>> Cc: Lucas De Marchi <lucas.demarchi@intel.com>
>>> Cc: Matthew Brost <matthew.brost@intel.com>
>>> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
>>> Cc: John Harrison <John.C.Harrison@Intel.com>
>>>
>>> Michal Wajdeczko (7):
>>> drm/xe/guc: Promote GuC ABI headers to shared location
>>> Documentation/gpu: Separate GuC ABI section
>>> Documentation/gpu: Switch to shared GuC ABI definitions
>>> drm/intel/guc: Update CTB communication ABI
>>> drm/intel/guc: Add new KLV definitions
>>> drm/i915: Use shared GuC ABI definitions
>>> drm/xe: Promote SR-IOV GuC ABI definitions to shared location
>>>
>>> Documentation/gpu/drivers.rst | 1 +
>>> Documentation/gpu/guc.rst | 23 ++
>>> Documentation/gpu/i915.rst | 9 -
>>> drivers/gpu/drm/i915/Makefile | 5 +
>>> .../gt/uc/abi/guc_communication_ctb_abi.h | 170 -----------
>>> .../gt/uc/abi/guc_communication_mmio_abi.h | 49 ----
>>> drivers/gpu/drm/i915/gt/uc/abi/guc_klvs_abi.h | 112 --------
>>> .../gpu/drm/i915/gt/uc/abi/guc_messages_abi.h | 264 ------------------
>>> .../guc}/abi/guc_actions_sriov_abi.h | 0
>>> .../guc}/abi/guc_communication_ctb_abi.h | 2 +
>>> .../guc}/abi/guc_communication_mmio_abi.h | 0
>>> .../drm/{xe => intel/guc}/abi/guc_klvs_abi.h | 18 +-
>>> .../{xe => intel/guc}/abi/guc_messages_abi.h | 0
>>> .../guc}/abi/guc_relay_actions_abi.h | 0
>>> .../guc}/abi/guc_relay_communication_abi.h | 0
>>> drivers/gpu/drm/xe/Makefile | 5 +
>>> 16 files changed, 49 insertions(+), 609 deletions(-)
>>> create mode 100644 Documentation/gpu/guc.rst
>>> delete mode 100644
>>> drivers/gpu/drm/i915/gt/uc/abi/guc_communication_ctb_abi.h
>>> delete mode 100644
>>> drivers/gpu/drm/i915/gt/uc/abi/guc_communication_mmio_abi.h
>>> delete mode 100644 drivers/gpu/drm/i915/gt/uc/abi/guc_klvs_abi.h
>>> delete mode 100644 drivers/gpu/drm/i915/gt/uc/abi/guc_messages_abi.h
>>> rename drivers/gpu/drm/{xe => intel/guc}/abi/guc_actions_sriov_abi.h
>>> (100%)
>>> rename drivers/gpu/drm/{xe =>
>>> intel/guc}/abi/guc_communication_ctb_abi.h (98%)
>>> rename drivers/gpu/drm/{xe =>
>>> intel/guc}/abi/guc_communication_mmio_abi.h (100%)
>>> rename drivers/gpu/drm/{xe => intel/guc}/abi/guc_klvs_abi.h (97%)
>>> rename drivers/gpu/drm/{xe => intel/guc}/abi/guc_messages_abi.h (100%)
>>> rename drivers/gpu/drm/{xe => intel/guc}/abi/guc_relay_actions_abi.h
>>> (100%)
>>> rename drivers/gpu/drm/{xe =>
>>> intel/guc}/abi/guc_relay_communication_abi.h (100%)
>>>
prev parent reply other threads:[~2024-06-12 21:18 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-11 14:30 [RFC 0/7] Promote GuC ABI headers to shared location Michal Wajdeczko
2024-06-11 14:30 ` [RFC 1/7] drm/xe/guc: " Michal Wajdeczko
2024-06-11 14:30 ` [RFC 2/7] Documentation/gpu: Separate GuC ABI section Michal Wajdeczko
2024-06-11 14:30 ` [RFC 3/7] Documentation/gpu: Switch to shared GuC ABI definitions Michal Wajdeczko
2024-06-11 14:30 ` [RFC 4/7] drm/intel/guc: Update CTB communication ABI Michal Wajdeczko
2024-06-11 14:30 ` [RFC 5/7] drm/intel/guc: Add new KLV definitions Michal Wajdeczko
2024-06-11 14:30 ` [RFC 6/7] drm/i915: Use shared GuC ABI definitions Michal Wajdeczko
2024-06-11 14:30 ` [RFC 7/7] drm/xe: Promote SR-IOV GuC ABI definitions to shared location Michal Wajdeczko
2024-06-11 14:36 ` ✓ CI.Patch_applied: success for Promote GuC ABI headers " Patchwork
2024-06-11 14:36 ` ✗ CI.checkpatch: warning " Patchwork
2024-06-11 14:37 ` ✓ CI.KUnit: success " Patchwork
2024-06-11 14:49 ` ✓ CI.Build: " Patchwork
2024-06-11 14:51 ` ✗ CI.Hooks: failure " Patchwork
2024-06-11 14:52 ` ✗ CI.checksparse: warning " Patchwork
2024-06-11 15:13 ` [RFC 0/7] " Lucas De Marchi
2024-06-11 15:34 ` ✓ CI.BAT: success for " Patchwork
2024-06-11 17:08 ` ✗ CI.FULL: failure " Patchwork
2024-06-11 20:32 ` [RFC 0/7] " John Harrison
2024-06-11 21:45 ` Michal Wajdeczko
2024-06-11 22:12 ` Rodrigo Vivi
2024-06-12 4:12 ` Lucas De Marchi
2024-06-12 21:18 ` John Harrison [this message]
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=3a3a5560-ee3a-47e2-a609-faa64ae78475@intel.com \
--to=john.c.harrison@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=jani.nikula@intel.com \
--cc=lucas.demarchi@intel.com \
--cc=matthew.brost@intel.com \
--cc=michal.wajdeczko@intel.com \
--cc=rodrigo.vivi@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