All of lore.kernel.org
 help / color / mirror / Atom feed
From: John Harrison <John.C.Harrison@Intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, Intel-GFX@Lists.FreeDesktop.Org
Subject: Re: [PATCH 03/55] drm/i915: i915_add_request must not fail
Date: Tue, 23 Jun 2015 11:47:24 +0100	[thread overview]
Message-ID: <5589393C.5080901@Intel.com> (raw)
In-Reply-To: <20150623101654.GL31353@nuc-i3427.alporthouse.com>

On 23/06/2015 11:16, Chris Wilson wrote:
> On Fri, May 29, 2015 at 05:43:24PM +0100, John.C.Harrison@Intel.com wrote:
>> From: John Harrison <John.C.Harrison@Intel.com>
>>
>> The i915_add_request() function is called to keep track of work that has been
>> written to the ring buffer. It adds epilogue commands to track progress (seqno
>> updates and such), moves the request structure onto the right list and other
>> such house keeping tasks. However, the work itself has already been written to
>> the ring and will get executed whether or not the add request call succeeds. So
>> no matter what goes wrong, there isn't a whole lot of point in failing the call.
>>
>> At the moment, this is fine(ish). If the add request does bail early on and not
>> do the housekeeping, the request will still float around in the
>> ring->outstanding_lazy_request field and be picked up next time. It means
>> multiple pieces of work will be tagged as the same request and driver can't
>> actually wait for the first piece of work until something else has been
>> submitted. But it all sort of hangs together.
>>
>> This patch series is all about removing the OLR and guaranteeing that each piece
>> of work gets its own personal request. That means that there is no more
>> 'hoovering up of forgotten requests'. If the request does not get tracked then
>> it will be leaked. Thus the add request call _must_ not fail. The previous patch
>> should have already ensured that it _will_ not fail by removing the potential
>> for running out of ring space. This patch enforces the rule by actually removing
>> the early exit paths and the return code.
>>
>> Note that if something does manage to fail and the epilogue commands don't get
>> written to the ring, the driver will still hang together. The request will be
>> added to the tracking lists. And as in the old case, any subsequent work will
>> generate a new seqno which will suffice for marking the old one as complete.
>>
>> v2: Improved WARNings (Tomas Elf review request).
> Nak. Daniel please revert this mess.
>
> Even in the current code it has a failure mode it cannot handle.
> -Chris
>

Can you explain further?

Is this a new failure mode? This patch was originally posted back in 
March and there were no comments then to say that is was not the right 
direction to take. The discussion was that this might not be the 
prettiest way to do things but it is the best solution given where the 
driver is at and what we are trying to do.

John.

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2015-06-23 10:47 UTC|newest]

Thread overview: 120+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-29 16:43 [PATCH 00/55] Remove the outstanding_lazy_request John.C.Harrison
2015-05-29 16:43 ` [PATCH 01/55] drm/i915: Re-instate request->uniq becuase it is extremely useful John.C.Harrison
2015-06-03 11:14   ` Tomas Elf
2015-05-29 16:43 ` [PATCH 02/55] drm/i915: Reserve ring buffer space for i915_add_request() commands John.C.Harrison
2015-06-02 18:14   ` Tomas Elf
2015-06-04 12:06   ` John.C.Harrison
2015-06-09 16:00     ` Tomas Elf
2015-06-18 12:10       ` John.C.Harrison
2015-06-17 14:04     ` Daniel Vetter
2015-06-18 10:43       ` John Harrison
2015-06-19 16:34   ` John.C.Harrison
2015-06-22 20:12     ` Daniel Vetter
2015-06-23 11:38       ` John Harrison
2015-06-23 13:24         ` Daniel Vetter
2015-06-23 15:43           ` John Harrison
2015-06-23 20:00             ` Daniel Vetter
2015-06-24 12:18               ` John Harrison
2015-06-24 12:45                 ` Daniel Vetter
2015-06-24 17:05                   ` John Harrison
2015-05-29 16:43 ` [PATCH 03/55] drm/i915: i915_add_request must not fail John.C.Harrison
2015-06-02 18:16   ` Tomas Elf
2015-06-04 14:07     ` John Harrison
2015-06-05 10:55       ` Tomas Elf
2015-06-23 10:16   ` Chris Wilson
2015-06-23 10:47     ` John Harrison [this message]
2015-05-29 16:43 ` [PATCH 04/55] drm/i915: Early alloc request in execbuff John.C.Harrison
2015-05-29 16:43 ` [PATCH 05/55] drm/i915: Set context in request from creation even in legacy mode John.C.Harrison
2015-05-29 16:43 ` [PATCH 06/55] drm/i915: Merged the many do_execbuf() parameters into a structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 07/55] drm/i915: Simplify i915_gem_execbuffer_retire_commands() parameters John.C.Harrison
2015-05-29 16:43 ` [PATCH 08/55] drm/i915: Update alloc_request to return the allocated request John.C.Harrison
2015-05-29 16:43 ` [PATCH 09/55] drm/i915: Add request to execbuf params and add explicit cleanup John.C.Harrison
2015-05-29 16:43 ` [PATCH 10/55] drm/i915: Update the dispatch tracepoint to use params->request John.C.Harrison
2015-05-29 16:43 ` [PATCH 11/55] drm/i915: Update move_to_gpu() to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 12/55] drm/i915: Update execbuffer_move_to_active() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 13/55] drm/i915: Add flag to i915_add_request() to skip the cache flush John.C.Harrison
2015-06-02 18:19   ` Tomas Elf
2015-05-29 16:43 ` [PATCH 14/55] drm/i915: Update i915_gpu_idle() to manage its own request John.C.Harrison
2015-05-29 16:43 ` [PATCH 15/55] drm/i915: Split i915_ppgtt_init_hw() in half - generic and per ring John.C.Harrison
2015-06-18 12:11   ` John.C.Harrison
2015-05-29 16:43 ` [PATCH 16/55] drm/i915: Moved the for_each_ring loop outside of i915_gem_context_enable() John.C.Harrison
2015-05-29 16:43 ` [PATCH 17/55] drm/i915: Don't tag kernel batches as user batches John.C.Harrison
2015-05-29 16:43 ` [PATCH 18/55] drm/i915: Add explicit request management to i915_gem_init_hw() John.C.Harrison
2015-06-02 18:20   ` Tomas Elf
2015-05-29 16:43 ` [PATCH 19/55] drm/i915: Update ppgtt_init_ring() & context_enable() to take requests John.C.Harrison
2015-05-29 16:43 ` [PATCH 20/55] drm/i915: Update i915_switch_context() to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 21/55] drm/i915: Update do_switch() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 22/55] drm/i915: Update deferred context creation to do explicit request management John.C.Harrison
2015-06-02 18:22   ` Tomas Elf
2015-05-29 16:43 ` [PATCH 23/55] drm/i915: Update init_context() to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 24/55] drm/i915: Update render_state_init() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 25/55] drm/i915: Update i915_gem_object_sync() " John.C.Harrison
2015-06-02 18:26   ` Tomas Elf
2015-06-04 12:57     ` John Harrison
2015-06-18 12:14       ` John.C.Harrison
2015-06-18 12:21         ` Chris Wilson
2015-06-18 12:59           ` John Harrison
2015-06-18 14:24             ` Daniel Vetter
2015-06-18 15:39               ` Chris Wilson
2015-06-18 16:16                 ` John Harrison
2015-06-22 20:03                   ` Daniel Vetter
2015-06-22 20:14                     ` Chris Wilson
2015-06-18 16:36         ` 3.16 backlight kernel options Stéphane ANCELOT
2015-05-29 16:43 ` [PATCH 26/55] drm/i915: Update overlay code to do explicit request management John.C.Harrison
2015-05-29 16:43 ` [PATCH 27/55] drm/i915: Update queue_flip() to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 28/55] drm/i915: Update add_request() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 29/55] drm/i915: Update [vma|object]_move_to_active() to take request structures John.C.Harrison
2015-05-29 16:43 ` [PATCH 30/55] drm/i915: Update l3_remap to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 31/55] drm/i915: Update mi_set_context() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 32/55] drm/i915: Update a bunch of execbuffer helpers to take request structures John.C.Harrison
2015-05-29 16:43 ` [PATCH 33/55] drm/i915: Update workarounds_emit() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 34/55] drm/i915: Update flush_all_caches() " John.C.Harrison
2015-05-29 16:43 ` [PATCH 35/55] drm/i915: Update switch_mm() to take a request structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 36/55] drm/i915: Update ring->flush() to take a requests structure John.C.Harrison
2015-05-29 16:43 ` [PATCH 37/55] drm/i915: Update some flush helpers to take request structures John.C.Harrison
2015-05-29 16:43 ` [PATCH 38/55] drm/i915: Update ring->emit_flush() to take a request structure John.C.Harrison
2015-05-29 16:44 ` [PATCH 39/55] drm/i915: Update ring->add_request() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 40/55] drm/i915: Update ring->emit_request() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 41/55] drm/i915: Update ring->dispatch_execbuffer() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 42/55] drm/i915: Update ring->emit_bb_start() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 43/55] drm/i915: Update ring->sync_to() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 44/55] drm/i915: Update ring->signal() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 45/55] drm/i915: Update cacheline_align() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 46/55] drm/i915: Update intel_ring_begin() " John.C.Harrison
2015-06-23 10:24   ` Chris Wilson
2015-06-23 10:37     ` John Harrison
2015-06-23 13:25       ` Daniel Vetter
2015-06-23 15:27         ` John Harrison
2015-06-23 15:34           ` Daniel Vetter
2015-05-29 16:44 ` [PATCH 47/55] drm/i915: Update intel_logical_ring_begin() " John.C.Harrison
2015-05-29 16:44 ` [PATCH 48/55] drm/i915: Add *_ring_begin() to request allocation John.C.Harrison
2015-06-17 13:31   ` Daniel Vetter
2015-06-17 14:27     ` Chris Wilson
2015-06-17 14:54       ` Daniel Vetter
2015-06-17 15:52         ` Chris Wilson
2015-06-18 11:21           ` John Harrison
2015-06-18 13:29             ` Daniel Vetter
2015-06-19 16:34               ` John Harrison
2015-05-29 16:44 ` [PATCH 49/55] drm/i915: Remove the now obsolete intel_ring_get_request() John.C.Harrison
2015-05-29 16:44 ` [PATCH 50/55] drm/i915: Remove the now obsolete 'outstanding_lazy_request' John.C.Harrison
2015-05-29 16:44 ` [PATCH 51/55] drm/i915: Move the request/file and request/pid association to creation time John.C.Harrison
2015-06-03 11:15   ` Tomas Elf
2015-05-29 16:44 ` [PATCH 52/55] drm/i915: Remove 'faked' request from LRC submission John.C.Harrison
2015-05-29 16:44 ` [PATCH 53/55] drm/i915: Update a bunch of LRC functions to take requests John.C.Harrison
2015-05-29 16:44 ` [PATCH 54/55] drm/i915: Remove the now obsolete 'i915_gem_check_olr()' John.C.Harrison
2015-06-02 18:27   ` Tomas Elf
2015-06-23 10:23   ` Chris Wilson
2015-06-23 10:39     ` John Harrison
2015-05-29 16:44 ` [PATCH 55/55] drm/i915: Rename the somewhat reduced i915_gem_object_flush_active() John.C.Harrison
2015-06-02 18:27   ` Tomas Elf
2015-06-17 14:06   ` Daniel Vetter
2015-06-17 14:21     ` Chris Wilson
2015-06-18 11:03       ` John Harrison
2015-06-18 11:10         ` Chris Wilson
2015-06-18 11:27           ` John Harrison
2015-06-18 10:57     ` John Harrison
2015-06-04 18:23 ` [PATCH 14/56] drm/i915: Make retire condition check for requests not objects John.C.Harrison
2015-06-04 18:24   ` John Harrison
2015-06-09 15:56   ` Tomas Elf
2015-06-17 15:01     ` Daniel Vetter
2015-06-22 21:04 ` [PATCH 00/55] Remove the outstanding_lazy_request Daniel Vetter

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=5589393C.5080901@Intel.com \
    --to=john.c.harrison@intel.com \
    --cc=Intel-GFX@Lists.FreeDesktop.Org \
    --cc=chris@chris-wilson.co.uk \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.