From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, intel-gfx@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup
Date: Fri, 10 Jul 2020 12:27:25 +0100 [thread overview]
Message-ID: <0d4526e4-3e3b-a3f1-a014-7f08360b210f@linux.intel.com> (raw)
In-Reply-To: <20200706061926.6687-6-chris@chris-wilson.co.uk>
On 06/07/2020 07:19, Chris Wilson wrote:
> As a prelude to the next step where we want to perform all the object
> allocations together under the same lock, we first must delay the
> i915_vma_pin() as that implicitly does the allocations for us, one by
> one. As it only does the allocations one by one, it is not allowed to
> wait/evict, whereas pulling all the allocations together the entire set
> can be scheduled as one.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
> .../gpu/drm/i915/gem/i915_gem_execbuffer.c | 70 +++++++++++--------
> 1 file changed, 39 insertions(+), 31 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> index bf8193d9e279..35a57c1fc9c3 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> @@ -33,6 +33,8 @@ struct eb_vma {
>
> /** This vma's place in the execbuf reservation list */
> struct drm_i915_gem_exec_object2 *exec;
> +
> + struct list_head bind_link;
> struct list_head unbound_link;
> struct list_head reloc_link;
>
> @@ -240,8 +242,8 @@ struct i915_execbuffer {
> /** actual size of execobj[] as we may extend it for the cmdparser */
> unsigned int buffer_count;
>
> - /** list of vma not yet bound during reservation phase */
> - struct list_head unbound;
> + /** list of all vma required to bound for this execbuf */
> + struct list_head bind_list;
>
> /** list of vma that have execobj.relocation_count */
> struct list_head relocs;
> @@ -565,6 +567,8 @@ eb_add_vma(struct i915_execbuffer *eb,
> eb->lut_size)]);
> }
>
> + list_add_tail(&ev->bind_link, &eb->bind_list);
> +
> if (entry->relocation_count)
> list_add_tail(&ev->reloc_link, &eb->relocs);
>
> @@ -586,16 +590,6 @@ eb_add_vma(struct i915_execbuffer *eb,
>
> eb->batch = ev;
> }
> -
> - if (eb_pin_vma(eb, entry, ev)) {
> - if (entry->offset != vma->node.start) {
> - entry->offset = vma->node.start | UPDATE;
> - eb->args->flags |= __EXEC_HAS_RELOC;
> - }
> - } else {
> - eb_unreserve_vma(ev);
> - list_add_tail(&ev->unbound_link, &eb->unbound);
> - }
> }
>
> static int eb_reserve_vma(const struct i915_execbuffer *eb,
> @@ -670,13 +664,31 @@ static int wait_for_timeline(struct intel_timeline *tl)
> } while (1);
> }
>
> -static int eb_reserve(struct i915_execbuffer *eb)
> +static int eb_reserve_vm(struct i915_execbuffer *eb)
> {
> - const unsigned int count = eb->buffer_count;
> unsigned int pin_flags = PIN_USER | PIN_NONBLOCK;
> - struct list_head last;
> + struct list_head last, unbound;
> struct eb_vma *ev;
> - unsigned int i, pass;
> + unsigned int pass;
> +
> + INIT_LIST_HEAD(&unbound);
> + list_for_each_entry(ev, &eb->bind_list, bind_link) {
> + struct drm_i915_gem_exec_object2 *entry = ev->exec;
> + struct i915_vma *vma = ev->vma;
> +
> + if (eb_pin_vma(eb, entry, ev)) {
> + if (entry->offset != vma->node.start) {
> + entry->offset = vma->node.start | UPDATE;
> + eb->args->flags |= __EXEC_HAS_RELOC;
> + }
> + } else {
> + eb_unreserve_vma(ev);
> + list_add_tail(&ev->unbound_link, &unbound);
> + }
> + }
> +
> + if (list_empty(&unbound))
> + return 0;
>
> /*
> * Attempt to pin all of the buffers into the GTT.
> @@ -699,7 +711,7 @@ static int eb_reserve(struct i915_execbuffer *eb)
> if (mutex_lock_interruptible(&eb->i915->drm.struct_mutex))
> return -EINTR;
>
> - list_for_each_entry(ev, &eb->unbound, unbound_link) {
> + list_for_each_entry(ev, &unbound, unbound_link) {
> err = eb_reserve_vma(eb, ev, pin_flags);
> if (err)
> break;
> @@ -710,13 +722,11 @@ static int eb_reserve(struct i915_execbuffer *eb)
> }
>
> /* Resort *all* the objects into priority order */
> - INIT_LIST_HEAD(&eb->unbound);
> + INIT_LIST_HEAD(&unbound);
> INIT_LIST_HEAD(&last);
> - for (i = 0; i < count; i++) {
> - unsigned int flags;
> + list_for_each_entry(ev, &eb->bind_list, bind_link) {
> + unsigned int flags = ev->flags;
>
> - ev = &eb->vma[i];
> - flags = ev->flags;
> if (flags & EXEC_OBJECT_PINNED &&
> flags & __EXEC_OBJECT_HAS_PIN)
> continue;
> @@ -725,17 +735,17 @@ static int eb_reserve(struct i915_execbuffer *eb)
>
> if (flags & EXEC_OBJECT_PINNED)
> /* Pinned must have their slot */
> - list_add(&ev->unbound_link, &eb->unbound);
> + list_add(&ev->unbound_link, &unbound);
> else if (flags & __EXEC_OBJECT_NEEDS_MAP)
> /* Map require the lowest 256MiB (aperture) */
> - list_add_tail(&ev->unbound_link, &eb->unbound);
> + list_add_tail(&ev->unbound_link, &unbound);
> else if (!(flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS))
> /* Prioritise 4GiB region for restricted bo */
> list_add(&ev->unbound_link, &last);
> else
> list_add_tail(&ev->unbound_link, &last);
> }
> - list_splice_tail(&last, &eb->unbound);
> + list_splice_tail(&last, &unbound);
> mutex_unlock(&eb->i915->drm.struct_mutex);
>
> if (err == -EAGAIN) {
> @@ -891,8 +901,8 @@ static int eb_lookup_vmas(struct i915_execbuffer *eb)
> unsigned int i;
> int err = 0;
>
> + INIT_LIST_HEAD(&eb->bind_list);
> INIT_LIST_HEAD(&eb->relocs);
> - INIT_LIST_HEAD(&eb->unbound);
>
> for (i = 0; i < eb->buffer_count; i++) {
> struct i915_vma *vma;
> @@ -1539,11 +1549,9 @@ static int eb_relocate(struct i915_execbuffer *eb)
> if (err)
> return err;
>
> - if (!list_empty(&eb->unbound)) {
> - err = eb_reserve(eb);
> - if (err)
> - return err;
> - }
> + err = eb_reserve_vm(eb);
> + if (err)
> + return err;
>
> /* The objects are in their final locations, apply the relocations. */
> if (eb->args->flags & __EXEC_HAS_RELOC) {
>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2020-07-10 11:27 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-06 6:19 [Intel-gfx] s/obj->mm.lock// Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories Chris Wilson
2020-07-06 18:15 ` Matthew Auld
2020-07-06 18:20 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 02/20] drm/i915: Switch to object allocations for page directories Chris Wilson
2020-07-06 19:06 ` Matthew Auld
2020-07-06 19:31 ` Chris Wilson
2020-07-06 20:01 ` Chris Wilson
2020-07-06 21:08 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf Chris Wilson
2020-07-08 16:54 ` Tvrtko Ursulin
2020-07-08 18:08 ` Chris Wilson
2020-07-09 10:52 ` Tvrtko Ursulin
2020-07-09 10:57 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link Chris Wilson
2020-07-10 11:26 ` Tvrtko Ursulin
2020-07-06 6:19 ` [Intel-gfx] [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup Chris Wilson
2020-07-10 11:27 ` Tvrtko Ursulin [this message]
2020-07-06 6:19 ` [Intel-gfx] [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 08/20] drm/i915: Always defer fenced work to the worker Chris Wilson
2020-07-08 12:18 ` Tvrtko Ursulin
2020-07-08 12:25 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 09/20] drm/i915/gem: Assign context id for async work Chris Wilson
2020-07-08 12:26 ` Tvrtko Ursulin
2020-07-08 12:42 ` Chris Wilson
2020-07-08 14:24 ` Tvrtko Ursulin
2020-07-08 15:36 ` Chris Wilson
2020-07-09 11:01 ` Tvrtko Ursulin
2020-07-09 11:07 ` Chris Wilson
2020-07-09 11:59 ` Tvrtko Ursulin
2020-07-09 12:07 ` Chris Wilson
2020-07-13 12:22 ` Tvrtko Ursulin
2020-07-14 14:01 ` Chris Wilson
2020-07-08 12:45 ` Tvrtko Ursulin
2020-07-06 6:19 ` [Intel-gfx] [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire() Chris Wilson
2020-07-09 14:36 ` Maarten Lankhorst
2020-07-10 12:24 ` Tvrtko Ursulin
2020-07-10 12:32 ` Maarten Lankhorst
2020-07-13 14:29 ` Tvrtko Ursulin
2020-07-06 6:19 ` [Intel-gfx] [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list Chris Wilson
2020-07-13 14:53 ` Tvrtko Ursulin
2020-07-14 14:10 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding Chris Wilson
2020-07-14 9:02 ` Tvrtko Ursulin
2020-07-14 15:05 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf Chris Wilson
2020-07-14 12:19 ` Tvrtko Ursulin
2020-07-14 15:21 ` Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning Chris Wilson
2020-07-14 12:48 ` Tvrtko Ursulin
2020-07-06 6:19 ` [Intel-gfx] [PATCH 15/20] drm/i915/gem: Include secure batch " Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing Chris Wilson
2020-07-09 15:39 ` Tvrtko Ursulin
2020-07-06 6:19 ` [Intel-gfx] [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2 Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section Chris Wilson
2020-07-06 6:19 ` [Intel-gfx] [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class Chris Wilson
2020-07-09 14:06 ` Maarten Lankhorst
2020-07-06 6:19 ` [Intel-gfx] [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter Chris Wilson
2020-07-06 6:28 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories Patchwork
2020-07-06 6:29 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-07-06 6:51 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2020-07-06 7:55 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " Patchwork
2020-07-27 18:53 ` [Intel-gfx] s/obj->mm.lock// Thomas Hellström (Intel)
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=0d4526e4-3e3b-a3f1-a014-7f08360b210f@linux.intel.com \
--to=tvrtko.ursulin@linux.intel.com \
--cc=chris@chris-wilson.co.uk \
--cc=intel-gfx@lists.freedesktop.org \
/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