From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Subject: [Intel-gfx] [PATCH v5 26/63] drm/i915: Make intel_init_workaround_bb more compatible with ww locking.
Date: Wed, 25 Nov 2020 11:39:34 +0100 [thread overview]
Message-ID: <20201125104011.606641-27-maarten.lankhorst@linux.intel.com> (raw)
In-Reply-To: <20201125104011.606641-1-maarten.lankhorst@linux.intel.com>
Make creation separate from pinning, in order to take the lock only
once, and pin the mapping with the lock held.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
drivers/gpu/drm/i915/gt/intel_lrc.c | 43 ++++++++++++++++++++++-------
1 file changed, 33 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 821f419d3108..011b9fa090e3 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -3982,7 +3982,7 @@ gen10_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
#define CTX_WA_BB_OBJ_SIZE (PAGE_SIZE)
-static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
+static int lrc_init_wa_ctx(struct intel_engine_cs *engine)
{
struct drm_i915_gem_object *obj;
struct i915_vma *vma;
@@ -3998,10 +3998,6 @@ static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
goto err;
}
- err = i915_ggtt_pin(vma, NULL, 0, PIN_HIGH);
- if (err)
- goto err;
-
engine->wa_ctx.vma = vma;
return 0;
@@ -4010,9 +4006,16 @@ static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
return err;
}
-static void lrc_destroy_wa_ctx(struct intel_engine_cs *engine)
+static void lrc_destroy_wa_ctx(struct intel_engine_cs *engine, bool unpin)
{
- i915_vma_unpin_and_release(&engine->wa_ctx.vma, 0);
+ if (!engine->wa_ctx.vma)
+ return;
+
+ if (unpin)
+ i915_vma_unpin(engine->wa_ctx.vma);
+
+ i915_vma_put(engine->wa_ctx.vma);
+ engine->wa_ctx.vma = NULL;
}
typedef u32 *(*wa_bb_func_t)(struct intel_engine_cs *engine, u32 *batch);
@@ -4024,6 +4027,7 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
&wa_ctx->per_ctx };
wa_bb_func_t wa_bb_fn[2];
void *batch, *batch_ptr;
+ struct i915_gem_ww_ctx ww;
unsigned int i;
int ret;
@@ -4051,13 +4055,21 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
return 0;
}
- ret = lrc_setup_wa_ctx(engine);
+ ret = lrc_init_wa_ctx(engine);
if (ret) {
drm_dbg(&engine->i915->drm,
"Failed to setup context WA page: %d\n", ret);
return ret;
}
+ i915_gem_ww_ctx_init(&ww, true);
+retry:
+ ret = i915_gem_object_lock(wa_ctx->vma->obj, &ww);
+ if (!ret)
+ ret = i915_ggtt_pin(wa_ctx->vma, &ww, 0, PIN_HIGH);
+ if (ret)
+ goto err;
+
batch = i915_gem_object_pin_map(wa_ctx->vma->obj, I915_MAP_WB);
/*
@@ -4081,8 +4093,19 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
__i915_gem_object_flush_map(wa_ctx->vma->obj, 0, batch_ptr - batch);
__i915_gem_object_release_map(wa_ctx->vma->obj);
+
+ if (ret)
+ i915_vma_unpin(wa_ctx->vma);
+
+err:
+ if (ret == -EDEADLK) {
+ ret = i915_gem_ww_ctx_backoff(&ww);
+ if (!ret)
+ goto retry;
+ }
+ i915_gem_ww_ctx_fini(&ww);
if (ret)
- lrc_destroy_wa_ctx(engine);
+ lrc_destroy_wa_ctx(engine, false);
return ret;
}
@@ -5159,7 +5182,7 @@ static void execlists_release(struct intel_engine_cs *engine)
execlists_shutdown(engine);
intel_engine_cleanup_common(engine);
- lrc_destroy_wa_ctx(engine);
+ lrc_destroy_wa_ctx(engine, true);
}
static void
--
2.29.2.222.g5d2a92d10f8
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2020-11-25 10:41 UTC|newest]
Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-25 10:39 [Intel-gfx] [PATCH v5 00/63] drm/i915: Remove obj->mm.lock! Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 01/63] drm/i915: Do not share hwsp across contexts any more, v6 Maarten Lankhorst
2020-11-25 10:44 ` Chris Wilson
2020-11-27 10:06 ` Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 02/63] drm/i915: Pin timeline map after first timeline pin, v3 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 03/63] drm/i915: Move cmd parser pinning to execbuffer Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 04/63] drm/i915: Add missing -EDEADLK handling to execbuf pinning, v2 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 05/63] drm/i915: Ensure we hold the object mutex in pin correctly Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 06/63] drm/i915: Add gem object locking to madvise Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 07/63] drm/i915: Move HAS_STRUCT_PAGE to obj->flags Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 08/63] drm/i915: Rework struct phys attachment handling Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 09/63] drm/i915: Convert i915_gem_object_attach_phys() to ww locking, v2 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 10/63] drm/i915: make lockdep slightly happier about execbuf Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 11/63] drm/i915: Disable userptr pread/pwrite support Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 12/63] drm/i915: No longer allow exporting userptr through dma-buf Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 13/63] drm/i915: Reject more ioctls for userptr Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 14/63] drm/i915: Reject UNSYNCHRONIZED for userptr, v2 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 15/63] drm/i915: Make compilation of userptr code depend on MMU_NOTIFIER Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 16/63] drm/i915: Fix userptr so we do not have to worry about obj->mm.lock, v5 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 17/63] drm/i915: Flatten obj->mm.lock Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 18/63] drm/i915: Populate logical context during first pin Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 19/63] drm/i915: Make ring submission compatible with obj->mm.lock removal, v2 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 20/63] drm/i915: Handle ww locking in init_status_page Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 21/63] drm/i915: Rework clflush to work correctly without obj->mm.lock Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 22/63] drm/i915: Pass ww ctx to intel_pin_to_display_plane Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 23/63] drm/i915: Add object locking to vm_fault_cpu Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 24/63] drm/i915: Move pinning to inside engine_wa_list_verify() Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 25/63] drm/i915: Take reservation lock around i915_vma_pin Maarten Lankhorst
2020-11-25 10:39 ` Maarten Lankhorst [this message]
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 27/63] drm/i915: Make __engine_unpark() compatible with ww locking Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 28/63] drm/i915: Take obj lock around set_domain ioctl Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 29/63] drm/i915: Defer pin calls in buffer pool until first use by caller Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 30/63] drm/i915: Fix pread/pwrite to work with new locking rules Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 31/63] drm/i915: Fix workarounds selftest, part 1 Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 32/63] drm/i915: Prepare for obj->mm.lock removal Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 33/63] drm/i915: Add igt_spinner_pin() to allow for ww locking around spinner Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 34/63] drm/i915: Add ww locking around vm_access() Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 35/63] drm/i915: Increase ww locking for perf Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 36/63] drm/i915: Lock ww in ucode objects correctly Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 37/63] drm/i915: Add ww locking to dma-buf ops Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 38/63] drm/i915: Add missing ww lock in intel_dsb_prepare Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 39/63] drm/i915: Fix ww locking in shmem_create_from_object Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 40/63] drm/i915: Use a single page table lock for each gtt Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 41/63] drm/i915/selftests: Prepare huge_pages testcases for obj->mm.lock removal Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 42/63] drm/i915/selftests: Prepare client blit " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 43/63] drm/i915/selftests: Prepare coherency tests " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 44/63] drm/i915/selftests: Prepare context " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 45/63] drm/i915/selftests: Prepare dma-buf " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 46/63] drm/i915/selftests: Prepare execbuf " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 47/63] drm/i915/selftests: Prepare mman testcases " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 48/63] drm/i915/selftests: Prepare object tests " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 49/63] drm/i915/selftests: Prepare object blit " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 50/63] drm/i915/selftests: Prepare igt_gem_utils " Maarten Lankhorst
2020-11-25 10:39 ` [Intel-gfx] [PATCH v5 51/63] drm/i915/selftests: Prepare context selftest " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 52/63] drm/i915/selftests: Prepare hangcheck " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 53/63] drm/i915/selftests: Prepare execlists " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 54/63] drm/i915/selftests: Prepare mocs tests " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 55/63] drm/i915/selftests: Prepare ring submission " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 56/63] drm/i915/selftests: Prepare timeline tests " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 57/63] drm/i915/selftests: Prepare i915_request " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 58/63] drm/i915/selftests: Prepare memory region " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 59/63] drm/i915/selftests: Prepare cs engine " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 60/63] drm/i915/selftests: Prepare gtt " Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 61/63] drm/i915: Finally remove obj->mm.lock Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 62/63] drm/i915: Keep userpointer bindings if seqcount is unchanged, v2 Maarten Lankhorst
2020-11-25 10:40 ` [Intel-gfx] [PATCH v5 63/63] drm/i915: Move gt_revoke() slightly Maarten Lankhorst
2020-11-25 12:30 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915: Remove obj->mm.lock! (rev11) Patchwork
2020-11-25 12:31 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-11-25 12:34 ` [Intel-gfx] ✗ Fi.CI.DOCS: " Patchwork
2020-11-25 13:00 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2020-11-25 14:49 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " Patchwork
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=20201125104011.606641-27-maarten.lankhorst@linux.intel.com \
--to=maarten.lankhorst@linux.intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=thomas.hellstrom@linux.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