public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 8/9] drm/i915/gtt: Make swapping the pd entry generic
Date: Tue, 11 Jun 2019 20:27:30 +0300	[thread overview]
Message-ID: <20190611172731.19174-8-mika.kuoppala@linux.intel.com> (raw)
In-Reply-To: <20190611172731.19174-1-mika.kuoppala@linux.intel.com>

Swapping a pd entry is same across the page directories, if
we succeed we need to increment the count and write the phys page
entry. Make a common function for it.

Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 41 +++++++++++++++++++----------
 1 file changed, 27 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f1d7874834e2..9b0d0a077e31 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -752,6 +752,27 @@ static void __set_pd_entry(struct i915_page_directory * const pd,
 		       gen8_pde_encode(px_dma(to), I915_CACHE_LLC));	\
 })
 
+static void *__swap_pd_entry(struct i915_page_directory * const pd,
+			     const unsigned short pde,
+			     void * const old_val,
+			     void * const new_val,
+			     const u64 encoded_entry)
+{
+	void * const old = cmpxchg(&pd->entry[pde], old_val, new_val);
+
+	if (likely(old == old_val)) {
+		atomic_inc(&pd->used);
+		if (likely(pd_has_phys_page(pd)))
+			__set_pd_entry(pd, pde, encoded_entry);
+	}
+
+	return old;
+}
+
+#define swap_pd_entry(pd, pde, old, to) \
+	__swap_pd_entry((pd), (pde), (old), (to), \
+			gen8_pde_encode(px_dma(to), I915_CACHE_LLC))
+
 /*
  * PDE TLBs are a pain to invalidate on GEN8+. When we modify
  * the page table structures, we mark them dirty so that
@@ -1326,11 +1347,8 @@ static int gen8_ppgtt_alloc_pd(struct i915_address_space *vm,
 			if (count < GEN8_PTES || intel_vgpu_active(vm->i915))
 				gen8_initialize_pt(vm, pt);
 
-			old = cmpxchg(&pd->entry[pde], vm->scratch_pt, pt);
-			if (old == vm->scratch_pt) {
-				set_pd_entry(pd, pde, pt);
-				atomic_inc(&pd->used);
-			} else {
+			old = swap_pd_entry(pd, pde, vm->scratch_pt, pt);
+			if (unlikely(old != vm->scratch_pt)) {
 				free_pt(vm, pt);
 				pt = old;
 			}
@@ -1371,11 +1389,8 @@ static int gen8_ppgtt_alloc_pdp(struct i915_address_space *vm,
 
 			init_pd(vm, pd, vm->scratch_pt);
 
-			old = cmpxchg(&pdp->entry[pdpe], vm->scratch_pd, pd);
-			if (old == vm->scratch_pd) {
-				set_pd_entry(pdp, pdpe, pd);
-				atomic_inc(&pdp->used);
-			} else {
+			old = swap_pd_entry(pdp, pdpe, vm->scratch_pd, pd);
+			if (unlikely(old != vm->scratch_pd)) {
 				free_pd(vm, pd);
 				pd = old;
 			}
@@ -1440,10 +1455,8 @@ static int gen8_ppgtt_alloc_4lvl(struct i915_address_space *vm,
 
 			init_pd(vm, pdp, vm->scratch_pd);
 
-			old = cmpxchg(&pml4->entry[pml4e], vm->scratch_pdp, pdp);
-			if (old == vm->scratch_pdp) {
-				set_pd_entry(pml4, pml4e, pdp);
-			} else {
+			old = swap_pd_entry(pml4, pml4e, vm->scratch_pdp, pdp);
+			if (unlikely(old != vm->scratch_pdp)) {
 				free_pd(vm, pdp);
 				pdp = old;
 			}
-- 
2.17.1

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

  parent reply	other threads:[~2019-06-11 17:27 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-11 17:27 [PATCH 1/9] drm/i915/gtt: No need to zero the table for page dirs Mika Kuoppala
2019-06-11 17:27 ` [PATCH 2/9] drm/i915/gtt: Use a common type for page directories Mika Kuoppala
2019-06-11 19:41   ` Chris Wilson
2019-06-11 19:52     ` Chris Wilson
2019-06-11 19:43   ` Chris Wilson
2019-06-11 17:27 ` [PATCH 3/9] drm/i915/gtt: Introduce init_pd_with_page Mika Kuoppala
2019-06-11 19:45   ` Chris Wilson
2019-06-11 17:27 ` [PATCH 4/9] drm/i915/gtt: Introduce init_pd Mika Kuoppala
2019-06-11 17:27 ` [PATCH 5/9] drm/i915/gtt: Generalize alloc_pd Mika Kuoppala
2019-06-11 17:27 ` [PATCH 6/9] drm/i915/gtt: pde entry encoding is identical Mika Kuoppala
2019-06-11 17:27 ` [PATCH 7/9] drm/i915/gtt: Check for physical page for pd entry always Mika Kuoppala
2019-06-11 17:27 ` Mika Kuoppala [this message]
2019-06-11 19:50   ` [PATCH 8/9] drm/i915/gtt: Make swapping the pd entry generic Chris Wilson
2019-06-11 21:46     ` Chris Wilson
2019-06-11 17:27 ` [PATCH 9/9] drm/i915/gtt: Tear down setup and cleanup macros for page dma Mika Kuoppala
2019-06-11 17:53 ` ✗ Fi.CI.CHECKPATCH: warning for series starting with [1/9] drm/i915/gtt: No need to zero the table for page dirs Patchwork
2019-06-11 17:57 ` ✗ Fi.CI.SPARSE: " Patchwork
2019-06-11 18:18 ` ✓ Fi.CI.BAT: success " Patchwork
2019-06-11 19:37 ` [PATCH 1/9] " Chris Wilson
2019-06-12 21:32 ` ✗ Fi.CI.IGT: failure for series starting with [1/9] " 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=20190611172731.19174-8-mika.kuoppala@linux.intel.com \
    --to=mika.kuoppala@linux.intel.com \
    --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