linux-kernel - [30/89] drm/i915: paper over missed irq issues with force wake voodoo

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 01 Feb 2012 12:59:54 -0800
From:	Greg KH <gregkh@...uxfoundation.org>
To:	linux-kernel@...r.kernel.org, stable@...r.kernel.org
Cc:	torvalds@...ux-foundation.org, akpm@...ux-foundation.org,
	alan@...rguk.ukuu.org.uk, Eric Anholt <eric@...olt.net>,
	Kenneth Graunke <kenneth@...tecape.org>,
	Eugeni Dodonov <eugeni.dodonov@...el.com>,
	Daniel Vetter <daniel.vetter@...ll.ch>,
	Keith Packard <keithp@...thp.com>
Subject: [30/89] drm/i915: paper over missed irq issues with force wake voodoo

3.2-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Daniel Vetter <daniel.vetter@...ll.ch>

commit 4cd53c0c8b01fc05c3ad5b2acdad02e37d3c2f55 upstream.

Two things seem to do the trick on my ivb machine here:
- prevent the gt from powering down while waiting for seqno
  notification interrupts by grabbing the force_wake in get_irq (and
  dropping it in put_irq again).
- ordering writes from the ring's CS by reading a CS register, ACTHD
  seems to work.

Only the blt&bsd ring on ivb seem to be massively affected by this,
but for paranoia do this dance also on the render ring and on snb
(i.e. all gpus with forcewake).

Tested with Eric's glCopyPixels loop which without this patch scores a
missed irq every few seconds.

This patch needs my forcewake rework to use a spinlock instead of
dev->struct_mutex.

After crawling through docs a lot I've found the following nugget:

Internal doc "SNB GT PM Programming Guide", Section 4.3.1:

"GT does not generate interrupts while in RC6 (by design)"

So it looks like rc6 and irq generation are indeed related.

v2: Improve the comment per Eugeni Dodonov's suggestion.

v3: Add the documentation snipped. Also restrict the w/a to ivb only
for -fixes, as suggested by Keith Packard.

Cc: Eric Anholt <eric@...olt.net>
Cc: Kenneth Graunke <kenneth@...tecape.org>
Cc: Eugeni Dodonov <eugeni.dodonov@...el.com>
Tested-by: Eugeni Dodonov <eugeni.dodonov@...el.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@...el.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@...ll.ch>
Signed-off-by: Keith Packard <keithp@...thp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>

---
 drivers/gpu/drm/i915/intel_ringbuffer.c |   27 +++++++++++++++++++++++++--
 1 file changed, 25 insertions(+), 2 deletions(-)

--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -631,6 +631,19 @@ render_ring_add_request(struct intel_rin
 }
 
 static u32
+gen6_ring_get_seqno(struct intel_ring_buffer *ring)
+{
+	struct drm_device *dev = ring->dev;
+
+	/* Workaround to force correct ordering between irq and seqno writes on
+	 * ivb (and maybe also on snb) by reading from a CS register (like
+	 * ACTHD) before reading the status page. */
+	if (IS_GEN7(dev))
+		intel_ring_get_active_head(ring);
+	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
+}
+
+static u32
 ring_get_seqno(struct intel_ring_buffer *ring)
 {
 	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
@@ -795,6 +808,12 @@ gen6_ring_get_irq(struct intel_ring_buff
 	if (!dev->irq_enabled)
 	       return false;
 
+	/* It looks like we need to prevent the gt from suspending while waiting
+	 * for an notifiy irq, otherwise irqs seem to get lost on at least the
+	 * blt/bsd rings on ivb. */
+	if (IS_GEN7(dev))
+		gen6_gt_force_wake_get(dev_priv);
+
 	spin_lock(&ring->irq_lock);
 	if (ring->irq_refcount++ == 0) {
 		ring->irq_mask &= ~rflag;
@@ -819,6 +838,9 @@ gen6_ring_put_irq(struct intel_ring_buff
 		ironlake_disable_irq(dev_priv, gflag);
 	}
 	spin_unlock(&ring->irq_lock);
+
+	if (IS_GEN7(dev))
+		gen6_gt_force_wake_put(dev_priv);
 }
 
 static bool
@@ -1316,7 +1338,7 @@ static const struct intel_ring_buffer ge
 	.write_tail		= gen6_bsd_ring_write_tail,
 	.flush			= gen6_ring_flush,
 	.add_request		= gen6_add_request,
-	.get_seqno		= ring_get_seqno,
+	.get_seqno		= gen6_ring_get_seqno,
 	.irq_get		= gen6_bsd_ring_get_irq,
 	.irq_put		= gen6_bsd_ring_put_irq,
 	.dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
@@ -1451,7 +1473,7 @@ static const struct intel_ring_buffer ge
 	.write_tail		= ring_write_tail,
 	.flush			= blt_ring_flush,
 	.add_request		= gen6_add_request,
-	.get_seqno		= ring_get_seqno,
+	.get_seqno		= gen6_ring_get_seqno,
 	.irq_get		= blt_ring_get_irq,
 	.irq_put		= blt_ring_put_irq,
 	.dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
@@ -1474,6 +1496,7 @@ int intel_init_render_ring_buffer(struct
 		ring->flush = gen6_render_ring_flush;
 		ring->irq_get = gen6_render_ring_get_irq;
 		ring->irq_put = gen6_render_ring_put_irq;
+		ring->get_seqno = gen6_ring_get_seqno;
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->get_seqno = pc_render_get_seqno;


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/