[PATCH v2] drm/i915: Always normalize return timeout for wait_timeout_ioctl

[PATCH v2] drm/i915: Always normalize return timeout for wait_timeout_ioctl

[ville.syrjala]

From: Chris Wilson <chris@chris-wilson.co.uk>

As we recompute the remaining timeout after waiting, there is a
potential for that timeout to be less than zero and so need sanitizing.
The timeout is always returned to userspace and validated, so we should
always perform the sanitation.

v2 [vsyrjala]: Only normalize the timespec if it's invalid

Cc: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 6be940e..a469298 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1045,6 +1045,8 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 	if (timeout) {
 		struct timespec sleep_time = timespec_sub(now, before);
 		*timeout = timespec_sub(*timeout, sleep_time);
+		if (!timespec_valid(timeout))
+			set_normalized_timespec(timeout, 0, 0);
 	}
 
 	switch (end) {
@@ -1053,8 +1055,6 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 	case -ERESTARTSYS: /* Signal */
 		return (int)end;
 	case 0: /* Timeout */
-		if (timeout)
-			set_normalized_timespec(timeout, 0, 0);
 		return -ETIME;
 	default: /* Completed */
 		WARN_ON(end < 0); /* We're not aware of other errors */
-- 
1.8.1.5

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

[PATCH v3] drm/i915: Always normalize return timeout for wait_timeout_ioctl

[ville.syrjala]

From: Chris Wilson <chris@chris-wilson.co.uk>

As we recompute the remaining timeout after waiting, there is a
potential for that timeout to be less than zero and so need sanitizing.
The timeout is always returned to userspace and validated, so we should
always perform the sanitation.

v2 [vsyrjala]: Only normalize the timespec if it's invalid
v3: Add a comment to clarify the situation and remove the now
    useless WARN_ON() (ickle)

Cc: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 6be940e..6165535 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1045,6 +1045,8 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 	if (timeout) {
 		struct timespec sleep_time = timespec_sub(now, before);
 		*timeout = timespec_sub(*timeout, sleep_time);
+		if (!timespec_valid(timeout)) /* i.e. negative time remains */
+			set_normalized_timespec(timeout, 0, 0);
 	}
 
 	switch (end) {
@@ -1053,8 +1055,6 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 	case -ERESTARTSYS: /* Signal */
 		return (int)end;
 	case 0: /* Timeout */
-		if (timeout)
-			set_normalized_timespec(timeout, 0, 0);
 		return -ETIME;
 	default: /* Completed */
 		WARN_ON(end < 0); /* We're not aware of other errors */
@@ -2377,10 +2377,8 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	mutex_unlock(&dev->struct_mutex);
 
 	ret = __wait_seqno(ring, seqno, reset_counter, true, timeout);
-	if (timeout) {
-		WARN_ON(!timespec_valid(timeout));
+	if (timeout)
 		args->timeout_ns = timespec_to_ns(timeout);
-	}
 	return ret;
 
 out:
-- 
1.8.1.5

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

Re: [PATCH v3] drm/i915: Always normalize return timeout for wait_timeout_ioctl

[Ben Widawsky]

On Fri, Apr 26, 2013 at 04:22:46PM +0300, ville.syrjala@linux.intel.com wrote:
> From: Chris Wilson <chris@chris-wilson.co.uk>
> 
> As we recompute the remaining timeout after waiting, there is a
> potential for that timeout to be less than zero and so need sanitizing.
> The timeout is always returned to userspace and validated, so we should
> always perform the sanitation.
> 
> v2 [vsyrjala]: Only normalize the timespec if it's invalid
> v3: Add a comment to clarify the situation and remove the now
>     useless WARN_ON() (ickle)
> 
> Cc: Ben Widawsky <ben@bwidawsk.net>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

This does an excellent job of removing the warnings I'd sometimes see.

Unfortunately I've confirmed I can still reproduce the hang, and now
with no errors to leave any clues. I can't promise, but I'm 75%ish
certain I'd not always see a WARN before too (in fact I think that was
the case when Chris wrote the original patch)

The frequency with which I see the hangs seem to be increasing as well.
Does nobody else see this? I will have to take a shot at getting more
info.

[snip]

-- 
Ben Widawsky, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH v3] drm/i915: Always normalize return timeout for wait_timeout_ioctl

[Daniel Vetter]

On Sat, Apr 27, 2013 at 12:51:33PM -0700, Ben Widawsky wrote:
> On Fri, Apr 26, 2013 at 04:22:46PM +0300, ville.syrjala@linux.intel.com wrote:
> > From: Chris Wilson <chris@chris-wilson.co.uk>
> > 
> > As we recompute the remaining timeout after waiting, there is a
> > potential for that timeout to be less than zero and so need sanitizing.
> > The timeout is always returned to userspace and validated, so we should
> > always perform the sanitation.
> > 
> > v2 [vsyrjala]: Only normalize the timespec if it's invalid
> > v3: Add a comment to clarify the situation and remove the now
> >     useless WARN_ON() (ickle)
> > 
> > Cc: Ben Widawsky <ben@bwidawsk.net>
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> This does an excellent job of removing the warnings I'd sometimes see.
> 
> Unfortunately I've confirmed I can still reproduce the hang, and now
> with no errors to leave any clues. I can't promise, but I'm 75%ish
> certain I'd not always see a WARN before too (in fact I think that was
> the case when Chris wrote the original patch)
> 
> The frequency with which I see the hangs seem to be increasing as well.
> Does nobody else see this? I will have to take a shot at getting more
> info.

Picked up for -fixes, thanks for the patch.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch