drm/i915: Don't allow overuse of __intel_wait_for_register_fw()
authorMichal Wajdeczko <michal.wajdeczko@intel.com>
Mon, 10 Apr 2017 12:17:47 +0000 (12:17 +0000)
committerChris Wilson <chris@chris-wilson.co.uk>
Mon, 10 Apr 2017 14:24:52 +0000 (15:24 +0100)
This function should not be called with long timeouts in atomic context.
Annotate it as might_sleep if timeout is longer than 10us.

v2: fix comment (Michal)

Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170410121747.209200-1-michal.wajdeczko@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drivers/gpu/drm/i915/intel_uncore.c

index 1deb1a4342baa618e34add4a62d4d5f4e9ea1bc4..eb38392a2435d079d7b92d3756bc62bb01739ac7 100644 (file)
@@ -1600,6 +1600,8 @@ static int gen6_reset_engines(struct drm_i915_private *dev_priv,
  *     (I915_READ_FW(reg) & mask) == value
  *
  * Otherwise, the wait will timeout after @slow_timeout_ms milliseconds.
+ * For atomic context @slow_timeout_ms must be zero and @fast_timeout_us
+ * must be not larger than 10 microseconds.
  *
  * Note that this routine assumes the caller holds forcewake asserted, it is
  * not suitable for very long waits. See intel_wait_for_register() if you
@@ -1620,6 +1622,9 @@ int __intel_wait_for_register_fw(struct drm_i915_private *dev_priv,
 #define done (((reg_value = I915_READ_FW(reg)) & mask) == value)
        int ret;
 
+       /* Catch any overuse of this function */
+       might_sleep_if(fast_timeout_us > 10 || slow_timeout_ms);
+
        if (fast_timeout_us > 10)
                ret = _wait_for(done, fast_timeout_us, 10);
        else