drm: give up on edid retries when i2c bus is not responding
authorEugeni Dodonov <eugeni.dodonov@intel.com>
Thu, 5 Jan 2012 11:34:28 +0000 (09:34 -0200)
committerBen Hutchings <ben@decadent.org.uk>
Thu, 3 Jan 2013 03:32:44 +0000 (03:32 +0000)
commit 9292f37e1f5c79400254dca46f83313488093825 upstream.

This allows to avoid talking to a non-responding bus repeatedly until we
finally timeout after 15 attempts. We can do this by catching the -ENXIO
error, provided by i2c_algo_bit:bit_doAddress call.

Within the bit_doAddress we already try 3 times to get the edid data, so
if the routine tells us that bus is not responding, it is mostly pointless
to keep re-trying those attempts over and over again until we reach final
number of retries.

This change should fix https://bugs.freedesktop.org/show_bug.cgi?id=41059
and improve overall edid detection timing by 10-30% in most cases, and by
a much larger margin in case of phantom outputs (up to 30x in one worst
case).

Timing results for i915-powered machines for 'time xrandr' command:
Machine 1: from 0.840s to 0.290s
Machine 2: from 0.315s to 0.280s
Machine 3: from +/- 4s to 0.184s

Timing results for HD5770 with 'time xrandr' command:
Machine 4: from 3.210s to 1.060s

Reviewed-by: Chris Wilson <chris@hchris-wilson.co.uk>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Sean Finney <seanius@seanius.net>
Tested-by: Soren Hansen <soren@linux2go.dk>
Tested-by: Hernando Torque <sirius@sonnenkinder.org>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41059
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
drivers/gpu/drm/drm_edid.c

index 0c1a99b..bb95d59 100644 (file)
@@ -274,6 +274,11 @@ drm_do_probe_ddc_edid(struct i2c_adapter *adapter, unsigned char *buf,
                        }
                };
                ret = i2c_transfer(adapter, msgs, 2);
+               if (ret == -ENXIO) {
+                       DRM_DEBUG_KMS("drm: skipping non-existent adapter %s\n",
+                                       adapter->name);
+                       break;
+               }
        } while (ret != 2 && --retries);
 
        return ret == 2 ? 0 : -1;