On Wed, 2014-04-02 at 00:36 +0300, Imre Deak wrote: > On Tue, 2014-04-01 at 18:04 -0300, Paulo Zanoni wrote: > > 2014-04-01 17:52 GMT-03:00 Daniel Vetter <daniel@xxxxxxxx>: > > > On Tue, Apr 01, 2014 at 05:48:15PM -0300, Paulo Zanoni wrote: > > >> 2014-04-01 17:37 GMT-03:00 Daniel Vetter <daniel@xxxxxxxx>: > > >> > On Tue, Apr 01, 2014 at 02:55:09PM -0300, Paulo Zanoni wrote: > > >> >> From: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx> > > >> >> > > >> >> ... at edp_have_panel_vdd. Just return false, saying we don't have the > > >> >> panel VDD since the device is suspended. > > >> >> > > >> >> We started getting WARNs about this problem since the patch that > > >> >> started checking if we're suspended while reading registers. > > >> >> > > >> >> Testcase: igt/pm_pc8 > > >> >> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx> > > >> > > > >> > Hm, where's an example backtrace for this? I wonder whether we simply need > > >> > to extend the range where we hold the runtime pm ref a bit ... > > >> > > >> There are tons of WARNs, but here is one example: > > >> > > >> [ 63.572201] [drm:hsw_enable_pc8] Enabling package C8+ > > >> [ 63.581831] [drm:i915_runtime_suspend] Device suspended > > >> [ 63.664798] ------------[ cut here ]------------ > > >> [ 63.664824] WARNING: CPU: 3 PID: 828 at > > >> drivers/gpu/drm/i915/intel_uncore.c:47 > > >> assert_device_not_suspended.isra.7+0x32/0x40 [i915]() > > >> [ 63.664826] Device suspended > > >> [ 63.664828] Modules linked in: ccm fuse ip6table_filter ip6_tables > > >> ebtable_nat ebtables arc4 ath9k_htc ath9k_common ath9k_hw mac80211 ath > > >> cfg80211 iTCO_wdt iTCO_vendor_support x86_pkg_temp_thermal coretemp > > >> microcode i2c_i801 e1000e pcspkr serio_raw lpc_ich ptp pps_core mei_me > > >> mei mfd_core dm_crypt i915 crc32_pclmul crc32c_intel > > >> ghash_clmulni_intel i2c_algo_bit drm_kms_helper drm video > > >> [ 63.664867] CPU: 3 PID: 828 Comm: kworker/3:3 Not tainted 3.14.0+ #153 > > >> [ 63.664869] Hardware name: Intel Corporation Shark Bay Client > > >> platform/WhiteTip Mountain 1, BIOS HSWLPTU1.86C.0133.R00.1309172123 > > >> 09/17/2013 > > >> [ 63.664887] Workqueue: events edp_panel_vdd_work [i915] > > >> [ 63.664889] 0000000000000009 ffff88009d745c28 ffffffff8167ec6f > > >> ffff88009d745c70 > > >> [ 63.664895] ffff88009d745c60 ffffffff8106c8ed ffff880036278000 > > >> 00000000000c7204 > > >> [ 63.664900] ffff88014f2d3040 ffff880036278070 0000000000000001 > > >> ffff88009d745cc0 > > >> [ 63.664905] Call Trace: > > >> [ 63.664911] [<ffffffff8167ec6f>] dump_stack+0x4d/0x66 > > >> [ 63.664916] [<ffffffff8106c8ed>] warn_slowpath_common+0x7d/0xa0 > > >> [ 63.664920] [<ffffffff8106c95c>] warn_slowpath_fmt+0x4c/0x50 > > >> [ 63.664926] [<ffffffff810bd6be>] ? mark_held_locks+0xae/0x130 > > >> [ 63.664941] [<ffffffffa00d80d2>] > > >> assert_device_not_suspended.isra.7+0x32/0x40 [i915] > > >> [ 63.664956] [<ffffffffa00d99d2>] gen6_read32+0x32/0x120 [i915] > > >> [ 63.664969] [<ffffffffa00d99a0>] ? gen6_read8+0x120/0x120 [i915] > > >> [ 63.664985] [<ffffffffa0106f8f>] edp_have_panel_vdd+0x3f/0x50 [i915] > > >> [ 63.665000] [<ffffffffa01074e8>] edp_panel_vdd_off_sync+0x58/0x1c0 [i915] > > >> [ 63.665004] [<ffffffff8108a06c>] ? process_one_work+0x18c/0x560 > > >> [ 63.665018] [<ffffffffa0107684>] edp_panel_vdd_work+0x34/0x50 [i915] > > >> [ 63.665022] [<ffffffff8108a0d7>] process_one_work+0x1f7/0x560 > > >> [ 63.665026] [<ffffffff8108a06c>] ? process_one_work+0x18c/0x560 > > >> [ 63.665031] [<ffffffff8108ae2b>] worker_thread+0x11b/0x3a0 > > > > > > Ah, that's the async edp worker thread. I guess we need to grab a runtim > > > pm ref when we start it and drop it again in the worker instead? > > > > IMHO it doesn't make sense to keep the HW awake just so we can read a > > register to find out things are disabled, so I prefer the current > > solution. If we want a "better" solution, IMHO we should track the > > sate of the VDD in software, like intel_dp->has_panel_vdd. This would > > avoid many more register reads. > > Chiming in, as I was wondering what could cause the inbalance between > the edp_panel_vdd_on() and off() calls. One possibility is that > intel_dp_probe_oui() calls edp_panel_vdd_on() and then > edp_panel_vdd_off(..., false) which schedules the work and then > intel_enable_dp() calls edp_panel_vdd_on() and edp_panel_vdd_off(..., > true) and disables VDD synchronously, dropping the RPM ref. Note that in > this case we won't get the "eDP VDD not forced on" WARN either. > > Whether or not this was the case for you, I think for good measure we > should flush any pending vdd_work in _edp_panel_vdd_on() before setting > intel_dp->want_panel_vdd = true; Actually not flush but cancel the work. --Imre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
