| Summary: | [ilk DP] pipe off timeout | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | DRI | Reporter: | Chris Wilson <chris> | ||||||
| Component: | DRM/Intel | Assignee: | Daniel Vetter <daniel> | ||||||
| Status: | CLOSED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||
| Severity: | minor | ||||||||
| Priority: | medium | CC: | bingx.a.yan, cancan.feng, chris, consume.noise, daniel, jbarnes, jinxianx.guo, lei.a.liu, martin.x.andersen, mroos, patrik.lundquist, qingshuai.tian, Roel.Teuwen, yangweix.shui | ||||||
| Version: | unspecified | ||||||||
| Hardware: | Other | ||||||||
| OS: | All | ||||||||
| Whiteboard: | |||||||||
| i915 platform: | i915 features: | ||||||||
| Attachments: |
|
||||||||
|
Description
Chris Wilson
2012-09-09 11:19:57 UTC
I made those waits 5s long and they still timeout. It would seem that we are not properly turning the pipe off... Do you still see those timeouts? Also can you attach verbose dmesg to figure out which state transition we're looking at here ... Wasn't that obvious from the trace? Take your pick, disabling (either DPMS off or modeset) on both of my ilk laptops complain with DP, HDMI and VGA. As you can probably guess I haven't used LVDS on those for sometime... Checking the modeset sequence again, we're supposed to disable pch outputs _after_ having disable the cpu pipe, but before disabling pch transcoder stuff and the fdi link. I.e.
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 82267b2..6b2b8a8 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3490,9 +3490,6 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
if (!intel_crtc->active)
return;
- for_each_encoder_on_crtc(dev, crtc, encoder)
- encoder->disable(encoder);
-
intel_crtc_wait_for_pending_flips(crtc);
drm_vblank_off(dev, pipe);
intel_crtc_update_cursor(crtc, false);
@@ -3509,6 +3506,9 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
I915_WRITE(PF_WIN_SZ(pipe), 0);
for_each_encoder_on_crtc(dev, crtc, encoder)
+ encoder->disable(encoder);
+
+ for_each_encoder_on_crtc(dev, crtc, encoder)
if (encoder->post_disable)
encoder->post_disable(encoder);
Warning persists after applying patch to reorder encoder->disable() *** Bug 58790 has been marked as a duplicate of this bug. *** This issue was also exist in this commit.(latest regression) Kernel: (drm-intel-next-queued)fa22c28522f53b63c1b287da1b69dc5700eaa09f Some additional commit info: Author: Ben Widawsky <ben@bwidawsk.net> Date: Fri Jan 18 19:18:19 2013 -0800 drm/i915: re-add dropped dma_mask configuration Well the good news, for me at least, I'm not seeing this on dinq + the various hacks. I'll check vanilla sometime soon, sooner if the machine crashes ;-) *** Bug 63453 has been marked as a duplicate of this bug. *** *** Bug 63912 has been marked as a duplicate of this bug. *** I have a clue: [ 0.639300] WARNING: at drivers/gpu/drm/i915/intel_display.c:1173 assert_cursors_disabled+0x40/0x50() [ 0.639300] cursor assertion failure, should be off on pipe A but is still active [ 0.639301] Modules linked in: [ 0.639304] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.10.0-rc5+ #21 [ 0.639304] Hardware name: LENOVO 514328U/514328U, BIOS 6QET44WW (1.14 ) 04/20/2010 [ 0.639306] 0000000000000009 ffff880232897448 ffffffff814bb243 ffff880232897480 [ 0.639307] ffffffff81038f80 0000000000000000 0000000000000000 0000000000000000 [ 0.639308] 0000000000000000 ffff880232857800 ffff8802328974e0 ffffffff8103906c [ 0.639309] Call Trace: [ 0.639315] [<ffffffff814bb243>] dump_stack+0x19/0x1b [ 0.639319] [<ffffffff81038f80>] warn_slowpath_common+0x70/0xa0 [ 0.639320] [<ffffffff8103906c>] warn_slowpath_fmt+0x4c/0x50 [ 0.639321] [<ffffffff812f3850>] assert_cursors_disabled+0x40/0x50 [ 0.639325] [<ffffffff812f80ab>] intel_disable_pipe+0x4b/0xc0 [ 0.639326] [<ffffffff812fa454>] ironlake_crtc_disable+0xd4/0x880 [ 0.639328] [<ffffffff812f68e2>] ? intel_dump_pipe_config.isra.43+0x1e2/0x1f0 [ 0.639330] [<ffffffff812feb41>] __intel_set_mode+0x311/0x1240 [ 0.639333] [<ffffffff813021b6>] intel_set_mode+0x16/0x30 [ 0.639334] [<ffffffff81302992>] intel_crtc_set_config+0x7c2/0x980 [ 0.639339] [<ffffffff812bee2e>] drm_mode_set_config_internal+0x2e/0x60 [ 0.639343] [<ffffffff812ab211>] drm_fb_helper_set_par+0x71/0xf0 [ 0.639345] [<ffffffff81248044>] fbcon_init+0x514/0x590 [ 0.639349] [<ffffffff81296e18>] visual_init+0xb8/0x120 [ 0.639351] [<ffffffff81299443>] do_bind_con_driver+0x163/0x320 [ 0.639352] [<ffffffff812996d1>] do_take_over_console+0x61/0x70 [ 0.639355] [<ffffffff812464f3>] do_fbcon_takeover+0x63/0xc0 [ 0.639356] [<ffffffff8124a0d5>] fbcon_event_notify+0x705/0x810 [ 0.639360] [<ffffffff81062dcd>] notifier_call_chain+0x4d/0x70 [ 0.639362] [<ffffffff81062fc8>] __blocking_notifier_call_chain+0x58/0x80 [ 0.639363] [<ffffffff81063006>] blocking_notifier_call_chain+0x16/0x20 [ 0.639365] [<ffffffff8123e29b>] fb_notifier_call_chain+0x1b/0x20 [ 0.639367] [<ffffffff8123fbb6>] register_framebuffer+0x1d6/0x310 [ 0.639368] [<ffffffff812aaee1>] drm_fb_helper_initial_config+0x371/0x540 [ 0.639372] [<ffffffff812cd833>] ? i915_write32+0x83/0x160 [ 0.639375] [<ffffffff811104a3>] ? __kmalloc+0x183/0x1d0 [ 0.639377] [<ffffffff812a9f7e>] ? drm_fb_helper_init+0xfe/0x200 [ 0.639378] [<ffffffff8111064f>] ? kmem_cache_alloc_trace+0x15f/0x180 [ 0.639382] [<ffffffff813236f1>] intel_fbdev_initial_config+0x21/0x30 [ 0.639384] [<ffffffff812d171e>] i915_driver_load+0xe3e/0xe70 [ 0.639386] [<ffffffff812ba4f6>] drm_get_pci_dev+0x176/0x2a0 [ 0.639388] [<ffffffff812cc83c>] i915_pci_probe+0x2c/0x70 [ 0.639390] [<ffffffff8123053c>] pci_device_probe+0x9c/0xe0 [ 0.639394] [<ffffffff81338ae7>] driver_probe_device+0x77/0x240 [ 0.639395] [<ffffffff81338d5b>] __driver_attach+0xab/0xb0 [ 0.639397] [<ffffffff81338cb0>] ? driver_probe_device+0x240/0x240 [ 0.639399] [<ffffffff81336d7d>] bus_for_each_dev+0x5d/0xa0 [ 0.639400] [<ffffffff813385ce>] driver_attach+0x1e/0x20 [ 0.639402] [<ffffffff8133813e>] bus_add_driver+0xfe/0x250 [ 0.639403] [<ffffffff813393f7>] driver_register+0x77/0x170 [ 0.639407] [<ffffffff818b1988>] ? ftrace_define_fields_drm_vblank_event+0x69/0x69 [ 0.639411] [<ffffffff8122f69b>] __pci_register_driver+0x4b/0x50 [ 0.639413] [<ffffffff812ba735>] drm_pci_init+0x115/0x130 [ 0.639414] [<ffffffff818b1988>] ? ftrace_define_fields_drm_vblank_event+0x69/0x69 [ 0.639416] [<ffffffff818b19ee>] i915_init+0x66/0x68 [ 0.639418] [<ffffffff81000312>] do_one_initcall+0x112/0x170 [ 0.639421] [<ffffffff81886e47>] kernel_init_freeable+0x101/0x18e [ 0.639423] [<ffffffff8188680b>] ? do_early_param+0x8a/0x8a [ 0.639426] [<ffffffff814abc40>] ? rest_init+0x90/0x90 [ 0.639427] [<ffffffff814abc4e>] kernel_init+0xe/0xf0 [ 0.639429] [<ffffffff814c105c>] ret_from_fork+0x7c/0xb0 [ 0.639430] [<ffffffff814abc40>] ? rest_init+0x90/0x90 [ 0.639436] ---[ end trace 2caf42eeeb901cd0 ]--- False alarm. *** Bug 62791 has been marked as a duplicate of this bug. *** Note, I am not in a position to test this anymore, so please could one of the other reporters provide updates and testing feedback? (In reply to comment #14) > Note, I am not in a position to test this anymore, so please could one of > the other reporters provide updates and testing feedback? Yeah, we are tracking DP issues on ILK for a long time, and all of them duplicated to this bug. We will handle the testing feedback from now on. shui yangwei, what's the status? Created attachment 90911 [details] dmesg of I-G-T/display DP calltrace I finished a round of IGT/display test with the latest drm-intel-next-queued kernel(820c19). We can still see the Calltrace as follows in the dmesg: [ 259.358172] WARNING: CPU: 3 PID: 3855 at drivers/gpu/drm/i915/intel_display.c:857 intel_wait_for_pipe_off+0x131/0x138 [i915]() [ 259.358175] pipe_off wait timed out [ 259.358176] Modules linked in: dm_mod snd_hda_codec_hdmi snd_hda_codec_idt iTCO_wdt iTCO_vendor_support dcdbas snd_hda_intel firewire_ohci pcspkr snd_hda_codec lpc_ich snd_hwdep i2c_i801 mfd_core snd_pcm firewire_core snd_page_alloc crc_itu_t snd_timer snd soundcore wmi battery ppdev parport_pc parport ac acpi_cpufreq joydev i915 video button drm_kms_helper drm [ 259.358209] CPU: 3 PID: 3855 Comm: testdisplay Tainted: G W 3.13.0-rc3_drm-intel-next-queued_820c19_20131218+ #3845 [ 259.358212] Hardware name: Dell Inc. Latitude E6510/0JKDHD, BIOS A05 08/10/2010 And the 1920x1080i mode still can't be showed correctly as reported in Bug 58790. CRTS(3):[0] 1920x1080i 60 1920 2008 2052 2200 1080 1084 1094 1125 0x15 0x40 74250 So I thing this situation may not have been improved a lot for now. You can check the dmesg in the attachment. Hope it's helpful. :) (In reply to comment #14) > Note, I am not in a position to test this anymore, so please could one of > the other reporters provide updates and testing feedback? I can see it with drm-intel-nightly (@ccd4562): http://lists.freedesktop.org/archives/intel-gfx/2014-March/041202.html on every xrandr --output DP2 --off I'm seeing a similar issue, and have been for many releases now. See https://bugzilla.kernel.org/show_bug.cgi?id=62251 and https://bugzilla.redhat.com/show_bug.cgi?id=929220 (where another reporter is also seeing it). For me, this occurs on every boot of my affected system (a 2010 model Sony Vaio Z, VGA adapter 8086:0046), no external monitor has to be involved (and indeed, there usually isn't one). My latest trace with 3.14.1-200.fc20.x86_64: WARNING: CPU: 0 PID: 283 at drivers/gpu/drm/i915/intel_display.c:851 intel_wait_for_pipe_off+0x1db/0x1f0 [i915]() pipe_off wait timed out Modules linked in: crct10dif_pclmul crc32_pclmul crc32c_intel i915(+) ghash_clmulni_intel e1000e i2c_algo_bit drm_kms_helper ptp drm pps_core i2c_core video CPU: 0 PID: 283 Comm: systemd-udevd Not tainted 3.13.10-200.fc20.x86_64 #1 Hardware name: Sony Corporation VPCZ112GD/VAIO, BIOS R3031C3 10/04/2011 0000000000000009 ffff88024ce172e8 ffffffff81687eac ffff88024ce17330 ffff88024ce17320 ffffffff8106d4dd ffff880036964000 0000000000070008 00000000fffb7b27 0000000000000000 ffff88024ccdd320 ffff88024ce17380 Call Trace: [<ffffffff81687eac>] dump_stack+0x45/0x56 [<ffffffff8106d4dd>] warn_slowpath_common+0x7d/0xa0 [<ffffffff8106d54c>] warn_slowpath_fmt+0x4c/0x50 [<ffffffffa015d9c8>] ? gen5_read32+0x48/0xc0 [i915] [<ffffffffa011e2ab>] intel_wait_for_pipe_off+0x1db/0x1f0 [i915] [<ffffffffa011e360>] intel_disable_pipe+0xa0/0xb0 [i915] [<ffffffffa011f434>] ironlake_crtc_disable+0xe4/0x940 [i915] [<ffffffffa011b39e>] ? intel_dump_pipe_config.isra.46+0x35e/0x370 [i915] [<ffffffffa0126374>] __intel_set_mode+0x7e4/0x14d0 [i915] [<ffffffffa0129456>] intel_set_mode+0x16/0x30 [i915] [<ffffffffa0129d0b>] intel_crtc_set_config+0x7ab/0x9a0 [i915] [<ffffffffa0031a5d>] drm_mode_set_config_internal+0x5d/0xe0 [drm] [<ffffffffa0088f41>] drm_fb_helper_set_par+0x71/0xf0 [drm_kms_helper] [<ffffffff8137b534>] fbcon_init+0x4f4/0x570 [<ffffffff813eb3b8>] visual_init+0xb8/0x120 [<ffffffff813edcd3>] do_bind_con_driver+0x163/0x330 [<ffffffff813ee484>] do_take_over_console+0x114/0x1c0 [<ffffffff81376df3>] do_fbcon_takeover+0x63/0xd0 [<ffffffff8137c09d>] fbcon_event_notify+0x60d/0x720 [<ffffffff81692a1c>] notifier_call_chain+0x4c/0x70 [<ffffffff81093a0d>] __blocking_notifier_call_chain+0x4d/0x70 [<ffffffff81093a46>] blocking_notifier_call_chain+0x16/0x20 [<ffffffff8136db4b>] fb_notifier_call_chain+0x1b/0x20 [<ffffffff8136fb2b>] register_framebuffer+0x1db/0x320 [<ffffffffa0088bef>] drm_fb_helper_initial_config+0x32f/0x500 [drm_kms_helper] [<ffffffffa015d9c8>] ? gen5_read32+0x48/0xc0 [i915] [<ffffffffa015dbfb>] ? gen5_write32+0x4b/0xb0 [i915] [<ffffffffa0163ff1>] intel_fbdev_initial_config+0x21/0x30 [i915] [<ffffffffa00ee2ac>] i915_driver_load+0xe0c/0xe40 [i915] [<ffffffffa002ba92>] drm_dev_register+0xa2/0x1e0 [drm] [<ffffffffa002d962>] drm_get_pci_dev+0x92/0x140 [drm] [<ffffffffa00ea67c>] i915_pci_probe+0x3c/0x90 [i915] [<ffffffff81352295>] local_pci_probe+0x45/0xa0 [<ffffffff81093a18>] ? __blocking_notifier_call_chain+0x58/0x70 [<ffffffff81353611>] pci_device_probe+0xd1/0x130 [<ffffffff81418305>] driver_probe_device+0x125/0x3a0 [<ffffffff81418653>] __driver_attach+0x93/0xa0 [<ffffffff814185c0>] ? __device_attach+0x40/0x40 [<ffffffff81416293>] bus_for_each_dev+0x63/0xa0 [<ffffffff81417cbe>] driver_attach+0x1e/0x20 [<ffffffff814178a0>] bus_add_driver+0x180/0x250 [<ffffffffa01aa000>] ? 0xffffffffa01a9fff [<ffffffff81418c94>] driver_register+0x64/0xf0 [<ffffffffa01aa000>] ? 0xffffffffa01a9fff [<ffffffff81351c3c>] __pci_register_driver+0x4c/0x50 [<ffffffffa002db2a>] drm_pci_init+0x11a/0x130 [drm] [<ffffffffa01aa000>] ? 0xffffffffa01a9fff [<ffffffffa01aa066>] i915_init+0x66/0x68 [i915] [<ffffffff8100214a>] do_one_initcall+0xfa/0x1b0 [<ffffffff810548d3>] ? set_memory_nx+0x43/0x50 [<ffffffff810e07ed>] load_module+0x1e1d/0x25b0 [<ffffffff810dc500>] ? store_uevent+0x40/0x40 [<ffffffff810e10f6>] SyS_finit_module+0x86/0xb0 [<ffffffff81696e69>] system_call_fastpath+0x16/0x1b oh, hum, that's clearly from 3.13.10. Still seems to occur with 3.14.0, though. *** Bug 78763 has been marked as a duplicate of this bug. *** Please test this patch here, should apply to any recent-ish upstream: http://patchwork.freedesktop.org/patch/24864/ It fixes pipe_off wait timed out,but another Call Trace was caused. [ 1.580745] [drm:intel_crtc_mode_set] [ENCODER:11:DAC-11] set [MODE:0:1280x1024] [ 1.581247] ------------[ cut here ]------------ [ 1.581274] WARNING: CPU: 0 PID: 1256 at drivers/gpu/drm/i915/intel_display.c:1151 ironlake_fdi_link_train+0x5b/0x331 [i915]() [ 1.581275] plane A assertion failure (expected on, current off) [ 1.581277] Modules linked in: i915(+) video button drm_kms_helper drm [ 1.581279] CPU: 0 PID: 1256 Comm: udevd Not tainted 3.14.0_kcloud_a46ade_20140526+ #54 [ 1.581280] Hardware name: Gigabyte Technology Co., Ltd. H55M-UD2H/H55M-UD2H, BIOS F4 12/02/2009 [ 1.581281] 0000000000000000 ffffffff816fd068 ffff8801128c5598 ffffffff81033e44 [ 1.581282] ffffffffa0096071 ffff880002d6c000 ffff8801128c55e8 00000000000f0018 [ 1.581283] 0000000000000000 ffffffff81033ea0 ffffffffa00dea48 0000000000000030 [ 1.581284] Call Trace: [ 1.581288] [<ffffffff816fd068>] ? dump_stack+0x41/0x51 [ 1.581291] [<ffffffff81033e44>] ? warn_slowpath_common+0x6f/0x84 [ 1.581308] [<ffffffffa0096071>] ? ironlake_fdi_link_train+0x5b/0x331 [i915] [ 1.581309] [<ffffffff81033ea0>] ? warn_slowpath_fmt+0x47/0x49 [ 1.581325] [<ffffffffa008c179>] ? gen5_read32+0x3a/0x41 [i915] [ 1.581342] [<ffffffffa0096071>] ? ironlake_fdi_link_train+0x5b/0x331 [i915] [ 1.581358] [<ffffffffa008c248>] ? gen5_write64+0x45/0x45 [i915] [ 1.581375] [<ffffffffa009a759>] ? ironlake_crtc_enable+0x31f/0xa79 [i915] [ 1.581392] [<ffffffffa009d4cd>] ? __intel_set_mode+0xbf3/0x118c [i915] [ 1.581410] [<ffffffffa009f682>] ? intel_set_mode+0xb/0x22 [i915] [ 1.581427] [<ffffffffa00a0476>] ? intel_crtc_set_config+0x95c/0x9f6 [i915] [ 1.581433] [<ffffffffa000be6c>] ? drm_mode_set_config_internal+0x48/0xc0 [drm] [ 1.581435] [<ffffffffa00484cf>] ? drm_fb_helper_restore_fbdev_mode+0x90/0xa5 [drm_kms_helper] [ 1.581436] [<ffffffffa0049d7f>] ? drm_fb_helper_set_par+0x42/0x69 [drm_kms_helper] [ 1.581439] [<ffffffff8130717b>] ? fbcon_init+0x2fa/0x400 [ 1.581441] [<ffffffff81352b89>] ? visual_init+0xaf/0x102 [ 1.581443] [<ffffffff8135448a>] ? do_bind_con_driver+0x19e/0x2c2 [ 1.581444] [<ffffffff81354a4b>] ? do_take_over_console+0x125/0x152 [ 1.581445] [<ffffffff81306885>] ? do_fbcon_takeover+0x4e/0x8e [ 1.581446] [<ffffffff817053ec>] ? notifier_call_chain+0x32/0x52 [ 1.581448] [<ffffffff8104cb4e>] ? __blocking_notifier_call_chain+0x43/0x5b [ 1.581450] [<ffffffff812fea77>] ? register_framebuffer+0x265/0x29b [ 1.581451] [<ffffffffa004996f>] ? drm_fb_helper_initial_config+0x3b9/0x45f [drm_kms_helper] [ 1.581468] [<ffffffffa008c268>] ? gen5_write32+0x20/0x45 [i915] [ 1.581484] [<ffffffffa008c268>] ? gen5_write32+0x20/0x45 [i915] [ 1.581503] [<ffffffffa00c36cd>] ? i915_driver_load+0xb7b/0xd70 [i915] [ 1.581508] [<ffffffffa000744d>] ? drm_dev_register+0x72/0xe3 [drm] [ 1.581512] [<ffffffffa00093ae>] ? drm_get_pci_dev+0xfc/0x1ba [drm] [ 1.581514] [<ffffffff812eea09>] ? local_pci_probe+0x34/0x77 [ 1.581515] [<ffffffff812eeb0d>] ? pci_device_probe+0xc1/0xe5 [ 1.581517] [<ffffffff8136feaa>] ? driver_probe_device+0x92/0x1b3 [ 1.581518] [<ffffffff8137004f>] ? __driver_attach+0x53/0x73 Are you sure the calltrace in comment #23 is from the patch? Please double-check that it doesn't exist with plain -nightly. i check on plain -nightly , and didn't find any Call Trace. I retest use patch , and the Call Trace is there. Does not apply against current upstream v3.15-rc7-58-gf2159d1 - should I test against some other tree? Daniel/Chris, can we move on this aging bug? I was about to open a bugreport for this when I notice it was initially created back in 2012. 8) Still occurs with 3.16.0. Surely there must be some fix for it? (In reply to comment #28) > I was about to open a bugreport for this when I notice it was initially > created back in 2012. 8) Still occurs with 3.16.0. Surely there must be some > fix for it? Using latest nightly, "pipe_off wait time out" issue still exists. Because some of IP environment reason , I can't get its dmesg. If it's necessary i will attach it later. reproduce steps: ---------------------- plugin DP ./testdisplay -a or hot plug DP (I ever reproduced this issue by this way) Created attachment 105280 [details]
dmesg output with drm.debug=0xe on 3.17-rc2
Still happens on 3.17-rc2.
Please test this patch http://patchwork.freedesktop.org/patch/32012/ The patch didn't help. Still hangs and time outs. I can confirm that Ville's patch resolves the ' pipe_off wait timed out' issue for me (Ironlake PCI ID: 0042). Did this patch make it into 3.17 mainline (assume not, as I see this on 3.17.0-0317009), or any of the intel-drm-nightlies? I'm seeing this error on a number of IVB workstations, so it would be good to have a solution for it. Never mind; I see the 'Move DP port disable to post_disable for pch platforms' patch is now present in 3.18-rc1. Fix landed in 3.18-rc: commit 08aff3fe26ae7a0d6f302ac2e1b7e2eb9933cd42 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Aug 18 22:16:09 2014 +0300 drm/i915: Move DP port disable to post_disable for pch platforms Closing resolved+fixed after 2 years of no updates. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.