From: Tomi Valkeinen <tomi.valkeinen@ti.com>
To: Joe Woodward <jw@terrafix.co.uk>, Kevin Hilman <khilman@ti.com>
Cc: linux-omap@vger.kernel.org, Archit Taneja <archit@ti.com>
Subject: Re: DSS2/PM on 3.2 broken?
Date: Thu, 12 Jan 2012 18:42:31 +0200 [thread overview]
Message-ID: <1326386551.1896.41.camel@deskari> (raw)
In-Reply-To: <WC20120109124643.880164@terrafix.co.uk>
[-- Attachment #1: Type: text/plain, Size: 9090 bytes --]
On Mon, 2012-01-09 at 12:46 +0000, Joe Woodward wrote:
> I'm running on a Gumstix Overo (OMAP3530) with an 24-bit LCD panel connected via the DPI interface (using the generic panel driver).
>
> Entering standby used to work just fine on 3.0, but on 3.2 I get the following:
I've been debugging this, but I'm at loss. I added some debug prints,
which I show below, and I also pushed them to
"git://gitorious.org/linux-omap-dss2/linux.git pm-test-prints" if
somebody wants to see exactly what they print.
Kevin, perhaps you have an idea what could be wrong here. Long version
below, short version: doing system suspend breaks omapdss, if omapdss
uses pm_runtime_put, but works with pm_runtime_put_sync.
First, as a test case, here's what happens if I manually disable a
display (with my comments beginning with #):
echo 0 > /sys/devices/platform/omapdss/display0/enabled
# here dpi.c calls dispc_runtime_put
[ 44.528015] dispc_runtime_put
[ 44.531158] dispc disable depth: 0
[ 44.535308] dispc disable depth: 0
[ 44.538879] dispc_runtime_put end 0
# here dpi.c calls dss_runtime_put
[ 44.542633] dss_runtime_put
[ 44.545593] dss disable depth: 0
[ 44.548980] dss disable depth: 0
[ 44.552429] dss_runtime_put end 0
# then the runtime PM runs dispc suspend in its workqueue
[ 44.557861] dispc_runtime_suspend start
# here dispc_runtime_suspend uses dss_runtime_put
[ 44.561981] dss_runtime_put
[ 44.564910] dss disable depth: 0
[ 44.568420] dss disable depth: 0
[ 44.571838] dss_runtime_put end 0
[ 44.575378] dispc_runtime_suspend end
# then the runtime PM runs dss suspend in its workqueue
[ 44.579315] dss_runtime_suspend start
[ 44.583221] dss_runtime_suspend end
And similarly when enabling the display:
echo 1 > /sys/devices/platform/omapdss/display0/enabled
[ 442.109222] dss_runtime_get
[ 442.112304] dss disable depth: 0
[ 442.115753] dss_runtime_resume start
[ 442.119537] dss_runtime_resume end
[ 442.123199] dss disable depth: 0
[ 442.126586] dss_runtime_get end 0
[ 442.130126] dispc_runtime_get
[ 442.133270] dispc disable depth: 0
[ 442.136932] dispc_runtime_resume start
[ 442.140869] dss_runtime_get
[ 442.143890] dss disable depth: 0
[ 442.147308] dss disable depth: 0
[ 442.150695] dss_runtime_get end 1
[ 442.154235] dispc_runtime_resume end
[ 442.158020] dispc disable depth: 0
[ 442.161651] dispc_runtime_get end 0
This all works fine. Now, if I suspend the system:
echo mem > /sys/power/state
[ 482.097930] PM: Syncing filesystems ... done.
[ 482.105041] PM: Preparing system for mem sleep
[ 482.116394] Freezing user space processes ... (elapsed 0.02 seconds) done.
[ 482.146331] Freezing remaining freezable tasks ... (elapsed 0.02 seconds) done.
[ 482.177520] PM: Entering mem sleep
# dss core suspend disables the displays, as was done manually above
[ 482.197570] core suspend 2
[ 482.213867] dispc_runtime_put
[ 482.217010] dispc disable depth: 0
[ 482.220886] dispc disable depth: 0
[ 482.224578] dispc_runtime_put end 0
[ 482.228271] dss_runtime_put
[ 482.231262] dss disable depth: 0
[ 482.234649] dss disable depth: 0
[ 482.238128] dss_runtime_put end 0
[ 482.242218] core suspend end 0
# so far looks exactly the same as above
[ 482.246795] PM: suspend of devices complete after 60.260 msecs
# now runtime PM runs the suspends in its workqueue
[ 482.256744] dispc_runtime_suspend start
[ 482.260955] dss_runtime_put
# dss's disable_depth is 1?? Which, afaik, means something like runtime PM is disabled for DSS.
[ 482.263916] dss disable depth: 1
[ 482.267303] dss disable depth: 1
[ 482.270782] dss_runtime_put end 0
[ 482.274261] dispc_runtime_suspend end
[ 482.278381] dss_runtime_suspend start
[ 482.282257] dss_runtime_suspend end
[ 482.286224] PM: late suspend of devices complete after 33.209 msecs
[ 482.292968] Disabling non-boot CPUs ...
Except the disable_depth, the suspend went as in manual case. But on resume things are rather broken:
[ 696.654754] Successfully put all powerdomains to target state
# Here the PM framework calls dss's runtime_resume callback, even if nobody has called pm_rutime_get for dss!
[ 696.661254] dss_runtime_resume start
[ 696.665100] dss_runtime_resume end
[ 696.668731] dispc_runtime_resume start
[ 696.672760] dss_runtime_get
[ 696.675689] dss disable depth: 1
[ 696.679168] EACCESS RPM_RESUME
[ 696.682373] dss disable depth: 1
# And here pm_runtime_get fails for DSS, because dss disable_depth is 1...
[ 696.685791] ------------[ cut here ]------------
[ 696.690734] WARNING: at drivers/video/omap2/dss/dss.c:716 dss_runtime_get+0x74/0x8c()
[ 696.699035] Modules linked in:
[ 696.702392] [<c001b3c4>] (unwind_backtrace+0x0/0xf0) from [<c005248c>] (warn_slowpath_common+0x4c
/0x64)
[ 696.712371] [<c005248c>] (warn_slowpath_common+0x4c/0x64) from [<c00524c0>] (warn_slowpath_null+0
x1c/0x24)
[ 696.722625] [<c00524c0>] (warn_slowpath_null+0x1c/0x24) from [<c0273cd8>] (dss_runtime_get+0x74/0
x8c)
[ 696.732421] [<c0273cd8>] (dss_runtime_get+0x74/0x8c) from [<c0274610>] (dispc_runtime_resume+0x10
/0x149c)
[ 696.742584] [<c0274610>] (dispc_runtime_resume+0x10/0x149c) from [<c02d09c0>] (pm_generic_runtime
_resume+0x2c/0x40)
[ 696.753662] [<c02d09c0>] (pm_generic_runtime_resume+0x2c/0x40) from [<c003af70>] (_od_resume_noir
q+0x48/0x54)
[ 696.764190] [<c003af70>] (_od_resume_noirq+0x48/0x54) from [<c02d1c54>] (pm_noirq_op.clone.5+0x9c
/0x164)
[ 696.774291] [<c02d1c54>] (pm_noirq_op.clone.5+0x9c/0x164) from [<c02d1d7c>] (dpm_resume_noirq+0x6
0/0x1f4)
[ 696.784454] [<c02d1d7c>] (dpm_resume_noirq+0x60/0x1f4) from [<c0098e2c>] (suspend_devices_and_ent
er+0x114/0x2d0)
[ 696.795227] [<c0098e2c>] (suspend_devices_and_enter+0x114/0x2d0) from [<c0099128>] (enter_state+0
x140/0x180)
[ 696.805694] [<c0099128>] (enter_state+0x140/0x180) from [<c0097f54>] (state_store+0xd0/0x178)
[ 696.814758] [<c0097f54>] (state_store+0xd0/0x178) from [<c02460a4>] (kobj_attr_store+0x14/0x20)
[ 696.824005] [<c02460a4>] (kobj_attr_store+0x14/0x20) from [<c01626c0>] (sysfs_write_file+0x100/0x
184)
[ 696.833801] [<c01626c0>] (sysfs_write_file+0x100/0x184) from [<c0101e20>] (vfs_write+0xb4/0x148)
[ 696.843139] [<c0101e20>] (vfs_write+0xb4/0x148) from [<c01020a8>] (sys_write+0x40/0x70)
[ 696.851654] [<c01020a8>] (sys_write+0x40/0x70) from [<c0013b60>] (ret_fast_syscall+0x0/0x3c)
[ 696.860626] ---[ end trace 14e40b198c698a6f ]---
[ 696.865478] dss_runtime_get end -13
[ 696.869201] dispc_runtime_resume end fail
[ 696.875549] PM: early resume of devices complete after 214.509 msecs
[ 696.884704] dispc_runtime_suspend start
[ 696.888824] dss_runtime_put
[ 696.891845] dss disable depth: 0
[ 696.895568] dss disable depth: 0
[ 696.899047] dss_runtime_put end 0
[ 696.902526] dispc_runtime_suspend end
[ 696.906616] core resume
[ 696.910339] dss_runtime_get
[ 696.913391] dss disable depth: 0
[ 696.916809] dss disable depth: 0
[ 696.920196] dss_runtime_get end 1
[ 696.923767] dispc_runtime_get
[ 696.926879] dispc disable depth: 0
[ 696.930603] dispc_runtime_resume start
[ 696.934539] dss_runtime_get
[ 696.937561] dss disable depth: 0
[ 696.940979] dss disable depth: 0
[ 696.944366] dss_runtime_get end 1
[ 696.947906] dispc_runtime_resume end
[ 696.951751] dispc disable depth: 0
[ 696.955383] dispc_runtime_get end 0
[ 696.961212] core resume end 0
[ 696.973419] PM: resume of devices complete after 89.477 msecs
[ 696.984893] PM: Finishing wakeup.
[ 696.988403] Restarting tasks ... done.
After changing pm_runtime_put calls in dss and dispc to sync versions:
echo mem > /sys/power/state
[ 21.312133] PM: Syncing filesystems ... done.
[ 21.319000] PM: Preparing system for mem sleep
[ 21.331390] Freezing user space processes ... (elapsed 0.02 seconds) done.
[ 21.361297] Freezing remaining freezable tasks ... (elapsed 0.02 seconds) done.
[ 21.392517] PM: Entering mem sleep
[ 21.412506] core suspend 2
[ 21.423370] dispc_runtime_put
[ 21.426513] dispc disable depth: 0
[ 21.430236] dispc_runtime_suspend start
[ 21.434326] dss_runtime_put
[ 21.437347] dss disable depth: 0
[ 21.440734] dss disable depth: 0
[ 21.444152] dss_runtime_put end 0
[ 21.447692] dispc_runtime_suspend end
[ 21.451660] dispc disable depth: 0
[ 21.455322] dispc_runtime_put end 0
[ 21.458984] dss_runtime_put
[ 21.462005] dss disable depth: 0
[ 21.465423] dss_runtime_suspend start
[ 21.469329] dss_runtime_suspend end
[ 21.473052] dss disable depth: 0
[ 21.476531] dss_runtime_put end 0
[ 21.480560] core suspend end 0
[ 21.485137] PM: suspend of devices complete after 83.606 msecs
[ 21.495483] PM: late suspend of devices complete after 4.089 msecs
[ 21.502166] Disabling non-boot CPUs ...
Looks quite similar, but everything is done before "PM: suspend of
devices..." line. And resume works fine also.
Tomi
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
next prev parent reply other threads:[~2012-01-12 16:42 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-01-09 12:46 DSS2/PM on 3.2 broken? Joe Woodward
2012-01-09 21:08 ` NeilBrown
2012-01-10 9:58 ` Joe Woodward
2012-01-11 13:43 ` Paul Walmsley
2012-01-11 14:22 ` Archit
2012-01-11 15:15 ` Joe Woodward
2012-01-11 15:52 ` Archit
2012-01-11 16:13 ` Joe Woodward
2012-01-11 16:54 ` Archit
2012-01-12 9:28 ` Tomi Valkeinen
2012-01-12 9:30 ` Tomi Valkeinen
2012-01-12 9:51 ` Tomi Valkeinen
2012-01-11 22:59 ` NeilBrown
2012-01-13 10:05 ` Paul Walmsley
2012-01-13 11:20 ` NeilBrown
2012-01-13 11:31 ` Paul Walmsley
2012-01-13 23:09 ` NeilBrown
2012-01-13 23:35 ` Paul Walmsley
2012-01-17 21:24 ` NeilBrown
2012-01-22 0:07 ` Paul Walmsley
2012-01-22 11:30 ` NeilBrown
2012-01-24 10:37 ` OMAP HDQ: was " NeilBrown
2012-01-26 14:19 ` Paul Walmsley
2012-01-27 22:35 ` NeilBrown
2012-01-27 22:58 ` Paul Walmsley
2012-01-28 0:40 ` NeilBrown
2012-01-28 6:02 ` Paul Walmsley
2012-02-01 7:51 ` NeilBrown
2012-02-01 18:36 ` Paul Walmsley
2012-01-18 7:13 ` Tomi Valkeinen
2012-01-18 11:15 ` NeilBrown
2012-01-18 11:42 ` Tomi Valkeinen
2012-01-18 20:30 ` NeilBrown
2012-01-19 10:17 ` Joe Woodward
2012-01-19 10:40 ` Tomi Valkeinen
2012-01-19 11:29 ` Joe Woodward
2012-01-19 11:36 ` Tomi Valkeinen
2012-01-19 12:21 ` Joe Woodward
2012-01-19 14:52 ` Tomi Valkeinen
2012-01-19 19:37 ` Kevin Hilman
2012-01-19 21:05 ` NeilBrown
2012-01-20 0:22 ` Kevin Hilman
2012-01-21 12:12 ` NeilBrown
2012-01-23 22:11 ` Kevin Hilman
2012-01-25 0:32 ` NeilBrown
2012-01-13 11:34 ` Govindraj
2012-01-13 13:23 ` Paul Walmsley
2012-01-13 19:21 ` Kevin Hilman
2012-01-13 22:37 ` Kevin Hilman
2012-01-13 23:06 ` Paul Walmsley
2012-01-13 23:34 ` Paul Walmsley
2012-01-14 1:17 ` NeilBrown
2012-01-14 1:28 ` Paul Walmsley
2012-01-13 23:39 ` Paul Walmsley
2012-01-13 11:19 ` Paul Walmsley
2012-01-11 13:32 ` Paul Walmsley
2012-01-12 16:42 ` Tomi Valkeinen [this message]
2012-01-12 22:40 ` Kevin Hilman
2012-01-13 5:29 ` Tomi Valkeinen
2012-01-13 19:30 ` Kevin Hilman
2012-01-16 11:11 ` Tomi Valkeinen
2012-01-19 19:24 ` Kevin Hilman
2012-01-20 7:16 ` Tomi Valkeinen
2012-01-20 18:06 ` Kevin Hilman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1326386551.1896.41.camel@deskari \
--to=tomi.valkeinen@ti.com \
--cc=archit@ti.com \
--cc=jw@terrafix.co.uk \
--cc=khilman@ti.com \
--cc=linux-omap@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox