Re: 3.5-rc3: PM/DSS broken (was vdd_mpu_iva warnings)

[Archit Taneja <a0393947@ti.com>]

On Tuesday 03 July 2012 02:07 PM, Joe Woodward wrote:
> -----Original Message-----
> From: Kevin Hilman<khilman@ti.com>
> To: "Joe Woodward"<jw@terrafix.co.uk>
> Cc: "linux-omap\@vger.kernel.org"<linux-omap@vger.kernel.org>
> Date: Mon, 02 Jul 2012 14:19:49 -0700
> Subject: Re: 3.5-rc3: vdd_mpu_iva warnings
>
>> "Joe Woodward"<jw@terrafix.co.uk>  writes:
>>
>>> I have a GUMSTIX Overo AirSTORM module (AM3703-based).
>>>
>>> When booting the kernel the following features are listed:
>>> OMAP3630 ES1.2 (l2cache neon isp 192mhz_clk )
>>>
>>> After booting I get the following (repeating every few seconds):
>>>
>>> [   81.122558] voltdm_scale: No voltage scale API registered for
>> vdd_mpu_iva
>>> [   81.130340] platform mpu.0: omap_target: unable to scale voltage
>> up.
>>>
>>
>> BTW, after boot, can you run the shell snippet below.  This should run
>> the board through MPU DVFS for all the available OPPs:
>>
>> Thanks,
>>
>> Kevin
>>
>
> First, apologies...
>
> I had added a GPIO control to board-overo.c, and after cleaning and
> starting again I realised that in the process I had deleted an
> important line.
>
> So, the warnings I posted are not present (i.e. the power domains
> do get registered), sorry for the noise!
>
> But, I still have problems with 3.5-rc5 with system suspend and DSS.
>
> I've started again with a completely clean area, built using
> omap2plus_defconfig with the following changes (DSS built-in, and
> use of devtmpfs and SquashFS):
> CONFIG_OMAP2_DSS=y
> CONFIG_OMAP2_VRAM_SIZE=4
> CONFIG_FB_OMAP2=y
>
> CONFIG_DEVTMPFS=y
> CONFIG_DEVTMPFS_MOUNT=y
> CONFIG_SQUASHFS=y
>
> I also modified board-overo.c to switch from DVI to LCD as the default device:
> static struct omap_dss_board_info overo_dss_data = {
> 	.num_devices	= ARRAY_SIZE(overo_dss_devices),
> 	.devices	= overo_dss_devices,
> //	.default_device	=&overo_dvi_device,
> 	.default_device	=&overo_lcd43_device,
> };
>
> And when suspending I get:
>
> # echo mem>  /sys/power/state
> [   20.138214] PM: Syncing filesystems ... done.
> [   20.158477] Freezing user space processes ... (elapsed 0.01 seconds) done.
> [   20.184509] Freezing remaining freezable tasks ... (elapsed 0.02 seconds) done.
> [   20.215698] Suspending console(s) (use no_console_suspend to debug)
> [   20.804626] Modules linked in:
> [   20.401611] PM: suspend of devices complete after 175.963 msecs
> [   20.404388] PM: late suspend of devices complete after 2.777 msecs
> [   20.409301] PM: noirq suspend of devices complete after 4.882 msecs
> [   20.409362] Disabling non-boot CPUs ...
> [   20.428527] Powerdomain (core_pwrdm) didn't enter target state 1
> [   20.428558] Could not enter target state in pm_suspend
> [   20.431060] PM: noirq resume of devices complete after 2.319 msecs
> [   20.434173] PM: early resume of devices complete after 1.739 msecs
> [   20.803344] mmc1: error -110 during resume (card was removed?)
> [   20.804595] ------------[ cut here ]------------
> [   20.804626] WARNING: at drivers/video/omap2/dss/dispc.c:376 dispc_runtime_get+0x50/0x64()
> [   20.804687] [<c001b61c>] (unwind_backtrace+0x0/0xf0) from [<c00401fc>] (warn_slowpath_common+0x4c/0x64)
> [   20.804718] [<c00401fc>] (warn_slowpath_common+0x4c/0x64) from [<c0040230>] (warn_slowpath_null+0x1c/0x24)
> [   20.804718] [<c0040230>] (warn_slowpath_null+0x1c/0x24) from [<c02a72d4>] (dispc_runtime_get+0x50/0x64)
> [   20.804748] [<c02a72d4>] (dispc_runtime_get+0x50/0x64) from [<c02b2d80>] (omapdss_dpi_display_enable+0x48/0x230)
> [   20.804779] [<c02b2d80>] (omapdss_dpi_display_enable+0x48/0x230) from [<c02c9a28>] (generic_dpi_panel_power_on+0x2c/0x78)
> [   20.804779] [<c02c9a28>] (generic_dpi_panel_power_on+0x2c/0x78) from [<c02c9a80>] (generic_dpi_panel_resume+0xc/0x1c)
> [   20.804809] [<c02c9a80>] (generic_dpi_panel_resume+0xc/0x1c) from [<c02ae28c>] (dss_resume_device+0x28/0x40)
> [   20.804840] [<c02ae28c>] (dss_resume_device+0x28/0x40) from [<c02fc848>] (bus_for_each_dev+0x50/0x7c)
> [   20.804870] [<c02fc848>] (bus_for_each_dev+0x50/0x7c) from [<c02ff530>] (platform_pm_resume+0x2c/0x50)
> [   20.804901] [<c02ff530>] (platform_pm_resume+0x2c/0x50) from [<c0303aa4>] (dpm_run_callback.clone.7+0x30/0xb0)
> [   20.804901] [<c0303aa4>] (dpm_run_callback.clone.7+0x30/0xb0) from [<c0304828>] (device_resume+0xc8/0x188)
> [   20.804931] [<c0304828>] (device_resume+0xc8/0x188) from [<c0304c50>] (dpm_resume+0xfc/0x21c)
> [   20.804962] [<c0304c50>] (dpm_resume+0xfc/0x21c) from [<c0304f04>] (dpm_resume_end+0xc/0x18)
> [   20.804962] [<c0304f04>] (dpm_resume_end+0xc/0x18) from [<c007ffe4>] (suspend_devices_and_enter+0x15c/0x2d0)
> [   20.804992] [<c007ffe4>] (suspend_devices_and_enter+0x15c/0x2d0) from [<c00802e4>] (pm_suspend+0x18c/0x208)
> [   20.805023] [<c00802e4>] (pm_suspend+0x18c/0x208) from [<c007f588>] (state_store+0x120/0x134)
> [   20.805053] [<c007f588>] (state_store+0x120/0x134) from [<c026800c>] (kobj_attr_store+0x14/0x20)
> [   20.805053] [<c026800c>] (kobj_attr_store+0x14/0x20) from [<c0169ff4>] (sysfs_write_file+0x100/0x184)
> [   20.805084] [<c0169ff4>] (sysfs_write_file+0x100/0x184) from [<c01093fc>] (vfs_write+0xb4/0x148)
> [   20.805114] [<c01093fc>] (vfs_write+0xb4/0x148) from [<c0109684>] (sys_write+0x40/0x70)
> [   20.805145] [<c0109684>] (sys_write+0x40/0x70) from [<c0013d60>] (ret_fast_syscall+0x0/0x3c)
> [   20.805145] ---[ end trace c7f1420c2bbb60a2 ]---
> [   20.805511] dpm_run_callback(): platform_pm_resume+0x0/0x50 returns -13
> [   20.805541] PM: Device omapdss failed to resume: error -13
> [   20.857238] PM: resume of devices complete after 422.943 msecs
> [   21.162658] Restarting tasks ... done.
> sh: write error: Operation not permitted
>
> This returns immediately.
>
> Any ideas?

Could you try to just disable/enable the panel and see if you get the 
same error? disabling/enabling result in the same code flow from a DSS 
point of view:

echo 0 > /sys/devices/platform/omapdss/display0/enabled
echo 1 > /sys/devices/platform/omapdss/display0/enabled

The error returned from the runtime_resume call is -ENOACCES, which by 
looking at the code, is returned by rpm_resume() when:

	dev->power.disable_depth > 0

I don't know what that means, does someone have any idea?

Archit