Re: DRM MSM initialization problem

From: Sean Paul <sean-p7yTbzM4H96eqtR555YLDQ@public.gmane.org>
To: Dima Antonets
	<dantonets-KXo+hBOZUgcVW8aojLLWWJvRsg8Z6D3kAL8bYrjMMd8@public.gmane.org>
Cc: linux-arm-msm-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	freedreno-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org,
	"dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org"
	<dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org>
Subject: Re: DRM MSM initialization problem
Date: Wed, 29 Aug 2018 13:32:57 -0400	[thread overview]
Message-ID: <20180829173257.GC188300@art_vandelay> (raw)
In-Reply-To: <DM6PR15MB22011B589422366CFBCEBC55AA090-W/oHEmsxt5II9Q/2ysJlLYd3EbNNOtPMvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>

On Wed, Aug 29, 2018 at 04:34:13PM +0000, Dima Antonets wrote:
> Hi all,
> 

Hi Dmitriy,
I don't have an answer for you, but I've Cc'd the msm lists in case someone
there does.

Sean

> I need an advice for the following DRM MSM initialization problem:
> drm/msm/msm_drv.c  msm_pdev_probe() fails with the following message (screenshot below)
> 
> [   54.833779] irq: no irq domain found for /soc/mdss@1a00000 !
> [   54.839019] CPU: 6 PID: 1132 Comm: modprobe Tainted: G        W      3.18.31 #1
> [   54.846271] Hardware name: Qualcomm Technologies, Inc.  (DT)
> [   54.852868] Call trace:
> [   54.855307] [<ffffffc000089b14>] dump_backtrace+0x0/0x270
> [   54.860682] [<ffffffc000089d98>] show_stack+0x14/0x1c
> [   54.865718] [<ffffffc000d8706c>] dump_stack+0x80/0xa4
> [   54.870752] [<ffffffc0000fb428>] irq_create_of_mapping+0xb4/0x1a8
> [   54.876831] [<ffffffc0008bf9a0>] irq_of_parse_and_map+0x30/0x58
> [   54.882731] [<ffffffc0008bf9f8>] of_irq_to_resource+0x30/0xd8
> [   54.888459] [<ffffffc0008bfad8>] of_irq_to_resource_table+0x38/0x5c
> [   54.894711] [<ffffffc0008bc92c>] of_device_alloc+0x118/0x1a0
> [   54.900351] [<ffffffc0008bca08>] of_platform_device_create_pdata+0x54/0xd8
> [   54.907209] [<ffffffc0008bcd68>] of_platform_bus_create+0x2c0/0x398
> [   54.913459] [<ffffffc0008bcfbc>] of_platform_populate+0x70/0xac
> [   54.919521] [<ffffffbffc133148>] msm_pdev_probe+0x68/0x2e8 [msm]
> [   54.925355] [<ffffffc0004ba35c>] platform_drv_probe+0x38/0x84
> [   54.931081] [<ffffffc0004b8558>] driver_probe_device+0x164/0x370
> [   54.937071] [<ffffffc0004b8814>] __driver_attach+0x64/0x90
> [   54.942539] [<ffffffc0004b7400>] bus_for_each_dev+0x80/0xb0
> [   54.948094] [<ffffffc0004b7ee4>] driver_attach+0x20/0x28
> [   54.953390] [<ffffffc0004b7b28>] bus_add_driver+0x128/0x200
> [   54.958945] [<ffffffc0004b9228>] driver_register+0x100/0x184
> [   54.964588] [<ffffffc0004ba2a4>] __platform_driver_register+0x60/0x68
> [   54.971134] [<ffffffbffc172210>] msm_drm_register+0xb8/0x100 [msm]
> [   54.977175] [<ffffffc000082b38>] do_one_initcall+0x1a0/0x1bc
> [   54.982819] [<ffffffc0001217ac>] load_module+0xdb0/0x1430
> [   54.988199] [<ffffffc00012205c>] SyS_finit_module+0xd4/0xec
> 
> Actually, the of_platform_populate() calls from msm_pdev_probe()->add_display_components().
> 
> The MDSS IRQ domain is not initialized at this time because it will initialize later (drm/msm/mdp/mdp5/mdp5_mdss.c, mdss_irq_domain_init()), via call     component_master_add_with_match()  AFTER this error happen ..
> Is it a some "trick" for init the IRQ domain early?
> 
> Thanks,
> Dmitriy Antonets

> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Sean Paul, Software Engineer, Google / Chromium OS
_______________________________________________
Freedreno mailing list
Freedreno@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/freedreno