ACPI: Invalid Power Resource to register!

[Borislav Petkov <bp@amd64.org>]

Hi,

I'm getting the message in $Subject with latest linus:v3.5-8833-g2d534926205d

The dmesg printk is delayed for 5 seconds probably due to async scanning
(see trace below):

[   10.746148] [drm] vram apper at 0xD0000000
[   10.750680] [drm] size 786432
[   10.754050] [drm] fb depth is 8
[   10.757593] [drm]    pitch is 1024
[   10.761862] fbcon: radeondrmfb (fb0) is primary device

[    5.163363] ACPI: Invalid Power Resource to register			<----

[   10.949414] Console: switching to colour frame buffer device 128x48
[   10.977810] fb0: radeondrmfb frame buffer device
[   10.982601] drm: registered panic notifier
[   10.986892] [drm] Initialized radeon 2.17.0 20080528 for 0000:02:06.0 on minor 0

So I added some debugging code to see where exactly this happens:

diff --git a/drivers/acpi/power.c b/drivers/acpi/power.c
index 215ecd097408..d79e0f8c1f4e 100644
--- a/drivers/acpi/power.c
+++ b/drivers/acpi/power.c
@@ -432,11 +432,17 @@ int acpi_power_resource_register_device(struct device *dev, acpi_handle handle)
                return -ENODEV;
 
        ret = acpi_bus_get_device(handle, &acpi_dev);
-       if (ret)
+       if (ret) {
+               pr_warn(PREFIX "error getting device %s\n", dev_name(&acpi_dev->dev));
+               dump_stack();
                goto no_power_resource;
+       }
 
-       if (!acpi_dev->power.flags.power_resources)
+       if (!acpi_dev->power.flags.power_resources) {
+               pr_warn(PREFIX "no power resources %s\n", dev_name(&acpi_dev->dev));
+               dump_stack();
                goto no_power_resource;
+       }
 
        powered_device = kzalloc(sizeof(*powered_device), GFP_KERNEL);
        if (!powered_device)
---

and I got the following trace:

---
[    5.163287] ACPI: no power resources device:1a
[    5.163292] Pid: 1023, comm: kworker/u:4 Tainted: G        W    3.5.0+ #9
[    5.163293] Call Trace:
[    5.163300]  [<ffffffff81227866>] acpi_power_resource_register_device+0x8f/0x1bf
[    5.163305]  [<ffffffff81221a5f>] ? acpi_get_child+0x48/0x4e
[    5.163310]  [<ffffffff812e34db>] ata_acpi_bind+0x80/0x87
[    5.163313]  [<ffffffff812d891e>] ata_scsi_scan_host+0x91/0x184
[    5.163317]  [<ffffffff812d4f4f>] async_port_probe+0x4a/0x4f
[    5.163321]  [<ffffffff810568cc>] async_run_entry_fn+0x9a/0x152
[    5.163325]  [<ffffffff8104a103>] process_one_work+0x265/0x438
[    5.163328]  [<ffffffff8104a080>] ? process_one_work+0x1e2/0x438
[    5.163331]  [<ffffffff81475a01>] ? _raw_spin_lock_irq+0x17/0x46
[    5.163335]  [<ffffffff81056832>] ? async_schedule+0x17/0x17
[    5.163339]  [<ffffffff8104c47f>] worker_thread+0x260/0x329
[    5.163342]  [<ffffffff8107ac72>] ? trace_hardirqs_on+0xd/0xf
[    5.163346]  [<ffffffff8104c21f>] ? manage_workers.clone.15+0x1f6/0x1f6
[    5.163349]  [<ffffffff810502ee>] kthread+0x89/0x91
[    5.163353]  [<ffffffff8147dd64>] kernel_thread_helper+0x4/0x10
[    5.163356]  [<ffffffff814762d9>] ? retint_restore_args+0xe/0xe
[    5.163359]  [<ffffffff81050265>] ? __init_kthread_worker+0x5b/0x5b
[    5.163362]  [<ffffffff8147dd60>] ? gs_change+0xb/0xb


And as it looks like, acpi_dev->power.flags.power_resources is 0.

This thing is set in acpi_bus_get_power_flags() so I'd guess some of
those _PR* objects are probably not completely correct and they should
be fixed in the BIOS?

Any comments are appreciated.

Thanks.

-- 
Regards/Gruss,
Boris.

Advanced Micro Devices GmbH
Einsteinring 24, 85609 Dornach
GM: Alberto Bozzo
Reg: Dornach, Landkreis Muenchen
HRB Nr. 43632 WEEE Registernr: 129 19551