Samsung Series 9 seems to require acpi=nonvs to resume properly
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely? I ask because resume used to work with the 2.6.39 kernels, although it was flaky and would sometimes lock up.
This seems to be the necessary information for the blacklist, taken from dmidecode:
Handle 0x0001, DMI type 1, 27 bytes
System Information
Manufacturer: SAMSUNG ELECTRONICS CO., LTD. Product Name: 90X3A
Version: 0.1
Serial Number: CZPR93GB301834
UUID: 249F9300-1DD2-11B2-8000-89C120F0C461 Wake-up Type: APM Timer
SKU Number: System SKUNumber
Family: HuronRiver System
I can make this into a patch if necessary (drivers/acpi/sleep.c, right?). I am not /entirely/ sure it is the correct fix.
Signature · 4 lines
Knowledge Is Power Power Corrupts Study Hard Be Evil
Quoted 10 lines · Zan Lynx
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely?
Please ignore this about nonvs. nonvs has nothing to do with the suspend/resume problem on this laptop.
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
So now it appears that everything I ever tried to fix this only seemed to work because I rebooted after changing GRUB or module options.
Signature · 4 lines
Knowledge Is Power Power Corrupts Study Hard Be Evil
Quoted 15 lines · Zan Lynx
Quoted 10 lines · Zan Lynx
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely?
Please ignore this about nonvs. nonvs has nothing to do with the suspend/resume problem on this laptop.
Quoted 7 lines
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
So now it appears that everything I ever tried to fix this only seemed to work because I rebooted after changing GRUB or module options.
What graphics is there in the system?
Rafael
Quoted 21 lines · Rafael J. Wysocki
Quoted 20 lines · Zan Lynx
Quoted 10 lines · Zan Lynx
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely?
Please ignore this about nonvs. nonvs has nothing to do with the suspend/resume problem on this laptop.
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
Quoted 4 lines
Quoted 2 lines
So now it appears that everything I ever tried to fix this only seemed to work because I rebooted after changing GRUB or module options.
What graphics is there in the system?
It is the Intel i915 driver. It is an Intel i5-2537M system with the integrated video on the CPU.
lspci shows it as:
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])
Signature · 4 lines
Knowledge Is Power Power Corrupts Study Hard Be Evil
Quoted 27 lines · Zan Lynx
Quoted 21 lines · Rafael J. Wysocki
Quoted 20 lines · Zan Lynx
Quoted 10 lines · Zan Lynx
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely?
Please ignore this about nonvs. nonvs has nothing to do with the suspend/resume problem on this laptop.
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
Quoted 4 lines
Quoted 2 lines
So now it appears that everything I ever tried to fix this only seemed to work because I rebooted after changing GRUB or module options.
What graphics is there in the system?
Quoted 7 lines
It is the Intel i915 driver. It is an Intel i5-2537M system with the integrated video on the CPU.
lspci shows it as:
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])
Well, that should work. I'm afraid I can't help you with this issue.
Thanks,
Rafael
Quoted 2 lines · Zan Lynx
It is the Intel i915 driver. It is an Intel i5-2537M system with the integrated video on the CPU.
You might compare the output of intel_reg_dumper in the working/not-working case before the i915 driver loads.
Signature · 1 line
[email protected]
Quoted 12 lines · Zan Lynx
Quoted 10 lines · Zan Lynx
After being annoyed that the latest Fedora 2.6.40 (aka 3.0 something) kernels would not suspend/resume on my laptop I followed a Ubuntu forum suggestion to use acpi=nonvs. This worked, or at least it is working today with Fedora 15's 2.6.40.4-5.fc15.x86_64.
The system is also booting via UEFI if that makes any difference.
I understand that there is a no-NVS blacklist in the kernel. Should the Samsung Series 9 be added to this list? Or could it be a different suspend/resume problem entirely?
Quoted 7 lines
Please ignore this about nonvs. nonvs has nothing to do with the suspend/resume problem on this laptop.
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
Could you provide any error logs you found? That could give us a clue as to what is going wrong...
Quoted 4 lines
So now it appears that everything I ever tried to fix this only seemed to work because I rebooted after changing GRUB or module options.
Signature · 4 lines
Regards, Srivatsa S. Bhat <[email protected]> Linux Technology Center, IBM India Systems and Technology Lab
Re: Samsung Series 9 won't resume
Quoted 1 line · Srivatsa S. Bhat
On 10/01/2011 01:39 AM, Zan Lynx wrote:
[snip]
Quoted 7 lines
Quoted 4 lines
Instead it seems to involve reboots, which is crazy. From a cold power-on start, suspend resume will always fail. But if the laptop is rebooted first, suspend and resume appear to work. What the heck? I have no clue how to debug this.
Could you provide any error logs you found? That could give us a clue as to what is going wrong...
Sorry for the delayed response.
No, there were no error logs. I even tried the suspend debugging which is supposed to write into the clock registers.
The symptom is that suspend appears to work correctly. Everything powers down and the power light slowly flashes on and off. When the power button is pressed to trigger resume, the fans come on, the power light comes on full, but there is no hard drive activity and the screen does nothing.
After forcing a power-off and rebooting, when using the "echo 1
Quoted 1 line
/sys/power/pm_trace" trick, the system has nothing in the dmesg about
resume failures. I have to wonder if it even started executing the resume.
Other interesting info is that I have the laptop using UEFI and the default reboot causes a kernel panic, so I have the kernel option "reboot=k" set.
I just thought of this, could the EFI services be messing up the RTC programming that pm_trace uses?
I suppose that I should also test suspend resume without UEFI, in BIOS mode. Such a pain.
Signature · 4 lines
Knowledge Is Power Power Corrupts Study Hard Be Evil