Re: ACPI regression with 3.19+

[Jiang Liu]

On 2015/3/3 6:03, Rafael J. Wysocki wrote:
> On Monday, March 02, 2015 10:07:20 PM Hudd wrote:
>> On Sun, 2015-03-01 at 01:00 +0100, Rafael J. Wysocki wrote:
>>> On Saturday, February 28, 2015 10:35:21 AM Prakash Punnoor wrote:
>>>> This is a multi-part message in MIME format.
>>>> --------------080704070901080904040008
>>>> Content-Type: text/plain; charset=utf-8
>>>> Content-Transfer-Encoding: 7bit
>>>>
>>>> Hallo,
>>>>
>>>> my system won't boot with current GIT kernel (see attached screenshot).
>>>> The system seems somewhat frozen (cannot ssh into it), but magic sysrq
>>>> still works. I bisected the problem to commit
>>>> 593669c2ac0fe18baee04a3cd5539a148aa48574.
>>>
>>> Thanks for reporting, we're working on it.
>>
>> Hello folks,
>>
>> I've tried out 4.0-rc1, and it does not boot  on  my  system.   It  just
>> freezes without any messages (The only message I see is âLoading initial
>> ramdiskâ from GRUB).  I bisected the problem down to the same commit  as
>> in this  report.
>>
>> # first bad commit: [593669c2ac0fe18baee04a3cd5539a148aa48574]
>> x86/PCI/ACPI: Use common ACPI resource interfaces to simplify implementation
> 
> So at the moment I'm inclined to revert it.  Clearly, it wasn't thought
> through enough, so my suggestion would be to revert it for 4.0 and try again
> in the next cycle and be more careful this time.
> 
> Gerry, would it cause problems to happen if I reverted this one?
Hi Rafael,
	I will send out a patch today for this issue. On Thomas
Voegtle's system, the issue is caused by that BIOS report incorrect
length for ACPI address space descriptor. For example,
[    0.761553] acpi: address space [80000000-d0716ffe] len 50717000
[    0.767572] ACPI: invalid or unassigned resource mem
[0000000080000000 - 00000000d0716ffe] length [0000000050717000]

BIOS should report length of 0x50716fff instead of 0x50717000.
So seems we need to relax the length check code to workaround
such bios issues.

Will send out a patch for it soon.
Thanks!
Gerry
> 
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/