[git,pull,request] ACPI & driver patches for Linux-2.6.32-rc0

Commit Message

Alexey Starikovskiy Sept. 29, 2009, 8:25 a.m. UTC

Thomas Backlund пишет:
> Alexey Starikovskiy wrote:
>> Thomas Backlund пишет:
>>> Thomas Backlund wrote:
>>>> Alexey Starikovskiy wrote:
>>>>> Thomas Backlund пишет:
>>>>>> Now this is a Acer TravelMate 5720G laptop with the latest 1.35 bios.
>>>>>> Distro is Mandriva Linux Cooker, and arch is x86_64
>>>>>>
>>>>>> Attached is output of "grep ACPI /var/log/dmesg"
>>>>>> dmesg-acpi-20090903
>>>>>> - all acpi patches as of 2.6.32-rc1
>>>>> This one does not have errors you've mentioned earlier, could you 
>>>>> find dmesg with them?
>>>> Ah, sorry about that...
>>>> I should have done dmesg|grep ACPI to get the errors... it would 
>>>> have shown this too:
>>>> tg3 0000:02:00.0: wake-up capability disabled by ACPI
>>>> ACPI Exception: AE_TIME, Returned by Handler for [EmbeddedControl] 
>>>> (20090903/evregion-424)
>>>> ACPI Error (psparse-0537): Method parse/execution failed 
>>>> [\_SB_.PCI0.LPCB.EC0_.GBST] (Node ffff88013f81c5e0), AE_TIME
>>>> ACPI Error (psparse-0537): Method parse/execution failed 
>>>> [\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node ffff88013f81c720), AE_TIME
>>>> ACPI Exception: AE_TIME, Evaluating _BST (20090903/battery-393)
>>>>
>>>>
>>>>
>>>>>> dmesg-acpi-20090903-ec-2.6.31
>>>>>> - all acpi patches as of 2.6.32-rc1
>>>>>> - ec.c from 2.6.31.1
>>>>>>
>>>>>>
>>>>>> Should I start bisecting, or do you have any other suggestion?
>>>>> There is only one patch which touch ec.c, 
>>>>> 2a84cb9852f52c0cd1c48bca41a8792d44ad06cc. You may try to revert it, 
>>>>> and see.
>>>>>
>>>> Isn't that what pretty much what I did when I built a kernel with 
>>>> the 2.6.31.1 ec.c ?
>>>>
>>>> Anyway, after I applied your patch you just posted on acpi-devel:
>>>> ACPI: EC: Rewrite DMI checks
>>>>
>>>> to the 2.6.32-rc1 tree the:
>>>>
>>>>> -ACPI Error (psargs-0359): [\_PR_.CPU0.CSTX] Namespace lookup 
>>>>> failure, AE_NOT_FOUND
>>>>> -ACPI Error (psparse-0537): Method parse/execution failed 
>>>>> [\_SB_.PCI0.LPCB.EC0_._Q2F] (Node ffff88013f81c1a0), AE_NOT_FOUND
>>>> does not show up at boot anymore, and so far I haven't seen the 
>>>> other errors either...
>>>>
>>>> I'll post a follow-up if they show up again...
>>>>
>>> Bah...
>>> I spoke too soon:
>>>
>> Could you please turn on timing information in printk ? Please also 
>> uncomment "#define DEBUG" in ec.c
>>
> 
> That generated a big log......
> 
> This is now grep ACPI /var/log/syslog from the 2.6.32-rc1 + the above 
> "ACPI: EC: Rewrite DMI checks"
> 
> As it's big, I posted it here:
> http://tmb.mine.nu/acpi-ec-bug/acpi-ec-debug.bz2
Thanks,
please check if attached patch helps:
> 
> -- 
> Thomas

Patch

ACPI: EC: Restart command even if no interrupts from EC

From: Alexey Starikovskiy <[e-mail]>

Signed-off-by: Alexey Starikovskiy <[e-mail]>
---

 drivers/acpi/ec.c |    4 +---
 1 files changed, 1 insertions(+), 3 deletions(-)


diff --git a/drivers/acpi/ec.c b/drivers/acpi/ec.c
index da7da37..9c34515 100644
--- a/drivers/acpi/ec.c
+++ b/drivers/acpi/ec.c
@@ -234,10 +234,8 @@  static int ec_poll(struct acpi_ec *ec)
 			}
 			advance_transaction(ec, acpi_ec_read_status(ec));
 		} while (time_before(jiffies, delay));
-		if (!ec->curr->irq_count ||
-		    (acpi_ec_read_status(ec) & ACPI_EC_FLAG_IBF))
+		if (acpi_ec_read_status(ec) & ACPI_EC_FLAG_IBF)
 			break;
-		/* try restart command if we get any false interrupts */
 		pr_debug(PREFIX "controller reset, restart transaction\n");
 		spin_lock_irqsave(&ec->curr_lock, flags);
 		start_transaction(ec);