Acer Aspire 7720 shutdown problems and DSDT fixes

Acer Aspire 7720 shutdown problems and DSDT fixes

[Bernhard Rosenkränzer]

Hi,
I recently got an Acer Aspire 7720 using BIOS 1.43. Obviously, the first thing 
to do was install Linux (currently running 2.6.27-rc4).

Everything is working, but under heavy load, the machine tends to shut down 
without prior warning. Unlike the similar problem described at 
http://bugzilla.kernel.org/show_bug.cgi?id=9167, /proc/acpi/thermal_zone/TZ01/temperature 
does show the correct temperature all the time, and the fan does turn on.

I've looked at the DSDT and found a huge number of obvious bugs (_OSC taking 
the wrong number of arguments and actually having an off-by-one bug when 
looking at the arguments -- looks like someone didn't know ACPI arguments 
start at Arg0 and then increased the argument count by 1 so he could get the 
4th argument in Arg4; Funny comparisons like
If (LOr (LGreater (OSYS, 0x07D0), LLess (OSYS, 0x07D7)))
(not that that could ever not be true...); doesn't even compile.....), but 
fixing them doesn't help.

I've tried all sorts of things to get it to behave, including patching the 
DSDT to actually know about the fact that TZ01 is connected to the CPUs and 
throttling the CPUs may help:

            Method (_PSV, 0, Serialized)
            {
                Return (3552) // 3552 = (273.2+82.0)*10
            }

            Method (_PSL, Package() {
                \_PR.CPU0,
                \_PR.CPU1
            })

            Name (_TC1, 3)
            Name (_TC2, 4)
            Name (_TSP, 20)
            Name (_TZP, Zero)

(This part works nicely by the way)

But at some point, it still shuts down.

Looking at sensors output, the box seems to shut down when it hits 85.0°C, 
which is also what the coretemp module thinks is right (tjmax value) -- but 
the DSDT sets the trip point at 110°C:

            Method (_CRT, 0, Serialized)
            {
                Return (Add (0x0AAC, Multiply (0x6E, 0x0A)))
            }

Guess this explains why I'm getting a sudden shutdown instead of a clean 
shutdown through acpid - adjusting _CRT to return 84°C doesn't fix the 
problem, but does let acpid shut down the system gracefully instead of just 
going down.

However, given that's what the DSDT says, my guess is the box has actually 
been designed to run at 110°C and simply shouldn't be shutting down at 85°C 
without warning - any ideas what may be causing this or how to fix it?

The DSDTs are available at
http://arklinux.org/~bero/dsdt-acer.dsl (Original DSDT)
http://arklinux.org/~bero/dsdt-bugfixes.dsl (Bugfixed version)
http://arklinux.org/~bero/dsdt-newfeatures.dsl (Version with CPU throttling 
etc.)
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Acer Aspire 7720 shutdown problems and DSDT fixes

[Alan Jenkins]

Bernhard Rosenkränzer wrote:
> Hi,
> I recently got an Acer Aspire 7720 using BIOS 1.43. Obviously, the first thing 
> to do was install Linux (currently running 2.6.27-rc4).
>
> Everything is working, but under heavy load, the machine tends to shut down 
> without prior warning. Unlike the similar problem described at 
> http://bugzilla.kernel.org/show_bug.cgi?id=9167, /proc/acpi/thermal_zone/TZ01/temperature 
> does show the correct temperature all the time, and the fan does turn on.
>
> I've looked at the DSDT and found a huge number of obvious bugs (_OSC taking 
> the wrong number of arguments and actually having an off-by-one bug when 
> looking at the arguments -- looks like someone didn't know ACPI arguments 
> start at Arg0 and then increased the argument count by 1 so he could get the 
> 4th argument in Arg4; Funny comparisons like
> If (LOr (LGreater (OSYS, 0x07D0), LLess (OSYS, 0x07D7)))
> (not that that could ever not be true...); doesn't even compile.....), but 
> fixing them doesn't help.
>
> I've tried all sorts of things to get it to behave, including patching the 
> DSDT to actually know about the fact that TZ01 is connected to the CPUs and 
> throttling the CPUs may help:
>
>             Method (_PSV, 0, Serialized)
>             {
>                 Return (3552) // 3552 = (273.2+82.0)*10
>             }
>
>             Method (_PSL, Package() {
>                 \_PR.CPU0,
>                 \_PR.CPU1
>             })
>
>             Name (_TC1, 3)
>             Name (_TC2, 4)
>             Name (_TSP, 20)
>             Name (_TZP, Zero)
>
> (This part works nicely by the way)
>
> But at some point, it still shuts down.
>
> Looking at sensors output, the box seems to shut down when it hits 85.0°C, 
> which is also what the coretemp module thinks is right (tjmax value) -- but 
> the DSDT sets the trip point at 110°C:
>
>             Method (_CRT, 0, Serialized)
>             {
>                 Return (Add (0x0AAC, Multiply (0x6E, 0x0A)))
>             }
>
> Guess this explains why I'm getting a sudden shutdown instead of a clean 
> shutdown through acpid - adjusting _CRT to return 84°C doesn't fix the 
> problem, but does let acpid shut down the system gracefully instead of just 
> going down.
>
> However, given that's what the DSDT says, my guess is the box has actually 
> been designed to run at 110°C and simply shouldn't be shutting down at 85°C 
> without warning - any ideas what may be causing this or how to fix it?
>   
In theory they sell a version with Linux preinstalled, and are
interested in engaging with Linux generally.  Maybe you can find the
kernel version, patches & Kconfig and see if their kernel is any better.

Alan
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Acer Aspire 7720 shutdown problems and DSDT fixes

[Carlos Corbacho]

On Sunday 24 August 2008 18:48:57 Alan Jenkins wrote:
> In theory they sell a version with Linux preinstalled, and are
> interested in engaging with Linux generally.  Maybe you can find the
> kernel version, patches & Kconfig and see if their kernel is any better.

The "version" they ship is Linpus Linux, without even X configured, in the 
expectation that people are just going to wipe that and put Windows on.

>From what I've seen, Acer just aren't that interested in Linux[1].

-Carlos

[1] The Aspire One seems to be the first and only exception to this so far.
-- 
E-Mail: carlos@strangeworlds.co.uk
Web: strangeworlds.co.uk
GPG Key ID: 0x23EE722D