Bugzilla – Bug 104088
The asus_acpi module doesn't work on Asus A6726GU laptop
Last modified: 2007-06-05 09:57:53 UTC
The module doesn't handle the display properly. It being loaded results in - a lot of records in the log (/var/log/messages) - big CPU load - Messages blinking on the display This will be problem since the vendor-specific modules will be loaded by default (according to a mail from suse-dist). Timo already has a solution tested on affected laptop.
Here it goes. The upcoming patch adds support for the laptop to the ASUS ACPI driver. The patch is against our beta 1 kernel which includes the mainline ASUS ACPI driver. I'll send the patch to the driver developers, too. Please review.
Created attachment 45719 [details] Patch to add support for ASUS A6G models
I wonder why it was loaded at all before adding the A6G string? Or did it fall into some generic A6x case which does not comfort the A6G?
It has not returned -ENODEV because it detects that it is a ASUS laptop. Hence, the module was running with "default" settings: if (hotk->model == END_MODEL) { printk("unsupported, trying default values, supply the " "developers with your DSDT\n"); hotk->model = M2E; This might work for other models. If we don't like it we could return -ENODEV. But in my opinion this is way to conservative. If one runs into problems we're happy to know and fix it (like I did for Jiri's machine).
Sounds sane. I build a kernel (there were also other problems) for Jiri (queen-trenn-12). We should add it if it works for him.
Thomas, I will be on vacation on Monday. So if testing shows that the fix is fine, please submit to the kernel CVS or have some other hacker do it. Timo, one remark: When submitting patches into bugzilla, please mark them as "patch" (there is an explicit checkbox for that). Makes life for many people easier as one does not need to download the patch before being able to look at it.
(In reply to comment #6) > Timo, one remark: When submitting patches into bugzilla, please mark them as "patch" > (there is an explicit checkbox for that). Makes life for many people easier as > one does not need to download the patch before being able to look at it. Yes, will do next time. I found that option right after attaching the patch last time... If we are accepting this patch, I'd like to rework it once more to support another three additional ASUS laptops. I will look for people having these models for testing, otherwise I won't add those models to the upcoming patch.
Please note that this bug also exists with my Asus V6800V (V6V). The symptoms are exactly the same. The driver uses some generic support, since this is an unsupported model, and this creates chaos ;) Hope it gets fixed also for my V6. Note: I didn't try Beta2, but I'll sure download Beta3 when it's out.
Great! Good to have someone with a V6. If submitted an updated patch which also includes support for the V6. Though, this patch will _not_ yet be included in Beta 3. It would be nice if you could either test Beta 4 (which hopefully will be released with the updated patch) or I can attach the updated patch to bugzilla so you can compile and test the module yourself.
As I'm completely new to the whole SuSE world, I'm not sure I'll dare re-compiling the whole kernel from SRPMS, which seems way more complicated than just downloading a tarball off kernel.org and building it... But anyway, as I really hope SuSE 10.0 works perfect on this laptop, I'll be sure to try all the future betas, and 'abuse' bugzilla acordingly, trust me... ;)
DAnny, please explain the kernel-of-the-day...
(In reply to comment #10) > As I'm completely new to the whole SuSE world, I'm not sure I'll dare > re-compiling the whole kernel from SRPMS There will be no need for that. We have KOTDs (kernel of the day) you can simply install. No need to compile anything. I'll let you know once we got the patch in our tree.
Reassign to thoenig. Assign it to a kernel hacker with kernel cvs write rights (mantel/pavel/ak) if you need a patch in the kernel.
Jean-Christophe: Could you please install Beta 4 an let me know whether the module works for you now?
Is there a link with https://bugzilla.novell.com/show_bug.cgi?id=114960 ?
Basically, yes. The current problem is that the driver will load on unsupported ASUS laptop with some default values which are not sane. Thus, reading the LCD brightness triggers an error and causes the writes errors syslog. Since kmilo of KDE polls the LCD brightness at a insane rate the syslog gets bloated. We will try to support as many ASUS laptop by the driver and drop the fallback option of the acpi_asus module to load with default values for unsupported systems.
Jiri, can we close this one?
Yes, I think we can close this one (as it si not related to the ec-burst mode which still doesnt' work :-( )
You're referring to the delay of hotkeys on ASUS laptops?
Yes, exactly.