Talk:Asus A6T

From Gentoo Linux Wiki

(Redirected from Talk:HARDWARE Asus A6T)
Jump to: navigation, search

hi all.

any ide for the make.conf? and use flag?

Use any flags you want (don’t forget to set ‘nvidia’). Set ‘-j3’ in ‘MAKEOPTS’.
I use these ‘CFLAGS’: ‘-O2 -march=k8 -mmmx -msse -msse2 -msse3 -m3dnow -mfpmath=sse -momit-leaf-frame-pointer -minline-all-stringops’
--Sten 23:16, 12 November 2006 (UTC)

Contents

[edit] BIOS versions?

Now about BIOSes.

<04xx series have stability problems.

<08xx revisions have serious problems with USB

=0802 has more ACPI bugs.

Which one are you using? is there some bios working without problems? I'm also trying to get GeForce and Realtek8168 on different IRQs - usually share IRQ 17 or 19. I guess IRQ sharing is causing network hangs&x11 crashes.

[edit] USB problem

Any ide about this?


irq 7: nobody cared (try booting with the "irqpoll" option)

Call Trace: 

[<ffffffff801b07ec>] __report_bad_irq+0x30/0x7d
[<ffffffff801b0a26>] note_interrupt+0x1ed/0x22e
[<ffffffff801aff33>] __do_IRQ+0xc7/0x105
[<ffffffff80165a1a>] do_IRQ+0x65/0x73
[<ffffffff80163c09>] default_idle+0x0/0x50
[<ffffffff8015a04d>] ret_from_intr+0x0/0xa

handlers: [<ffffffff88047b90>] (usb_hcd_irq+0x0/0x55 [usbcore]) Disabling IRQ #7


if boot whit "irqpoll" the system frezze

This means IRQ #7 is handled by USB (‘ehci_hcd’ module to be exact) but the IRQ was not generated by USB EHCI and the USB core reported that it does not belong to USB. Just ignore it. It is probably caused by the webcam or IrDA or something else not working or working partially.
IRQ polling does not work because the APIC is broken.
--Sten 23:16, 12 November 2006 (UTC)
Can this be the reason my usb-sticks are not working? USB modules enabled as documented (and working on other systems),"/dev/sda1 does not exist"? I actually could mount my USB-stick when i used vanilla, after i stepped over to suspend2-sources not anymore (not saying that that was the reason).
thanks, bart
This concerns Linux <= 2.6.18 with only the 'noapic' option. I don't know if 'pci=assign-busses' addresses the problem.
Anyway, with 'acpi=noirq' instead of 'noapic', USB works great.
So... what about changing 'noapic' by 'acpi=noirq' in the wiki?
--Lcld 00:44, 22 December 2006 (UTC)

[edit] Ethernet problem

Thanks for the guide!
After editing 'src/r1000_n.c' I still get an error:
any idea?

$ make modules clean
ke -C src/ modules
make[1]: Entering directory `/root/r1000_v1.04/src'
make -C /lib/modules/2.6.17.7/build SUBDIRS=/root/r1000_v1.04/src modules
make[2]: Entering directory `/usr/src/linux-2.6.17.7'
make[3]: Warning: File `/root/r1000_v1.04/src/r1000_n.c' has modification time 2.9e+03 s in the future
CC [M] /root/r1000_v1.04/src/r1000_n.o
/root/r1000_v1.04/src/r1000_n.c: In function '__check_speed':
/root/r1000_v1.04/src/r1000_n.c:51: warning: return from incompatible pointer type
/root/r1000_v1.04/src/r1000_n.c: In function '__check_duplex':
/root/r1000_v1.04/src/r1000_n.c:53: warning: return from incompatible pointer type
/root/r1000_v1.04/src/r1000_n.c: In function '__check_autoneg':
/root/r1000_v1.04/src/r1000_n.c:55: warning: return from incompatible pointer type
CC [M] /root/r1000_v1.04/src/r1000_ioctl.o
LD [M] /root/r1000_v1.04/src/r1000.o
make[3]: warning: Clock skew detected. Your build may be incomplete.
Building modules, stage 2.
MODPOST
CC /root/r1000_v1.04/src/r1000.mod.o
LD [M] /root/r1000_v1.04/src/r1000.ko
make[2]: Leaving directory `/usr/src/linux-2.6.17.7'
make[1]: Leaving directory `/root/r1000_v1.04/src'
make -C src/ clean
make[1]: Entering directory `/root/r1000_v1.04/src'
rm -f *.o *.ko *~ core* .dep* .*.d .*.cmd *.mod.c *.a *.s .*.flags
make[1]: Leaving directory `/root/r1000_v1.04/src'

That’s warnings only. Ignore them and just run ‘make’ and ‘make install’.
--Sten 10:37, 13 November 2006 (UTC)
ok, i forgot 'make' and went straight to 'make install'. 'make' gives me the error:
install: target '/lib/modules/2.6.17.7/kernel/drivers/net/' is not a directory: No such file or directory
so i make this net directory. Then run make and make install without errors. r1000 loads on boot (added to make.conf)
Its's working now! so i can start with the rest of the installation :)
thanks!, bart
I don’t think making that directory is a good idea. You should install your kernel sources (eg. sys-kernel/gentoo-sources) and run ‘make oldconfig’ in ‘/usr/src/linux’ otherwise some module dependencies may be missing (especially when using non-Gentoo-default kernel). But if ‘depmod -a’ returned no error and the module loads, it should work.
--Sten 13:22, 13 November 2006 (UTC)
In that case i think i was lucky :). Because i did not understand how the livedvd installed the kernel (although now i think it was genkernel), and since i'm used to building the kernel myself (on my old ppc laptop), i emerged vanilla-sources (by first downloading the source files and putting these in the distfiles). forgot totally about make oldconfig, but 'depmod -a' gives (luckily) no errors.

[edit] fan problem

Hi there! I've got a problem with my fan: When I boot my cold laptop then fan keeps running all the time, even if CPU temperature stays under 50 degrees celcius. However, when my laptop once run hot and cools down again, the fan stops turning at about 53 degrees. But often my laptop doesn't run that hot and the fan keeps running all the time, even with the CPU temperature at 43 degrees. Did you experience the same?

[edit] STR

I use gentoo-sources 2.6.19 and have them patched with Suspend2. I use nVidia drivers 1.0.9742. However I cannot resume (them same problem as I had on older kernels) even when suspending from console in single-user mode :( Using ‘noapic’ does not help. Has anyone of them who make it work modified something somewhere in order to make it work? --Sten 02:22, 6 December 2006 (UTC)

[edit] TV Out

What about tv-out? Is it working?

Yes, without problems. Just fill the appropriate configuration as described in nvidia-drivers documentation (don't forget to specify TV standard otherwise you'll probably end up without colors). I think xv does not work on it but I'm not sure. And try using lower resolutions, 1024×768 is a bit fuzzy on TV screen. ---Sten 23:26, 17 April 2007 (UTC)
I have checked it: xv works on it in all cases: TwinView Clonning, TwinView Dual Monitor, and Xinerama Dual Monitor (ie. w/o TwinView). With both TwinView also OpenGL works (not tested w/o it). And the best resolution for PAL TV is 800×600. --Sten 18:57, 1 May 2007 (UTC)

[edit] Additional information and common reply to the 3 previous questions

I've written a page about my laptop (A6Tc-AP014H under Debian/Sid) at http://jmuchemb.no-ip.com/~jm/trac/wiki/A6Tc-AP014H

Major differences with the wiki:

  • wifi: The bcm43xx driver doesn't perform as bad as written in the wiki and I would recommend it over ndiswrapper. I've patched my kernel to address 2 issues:
    • The driver doesn't adapt the speed according the quality of the signal. I change the default speed to the minimum one in order to connect when the signal is weak. The data rate can still be increased afterwards.
    • When too many packets are lost, the driver stops to work and makes the system unstable. My patch solves that.
  • lid: The wiki doesn't mention it but it is ploblematic on my laptop.
  • display: 1280x800 console is possible with the 'nvidiafb' driver (incompatible with 'nvidia' though). nvidia-stable-9629 works with Suspend2.
  • STR: It still crashes on resuming with Linux 2.6.19, but with different symptoms than with previous kernels (the hard disk led isn't on anymore). The config file of a laptop that succeeds to resume might help me.
  • Hard disk: I don't have a Fujitsu one...

Previous question about:

--Lcld 01:21, 22 December 2006 (UTC)

[edit] Strange crashes, maybe a kernel bug (or something with r1000?)

Many people with A6TCs experience strange crashes - Everything just freezes and you can only move the mouse. No debug info, nothing. I tried several methods to take this crash from console (to see kernel lines or so..). I suspect r1000 driver (because the network usually makes problems just before it crashes). This bug is unpredictable, unreproducible, usually happens after some higher loads.

The second one, happens just after grub loads the kernel - system fails to switch to the framebuffer, or fails just after - screen is black and nothing more happens. This is an apic problem, not related to the first, if you boot without the fb and without quiet you'll see the error message. I sometimes get this problem after a restart, and never after a first boot

Kernels I tested are 2.6.17-gentoo-r3, 2.6.18-gentoo, 2.6.19-gentoo, 2.6.19-suspend2, 2.6.20. Every of them crashes.

If anyone had this solved, please help. thx. [exa]

If you mean you can move cursor but everything else is frozen, then it is a problem with X.org (probably DRI or something else that does not affect hardware cursor). The non-starting framebuffer might be either the same problem or some problem with ACPI/APIC. Check if your BIOS is newer then 04xx. Also try using non-Gentoo kernel (emerge vanilla-sources and patch them with patches you’d like, eg. fbsplash and suspend2). Gentoo patchsets have many strange issues on AMD64 SMP. --Sten 11:51, 5 January 2007 (UTC)

- I've done some testing and it appears that no matter which ethernet driver you choose (r1000 from realtek or r8169 from the kernel) the same thing happens. When you reach 100% of UPSTREAM bandwidth (problem does not happen when D/S does.. whatever) of your ethernet card the module just crashes -- and usually since nvidia is on the same IRQ everything ( = xorg) dies too. Strange thing is that I this 100% upstream can be 10Mbps or 100Mbps (1Gbps not tested). -- ZaR

--- Are you sure that you're really crashing? I noticed that waiting some minutes, clicking around, i get the system to work again. Also, you may log from a remote shell (ssh) and kill things. - Ooops, i have this problem too.

I solved by unlinking the dvd drive from the file system (unlink /dev/hda) --


I experienced this also (among problems above after the freeze the LCD won’t come up after turning off and the xv video overlay and SDL stopped working while log reported something about problems on AGP and therefore switching to PCI) and fixed it by using ‘acpi=noirq’. I also switched back to agpgart but I’m not sure whether it is required for the fix. I’m using the newest mm kernel patched with Suspend2 and fbsplash and the newest nVidia drivers and everything works just fine now. --Sten 13:14, 3 April 2007 (UTC)

Solved here too with ‘acpi=noirq’ on gentoo kernel 2.6.20. agpgart changes not required for the fix.

I noticed that ethernet driver can cause random hard lockups (when I'm not connected to the net, the system is pretty stable). Sometimes after that the power button on the notebook blinks, sometimes not. I use vanilla-sources-2.6.23, nvidia-drivers-169.04, bios 0702, the kernel commandline is: acpi=noirq pci=assign-busses (no idle=poll option, because it increases cpu temperature by 30 degrees) -flesh1911

OK, I solved the crashes that I described here (above) a long time ago, so I could finally share it. Let's just modify the kernel in the way Realtek shows, in file drivers/net/r8169.c:

  1. define NUM_TX_DESC 1024 /* Number of Tx descriptor registers */
  2. define NUM_RX_DESC 1024 /* Number of Rx descriptor registers */

...it is not a real solution, but totally avoids the crash condition. You need to run ~8 glxgears and do some 30MB/s upload to crash the thing again. (which is mostly unspeakable condition, agreed? ;D )

For kernel devs, it's a classical kernel oops, too bad no one is able to get any panic/oops info because nvidia driver locks the display hard. Maybe testing with nv or noveau drivers could show something.

-- [exa] @ 2008-09-03 ...and well im adding this to wiki because it helped in 100% cases.

[edit] NOTE NOTE NOTE NOTE NOTE

ASUS a6t Crash using some rams Attention RAM: im change the original 1gb of ram to 2gb "corsair" on my asus a6t ... and have very much problem.. frezzes crashes.. with the original ram i don't have any problem im have test the 2gb "corsair" using memtest86 for 48hs result 0 errors. the ram work fine on other laptop this asus have hardware incompatibility with some rams. now i testing 2gb "S3"

The manuals says: 2 x 1GB chips max. So is that corsair or s3 a single 2GB chip? -- not a very good idea.

You're not the only one with problems after having added memory. See [1]. I have the same problem, and I asked the shop to upgrade the memory to 2GB when I bought the laptop, so I don't even know what memory they used. As memtest runs fine and the crashing happens when I use the network, USB or sound card (if I don't, it will crash very rarely), I thought it was a driver and/or BIOS problem. I have upgraded the BIOS to 0802 (the original was 0403), to no avail (I've tried various kernels and kernel parameters too). Then thinking that this is a BIOS problem I tried booting with "mem=1024M" and can't say for sure yet whether it works, but at least it has now survived an amount of network use that would always crash before. The crashes are total lockups, reboots or at times the system is in a half-responsive state due to the hard drive becoming inaccessible (it reports being busy and kernel keeps trying to reset it). A few times at boot the BIOS has even said that its checksum is incorrect and reset the settings. -- 130.233.24.129 07:45, 1 March 2007 (UTC)
"mem=1024M" didn't fix it, it was just one in a long line of attempts that make only a slight difference. What I eventually ended up with was "pci=assign-busses no-hlt". "pci=assign-busses" alone is not enough. I haven't tried "no-hlt" alone. I use 32-bit 2.6.20 on a6tc. I first thought that "no-hlt" would run too hot, but that's not the case unless you also disable acpi. -- 130.233.24.129 08:17, 24 March 2007 (UTC)
Another update: It was not the hard drive but the cd-rom that becomes inaccessible sometimes, and "no-hlt" does not help with that particular problem[2], although the more serious ones disappear. -- 130.233.24.129 14:25, 25 March 2007 (UTC)
True True: After time I ended up with "noapic irqpoll" but now I'm faceing the terrible consequences of cdrom failures... don't really have to wait long for a dmesg flood of:

hda: packet command error: status=0xd0 { Busy } ide: failed opcode was: unknown hda: cdrom_pc_intr: The drive appears confused (ireason = 0x01). Trying to recover by ending request. hda: status error: status=0x58 { DriveReady SeekComplete DataRequest } ide: failed opcode was: unknown hda: drive not ready for command hda: status error: status=0x58 { DriveReady SeekComplete DataRequest } ide: failed opcode was: unknown hda: drive not ready for command hda: status error: status=0x58 { DriveReady SeekComplete DataRequest } ide: failed opcode was: unknown hda: drive not ready for command ...the last one keeps coming and coming, plus ofcourse sometimes same thing happens with the hdd, but that's a quick death - so no dmesg. Also with noapic there is no extended irq range and dmesg shows a warning about sporius IRQ 15 (thats the second IDE controller - the one where I have my hdd - not the cdrom) -- Zar

[edit] Experiences with A6Tc

I have A6Tc, with BIOS upgraded to latest version, kernel 2.6.19-gentoo, 64-bit. STR and STD works, but only with SWSUSP1 and latest nVidia drivers, SWSUSP2 is totally broken. I was able to get around problems with timers by adding "acpi_use_timer_override pci=assign-busses" to kernel boot line, now, it never happens that laptop dies with timer error. But I suffer from following:

bcm43xx - is just great, but can't be used to connect to APs with low signal, capped at 11Mbitps

ndiswrapper - offers superior speed with 64-bit drivers, but randomly after transferring a lot of data, puts the kernel to panic, not regarding whether nvidia module is loaded or not. (i suspected nvidia, but it is innocent)

realtek gigabit - almost unusable, dies at random, it shares IRQ with nvidia, when nvidia is not loaded, drivers dies alone and can be reloaded to trasmit a few more megs, but keeps dying. when nvidia is loaded, it takes it down too, hard reboot is inevitable then.

[edit] Modem

The wiki page says that the built-in modem based on the Si3054 chip works using the hsfmodem driver. Did anybody really got it working? I tried the hsfmodem and failed to get it working. It's not surprising as the modem is not HSF-based. Si3054 is, in theory, supported by ALSA and slmodemd, but I didn't manage to get this setup working either: the device is detected as an ALSA device hw:0,6, but when I try to connect, I get NO CARRIER. Googling gave me nothing - I found only postings from people with similar problem but no solution whatsoever. Any hints on getting the modem working anybody?

[edit] 0806 BIOS

Where did you get the 0806 BIOS from? I’ve searched both ASUSTek and Google and found only 0701 and older. --Sten 22:52, 23 July 2007 (UTC)

You can get latest version of BIOS here: http://support.asus.com/download/download_item.aspx?model=A6Tc&product=3&type=BIOS&SLanguage=en-us

Does the BIOS for A6Tc also works in A6T?

[edit] Wifi

With 2.6.22-1-amd64 (under debian/sid), bcm43xx is working (using WPA-TKIP)

[edit] BIOS 0702 (A6T/Z92T users) - Worth a mention?

For those having problems and not heard of BIOS 0702 (its been around for 8 months), then continue, ... if you aren't having problems, don't!


Anyway, you may want to forget the A6Tc BIOSes (I'd be interested to hear if they work), instead, you might want to try BIOS 0702, yes I know, Asus were rather hush hush about it, but it gives you decent access to the ACPI. Being a Windows user (slowly going over to Linux, just that to get PowerNOW! to work the old way, the coding scared me a bit), I've only tried using the Windows power options, so far, so good, except, the only problem I had, was the Wifi sticking in XP64.

Linky: http://vip.asus.com/forum/view.aspx?id=20070315175249279&board_id=3&model=A6T&page=1&SLanguage=en-us

I wouldn't be supprised if everyone didn't know about it, you have to of gone through the several thousand messages of questions to see why power now doesn't work on Vista, oh and the link is in the forum (which can be touchy, German needed!)

If anyone had a complaint about Asus TS, I would have to agree with some lack of it. The hell to get Vista working on my Z92T was hell, when my Clevo M37EW was too damm easy. Although, we are comparing 2 different flavors as well, 32bit vista vs 64bit vista. Now to see what Linux looks like on both, I've tried Ubuntu a couple of months ago (with the old 0403 BIOS - yes, by the time I updated it, it was about 9 months out of date!) and it was favourable, trying Kubuntu now and I am glad the PowerNOW! works with this BIOS.

Solved everyones problem or can I expect a barrage of abuse for my near n00bishness? ;-)

- Wessex_Electric_Nutter

[edit] New nvidia drivers & Vanilla-sources

Notice that nvidia-drivers 169.04 (which are discussed at http://forums.gentoo.org/viewtopic-p-4538087-highlight-.html?sid=a339238f6d1a99ac4a87bcce55bb237c) seem to greatly improve stability without idle=poll kernel option. Gentoo-sources 2.6.23 still contains SMP bug, causing the system to freeze at high loads (for example, running kvm), that can be fixed with maxcpus=1 kernel option, but this leaves you with only half of the cpu performance. Vanilla sources 2.6.23 doesn't have that bug. So now I'm on vanilla-sources-2.6.23 and nvidia-drivers-169.04, without idle=poll kernel option. As a result, the CPU temp is lower than before (idle=poll option adds ~10C-25C to the CPU temperature, depending of the CPU frequency), so as the graphic card temperature: was 70C+, now 65C.

- flesh1911

[edit] DSDT

Should we add a section for different fixed dsdt's on the page? I have an asus A6T myself and fixed the dsdt. the fixed dsdt is here: http://incoming.kallenberg.dk/dsdt.tar.bz2

I had a problem with fixed DSDT (not yours), compiled into kernel - after loading the asus_acpi kernel driver the screen brightness becomes about 10% - I can hardly see anything. Did you had such issues? -flesh1911

Retrieved from "http://gentoo-wiki.com/Talk:Asus_A6T"
Personal tools