noticed this the other night while fiddling around with a test kernel. i kept getting timeout errors on my mouse and other USB devices. finally tracked it down to having APIC enabled in the kernel. i disabled APIC and viola!, everything worked fine. probably outta post to the kernel mailing list also. Platform: AthlonXP 1800+ on VIA VT8367(KT333) / VT8233A motherboard 512MB RAM, tulip network card driver, no IRQ conflicts
what kernel?
Let me know what kernel you had this problem on, and then test on the latest lolo-sources. Thanks.
I have the same problem but I localized it. It is NEW acpi code problem, when I manually reverse patch on kernel all works fine (exсept that from this time I loose all information in /proc/acpi, it's not normal because reverce patching completes without errors ) . I try 2.4.20-xfs_pre5 and 2.4.20-gentoo-r1. BTW, I found there is no way to turn off acpi patch in this kernels and the sentence about patching gentoo-sources with acpi only if USE contains acpi4linux is not true(i found that in patches.txt for gentoo-sources)! In addition KERNEL_EXCLUDE="acpi" brakes emerge on this two type of sources! Seems that you (gentoo kernel team) applying all patches cumulative rather than one by one for making availible KERNEL_EXCLUDE to work as it must.
*** Bug 14715 has been marked as a duplicate of this bug. ***
max: should i reopen this ?
martin: it's good that you ask :) I think yes, this bug need to be reopened.
the bug still here in xfs-sources-2.4.20-r2 i.e. with last ACPI update but KERNEL_EXCLUDE="acpi" now works
glad to hear kernel exclude works. however, some acpi implementations are in need of a bios update in order to fully support the latest acpi patches. if no such update exists from your hardware vendor and you run into acpi related problems you either boot with acpi=off or disable it in your config. Not much more I can offer since the issue really comes down to your hardware supporting the latest acpi spec. Jay
O! thats great! I've just update my BIOS but situation is the same. And how can you explain that ACPI worked with old linux ACPI implementation, worked on windows, and only new linux ACPI code not working with any modern chipset for AMD that i try? Is it Intel sponsorship program? I have this problems with ACPI on Epox 8K9A(KT400) and Asus A7N8X(NForce2). Can you name the motherboard for AMD processors that worked with new linux ACPI code?
here's an answer - cd to a kernel with updated acpi ie. /usr/src/linux-2.4.20-gentoo-r5/drivers/acpi/ or /usr/src/linux-2.4.21-pfeifer-r1_pre2/drivers/acpi grep -in intel * grep -in amd * that ought to let you know who is running the acpi show. Jay
so, if the new ACPI code supports only intel chipsets, why it is not documented anywhere? I write you about exactly next: if some hardware worked under one software version and become unworkable under next new software version than this means that something broken in software. isn't it? btw your comparision is also wrong because of I'm using non AMD chipset and ACPI support independed of the processor manufacturer
i never said it only supports intel only. i was just showing that intel is the company that is providing and backing the acpi4linux project. therefore, chipsets like nforce2, amd, and via may not always have the same support. and those companies may not implent it to spec in their chipset(s). comparison wrong? oh - ok - what did i compare? - nothing. also, if it worked on the older version but not on the newer - that is your problem. get hardware to support the new stuff or get your hardware manufacturer to work with the newer stuff. Who said ACPI is backwards compatible? And alot of manufacturers have incomplete or buggy ACPI support. Again, that is your issue. Jay
thanks a lot but I really don't need to write a bug report to understand that i have a problem! and thanks, now I know than noone in this bugtracking system, really, don't needed of mine reports, because it's not theyer problem. but I hope, that serious commercical organizations, which making the buisness with Linux, don't agree with your opinion. P.S. >grep -in intel * >grep -in amd * this is your words and you are wrong P.P.S. I don't want flaming any longer, you are describe your point of view clearly.
2Jay Hi, i think that it would be intresting for you Alan Cox in his last ac4 patch for 2.4.22 correct that what you claim buggy BIOS and you can be sure that he didn't send me any BIOS updates he is just make the code workable. Now everithing work fine. and only now that bug report can have status RESOLVED and FIXED. Best regards, Max