[ 3.814717] [Firmware Bug]: powernow-k8: Your BIOS does not provide ACPI _PSS objects in a way that Linux understands. Please report this to the Linux ACPI maintainers and complain to your BIOS vendor. [ 3.814803] [Firmware Bug]: powernow-k8: Your BIOS does not provide ACPI _PSS objects in a way that Linux understands. Please report this to the Linux ACPI maintainers and complain to your BIOS vendor. [ 3.814889] [Firmware Bug]: powernow-k8: Your BIOS does not provide ACPI _PSS objects in a way that Linux understands. Please report this to the Linux ACPI maintainers and complain to your BIOS vendor. [ 3.814974] [Firmware Bug]: powernow-k8: Your BIOS does not provide ACPI _PSS objects in a way that Linux understands. Please report this to the Linux ACPI maintainers and complain to your BIOS vendor. This message is shown with an ASUS M3N78-EM motherboard, there is a quad core cpu so the message is shown x4. Also there is a trouble with the linux acpi sub-sytem that I have explained here https://bugs.launchpad.net/bugs/392938 and also reported in nvidia forums but no one response. The only thing I haven't done is to contact with ASUS support. I have the 0604 BIOS version of this mobo. Thanks in advise.
This is a BIOS bug. And it had better be fixed by BIOS ugprading. We see the above warning message on a lot of boxes based on AMD chipset. Thanks.
(In reply to comment #1) > This is a BIOS bug. And it had better be fixed by BIOS ugprading. > We see the above warning message on a lot of boxes based on AMD chipset. > > Thanks. Thanks for the reply I supouse I have to ask to ASUS to support it or I will have to make my own DSDT, but, this is only about the [Firmware Bug] messages, what about the other issue? With acpi=on and nvidia driver no sound from interl or any dedicated sound card, but, when reentering to a console (control+alt+f1) sounds normal, only when the nvidia driver is outputing.
Hi, David I have no idea about the sound issue. You can open a new bug in drives/sound category. Maybe you can get help from there. Thanks.
This bug has been moved so i suppose is now needed to open a new one, or is it? Thanks for your support, in the other sites where I have reported the issue nobody response.
I have seen that the problem was that I had acpi=off due to the nvidia-acpi sound bug, and when acpi=on the messages claiming Firmware Bug disappear. Won't be better to check if acpi is deactivated and to show a message like acpi is off skiping powernow before trying to use acpi capabilities or warning the user about it? Thanks.
(In reply to comment #5) > I have seen that the problem was that I had acpi=off due to the nvidia-acpi > sound bug, and when acpi=on the messages claiming Firmware Bug disappear. > > Won't be better to check if acpi is deactivated and to show a message like > acpi > is off skiping powernow before trying to use acpi capabilities or warning the > user about it? > sorry I don't quite understand your question, do you mean the we should not print the Firmware Bug when acpi=off? please double check that the warning message only appears when acpi=on. because the powernow-k8 driver will not be loaded when acpi=off.
I have checked it, sorry my bad explanation. The thing is, when acpi=on no message at all and powernow seems ok. When acpi=off the Firmware Bug message appear. I have acpi to off due to the sound bug I commented, but recently, I think about acpi should be on to powernow to work and tested it, thats why I think the message shouldn't be shown when acpi=off because it's confusing.
Created attachment 22370 [details] patch: prevent powernow-k8 driver from being loaded when acpi=off please try this patch and see if it helps.
Great work man! When acpi=on powernow continue working: # cat dmesg.0 | grep powernow [ 0.000000] Linux version 2.6.28.9-powernow-patch (root@multimedia) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #1 SMP Thu Jul 16 12:50:28 CEST 2009 (Ubuntu 2.6.28-13.45-generic) [ 8.834888] powernow-k8: Found 1 AMD Phenom(tm) 9650 Quad-Core Processor processors (4 cpu cores) (version 2.20.00) [ 8.834924] powernow-k8: 0 : pstate 0 (2300 MHz) [ 8.834927] powernow-k8: 1 : pstate 1 (1200 MHz) When acpi=off, no message at all: # zcat dmesg.1.gz | grep powernow [ 0.000000] Linux version 2.6.28.9-powernow-patch (root@multimedia) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #1 SMP Thu Jul 16 12:50:28 CEST 2009 (Ubuntu 2.6.28-13.45-generic) Displaying the message confuses because acpi_pss objects are ok in the bios and the real problem for powernow to init was that acpi was off. I think if a message showing something like: [ 8.834888] powernow-k8: ACPI is off powernow won't start could be better for people who don't know that powernow needs acpi but that as you wish, now is perfect :) Only four things I want to mention: 1.- I have tested it with 2.6.28.9 kernel and I had to patch it by hand because with powernow-k7.c: # cat ~/Escritorio/patch-13695 | patch -p1 --dry-run patching file arch/x86/kernel/cpu/cpufreq/powernow-k7.c Hunk #1 FAILED at 652. 1 out of 1 hunk FAILED -- saving rejects to file arch/x86/kernel/cpu/cpufreq/powernow-k7.c.rej patching file arch/x86/kernel/cpu/cpufreq/powernow-k8.c Hunk #1 succeeded at 1312 with fuzz 2 (offset -124 lines). 2.- I have only test powernow-k8 because of the cpu but powernow-k7 should work with the patch. 3.- I don't know anything about kernel programming and I don't know the source structure but if there is somewhere where powernow-k7 and powernow-k8 are loaded acording the system cpu, shouldn't be better to check there if acpi is enabled? 4.- Thanks again for your work. I leave the status as NEEDINFO if you consider that someone has to test powernow-k7. I have a k7 athlon but is "off-line" now, anyway if is necessary I can assemble it and test.
The issue with the sound is now solved, the problem was the nvidia driver, the lastest beta one 190.16 from nvidia, works perfect. I only want to say with this that the alsa and the acpi sub-system in the linux kernel works perfect with this mobo. I haven't yet tested the powernow-k7 patch.
should we close this track?
The problem with nvida that required acpi=off is now gone. the remaining problem is that the powernow-k7 driver complains when loaded and acpi=off. this should be addressed by rui's patch in comment 8, but I'm re-assigning this to cpufreq sub-system, since the messages from their driver are their business.
Yes I forget to tell it but nvidia problem solved, with kernel 2.6.28 and nvidia driver 190.xx beta nowadays sound works but no with 180.xx or 185.xx and with kernel 2.6.30 or .31, I don't remember, doesn't matter the nvidia driver version, sound works will all of them.
Ah, also, there is a better patch, I copy here an email from Thomas Renninger <trenn@suse.de>: >On Thursday 23 July 2009 04:11:10 Zhang Rui wrote: >> when acpi=off, powernow-k8 driver is still loaded, and fails with >> a firmware bug warning message. >> http://bugzilla.kernel.org/show_bug.cgi?id=13695#c5 >> >> this is misleading, >> we should not load powernow-k8 driver when acpi=off. >> http://bugzilla.kernel.org/show_bug.cgi?id=13695 > >Maybe this one is better: > - It allows these special, deprecated PSB BIOS structure > use in acpi=off case > - It complains that powernow-k8 won't work without ACPI, instead > of complaining about the BIOS missing _PSS ACPI structures > (in acpi=off case). > >Compile tested on latest 2.6.31-rc2. > > Thomas > >---- >CPUFREQ: Avoid powernow-k8 BIOS complains when acpi=off > > - It still allows these special, deprecated PSB BIOS structure > use in acpi=off case > - It complains that powernow-k8 won't work without ACPI, instead > of complaining about the BIOS missing _PSS ACPI structures > (in acpi=off case). > >Signed-off-by: Thomas Renninger <trenn@suse.de> > >--- > arch/x86/kernel/cpu/cpufreq/powernow-k8.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > >Index: cpufreq_governor_cleanups/arch/x86/kernel/cpu/cpufreq/powernow-k8.c >=================================================================== >--- cpufreq_governor_cleanups.orig/arch/x86/kernel/cpu/cpufreq/powernow-k8.c >+++ cpufreq_governor_cleanups/arch/x86/kernel/cpu/cpufreq/powernow-k8.c >@@ -1286,7 +1286,11 @@ static int __cpuinit powernowk8_cpu_init > * an UP version, and is deprecated by AMD. > */ > if (num_online_cpus() != 1) { >- printk_once(ACPI_PSS_BIOS_BUG_MSG); >+ if (acpi_disabled) >+ printk_once(KERN_WARN PFX >+ "ACPI disabled, exiting\n"); >+ else >+ printk_once(ACPI_PSS_BIOS_BUG_MSG); > goto err_out; > } > if (pol->cpu != 0) { I don't know anything about powernow but if powernow-k7 can support the same conditions that powernow-k8, the patch for k7 also could be necessary to update the conditionals.
Finally which code has been applied for poernow-k8? powernow-k7 has also been fixed with the same conditions?
i don't see this patch upstream.
a0abd520fd69295f4a3735e29a9448a32e101d47 cpumask: fix powernow-k8: partial revert of 2fdf66b49 shipped in 2.6.29. is this issue fixed?