Most recent kernel where this bug did *NOT* occur: Distribution: Fedora Core 3 2.6.9-1.667 Hardware Environment:1161-43G Software Environment:linux Problem Description:hang on boot Steps to reproduce: boot on a 1161-43G See http://www.redhat.com/archives/redhat-install-list/2002-June/msg00337.html I discovered that this machine will boot up with Fedora Core 3 and successfully initialise the usb support while trying to resurrect this machine to run pvrusb2. I tried to build 2.6.19.1 by copying the config from the working image, running make oldconfig and taking all the defaults but it again hangs again.
Have you had any lack with installation lately? Do you still experience problems that you need to bring up? Thanks.
I found I could in install Fedora Core 3 specifying nousb, but obviously the install has no usb support.
Where does this hang? Are there any log messages? Can you enable CONFIG_USB_DEBUG and see if that gives us some more hints?
I can try but it wont be till next week.
PS I presume you would like me to try the latest kernel.
Yes please.
Sorry for the delay, id been trying upgrade it to FC4 and I had broken it. I reinstalled FC3 and the initial kernel is 2.6.9-1.667. I hadnt realised before but the usb works on this version. I ran yum update and the 2.16.12-1.1381_FC3 gets install and then the usb stops working. Still waiting for the latest kernel to build.
OK, I built the 2.6.22 kernel with USB_DEBUG=y and booted it. It hung at the same place. Where do I find the debug output?
You should probably use the serial console to get the messages from the failed boot - I suppose the hang happens before you get the prompt. It would be helpful to have your system's data such as lsusb, lspci, dmidecode, dmesg from successful boot, etc.
Okey doke.
Created attachment 13253 [details] Output of dmesg, lsusb -v, lspci -v, dmidecode dmidecode didnt say a lot.
Catch 22 - the machine doesnt have a serial port and I dont think the usb/serial adapter will work in this situation.
PS there is a photo of the console at hang time when booting ubuntu at http://www.uluga.ubuntuforums.org/showthread.php?t=572716 Just in case there is any info there that will be of help.
I don't have an account on the Ubuntu forum. Can you just attach the photo to this bug report? Do you have any USB devices attached or builtin which might interfere with the boot-up?
Created attachment 13254 [details] Image of console at hang (on ubuntu) I tried with and without a usb memory stick attached but same result. There isnt much builtin on this machine, its pretty basic. No wireless, no lan, no web cam. This is the spec ... http://www-307.ibm.com/pc/support/site.wss/document.do?sitestyle=lenovo&lndocid=MIGR-4WMLS5
It looks like this is not a USB problem but rather an APIC or ACPI problem. According to your screen image, the system guesses the IRQ line used by the USB controller, and apparently the guess is wrong. That's why it hangs. Natalie, this bug should be reassigned to someone else.
K, thanks.
From the logs it appears that this system doesn't have IO-APIC. Can you please attach dmesg from successful boot also, and /proc/interrupts? I will assign it to ACPI for now. Martin, have you tried booting with acpi=off or pci=noacpi?
Ive tried the acpi=off but not pci=noacpi wilco on the dmesg
Created attachment 13473 [details] dmesg output Tried acpi=off and pci=noacpi on FC6 with a 2.6.22 kernel and xubuntu but still the hang Thanks
Martin, would you please attach output of the following : 1) acpidump 2) lspci -xxxvv thanks.
Created attachment 13632 [details] Output from lspci -xxxvv acpidump wasnt happy ... [root@localhost pmtools-20070714]# acpidump -a 0xEAB80 ACPI tables were not found. If you know location of RSD PTR table (from dmesg, etc), supply it with either --addr or -a option acpidump alone gave the same thing, I used the address from DMESG ... ACPI: RSDP (v000 IBM ) @ 0x000eab80 ACPI: RSDT (v001 IBM AXXXXXX4 0x00000004 IBM 0x00000000) @ 0x0b7f0000 ACPI: FADT (v001 IBM AXXXXXX4 0x00000004 IBM 0x00000000) @ 0x0b7f008c ACPI: BOOT (v001 IBM AXXXXXX4 0x00000004 IBM 0x00000000) @ 0x0b7f0030 ACPI: DBGP (v001 IBM AXXXXXX4 0x00000004 IBM 0x00000000) @ 0x0b7f0058 ACPI: DSDT (v001 IBM AXXXXXX4 0x00000004 MSFT 0x0100000d) @ 0x00000000 ACPI: PM-Timer IO Port: 0xf108 Thanks
Martin, what if you just run the acpidump?
> acpidump alone gave the same thing, You missed the above comment, it was a bit hidden :-) I ran acpidump first, then tried the suggestion from acpidump.
Martin, there is a bug of pmtools-20070714, pls try the lastest acpidump at: http://ftp.ist.utl.pt/pub/kernel/people/lenb/acpi/utils/pmtools-20071116.tar.bz2
Created attachment 13935 [details] acpidump output Worked this time, thanks
It seems like a duplicate of bug 5637, please try the patch in commet #15 in bug 5637 and see if it helps. Thanks.
No didnt help. I applied the patch to 2.6.22 kernel. Tried with no command line options and command line pci=noacpi as per bug 5637 but it hangs initializing the hardware the console displays storage network audio and then hangs. Thanks
Just tried with the latest 2.6.22.9 kernel, same problem. However I did notice a new message when booting with out the pci=noacpi, it said something to the affect DMIBIOS has no date use acpi(force) .... btw thanks for persisting with this, your efforts much appreciated.
(Martin, for completeness please attach the boot trace like you did in #20, but without "quiet" boot parameter)
Created attachment 14001 [details] dmesg output dmesg output with out quiet set, this from the 2.6.9-1.667 boot
Please remove the boot option quite and attach the image of console at hang for 2.6.22.9 just like comment #15 in following four cases: 1.no boot option 2.pci=noacpi 3.acpi=off 4.pci=routeirq
Unfortunately they all look the same and not very verbose, not like comment #15. That was on ubuntu. Ive removed the quiet option but it doesnt seem to output any more info. I tried debug, that prints out a bit more at the start but the final screen is just same.
The screen capture of the failure in comment #15 is in acpi=off mode. ACPI had nothing to do with PCI guessing IRQ10 for USB. (though looking at PCI config space, that guess was probably correct) The 2.6.9 dmesg shows... > DMI not present. So probably your Fedora kernel has CONFIG_ACPI_BLACKLIST_YEAR=1999 or some other non-zero value. Try booting it with "acpi=force", or building a kernel.org kernel with CONFIG_ACPI_BLACKLIST_YEAR=0. Either of them should get you into ACPI mode like your working 2.6.9 kernel.
Sorry for the delay, been away for the weekend. yep, that was the message I noticed. There is a 2.6.23 kernel Ill try that.
Good news and bad news Good: pci=noacpi system boots Bad: acpi=off hangs at initializing hardware Worse: no option/pci=routeirq hangs earlier Thanks pictures to follow
Created attachment 14109 [details] acpi=off
Created attachment 14110 [details] no option/pci=routeirq The last 4/6 messages took 20-30 seconds to appear.
PS this was on 2.6.23 with apci year set to 0.
Created attachment 14112 [details] dmesg output from 2.6.23 boot
Hi, Martin Thanks for the info. Will you please try the patch in comment #15 of bug5637 and boot with no option? Thanks.
Created attachment 14113 [details] Attach the output of BIOS PRT and MPS table
Hi, Martin Will you please attach the output using the tool in comment #42? Thanks.
Created attachment 14123 [details] pirq_mps output
well done guys, with patch applied to 2.6.23 and no options the machine boots up.
Will you please attach the output of dmesg and /proc/interrupts ?
cat /proc/interrupts CPU0 0: 68535 XT-PIC-XT timer 1: 207 XT-PIC-XT i8042 2: 0 XT-PIC-XT cascade 8: 1 XT-PIC-XT rtc 9: 2 XT-PIC-XT acpi 10: 327 XT-PIC-XT ohci_hcd:usb1, yenta, eth0 11: 902 XT-PIC-XT ALI 5451 12: 3899 XT-PIC-XT i8042 14: 10722 XT-PIC-XT ide0 15: 6178 XT-PIC-XT ide1 NMI: 0 ERR: 0
Created attachment 14128 [details] dmesg output
Created attachment 14129 [details] dmesg output
It seems that the system can boot after applying the patch in bug 5637. So it is marked as a duplicate of bug5637. *** This bug has been marked as a duplicate of bug 5637 ***