Distribution: Ubuntu 7.10 Hardware Environment: ASUS P4C800 Motherboard, P4 2.8 GHz, 2x 512 mb Corsair memory Seagate ST3120023AS harddisk, SATA 120gb Nvidia GeForce 4Mx graphics card ps2 mouse and keyboard Plextor cd-writer samsung dvd-reader Problem Description: My computer hangs on boot if I have certain usb sticks attached to it while booting. With a Sandisk Cruzer Micro 512 mb the bug happens, with a 1gb Kingston it doesn't. The initial initializing of the kernel seems to go fine, but it seems likes it hangs when it comes to the ACPI part. On the mainline 2.6.25-rc8 kernel, the last message it outputs is just: ACPI: Core revision 20070126 On the Ubuntu 2.6.24-12 kernel, it also outputs ACPI: Looking for DSDT in initramfs... error, file /DSDT.aml not found The cursor still blinks, indicating the system is not completely locked up, but ctrl-alt-del does not work, I have to make a hard reset via the computers reset button. I'm guessing that this problem should be filed under ACPI rather than somewhere under USB, but please tell me if I should do otherwise. During boot, the BIOS detects and initializes the usb-stick as a mass storage device, maybe this has something do with the problem. I've not been able to turn this behavior off from the BIOS. Steps to reproduce: Put in the Sandisk Cruzer Micro 512mb usb stick into the usb port. Boot into linux It hangs
Created attachment 15629 [details] dmesg output
Created attachment 15630 [details] uname -r
Created attachment 15631 [details] dmidecode output
Created attachment 15632 [details] Video clip showing an unsuccesful boot
I've read my homework about debugging ACPI problems and done some additional testing: If I insert the usb-stick after the BIOS has initialized (when the GRUB screen is up) Linux boots fine. If the usb-stick is in before starting the computer (hence letting the bios initialize it), and I remove under the GRUB screen, Linux boots fine. When the usb-stick is in, I've tried some different boot options: With these, it boots: acpi=off acpit=ht With these, it hangs: pnpacpi=off pci=noacpi This time it gives an additional error message: ACPI: setting ELCR to ce20 (from cc20) noapic or nolapic Hangs with this error message: ACPI: setting ELCR to 0200 (from cc20) Please tell me if I should do any additional testing.
Please attach the acpidump output. Latest pmtools can be found at http://www.kernel.org/pub/linux/kernel/people/lenb/acpi/utils/
Please verify if it's a duplicate of bug #6208.
Created attachment 15659 [details] output of acpidump when doing acpidump > acpidump.log, I got these extra error messages: Wrong checksum for OEMB Wrong checksum for OEMB! I have inserted them at the right places in log file, and marked the lines as such.
First of all, thanks for your fast response. I will try and verify if this is a duplicate of bug 6208. It may take some time though, as I'm quite new to these things.
Will you please try the following commit and see whether the problem still exists? >commit cdc647a9b75741659bfc6acc44a6b3a646ad53bf >Author: David Brownell <david-b@pacbell.net> >Date: Wed Apr 2 13:40:20 2008 -0700 USB: another ehci_iaa_watchdog fix This patch, suggested by Alan Stern, fixes the hung USB issues on my notebook from suspend/resume cycles. It does so by eliminating some confusion about the internal state machine associated with unlinking from the EHCI async schedule ring, which caused a recent regression: http://bugzilla.kernel.org/show_bug.cgi?id=10345 Signed-off-by: Mark Lord <mlord@pobox.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
I've been looking at bug 6208. I'd like to try that acpi_osi boot option. Should I supply to GRUB menu for the kernel acpi_osi="Microsoft Windows" or just acpi_osi= ?
I today did a "git clone" Linus's tree, and as far as I can tell, my version is b87e81e5c6e64ae0eae3b4f61bf07bfeec856184 Tue Apr 15 14:34:49. I compiled and tested that. If I understand things correctly, this one includes the commit mentioned above. Sadly, it doesn't work any better. Then I did some more thorough testing. So far, it is only that SanDisk Cruzer usb-stick which causes the hang. I've tested with a HP printer, Canon digicam, Logitech wireless mouse/keyboard hub, and a USB mouse. Linux boots fine with all those plugged in. In bug 6208, it seemed that every usb-device would cause a hang. Does this mean they are not duplicates? I have tried tried supplying both of the acpi_osi options above but it doesn't help anything. This is the message I get now before it halts: ACPI: Core revision 20070126 Parsing all Control Methods Table [DSDT](id 0001) - 511 Objects with 48 Devices 150 Methods 14 Regions tbxface-0598 [02] tb_load_namespace : ACPI Tables successfully acquired I've also tried supplying acpi_dbg_level=0x1f and acpi_dbg_level=ffffffff, but this don't give any more debug output. I switched on ACPI verbose debugging during compile.
Then this is more a specific driver problem than a ACPI problem to me. what will happen if you plug the usb-stick AFTER the system boots? If it doesn't hang, please attach the dmesg output after the driver is loaded.
Thanks for the test and the analysis. Will you please catpure the picutre when your computer hangs on boot with the Sandisk Cruzer Micro 512 mb plugged? Will you please confirm whether the system can be booted with the boot option of "acpi=off"? (With Sandisk Cruzer Micro 512 plugged). If it can be booted, please attach the output of acpidump, lspci -vxxx, dmesg. Thanks.
Created attachment 15779 [details] dmesg output when usb-stick inserted after bios init Zhang: If I attach the usb-stick after the bios has initialized, but before pressing enter at the grub menu, it works fine. This way I avoid the bios initializing it. I'm also able to access the contents of the usb-stick fine when the system is up. This is the dmesg output from such a boot.
ykzhao: I'm able to boot with acpi=off when usb-stick plugged in. However, this leads to that my (ps2) mouse doesn't work, so it makes things a bit more complicated though (but that's a different story)... Attaching the files you requested.
Created attachment 15787 [details] acpidump when acpi=off
Created attachment 15788 [details] dmesg when acpi=off
Created attachment 15789 [details] lspci -vxxx when acpi=off
Created attachment 15790 [details] Video clip showing an unsuccesful boot with the git build used above
Will you please confirm whether the windows can be booted successfully if the Sandisk Cruzer Micro 512 is plugged?
From the info in comment #4 and #21 it seems that the system will print the following message and then hang if the Sandisk Cruzer Micro 512 is plugged. >Table [DSDT](id 0007) - >tbxface-0598 [00] tb_load_namespace : ACPI Tables successfully acquired Maye in such case it hangs in the course of transition from legacy mode to ACPI mode. Please confirm whether the windows can be booted successfully if the Sandisk Micro 512 is plugged? thanks.
Yes, I will try that. I do not currently have windows installed on it, but I think I can get a copy of it somehow. But it will probably take some time, though.
any updates?
Is there BIOS option about USB legacy support in the system? Try disablind it.
As there is no response for more than one month, the bug will be rejected. If the problem still exists, please reopen it again.