I built the kernel i got from the SVN around 12:30pm EST (new york time to be clear) and i just finished buildkernel and rebooted. I didnt get far into the boot before i got a kernel panic. I'm not totally sure what caused this because the computer froze with this: db> if need be i can build and install the kernel again, provided that someone tel me how i can save the boot message to a file. As i said before though, the book message screen froze so i couldnt even drop to a shell and i had to hard reset power off my computer to get it working and boot to kernel.old thank you, ~wes
Do you have a swap device or other kernel dump device configured? If so, please try setting debug.debugger_on_panic="0" in loader.conf and get a stack trace / etc from crashinfo. If not, next time 'bt' (backtrace) and a photo would be helpful. You might consider renaming /boot/kernel to /boot/kernel.bad; /boot/kernel.old to /boot/kernel; and then 'nextboot -k kernel.bad -e debug.debugger_on_panic=0 && reboot'. This way you only hit the bad kernel once (if your /boot is not ZFS) and don't need to manually use loader to find kernel.old. There's not much to go on for now, so please provide more information and reopen the bug. Thanks!
Created attachment 205225 [details] Kernel panic boot screen pic Here is a picture of my computer screen with the kernel panic. I would have liked to provide more than this but the kernel crashes before it initializes both the mouse and keyboard.
Looks like it is possibly related to base r349161.
Please send me acpidump -dt result.
Im building r249223 right now. If i get the same result i will post a photo and the acpidump -dt result.
I also got same panic. The reason is probably that although the function AcpiNsWalkNamespace returns AE_OK, the value of the argument ReturnValue is not changed, resulting in an invalid handle. After changing as below, my system is boot without panic. --- src/sys/contrib/dev/acpica/components/namespace/nswalk.c 2019-06-20 16:40:45.119908000 +0900 +++ src_org/sys/contrib/dev/acpica/components/namespace/nswalk.c 2019-06-20 15:45:56.709177000 +0900 @@ -331,9 +331,6 @@ AcpiNsWalkNamespace ( ChildType = ACPI_TYPE_ANY; Level = 1; - if (ReturnValue) - *ReturnValue = NULL; - /* * Traverse the tree of nodes until we bubble back up to where we * started. When Level is zero, the loop is done because we have
Created attachment 205230 [details] Proposed patch for fix this
(In reply to yoaucid from comment #6) Thank you! So what about do next.
Please style: if(acpi_disabled("usb")){ Into: if (acpi_disabled("usb")) { Before committing. --HPS
Created attachment 205235 [details] kernel panic 10:24am nyc time build from 6:51am new york time still fails
Comment on attachment 205235 [details] kernel panic 10:24am nyc time build from 6:51am new york time build from 6:51 still fails. i added a picture of the boot screen. i will also inclose a text file called 'new' for the acpidump -dt result as per requested
Created attachment 205238 [details] acpidump -dt results as promised here is the acpidump -dt for debugging purposes.
(In reply to Wes Wroten from comment #12) Did you try the patch I post while ago? https://bz-attachments.freebsd.org/attachment.cgi?id=205230
Boot successful, the bug has been fixed with the patch.
A commit references this bug: Author: takawata Date: Thu Jun 20 23:52:34 UTC 2019 New revision: 349251 URL: https://svnweb.freebsd.org/changeset/base/349251 Log: Fix the case where no root hub object while host controller object exist in ACPI namespace. Also you can disable ACPI support for USB by setting debug.acpi.disabled="usb" PR: 238711 Changes: head/sys/dev/usb/usb_hub_acpi.c