On Thu, Oct 13, 2022 at 4:48 AM Simon Lindhorst <s.lindhorst@wut.de> wrote:
Hello all,
when I updated my Kernel from version 5.4.155 to 5.4.215 I get an strange xhci error:
xhci-hcd f10f0000.usb3: ERROR unknown event type 37 xhci-hcd f10f0000.usb3: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 2 comp_code 13
After a lot of this messages, my hardware makes a reboot without any more outputs.
The error only occures when i add console=null to my kernel bootargs. When I add instead console=ttyS0,115200 no error occured.
Now I go back in kernelversions. The error occured first in version 5.4.159. Between patchlevel 158 and 159 there is a change:
--- linux-5.4.158/kernel/printk/printk.c 2021-11-06 13:59:45.000000000 +0100 +++ linux-5.4.159/kernel/printk/printk.c 2021-11-12 14:43:05.000000000 +0100 @@ -2193,8 +2193,15 @@ char *s, *options, *brl_options = NULL; int idx;
- if (str[0] == 0) + /* + * console="" or console=null have been suggested as a way to + * disable console output. Use ttynull that has been created + * for exacly this purpose. + */ + if (str[0] == 0 || strcmp(str, "null") == 0) { + __add_preferred_console("ttynull", 0, NULL, NULL); return 1; + }
if (_braille_console_setup(&str, &brl_options)) return 1;
I checked my kernelconfig and found that I have no ttynull device configured (CONFIG_NULL_TTY=n). Add CONFIG_NULL_TTY=y to my kernelconfig doesn't made a change.
When I undo the change above, everything works fine.
Does anybody know, what could be the main trigger for the error above?
while there have been lots of change to printk, the code you cite is still there. If that code is a candidate for the root-cause, and you can re-create the error on master, you are 1/2 the way to getting it fixed. Also, note latest: commit 3ef4ea3d84ca568dcd57816b9521e82e3bd94f08 Merge: 30d024b5058e 5eb17c1f458c Author: Linus Torvalds <torvalds@linux-foundation.org> Date: Wed Mar 23 10:54:27 2022 -0700 Merge tag 'printk-for-5.18' of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux Pull printk updates from Petr Mladek: - Make %pK behave the same as %p for kptr_restrict == 0 also with no_hash_pointers parameter - Ignore the default console in the device tree also when console=null or console="" is used on the command line - Document console=null and console="" behavior that last one is pertinent.