When debugging the state of kernel tasks, eg by sending sysctrl-T, there is too much information with a typical login session for the default log buffer size, thus if the process is started earlier, the information is lost (either from dmesg, or /var/log/messages) and the first part of the output truncated:
When debugging the state of kernel tasks, eg by sending sysctrl-T, there is too much information with a typical login session for the default log buffer size, thus if the process is started earlier, the information is lost (either from dmesg, or /var/log/messages) and the first part of the output truncated:
$ cat /var/log/messages 2aa>] ? __up_read+0x9a/0xc0 3f9>] do_vfs_ ioctl+0x79/ 0x370 f79>] ? up_read+0x9/0x10 fe4>] ? do_page_ fault+0x194/ 0x370 771>] sys_ioctl+0x81/0xa0
Sep 19 11:42:32 exige kernel: [<ffffffff81273
Sep 19 11:42:32 exige kernel: [ 1803.421542] [<ffffffff81129
Sep 19 11:42:32 exige kernel: [ 1803.421542] [<ffffffff81077
Sep 19 11:42:32 exige kernel: [ 1803.421542] [<ffffffff81525
Sep 19 11:42:32 exige kernel: [ 1803.421542] [<ffffffff81129
<snip>
I think it's important to be able to debug kernel problems fully, so this is a quick and safe fix.