tech-kern archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: 9.1: forcible coredump -> panic
> It's no longer panicking. But it doesn't actually drop core.
Turns out it does, just to an unexpected place.
If the process's corename setting is a relative path, it is relative to
not the process's cwd, but to the dumping lwp's cwd (ultimately because
namei_getstartdir() uses curlwp for self). So the coredump ends up in
/, because that's what the workqueue's kernel thread has as cwd.
I'm going to be trying to figure out what to do about this. At the
moment I'm leaning towards arranging for the dumper kthread to have the
same cwd as the victim process for the duration of the dump.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse%rodents-montreal.org@localhost
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
Home |
Main Index |
Thread Index |
Old Index