Date: Mon, 8 Feb 2021 20:33:08 +0100 From: Andrea Venturoli <ml@netfence.it> To: freebsd-hackers@freebsd.org Subject: Understanding kernel deadlock Message-ID: <eac9cc57-2589-4a52-882e-a502835a86de@netfence.it>
next in thread | raw e-mail | index | archive | help
Hello. First off, I apologize if I'm asking stupid questions, but I'm a noob. I'm reading a lot of docs but that seems to never be enough. I'm trying to analyze a locked up 12.2 kernel (from #244048) with kgdb, but I'm stuck due to too little knowledge on my part. What I think I'm seeing are some threads (including the one originated from mksnap_ffs) waiting for some spare buffer (i.e. in bufspace_wait). However the buf_daemon thread and its child, bufspace_daemon, are seemingly looping without finding any buffer available to free. Is that explanation reasonable? Possible at all? Or am I getting it all wrong? Is there a paper somewhere explaining how bufdaemon & co are supposed to work? Will they not allocate new buffers if needed? Is there a limit set somewhere that I should increase? Or are they just probably unable to work due to a LOR somewhere else? Thanks for bearing with me. bye av.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?eac9cc57-2589-4a52-882e-a502835a86de>