Re: multi threaded program locks up /proc, ps



On Fri, 06 Jan 2006 17:02:27 +0100 Kasper Dupont <36237715889427396252@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
| phil-news-nospam@xxxxxxxx wrote:
|>
|> Why would it matter if the process being examined is blocked in I/O or
|> not when ps examines it,
|
| I guess it shouldn't matter. Sounds more likely that some data
| structure related to the process is locked, and whoever did that
| isn't going to unlock it again. It could even be that the
| process itself is just blocked on the same lock.
|
| So why is a data structure left locked for apparently no good
| reason? It could be a race condition or some other kernel bug.

Indeed. Worth investigating. Hope he can reproduce it in the current
kernel if it hasn't been fixed already.

--
-----------------------------------------------------------------------------
| Phil Howard KA9WGN | http://linuxhomepage.com/ http://ham.org/ |
| (first name) at ipal.net | http://phil.ipal.org/ http://ka9wgn.ham.org/ |
-----------------------------------------------------------------------------
.



Relevant Pages

  • Re: multi threaded program locks up /proc, ps
    ... > | I guess it shouldn't matter. ... > | So why is a data structure left locked for apparently no good ... It could be a race condition or some other kernel bug. ... badly behaved USB device (or any other type, ...
    (comp.os.linux.development.system)
  • Re: Page aging broken in 2.6
    ... Evicting a page that is infrequently referenced by many ... > processes increases the chance that all runnable processes block waiting ... no matter how frequently. ... paging I/O being congested must be due to ...
    (Linux-Kernel)
  • Re: Multiple Threads vs. Asynchronous IO
    ... to perform their own I/O or otherwise block? ... That's the beautiful thing about threads. ... the same process is that if you ever block for any reason, you suffer. ... That means every bit of code, no matter how rarely used, no matter how ...
    (comp.programming.threads)
  • Re: [PATCH 3/3] ring-buffer: make cpu buffer entries counter atomic
    ... The signedness thing is just an unfortunate thing with atomic_t and probably ... doesn't matter. ... The change in size might or might not be a bug - that's for the person ... who didn't document the data structure to work out;) ...
    (Linux-Kernel)
  • Re: converting a character array to a float (newbie)
    ... Jim Bancroft wrote: ... no matter what the actual value being read back? ... Put a known float value on this I/O. ...
    (comp.lang.c)