leak of the vnodes

[Kostik Belousov]

On Wed, Apr 07, 2010 at 04:25:52PM +0200, Petr Salinger wrote:
> >>I used the attached diff, with hackish snooping
> >>on allocated/freed memory for vnodes. When the vp pointer have been
> >>logged as active1/active2, it is (much) later shown with
> >>dead_vnodeops in DUMP_VP().
> >Is there a lot of such /dev/ttyp* vnodes ? This indeed might be
> >suspicious. See below for description of how to check that the
> >vnodes are leaked or not.
> 
> In VP_DUMP, there have been 90% nodes of this type:
> 
>  40285 active 0xffffff00164855a0 tag devfs type 8 flags 00000080 use count 
>  1 hold count 1
>  19410 active 0xffffff0024522d20 tag devfs type 8 flags 00000080 use count 
>  1 hold count 1
> 258526 active 0xffffff000a93b960 tag devfs type 8 flags 00000080 use count 
> 1 hold count 1
> 257411 active 0xffffff00245d8b40 tag devfs type 8 flags 00000080 use count 
> 1 hold count 1
> 247494 active 0xffffff0016c7d780 tag devfs type 8 flags 00000080 use count 
> 1 hold count 1
> 
> None of other line have VI_DOOMED set.
> The tag in my output is devfs, due to commented out  vp->v_tag = "none";
> I do not know, what the previous name (like ptyp*) have been.

You cut off the most interesting question from my followup, are any
of these doomed vnodes referenced by struct file from any process
in the system. Can you, please, look at it ?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-hackers/attachments/20100407/626da2e3/attachment.pgp