pkg_delete and malloc debugging (Re: port register/unregister hogs up resources and takes forever)

Kirill Ponomarew krion at voodoo.bawue.com
Sun Oct 2 03:50:18 PDT 2005


On Sat, Oct 01, 2005 at 07:26:27PM -0400, Kris Kennaway wrote:
> > fstat(3, {st_mode=0150320, st_size=15046755950319947984, ...}) = 0
> > read(3, "@comment PKG_FORMAT_REVISION:1.1"..., 4096) = 1525
> > close(3)                                = 0
> > lstat("<D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0><D0>
> 
> This looks like the signature of malloc debugging exposing a memory
> use bug.  I wonder if some of the recent changes to pkg_tools might be
> to blame (CC'ing Kirill)?  Try disabling it with
> 
> ln -sf aj /etc/malloc.conf 
> 
> If it "fixes" it, it's only a workaround, but it points the way to the
> solution.

Anyway it would be interesting to look at "uname -a" output, because
I can't "reproduce" this bug here.

-Kirill
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20051002/93ca31cb/attachment.bin


More information about the freebsd-ports mailing list