From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Philipp Reisner To: axboe@suse.de Date: Thu, 27 Jan 2005 10:47:09 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200501271047.09852.philipp.reisner@linbit.com> Cc: Lars Marowsky-Bree , drbd-dev@lists.linbit.com Subject: [Drbd-dev] Re: bio clone must copy io_vec List-Id: Coordination of development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hi all, I am reffering to Jens' fix for md... > The way md uses bio_clones, it's possible for the original bio to be freed > before the clone is freed. This means the clone ->bi_io_vec points to fre= ed > memory potentially. Fix it by duplicating the io_vec as well. [...] In DRBD the thing is different.=20 We use __bio_clone() but we will never end IO for our original bio before=20 we have not finished IO for our cloned BIO. My fist thought was to have a own copy of __bio_clone() in place, which behaves like the __bio_clone() of Linux-2.6.10-vanilla. But on the other hand I think I can convert it over to use bio_clone() within a few hours. Well I will give it a try, and let you know if I end up with a patch... PS: Jens, when I am the owner of a BIO I can do with the private member what ever I want, right ? And I am the owner of a bio clones with bio_clone() right ? =2DPhilipp =2D-=20 : Dipl-Ing Philipp Reisner Tel +43-1-8178292-50 : : LINBIT Information Technologies GmbH Fax +43-1-8178292-82 : : Sch=F6nbrunnerstr 244, 1120 Vienna, Austria http://www.linbit.com :