From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <3905CD51.90ED385B@wanadoo.fr> Date: Tue, 25 Apr 2000 18:52:33 +0200 From: Martin Costabel MIME-Version: 1.0 To: Ethan Benson CC: Daniel Gonzalez , LinuxPPC Dev List Subject: Re: Bye HFS partition? References: <390516C2.1DD0D2D@pandora.be> <20000425010826.T7022@plato.localdomain.local> Content-Type: text/plain; charset=us-ascii Sender: owner-linuxppc-dev@lists.linuxppc.org List-Id: Ethan Benson wrote: > > On Mon, Apr 24, 2000 at 11:53:38PM -0400, Daniel Gonzalez wrote: > > > > Hi, > > > > I have just destroyed my hfs partition (I was trying to make > > a directory there and ... oooops!) > > do you mean you ran into the general unreliablity of the kernel's HFS > support? I don't think that HFS support is *generally* unreliable. It works well with 2.2.x kernels and is broken on recent 2.3.x kernels. From his earlier posts I recall that Daniel is using a 2.3.99 kernel. For me, mounting an HFS partition under a 2.3.99 kernel and doing more than just a file listing leads to almost instant system freeze and sometimes destruction of the HFS partition (I am trying this only on a ZIP, so I can easily reformat). The oops I get comes, in general, from the hfs_bnode_relse function in fs/hfs/bnode.c, where a waitqueue bug is detected. It would be nice (and rather vital for the survival of LinuxPPC under the imminent (?) 2.4 kernel) if some LinuxPPC developers woke up to this situation. (Seems to be above my capacities, sorry). The official maintainer of the HFS code doesn't do much, and some developers (I remember A Joshi and Tom Rini) gave the impression that they have a working HFS filesystem under 2.3.x. In the todo list for kernel 2.4, however, HFS is marked as broken (but no showstopper for the i386 crowd, of course). -- Martin ** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/