Messages in this thread | | | Date | Mon, 13 Mar 2000 10:53:18 -0500 (EST) | From | Alexander Viro <> | Subject | Re: patch: reiserfs for 2.3.49 |
| |
On Mon, 13 Mar 2000, Manfred Spraul wrote:
> "Yury Yu. Rupasov" wrote: > > > > Chris Mason wrote: > > > Hans, I think Alexander has told us exactly what he wants done. He wants > > > us to audit our entire VFS interface, and make sure that we are dealing > > > with boundary conditions, special cases, and normal cases the same way the > > > existing linux filesystems are. > > Al, > is Minix maintained? Minix is a tiny filesystem, and it might be simpler > for the reiserfs developers if they could mimic Minix instead of ext2.
Minixfs is maintained (as a part of the whole tree - I suspect that last changes that were not part of global ones and were not cosmetic cleanups happened _long_ ago; I'll check it when I'll get to my CVS repository), but in many respects it is worse model than ext2. Just one example: it has directory-as-array layout and that means that many things that go unpunished here will be rather bad for ext2, let alone reiserfs. Fresh example: check the change to ext2/dir.c in recent 2.2.15-pre. Race that gives a local DoS and is impossible on minixfs... Another thing being whether reiserfs wants to do the right thing and support removed-but-busy directories. Soon it will become a non-issue (code is going to migrate into VFS), but for now they ought to protect from that (tests for d_unhashed() in rmdir() and rename()) _or_ check for attempts to add an entry to such directory (should never be allowed). Minixfs chooses the former. There's also one nasty thing about minixfs - it's a mix of two filesystems (v1 and v2) and code is not the easiest to read...
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |