From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiggers@kernel.org (Eric Biggers) Date: Sun, 18 Aug 2019 10:29:38 -0700 Subject: [PATCH] erofs: move erofs out of staging In-Reply-To: <20190818162201.GA16269@infradead.org> References: <20190817233843.GA16991@hsiangkao-HP-ZHAN-66-Pro-G1> <1405781266.69008.1566116210649.JavaMail.zimbra@nod.at> <20190818084521.GA17909@hsiangkao-HP-ZHAN-66-Pro-G1> <1133002215.69049.1566119033047.JavaMail.zimbra@nod.at> <20190818090949.GA30276@kroah.com> <790210571.69061.1566120073465.JavaMail.zimbra@nod.at> <20190818151154.GA32157@mit.edu> <20190818155812.GB13230@infradead.org> <20190818161638.GE1118@sol.localdomain> <20190818162201.GA16269@infradead.org> Message-ID: <20190818172938.GA14413@sol.localdomain> On Sun, Aug 18, 2019@09:22:01AM -0700, Christoph Hellwig wrote: > On Sun, Aug 18, 2019@09:16:38AM -0700, Eric Biggers wrote: > > Ted's observation was about maliciously-crafted filesystems, though, so > > integrity-only features such as metadata checksums are irrelevant. Also the > > filesystem version is irrelevant; anything accepted by the kernel code (even if > > I think allowing users to mount file systems (any of ours) without > privilege is a rather bad idea. But that doesn't mean we should not be > as robust as we can. Optionally disabling support for legacy formats > at compile and/or runtime is something we should actively look into as > well. > > > it's legacy/deprecated) is open attack surface. > > > > I personally consider it *mandatory* that we deal with this stuff. But I can > > understand that we don't do a good job at it, so we shouldn't hold a new > > filesystem to an unfairly high standard relative to other filesystems... > > I very much disagree. We can't really force anyone to fix up old file > systems. But we can very much hold new ones to (slightly) higher > standards. Thats the only way to get the average quality up. Some as > for things like code style - we can't magically fix up all old stuff, > but we can and usually do hold new code to higher standards. (Often not > to standards as high as I'd personally prefer, btw). Not sure what you're even disagreeing with, as I *do* expect new filesystems to be held to a high standard, and to be written with the assumption that the on-disk data may be corrupted or malicious. We just can't expect the bar to be so high (e.g. no bugs) that it's never been attained by *any* filesystem even after years/decades of active development. If the developers were careful, the code generally looks robust, and they are willing to address such bugs as they are found, realistically that's as good as we can expect to get... - Eric