From: Kees Cook <email@example.com> To: Greg Kroah-Hartman <firstname.lastname@example.org> Cc: Michal Hocko <email@example.com>, Andrew Morton <firstname.lastname@example.org>, Alexey Dobriyan <email@example.com>, Lee Duncan <firstname.lastname@example.org>, Chris Leech <email@example.com>, Adam Nichols <firstname.lastname@example.org>, email@example.com, firstname.lastname@example.org, email@example.com, Uladzislau Rezki <firstname.lastname@example.org> Subject: Re: [PATCH v2] seq_file: Unconditionally use vmalloc for buffer Date: Thu, 18 Mar 2021 08:51:45 -0700 [thread overview] Message-ID: <202103180847.53EB96C@keescook> (raw) In-Reply-To: <YFMKUZ5p1QbqkabY@kroah.com> On Thu, Mar 18, 2021 at 09:07:45AM +0100, Greg Kroah-Hartman wrote: > On Wed, Mar 17, 2021 at 02:30:47PM -0700, Kees Cook wrote: > > On Wed, Mar 17, 2021 at 04:38:57PM +0100, Greg Kroah-Hartman wrote: > > > On Wed, Mar 17, 2021 at 04:20:52PM +0100, Michal Hocko wrote: > > > > On Wed 17-03-21 15:56:44, Greg KH wrote: > > > > > On Wed, Mar 17, 2021 at 03:44:16PM +0100, Michal Hocko wrote: > > > > > > On Wed 17-03-21 14:34:27, Greg KH wrote: > > > > > > > On Wed, Mar 17, 2021 at 01:08:21PM +0100, Michal Hocko wrote: > > > > > > > > Btw. I still have problems with the approach. seq_file is intended to > > > > > > > > provide safe way to dump values to the userspace. Sacrificing > > > > > > > > performance just because of some abuser seems like a wrong way to go as > > > > > > > > Al pointed out earlier. Can we simply stop the abuse and disallow to > > > > > > > > manipulate the buffer directly? I do realize this might be more tricky > > > > > > > > for reasons mentioned in other emails but this is definitely worth > > > > > > > > doing. > > > > > > > > > > > > > > We have to provide a buffer to "write into" somehow, so what is the best > > > > > > > way to stop "abuse" like this? > > > > > > > > > > > > What is wrong about using seq_* interface directly? > > > > > > > > > > Right now every show() callback of sysfs would have to be changed :( > > > > > > > > Is this really the case? Would it be too ugly to have an intermediate > > > > buffer and then seq_puts it into the seq file inside sysfs_kf_seq_show. > > > > > > Oh, good idea. > > > > > > > Sure one copy more than necessary but it this shouldn't be a hot path or > > > > even visible on small strings. So that might be worth destroying an > > > > inherently dangerous seq API (seq_get_buf). > > > > > > I'm all for that, let me see if I can carve out some time tomorrow to > > > try this out. > > > > The trouble has been that C string APIs are just so impossibly fragile. > > We just get too many bugs with it, so we really do need to rewrite the > > callbacks to use seq_file, since it has a safe API. > > > > I've been trying to write coccinelle scripts to do some of this > > refactoring, but I have not found a silver bullet. (This is why I've > > suggested adding the temporary "seq_show" and "seq_store" functions, so > > we can transition all the callbacks without a flag day.) > > > > > But, you don't get rid of the "ability" to have a driver write more than > > > a PAGE_SIZE into the buffer passed to it. I guess I could be paranoid > > > and do some internal checks (allocate a bunch of memory and check for > > > overflow by hand), if this is something to really be concerned about... > > > > Besides the CFI prototype enforcement changes (which I can build into > > the new seq_show/seq_store callbacks), the buffer management is the > > primary issue: we just can't hand drivers a string (even with a length) > > because the C functions are terrible. e.g. just look at the snprintf vs > > scnprintf -- we constantly have to just build completely new API when > > what we need is a safe way (i.e. obfuscated away from the caller) to > > build a string. Luckily seq_file does this already, so leaning into that > > is good here. > > But, is it really worth the churn here? > > Yes, strings in C is "hard", but this _should_ be a simple thing for any > driver to handle: > return sysfs_emit(buffer, "%d\n", my_dev->value); > > To change that to: > return seq_printf(seq, "%d\n", my_dev->value); > feels very much "don't we have other more valuable things we could be > doing?" > > So far we have found 1 driver that messed up and overflowed the buffer > that I know of. While reworking apis to make it "hard to get wrong" is > a great goal, the work involved here vs. any "protection" feels very > low. I haven't been keeping a list, but it's not the only one. The _other_ reason we need seq_file is so we can perform checks against f_cred for things like %p obfuscation (as was needed for modules that I hacked around) and is needed a proper bug fix for the kernel pointer exposure bug from the same batch. So now I'm up to 3 distinct reasons that the sysfs API is lacking -- I think it's worth the churn and time. > How about moving everyone to sysfs_emit() first? That way it becomes > much more "obvious" when drivers are doing stupid things with their > sysfs buffer. But even then, it would not have caught the iscsi issue > as that was printing a user-provided string so maybe I'm just feeling > grumpy about the potential churn here... I need to fix the prototypes for CFI sanity too. Switching to seq_file solves 2 problems, and if we have to change the prototype once for that, we can include the prototype fixes for CFI at the same time to avoid double the churn. -- Kees Cook
next prev parent reply other threads:[~2021-03-18 15:52 UTC|newest] Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-03-15 17:48 Kees Cook 2021-03-15 18:33 ` Al Viro 2021-03-15 20:43 ` Kees Cook 2021-03-16 7:24 ` Greg Kroah-Hartman 2021-03-16 12:43 ` Al Viro 2021-03-16 12:55 ` Greg Kroah-Hartman 2021-03-16 13:01 ` Michal Hocko 2021-03-16 19:18 ` Kees Cook 2021-03-17 10:44 ` Greg Kroah-Hartman 2021-03-16 8:31 ` Michal Hocko 2021-03-16 19:08 ` Kees Cook 2021-03-17 12:08 ` Michal Hocko 2021-03-17 13:34 ` Greg Kroah-Hartman 2021-03-17 14:44 ` Michal Hocko 2021-03-17 14:56 ` Greg Kroah-Hartman 2021-03-17 15:20 ` Michal Hocko 2021-03-17 15:38 ` Greg Kroah-Hartman 2021-03-17 15:48 ` Michal Hocko 2021-03-17 21:30 ` Kees Cook 2021-03-18 8:07 ` Greg Kroah-Hartman 2021-03-18 15:51 ` Kees Cook [this message] 2021-03-18 17:56 ` Greg Kroah-Hartman 2021-03-19 14:07 ` [seq_file] 5fd6060e50: stress-ng.eventfd.ops_per_sec -49.1% regression kernel test robot 2021-03-19 19:31 ` Kees Cook
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=202103180847.53EB96C@keescook \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --subject='Re: [PATCH v2] seq_file: Unconditionally use vmalloc for buffer' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).