Re: feedkeys() allowed in sandbox
Ciaran McCreesh wrote: 100 bytes is more than enough room to download and execute a file that contains the real malicious code. I actually agree that it is extremely unlikely that a length check would make modelines more secure, but I'm being argumentative because it's irritating to be authoritatively assured that a length check would have no benefit in the future. We just don't know whether some future vulnerability (perhaps using a currently-unknown new feature) might be avoided with a modeline length check. John
Re: feedkeys() allowed in sandbox
On Fri, 4 May 2007 14:20:22 +1000 "John Beckett" <[EMAIL PROTECTED]> wrote: > I mentioned that the first step for point 4 should (IMHO) be > rejecting any modeline beyond some fairly small maximum size. Most previous exploits have been exploitable with far below the line length that is reasonably used by sensible people. > What I'd really like would be a separate sanity check that > verifies that the syntax in the modeline is boringly standard > 'set' options for a declared whitelist of things that a modeline > is allowed to do. http://www.vim.org/scripts/script.php?script_id=1876 > For example, 100 bytes of malware might be able to erase my > files, but perhaps it couldn't do something more sophisticated > like launching a hidden infiltration of my network. 100 bytes is more than enough room to download and execute a file that contains the real malicious code. -- Ciaran McCreesh signature.asc Description: PGP signature
Re: feedkeys() allowed in sandbox
Matthew Winn wrote: My objection to your idea is that it won't improve security by even the tiniest bit. It's not defence in depth at all. We've probably slugged this out enough, but I'm glad to have another opportunity to promote the "safe modelines" message. Bram has made the point that despite repeated modeline vulnerabilities over several years, there are no known cases of a malicious attack. We have only seen PoC and jokes. I see the sense of that position - why put in a bunch of ugly checking which is going to reduce features and upset some users? Why do it if there are no known benefits? My answer is essentially an appeal to a higher moral purpose. There may never be in-the-wild exploits based on modelines, but that would make it all the easier to direct a specific attack against a targeted victim. The attacker would have a list of 10 or 20 "slight" security flaws in the victim's network. One of those would be the fact that the victim uses Vim. An attacker may use a Vim modeline as the coup de grace to fully own the victim's network. I find that situation offensive, and believe that modelines should be REALLY fixed. My claim is: 1. A modeline can execute untrusted code. 2. That is incredibly dangerous. 3. Any bugs in modeline handling should be fixed. 4. In addition, modelines should be sanity checked. I think we agree on points 1-3. I mentioned that the first step for point 4 should (IMHO) be rejecting any modeline beyond some fairly small maximum size. However, that was just the first part of my hoped-for sanity check. After that, I would like the modeline to be examined to determine whether there are any constructs that "look" dangerous. I would reject any modeline with more than ten backslashes, and would reject anything looking like an expression or 'call'. What I'd really like would be a separate sanity check that verifies that the syntax in the modeline is boringly standard 'set' options for a declared whitelist of things that a modeline is allowed to do. Note that this checking should NOT be done only in the code that executes the modeline. The checking should be an independent, prior step. That redundancy is likely to save someone's foot in future years, when extra features are added. My objection to your idea [to limit modeline length] is that it won't improve security by even the tiniest bit. You may be right. But if I were to accidentally execute malware, I would prefer that the malware was short, rather than of an essentially unlimited length. I agree that 100 bytes of malware could do more damage than I could bear, but I would still prefer that situation. For example, 100 bytes of malware might be able to erase my files, but perhaps it couldn't do something more sophisticated like launching a hidden infiltration of my network. I don't really know why I want to limit the modeline length. That's the whole point of proper security measures. Just because I can't think of a way that a long modeline might be bad, does not mean that some attacker won't find a way, particularly in five years after a bunch more stuff has been added to Vim. That means that my modelines are quite long. I'm a reasonable guy. Let's take your longest modeline and double it. That length should be the maximum allowed for a modeline unless some new "anything goes" option is enabled. Re Perl tainting: I think we essentially agree on this, although I don't think Vim needs to mark an executable expression read from a modeline as tainted. Vim should immediately reject any modeline that might execute code (unless some new "anything goes" option is enabled). John
Re: feedkeys() allowed in sandbox
On Tue, 1 May 2007 19:42:02 +1000, "John Beckett" <[EMAIL PROTECTED]> wrote: > Matthew Winn wrote: > > If there was a security problem in Vim do you really think it > > couldn't be exploited in 100 characters? That's a pretty shaky > > foundation on which to build your security. > > I am quite surprised at the lack of appreciation for the merits > of "defense in depth" here. I am not claiming that a length > limit would preclude damage, just that a modeline should be > sanity checked before execution, and a reasonable length would > be the first check. What constitutes a "reasonable length"? Vim has to load the entire document including its modeline into memory anyway, so there's no denial-of-service implications in allowing unlimited modelines. Your suggestion amounts to "I won't use a modeline longer than X, so nobody will use a modeline longer than X". My objection to your idea is that it won't improve security by even the tiniest bit. It's not defence in depth at all. It's a worthless measure that can't achieve anything useful and can only get in the way of legitimate uses. Any modeline long enough to be useful for a legitimate purpose must also be long enough to be useful for a hostile one. > It's sensational that Vim can process files with very long > lines, for the occasions we need that. But it would be absurd > for Vim to process a multi-megabyte modeline. Where do you draw the line between absurd and reasonable? When I write options I spell out the names in full so they're easier to understand for someone who doesn't know Vim well. That means that my modelines are quite long. But someone who wanted to save space could use the abbreviated form of an option. That means that if a modeline can be long enough to satisfy me it would give an attacker the ability to use several times as many options to craft their exploit as are needed for general use. > By all means abuse me for my cheeky suggestion to limit > modelines to 100 bytes, but while doing that you might agree > that some limit under 1MB should be enforced. Why? In some places there are good reasons for limiting sizes. For example, RFC2822 places a limit of 998 characters on the length of a line. The reason for this is so that RFC2822-conforming applications don't have to deal with data of arbitrary length and allocate unlimited buffers to handle it. They can allocate a buffer 1001 characters long and discard anything that won't fit in the buffer, thereby preventing the possibility of denial-of-service attacks from someone sending a line several hundred megabytes long. Vim doesn't have that issue because it must load the entire file into memory anyway. Vim already knows how to deal with long lines, so there's no extra penalty incurred by a multi-megabyte modeline. > > A web browser should be able to handle anything thrown at it > > in a way that doesn't compromise security. _Every_ application > > should be able to handle anything thrown at it in a way that > > doesn't compromise security. > > Even if a program is perfect now, a later change can introduce a > bug. Any program which can automatically execute untrusted code > should sanity-check the input as a separate step from > sandboxing. That is standard Security 101 stuff - not my idea. I've been working with computer security for over two decades. I know about standard security stuff. I also know that security that doesn't work is worse than no security at all, because it creates an illusion of protection where none exists. > > Perl and Vim have exactly the same requirements: the need to > > safely handle code taken from an untrustworthy source. It > > makes no difference whether it comes directly from a network > > or from a disk. (If, like me, you use Vim as your source > > viewer for web pages, the need for the same level of security > > is obvious.) > > It doesn't matter, but for the record, Perl's tainting system is > not related to the scenario you describe. Perl wants to make > sure that untrusted input is not later used as the basis for > some expression that could do harm, such as executing SQL code. That's what I meant, and that's exactly what Vim needs as well. Both applications read data from a source that can't be trusted, and both need to ensure that untrusted data can't be used in a situation where it could be dangerous. In Vim's case it needs to make sure that an expression used in an option set from a modeline can't be used later in a way that would cause harm, such as executing a command. Take a look at the original message. It sets foldmethod to something that triggers the execution of an external command after the modeline has been processed. Imagine you have a web page that contains the following (with the real command removed so it can't cause problems, just in case someone does view this in Vim; think of "rm -rf /"): Now imagine that someone uses Vim as their browser's "view source" application. That's _exactly_ the thing Perl's tainting mech
Re: feedkeys() allowed in sandbox
Tony Mechelynck wrote: > Bram Moolenaar wrote: > [...] > > Modelines are default off when you are root. The mail filetype plugin > > also switches it off. > [...] > > Are you sure? In a terminal logged-in as root, using vim 7.0.235: > > vim -u NONE -N > :set ml? mls? >modeline >modelines=5 > > Modelines default off when 'compatible' is set, they default on (and 5) when > 'nocompatible' is set. Root login changes nothing to that AFAICT. Sorry, my mistake. There is a recommendation that when working as root you switch 'modeline' off, but it's not done automatically. I do think that it's a good idea to make it automatic. -- He who laughs last, thinks slowest. /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
Bram Moolenaar wrote: [...] Modelines are default off when you are root. The mail filetype plugin also switches it off. [...] Are you sure? In a terminal logged-in as root, using vim 7.0.235: vim -u NONE -N :set ml? mls? modeline modelines=5 Modelines default off when 'compatible' is set, they default on (and 5) when 'nocompatible' is set. Root login changes nothing to that AFAICT. Best regards, Tony. -- "You mean there really is an answer?" "Yes! But you're not going to like it!" "Oh do please tell us!" "You're really not going to like it!" "but we MUST know - tell us" "Alright, the answer is" "yes..." "... is ..." "yes... come on!" "is 42!" (Douglas Adams - The Hitchhiker's Guide to the Galaxy)
Re: feedkeys() allowed in sandbox
Matthew Winn wrote: If there was a security problem in Vim do you really think it couldn't be exploited in 100 characters? That's a pretty shaky foundation on which to build your security. I am quite surprised at the lack of appreciation for the merits of "defense in depth" here. I am not claiming that a length limit would preclude damage, just that a modeline should be sanity checked before execution, and a reasonable length would be the first check. It's certainly true that a modeline of 100 bytes could wreak havoc on an unpatched Vim. But it is quite possible that a future vulnerability might allow code to be injected from a modeline, and limiting its length *might* make the attacker's job harder. It's sensational that Vim can process files with very long lines, for the occasions we need that. But it would be absurd for Vim to process a multi-megabyte modeline. By all means abuse me for my cheeky suggestion to limit modelines to 100 bytes, but while doing that you might agree that some limit under 1MB should be enforced. A web browser should be able to handle anything thrown at it in a way that doesn't compromise security. _Every_ application should be able to handle anything thrown at it in a way that doesn't compromise security. Even if a program is perfect now, a later change can introduce a bug. Any program which can automatically execute untrusted code should sanity-check the input as a separate step from sandboxing. That is standard Security 101 stuff - not my idea. What you're suggesting isn't much different from security through obscurity. Jargon is great, but not when it's misused. Perl and Vim have exactly the same requirements: the need to safely handle code taken from an untrustworthy source. It makes no difference whether it comes directly from a network or from a disk. (If, like me, you use Vim as your source viewer for web pages, the need for the same level of security is obvious.) It doesn't matter, but for the record, Perl's tainting system is not related to the scenario you describe. Perl wants to make sure that untrusted input is not later used as the basis for some expression that could do harm, such as executing SQL code. John
Re: feedkeys() allowed in sandbox
Bram Moolenaar wrote: N times as safe still isn't 100% safe. I am not claiming that sanity-checking a modeline before execution would make it 100% safe. But there have been many examples in other software where minor bugs have turned into security disasters because some simple point that could have been checked, wasn't. While code is working correctly, a simple check is redundant, and indeed is offensive because it lengthens and obscures the code. But a few simple checks may prevent disaster at some future time, when Vim is further developed. The Google test (searching for past instances of trouble with Vim's modeline) proves the case that future problems are likely. Modelines are default off when you are root. The mail filetype plugin also switches it off. Good grief - I didn't know that. So you *have* got sanity checks built in! I'll go and sit in the corner now, but thanks for confirming that multiple layers of defence are desirable. John
Re: feedkeys() allowed in sandbox
John Beckett wrote: > A.J.Mechelynck wrote: > >> Is folding really needed in a default modeline? > > Folding may be useful in a modeline. > > (Don't know what you call a "default" modeline.) > > By "default modeline" I mean I would like Vim to be changed so > that its default behaviour is aggressively safe. If wanted, > there could be a new option to enable clever features, and a > user could choose to allow modelines with folding or expression > evaluation, etc. This is not true. It just reduces the chance of a mistake being made by an unknown factor. It's still possible to allow an option to be set, thinking that it is OK, but we later find out that it was not OK. Just like carefully removing mistakes and screening the options for mistakes does help to make it safer. Thus it doesn't make an essential difference. N times as safe still isn't 100% safe. In other words: If we have an option "run insecure" nobody would set it. Vim must be secure as-is. > But the only long-term safe procedure is to have Vim *default* > to work with only very restricted modelines (set tab and other > options - no way to even get near executing code). As they sometimes joke: The best way to protect your computer from malicious software is to switch it off. Likewise, the only really safe way is to disable modelines. Obviously you pay a price: restricted functionality. Options to partly disable modelines make it more complicated and don't help much for security. > I am wondering what the lack of comment on this topic indicates. > Do you understand that another modeline vulnerability could > allow the next file you open to overwrite all files under your > home folder? Or it might overwrite all sectors on your disk, if > you have sufficient privilege. Don't forget that this requires someone who intentionally wants this evil thing to happen. So far the only examples seen are jokes and proof of concept. I have never seen a file with a modeline that intentionally causes harm. > How about if you go to another computer that you rarely use. > Would you be happy using Vim on that computer? > Network admins in secure environments should be prohibited > from using Vim. Modelines are default off when you are root. The mail filetype plugin also switches it off. > If I am overlooking something, or am overly alarmist, please > tell me. For anyone new to this, enter following in Google: > vim vulnerability modeline Thanks for the advertisement! :-). -- Give a man a computer program and you give him a headache, but teach him to program computers and you give him the power to create headaches for others for the rest of his life... R. B. Forest /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
On Sun, 29 Apr 2007 19:10:55 +1000, "John Beckett" <[EMAIL PROTECTED]> wrote: > Matthew Winn wrote: > > I don't like the idea of preventing modelines over 100 bytes. > > I imagine (haven't looked) that a modeline has no hard limit to > its length. So multi-megabyte modelines are probably handled by > Vim. That's potentially offering attackers extraordinary power. It doesn't matter how many bytes are accepted. Security that depends on the assumption that an exploit can't be written in less than an arbitrarily chosen number of bytes is no security at all. Take a look at some of the coding golf competitions that take place online, where the object is to perform some complex task in the minimum number of characters. If there was a security problem in Vim do you really think it couldn't be exploited in 100 characters? That's a pretty shaky foundation on which to build your security. > Would someone who wants a modeline longer than 100 bytes please > show us an example. How about a 200-byte limit? Oh, so nobody will need a long modeline? Just like they will never need more than [insert your favourite inaccurate prediction about the maximum amount of computing power anyone would ever need here]. > Modelines are enabled by default, and are very useful for things > like setting tabs. So most people, and all new installs, will > have modelines enabled. It's very poor security practice to > offer a rich auto-execution environment with a single line of > defence (the sandbox). > > This discussion reminds me of the days of the Code Red > vulnerability in IIS (Microsoft web server), and of the > years of repeated vulnerabilities in Internet Explorer. > > The IIS and IE developers just couldn't bring themselves to > build in limits to what their wonderful software could do. > "But a web site might need a 100KB URL with hundreds of '../' > paths!". A web browser should be able to handle anything thrown at it in a way that doesn't compromise security. _Every_ application should be able to handle anything thrown at it in a way that doesn't compromise security. What you're suggesting isn't much different from security through obscurity. You're relying on the hope that nobody would ever be able to craft an exploit in under 100 bytes. Security doesn't work like that. Security needs to be something people can rely on to work every time, not something that depends on "Well, let's hope nobody thinks of this". If Vim's modeline security is written correctly and securely then the length of modeline it can handle safely would depend only on the amount of memory it wants to allocate to hold it. If it isn't able to do that then there's no security at all. > > Furthermore, what am I supposed to do if I want a long, > > complicated but legitimate modeline? > > I would like a default "high security" setting for handling > modelines. If people want modelines that do complex stuff, I > would recommend setting a new "anything goes" option. ABSOLUTELY NOT! Are you honestly suggesting that to enter a long modeline you have to disable security? I wouldn't touch an editor like that. > > I like Perl's approach to untrustworthy data. It's flagged as > > tainted at the point it is read, and anything derived from it > > is also flagged as tainted. > > Perl has to have that system because part of its intended usage > is to handle data entered into web pages. It's pretty complex > and has taken years to get right. > > Vim is a text editor - it should not automatically execute code > in any old file that I might accidentally open. Perl and Vim have exactly the same requirements: the need to safely handle code taken from an untrustworthy source. It makes no difference whether it comes directly from a network or from a disk. (If, like me, you use Vim as your source viewer for web pages, the need for the same level of security is obvious.) -- Matthew Winn
Re: feedkeys() allowed in sandbox
A.J.Mechelynck wrote: Is folding really needed in a default modeline? Folding may be useful in a modeline. (Don't know what you call a "default" modeline.) By "default modeline" I mean I would like Vim to be changed so that its default behaviour is aggressively safe. If wanted, there could be a new option to enable clever features, and a user could choose to allow modelines with folding or expression evaluation, etc. But the only long-term safe procedure is to have Vim *default* to work with only very restricted modelines (set tab and other options - no way to even get near executing code). I am wondering what the lack of comment on this topic indicates. Do you understand that another modeline vulnerability could allow the next file you open to overwrite all files under your home folder? Or it might overwrite all sectors on your disk, if you have sufficient privilege. How about if you go to another computer that you rarely use. Would you be happy using Vim on that computer? Network admins in secure environments should be prohibited from using Vim. If I am overlooking something, or am overly alarmist, please tell me. For anyone new to this, enter following in Google: vim vulnerability modeline I just noticed that the fourth hit features Ciaran McCreesh who discovered a vulnerability in January 2005. John
Re: feedkeys() allowed in sandbox
John Beckett wrote: [...] Is folding really needed in a default modeline? John Folding may be useful in a modeline. (Don't know what you call a "default" modeline.) Depending on how the particular file is written, you may want to set foldmethod=marker (and which marker), foldmethod=syntax, foldmethod=indent, or default it to whatever (if anything) is set by the filetype-plugin. Best regards, Tony. -- Some of these fortunes are dated: I have an ADSL connection and a 96 gig HD, and I don't feel it's any special kind of achievement. -- hundred-and-one symptoms of being an internet addict: 208. Your goals for the future are obtaining an ISDN connection and a 6 gig hard drive.
Re: feedkeys() allowed in sandbox
Ciaran McCreesh wrote: > On Sat, 28 Apr 2007 21:52:07 +0200 > Bram Moolenaar <[EMAIL PROTECTED]> wrote: > > I don't like this solution. Opening some files would be OK in the > > sandbox, e.g., for reading. readfile() would be OK in the sandbox, > > right? > > Probably not. In a multi-user environment it can be used as a > privilege escalation by inserting the contents of a non-world-readable > file into a world-readable file when the latter is edited by a user > with elevated privileges. In the sandbox you can't insert text into a file or buffer. Anything that requires saving text for undo is blocked. You can also get the text from an already opened file with getbufline(). It's difficult to draw a line, but I think blocking everything that writes is good enough. -- `The Guide says there is an art to flying,' said Ford, `or at least a knack. The knack lies in learning how to throw yourself at the ground and miss.' He smiled weakly. -- Douglas Adams, "The Hitchhiker's Guide to the Galaxy" /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
On Sat, 28 Apr 2007 21:52:07 +0200 Bram Moolenaar <[EMAIL PROTECTED]> wrote: > I don't like this solution. Opening some files would be OK in the > sandbox, e.g., for reading. readfile() would be OK in the sandbox, > right? Probably not. In a multi-user environment it can be used as a privilege escalation by inserting the contents of a non-world-readable file into a world-readable file when the latter is edited by a user with elevated privileges. -- Ciaran McCreesh signature.asc Description: PGP signature
Re: feedkeys() allowed in sandbox
Matthew Winn wrote: I don't like the idea of preventing modelines over 100 bytes. I imagine (haven't looked) that a modeline has no hard limit to its length. So multi-megabyte modelines are probably handled by Vim. That's potentially offering attackers extraordinary power. Would someone who wants a modeline longer than 100 bytes please show us an example. How about a 200-byte limit? Modelines are enabled by default, and are very useful for things like setting tabs. So most people, and all new installs, will have modelines enabled. It's very poor security practice to offer a rich auto-execution environment with a single line of defence (the sandbox). This discussion reminds me of the days of the Code Red vulnerability in IIS (Microsoft web server), and of the years of repeated vulnerabilities in Internet Explorer. The IIS and IE developers just couldn't bring themselves to build in limits to what their wonderful software could do. "But a web site might need a 100KB URL with hundreds of '../' paths!". Furthermore, what am I supposed to do if I want a long, complicated but legitimate modeline? I would like a default "high security" setting for handling modelines. If people want modelines that do complex stuff, I would recommend setting a new "anything goes" option. I like Perl's approach to untrustworthy data. It's flagged as tainted at the point it is read, and anything derived from it is also flagged as tainted. Perl has to have that system because part of its intended usage is to handle data entered into web pages. It's pretty complex and has taken years to get right. Vim is a text editor - it should not automatically execute code in any old file that I might accidentally open. John
Re: feedkeys() allowed in sandbox
On Sat, 28 Apr 2007 22:43:23 +1000, "John Beckett" <[EMAIL PROTECTED]> wrote: > "Potentially unsafe" means we're pretty sure it IS safe, but > (for example), it's simply not worthwhile allowing a modeline > longer than 100 bytes because if another vulnerability were > ever found, we don't want to make it easy for the attacker. I don't like the idea of preventing modelines over 100 bytes. To start with, there's no real logic behind it: it's an arbitrary number pulled out of thin air, and I put it in the same category as saying "it's OK to use gets() so long as you use a long enough buffer that it'll never overflow". A modeline that's long enough to allow useful things to be done is long enough to allow unpleasant things to be done. Furthermore, what am I supposed to do if I want a long, complicated but legitimate modeline? I like Perl's approach to untrustworthy data. It's flagged as tainted at the point it is read, and anything derived from it is also flagged as tainted. Tainted information cannot be used in unsafe operations, ever. From what I've read in this thread Vim does something similar, but in a way that's less complete. That's the right way to go about it. Setting an arbitrary limit and hoping it'll have the effect of improving security is far too optimistic for my tastes. -- Matthew Winn
Re: feedkeys() allowed in sandbox
Andrew Maykov wrote: > On 4/28/07, Bram Moolenaar <[EMAIL PROTECTED]> wrote: > >It's better to make sure the sandbox works as it should. > Yet another function to disable in sandbox: > vi: fdm=expr fde=writefile([""],"phantom_was_here") Yep, you found another one. Seems some of the new functions added in Vim 7 were not properly checked for sandbox use. I think system() should also not write the "input" argument to a file. It's quite harmless, since you can't control the file name, but the shell command is going to fail anyway. None of the others appear to be harmful. > Proposal. Maybe it's sane to put security checks not just in > functions like f_writefile(), but also put it to the core of fileio, > e.g. if mch_fopen macro will check permissions before actual openning > file, then f_writefile() and freinds if any will fail to harm user. > > i.e. replace something like this: > =CUT > --- macros.h.orig 2007-04-29 00:57:16.0 +0700 > +++ macros.h2007-04-29 00:58:38.0 +0700 > @@ -149,7 +149,7 @@ > #ifdef VMS > # define mch_access(n, p) access(vms_fixfilename(n), (p)) > /* see mch_open() comment */ > -# define mch_fopen(n, p) fopen(vms_fixfilename(n), (p)) > +# define mch_fopen_impl(n, p) fopen(vms_fixfilename(n), (p)) > # define mch_fstat(n, p) fstat(vms_fixfilename(n), (p)) > /* VMS does not have lstat() */ > # define mch_stat(n, p)stat(vms_fixfilename(n), (p)) > @@ -158,7 +158,7 @@ > # define mch_access(n, p)access((n), (p)) > # endif > # if !(defined(FEAT_MBYTE) && defined(WIN3264)) > -# define mch_fopen(n, p) fopen((n), (p)) > +# define mch_fopen_impl(n, p) fopen((n), (p)) > # endif > # define mch_fstat(n, p) fstat((n), (p)) > # ifdef MSWIN /* has it's own mch_stat() function */ > @@ -174,6 +174,9 @@ > # endif > #endif > > + > +#define mch_fopen(n, p)( check_secure() ? NULL : mch_fopen_impl(n,p) > ) > + > #ifdef HAVE_LSTAT > # define mch_lstat(n, p) lstat((n), (p)) > #else > =/CUT=== I don't like this solution. Opening some files would be OK in the sandbox, e.g., for reading. readfile() would be OK in the sandbox, right? -- How To Keep A Healthy Level Of Insanity: 15. Five days in advance, tell your friends you can't attend their party because you're not in the mood. /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
On 4/28/07, Bram Moolenaar <[EMAIL PROTECTED]> wrote: >It's better to make sure the sandbox works as it should. Yet another function to disable in sandbox: vi: fdm=expr fde=writefile([""],"phantom_was_here") Proposal. Maybe it's sane to put security checks not just in functions like f_writefile(), but also put it to the core of fileio, e.g. if mch_fopen macro will check permissions before actual openning file, then f_writefile() and freinds if any will fail to harm user. i.e. replace something like this: =CUT --- macros.h.orig 2007-04-29 00:57:16.0 +0700 +++ macros.h2007-04-29 00:58:38.0 +0700 @@ -149,7 +149,7 @@ #ifdef VMS # define mch_access(n, p) access(vms_fixfilename(n), (p)) /* see mch_open() comment */ -# define mch_fopen(n, p) fopen(vms_fixfilename(n), (p)) +# define mch_fopen_impl(n, p) fopen(vms_fixfilename(n), (p)) # define mch_fstat(n, p) fstat(vms_fixfilename(n), (p)) /* VMS does not have lstat() */ # define mch_stat(n, p)stat(vms_fixfilename(n), (p)) @@ -158,7 +158,7 @@ # define mch_access(n, p)access((n), (p)) # endif # if !(defined(FEAT_MBYTE) && defined(WIN3264)) -# define mch_fopen(n, p) fopen((n), (p)) +# define mch_fopen_impl(n, p) fopen((n), (p)) # endif # define mch_fstat(n, p) fstat((n), (p)) # ifdef MSWIN /* has it's own mch_stat() function */ @@ -174,6 +174,9 @@ # endif #endif + +#define mch_fopen(n, p)( check_secure() ? NULL : mch_fopen_impl(n,p) ) + #ifdef HAVE_LSTAT # define mch_lstat(n, p) lstat((n), (p)) #else =/CUT===
Re: feedkeys() allowed in sandbox
Bram Moolenaar wrote: Perhaps modelines should only allow a VERY limited set of operations by default (even more restricted than now). Sure, simply use ":set nomodeline". I'm suggesting "defence in depth". My vimrc might have ':set nomodeline', but what if I make a mistake? What if I'm using some other machine where I'm not sure what's in vimrc? What if I use the -u option? And there are probably other scenarios where a simple oversight could cause me to execute a modeline in an untrusted file. What if I want modelines (but never to do more than set a few options like tabs). Even setting 'textwidth' to 2 may already be considered harmful, or at least annoying. I don't mind being occasionally irritated with a stupid modeline. But I really don't want an editor to execute code when I open a file. Change volatile settings - no problem. Execute - no thanks. Perhaps we can add yet-another-option to completely disable setting some options from the modeline. Modelines should default to be safe (safe by design, AND safe because of defence in depth). If another option were added, it should be to allow expressions or other features that are potentially unsafe (not disallow unsafe features). "Potentially unsafe" means we're pretty sure it IS safe, but (for example), it's simply not worthwhile allowing a modeline longer than 100 bytes because if another vulnerability were ever found, we don't want to make it easy for the attacker. It's better to make sure the sandbox works as it should. Of course, but bitter experience shows that defence in depth is the best strategy. Vim has a lot of very clever features that are too complex to ever be secure, when executing a modeline from an untrusted source. John
Re: feedkeys() allowed in sandbox
John Beckett wrote: > Bram Moolenaar wrote: > > That's pretty nasty. I'll make a patch right away. > > Thanks. However, perhaps the modeline concept needs > more safety - defence in depth. > > Perhaps modelines should only allow a VERY limited set > of operations by default (even more restricted than now). Sure, simply use ":set nomodeline". Even setting 'textwidth' to 2 may already be considered harmful, or at least annoying. > Googling for 'vim feedkeys joke' shows "April 1 joke" with > the following (I've replaced "vim" with "vvv"): > > vvv: foldmethod=expr:foldexpr=feedkeys( > "x3a%!cat\\x20-n\\x3a%s/./\:)/g\\ > x3aq!\\"): > > I'm too lazy to unobfuscate this, but one glance tells you > that modelines should not be "fixed" - going down that path > is likely to give a new vulnerability every year. > > Instead, modelines should be SEVERELY limited by default. > Examples: > Total length < 100 bytes. > No expressions; no function calls; no execution. > Treat a double-quoted string as if in single quotes. > Is folding really needed in a default modeline? Modelines are already limited. You can't put commands there (like some old versions of vi did). Options with an expression are a border case. They are executed in the sandbox when an option was set from the modeline. Perhaps we can add yet-another-option to completely disable setting some options from the modeline. But I think it won't help much, most users won't take the time or even know about the option. It's better to make sure the sandbox works as it should. -- How To Keep A Healthy Level Of Insanity: 11. Specify that your drive-through order is "to go". /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
Tomas Golembiovsky wrote: today somebody came to #vim, and pasted some modeline (containig joke or such). Thanks for raising that issue. I found the April 1 "joke" with Google. I actually noticed that posting (seeing "Vim" while browsing a security list caught my attention), but there's so much waffle out there that I didn't pay any attention to the point of the message. John
Re: feedkeys() allowed in sandbox
Bram Moolenaar wrote: That's pretty nasty. I'll make a patch right away. Thanks. However, perhaps the modeline concept needs more safety - defence in depth. Perhaps modelines should only allow a VERY limited set of operations by default (even more restricted than now). Googling for 'vim feedkeys joke' shows "April 1 joke" with the following (I've replaced "vim" with "vvv"): vvv: foldmethod=expr:foldexpr=feedkeys( "x3a%!cat\\x20-n\\x3a%s/./\:)/g\\ x3aq!\\"): I'm too lazy to unobfuscate this, but one glance tells you that modelines should not be "fixed" - going down that path is likely to give a new vulnerability every year. Instead, modelines should be SEVERELY limited by default. Examples: Total length < 100 bytes. No expressions; no function calls; no execution. Treat a double-quoted string as if in single quotes. Is folding really needed in a default modeline? John
Re: feedkeys() allowed in sandbox
Tony Mechelynck wrote: > Tomas Golembiovsky wrote: > > Greetings mortals, > > > > today somebody came to #vim, and pasted some modeline (containig joke or > > such). He muttered something about not knowing what that means and left > > before long. But (!) what I noticed is that feedkeys() was used as part of > > foldexpression and it turned out that feedkeys() is allowed in sandbox, > > which means malicious file can run arbitrary command via modeline like > > this: > > > > vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") > > > > I guess you can see the consequences. Is this known/intentional? > > > > IIUC, feedkeys() called from sandbox should execute as if in sandbox, > i.e., only (at most) key sequences acceptable in sandbox should be > able to be "fed". > > Now this is what I think it "ought" to do. How does it "actually" > behave? Did you try your example? Did it "touch" the file? That was the idea: The "sandbox" flag is checked when the keys are executed. However, the sandbox flag may have been reset by then, as the example shows. Thus feedkeys() needs to be disallowed in the sandbox. -- I have a drinking problem -- I don't have a drink! /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: feedkeys() allowed in sandbox
Tomas Golembiovsky wrote: > today somebody came to #vim, and pasted some modeline (containig joke or > such). He muttered something about not knowing what that means and left > before long. But (!) what I noticed is that feedkeys() was used as part of > foldexpression and it turned out that feedkeys() is allowed in sandbox, > which means malicious file can run arbitrary command via modeline like > this: > > vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") > > I guess you can see the consequences. Is this known/intentional? That's pretty nasty. I'll make a patch right away. -- Far back in the mists of ancient time, in the great and glorious days of the former Galactic Empire, life was wild, rich and largely tax free. Mighty starships plied their way between exotic suns, seeking adventure and reward among the furthest reaches of Galactic space. In those days, spirits were brave, the stakes were high, men were real men, women were real women and small furry creatures from Alpha Centauri were real small furry creatures from Alpha Centauri. And all dared to brave unknown terrors, to do mighty deeds, to boldly split infinitives that no man had split before -- and thus was the Empire forged. -- Douglas Adams, "The Hitchhiker's Guide to the Galaxy" /// Bram Moolenaar -- [EMAIL PROTECTED] -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org///
Re: Re: feedkeys() allowed in sandbox
Michael Henry wrote : > A.J.Mechelynck wrote: > > Tomas Golembiovsky wrote: > >> vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") > >> > >> I guess you can see the consequences. Is this known/intentional? > >> > > > > IIUC, feedkeys() called from sandbox should execute as if in sandbox, > > i.e., only (at most) key sequences acceptable in sandbox should be able > > to be "fed". Now this is what I think it "ought" to do. How does it > > "actually" behave? Did you try your example? Did it "touch" the file? Of course I have tried it. I'm not that stupit so as to embarrass myself in the very first e-mail (that's what second one is for ;). And what is the key limit, I haven't noticed anything like that in the help. Also I was able to feed about 2k keys and still did not hit the limit. And 2k of characters is pretty much to do a lot of nasty stuff, just imagine what small and stupid fork bomb can do here. > I placed Tomas's modeline into a file test.txt and ran it. On both > Ubuntu Linux and my mac, the file phantom_was_here was created in my > working directory, e.g.: > >[EMAIL PROTECTED]:~/tmp/test$ ls >test.txt >[EMAIL PROTECTED]:~/tmp/test$ cat test.txt >vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") >[EMAIL PROTECTED]:~/tmp/test$ vim test.txt > > >Press ENTER or type command to continue >[EMAIL PROTECTED]:~/tmp/test$ ls >phantom_was_here test.txt > > > I've no idea is this is how it ought to behave. I also can't explain > the "Press ENTER..." prompt, but I had to press the ENTER key before Vim > allowed me to edit the file after it already contained the modeline. > That's nothing unusual after running :! command. -- Best regards, Tomas Golembiovsky -- ||- - - | | Albrecht's Law | | Social innovations tend to the level of minimum tolerable well being. | |- - -
Re: feedkeys() allowed in sandbox
Tomas Golembiovsky wrote: Greetings mortals, today somebody came to #vim, and pasted some modeline (containig joke or such). He muttered something about not knowing what that means and left before long. But (!) what I noticed is that feedkeys() was used as part of foldexpression and it turned out that feedkeys() is allowed in sandbox, which means malicious file can run arbitrary command via modeline like this: vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") I guess you can see the consequences. Is this known/intentional? IIUC, feedkeys() called from sandbox should execute as if in sandbox, i.e., only (at most) key sequences acceptable in sandbox should be able to be "fed". Now this is what I think it "ought" to do. How does it "actually" behave? Did you try your example? Did it "touch" the file? Best regards, Tony. -- Of what you see in books, believe 75%. Of newspapers, believe 50%. And of TV news, believe 25% -- make that 5% if the anchorman wears a blazer.
feedkeys() allowed in sandbox
Greetings mortals, today somebody came to #vim, and pasted some modeline (containig joke or such). He muttered something about not knowing what that means and left before long. But (!) what I noticed is that feedkeys() was used as part of foldexpression and it turned out that feedkeys() is allowed in sandbox, which means malicious file can run arbitrary command via modeline like this: vim: fdm=expr fde=feedkeys("\\:!touch\ phantom_was_here\\") I guess you can see the consequences. Is this known/intentional? -- Best regards, Tomas Golembiovsky -- ||- - - | | Alan's Law of Research | | The theory is supported as long as the funds are. | |- - -