Monday, March 29, 2021

Remotely Exploitable 0day in Internet Explorer Gets a Free Micropatch


by Mitja Kolsek, the 0patch Team

 

[Note: This blog post was originally published on February 17, 2021 under URL https://blog.0patch.com/2021/02/remotely-exploitable-0day-in-internet.html but was for some reason deleted. Perhaps it was our fat fingers, perhaps evil forces - we'll never know. We have now reconstructed it from the Internet Archive which is an incredible service that you should donate to if you like this post, as we did.]

[Update 3/19/2021: This issue has been fixed by March 2021 Windows Updates. 0patch users had this critical issue, now assigned CVE-2021-26411, patched since February 17, full 20 days before an official patch became available. Since the official fix is available, this micropatch is no longer FREE and requires a PRO license.]

On February 4, 2021, security researchers at ENKI, a South Korean security consultancy, published a blog post detailing an unpatched vulnerability in Internet Explorer. This "0day" vulnerability was used in an attack campaign against various security researchers, including ENKI researchers, who noticed the attack and took the exploit apart to extract the vulnerability information. ENKI researchers kindly shared their proof of concept with us, so we could quickly start analyzing the vulnerability and create a micropatch for it.

The vulnerability is a "double free" bug that can be triggered with JavaScript code and causes memory corruption in Internet Explorer's process space. As is often the case, this memory corruption could be carefully managed and turned into arbitrary read/write memory access - which can then be leveraged to arbitrary code execution. Attackers delivered the exploit in an MHTML file to ensure recipients would open it in Internet Explorer (which is registered to open this file type). While this delivery method required recipients to confirm a security warning about executing active content, the exploit could be delivered without such warning if the victim visited a malicious web site with Internet Explorer. 

In such case, just opening the malicious web site would instantly, or a benign web site hosting a malicious ad, would result in malicious native code execution inside Internet Explorer's render process running (by default) in Low Integrity. Such code could read any data from the computer and network that the user running Internet Explorer could read, and silently send it to attacker. An additional vulnerability would be needed to escape the "Low Integrity sandbox" and achieve a long-term compromise of the computer.

Is anyone still browsing the web with Internet Explorer? While Internet Explorer is not widely used for browsing web sites anymore, it is installed on every Windows computer and (a) opens MHT/MHTML files by default, (b) is being used internally in many large organizations, and (c) executes HTML content inside various Windows applications.


The Vulnerability

Exploit and proof-of-concept have not been published yet and we won't be the first to do so, but the root of this vulnerability is not new - it's about tricking the browser to delete an object that has already been deleted in some unexpected way that existing sanitization checks don't notice. In this case, it's about deleting a node value of an HTML Attribute. The trick is to create an attribute, assign it a value that is not a string or a number, but an object (why is this even allowed?) - then when deleting this attribute, said object makes sure that the attribute is deleted before it gets deleted, so to speak.


The Micropatch

While Internet Explorer developers will probably fix the way the attribute node is deleted so that it doesn't actually get deleted while references to it still exist, we decided that such approach would simply require too much time for us and would introduce an unnecessary risk of breaking something. We thus rather decided to break the obscure browser functionality that allows setting an HTML Attribute value to an object. We assess this functionality to be useful to very few web developers whose apps are supposed to work with Internet Explorer.

Our micropatch gets applied inside the CAttribute::put_ie9_nodeValue function of mshtml.dll, where it checks the VARIANT type of the value that JavaScript code wants to assign to an attribute - and prevents that from happening if the type is 9 (VT_DISPATCH) - which is used for Object, Array, or Date.



The 64bit micropatch only has 5 CPU instructions, and the 32bit one has 6 CPU instructions.



MODULE_PATH "..\Affected_Modules\mshtml.dll_11.0.9600.19597_64bit\mshtml.dll"
PATCH_ID 560
PATCH_FORMAT_VER 2
VULN_ID 6943
PLATFORM win64

patchlet_start
 PATCHLET_ID 1
 PATCHLET_TYPE 2
 PATCHLET_OFFSET 0xbf34b4

 N_ORIGINALBYTES 5
 PIT mshtml.dll!0xbf359f ; address of exit block

 code_start

  ; we're going to check the incoming VARIANT's data type; if it's 9 (object), we're going
  ; to prevent it from being copied to the attribute.
  ; The incoming VARIANT is pointed to by rdx, and the type is in the first byte.

  mov r14, rcx         ; replicate the instruction we're injected in front of to make sure
                       ; rcx is stored in r14 in case we jump to the exit block (where rcx is
                       ; restored from r14)
  cmp byte [rdx], 0x09 ; is the incoming VARIANT data type 9 (object)?
  jne DO_NOTHING       ; if not, we don't interfere
 
  mov rbx, 0            ; return value - we simulate a successful operation
  jmp PIT_0xbf359f     ; jump to exit block
 
 DO_NOTHING:

 code_end
    
patchlet_end


Here's a video of the micropatch in action:




The micropatch applies to the following Windows versions (32bit and 64bit). 

Updated to February 2021:

  1. Windows 7 + ESU (first update from ESU year 2)
  2. Windows Server 2008 R2 + ESU (first update from ESU year 2)
  3. Windows 10 v1809, v1909, v2004, v20H2
  4. Windows Server 2016, 2019

Updated to January 2021:

  1. Windows 7 + ESU (last update from ESU year 1)
  2. Windows Server 2008 R2 + ESU (last update from ESU year 1)
  3. Windows 10 v1809, v1909, v2004, v20H2
  4. Windows Server 2016, 2019 

Updated to January 2020:

  1. Windows 7 without ESU (last free update without ESU)
  2. Windows Server 2008 R2 without ESU (last free update without ESU)
 

Online Test

 
We have prepared a simple online test to demonstrate how our micropatch changes the behavior of Internet Explorer. To perform this test, you have to use Internet Explorer 11 on one of the Windows systems listed above.

Step 1: With 0patch disabled, open https://0patch.com/poc/IE_Attribute_nodeValue_0day/test.html in Internet Explorer 11. The web page should look like the image below, indicating 6 FAILed tests.
 

Step 2: With 0patch enabled, press F5 to refresh the test page in Internet Explorer 11. The web page should look like the image below, indicating no failed tests.


 

According to our guidelines, this micropatch is free for everyone until Microsoft issues an official fix for it. By the time you're reading this the micropatch has already been distributed to all online 0patch Agents and also automatically applied except where Enterprise policies prevented that. If you're not a 0patch user and would like to use this micropatch on your computer(s), create an account in 0patch Central, install 0patch Agent and register it to your account. Note that no computer restart is needed for installing the agent or applying/un-applying any 0patch micropatch.
 
We'd like to thank ENKI researchers for their analysis of the vulnerability and an elegant proof-of-concept, which allowed us to create a micropatch.

While you're here: If your organization has Windows 7 or Server 2008 R2 machines with Extended Security Updates and wouldn't mind saving lots of money on less expensive security patches in 2021 that don't even need your machines to be restarted, proceed to our New Year's Resolution. The same applies if you're still using Office 2010 and want to keep patching critical vulnerabilities now that support has ended.

To learn more about 0patch, please visit our Help Center.  

Analyzing And Micropatching With Tetrane REVEN (Part 1, CVE-2021-26897)

 

by Mitja Kolsek, the 0patch Team


March 2021 Windows Updates included fixes for seven vulnerabilities in Windows DNS Server, two of which were marked by Microsoft as "Exploitation More Likely": CVE-2021-26877 and CVE-2021-26897. They were not known to be exploited and no details were publicly available until security researchers Eoin Carroll and Kevin McGrath published their analysis on McAfee Labs blog. Their article included enough information for us to reproduce both vulnerabilities, and then create micropatches for them.

This article will be about CVE-2021-26897, while CVE-2021-26877 will be covered in a parallel article.

These two vulnerabilities were the first we have ever analyzed with the help of Tetrane REVEN, an incredibly powerful reverse engineering tool that allows you to record a virtual machine and then browse or search through all recorded instructions in all processes and the kernel, or taint any data value forward or back in time, across processes and between user and kernel space (plus much more). I wanted to use this opportunity to show how REVEN helped us perform these analyses, which would otherwise have been done with WinDbg and countless re-launching of dns.exe process, having all interesting objects bouncing around on different memory addresses every time.


Analysis

CVE-2021-26897 is a buffer overflow issue, whereby a series of oversized "dynamic update" DNS queries with SIG (signature) records causes writing beyond the buffer boundary when these records are saved to file. DNS server periodically saves all received updates to file (so they don't get lost on restart or crash), and the issue gets triggered by simply waiting for this file write to happen after sending a number of requests, or by stopping the DNS Server service, which was a convenient time saver for us.

Our proof-of-concept (POC) sends ten malformed DNS requests, and upon stopping the DNS Server service, the dns.exe process crashes. Let's use REVEN to see what goes on.

We used a virtual machine with Windows Server 2019 and DNS Server role without March updates to keep it in the vulnerable state. It is important to record as few events (called "transitions") as possible, so "lightening" of a machine - stopping unnecessary services, and disabling Windows Defender and Windows Error Reporting - is generally a good idea. We did not stop any services but did the latter as error reporting gets triggered upon every process crash and, well, executes a lot of instructions. While it's easiest to start and stop recording manually, even one second of extra recording can create tens of millions of unneeded transitions that will just slow down post-processing of the recording. To optimize start and end of recording, REVEN provides a cool trick they call "ASM stubs", which is a fancy name for calling the int3 CPU instruction while having the ecx register set to some magic value. In other words, you can trigger starting and stopping of recording from within the virtual machine you're recording, which means you can start right before the interesting stuff happens, and stop right after it's done.

In our case, the mere sending of malformed DNS requests does not crash the process, but the stopping of the DNS Server service does. So we used our POC to send the requests, and then launched a batch script that started the recording and stopped the service. Before that, we have "abused" the Postmortem Debugging mechanism to make it launch a small executable that stops recording whenever a process crashes - instead of launching a debugger, which postmortem debugging was designed for.

Our recording generated around 1.25 billion transitions.Yes, that's billion. But it's really no problem because REVEN handles that effortlessly. The only price you pay for a larger recording is the time you have to wait for REVEN to "replay" it, which extracts all machine states and transitions, indexes them, extracts everything that was happening to the memory, downloads symbol files and a bunch of other things to assure a swift analysis thereafter. Our replay took just over an hour and generated 55 GB of data, upon which the analysis would then actually be done.

Granted, we could have tried to further reduce the recording and possibly succeed, but that doesn't make too much sense - one hour for preparation while you're having lunch is more than acceptable, and fiddling with recording optimization also takes time that can quickly exceed the time saved by a potentially quicker replay.

Proceeding to "Axion", the analysis user interface where the magic stuff happens. Axion has multiple widgets; prominently positioned in the middle are Transitions, displaying a small part of the entire recorded execution conveniently grouped in code blocks with one or more CPU instructions. Other widgets include Backtrace (the call stack for the current transition), CPU (register values before and after the current transition), Memory (at chosen address, before or after the current transition, along with the entire history of read and write accesses), Search (immensely useful, allows you to find calls to specific function, or all executions of a specific instruction), Bookmarks, and - my favorite - Taint. REVEN allows you to select any piece of data (e.g., the current value in register r8, or the value at some memory address) and taint it either forward or back in time, to see what this value affects or where it came from, respectively. This is a huge value-add for our analyses - although by no means the only one.

Now let's dive into the analysis. The first thing we need to do is find where dns.exe crashed. Scrolling through a billion+ transitions is obviously out of the question, but we can search for one of the functions that get called when an unhandled exception is thrown. KiUserExceptionDispatch is one such function.

 

 

Search finds a single hit, as expected, and here we are looking at the first executed instruction in this user-space function, as it was launched by kernel's KiExceptionDispatch. (The kernel code is on grey background because we filtered out only user-space execution.) Now we want to see why this exception was thrown. A simple press of the "%" key transfers our to the other end of a call-ret pair, or in our case, to the other end of the kernel call.

 

The "%" key landed us on the exact instruction that caused access violation in function Dns_SecurityKeyToBase64String. It was an attempted write to address r8+1, and the Memory view shows this address to be immediately after a valid memory block, where we see a bunch of A's that were likely written to this buffer in some loop we're probably currently in. We did not use REVEN-IDA integration here but if we did, we would immediately see the code graph for this function in IDA, with the current instruction selected in IDA. And we would see that we are, indeed, in a loop that copies base64-encoded signature value from our DNS request to this buffer that just got overflown, and uses r8 as the destination pointer that gets increased in every iteration. 

(Note that we enabled Page Heap for dns.exe to make it crash immediately on buffer overflow, otherwise the overflow could just corrupt the heap and eventually cause some random malfunction. With Page Heap, every heap buffer is allocated at the very end of a read-write memory block, followed by unallocated memory page - which means any typical buffer overflow will immediately trigger an access violation exception.)

Execution of a loop is shown in REVEN as a repetition of loop instructions, over and over again, but you can of course select any of these instructions and see how registers and memory looked like at that exact moment. What we want here, though, is to see where the buffer was allocated.

If we were in WinDbg, we'd have used the !heap command to get the call stack from the moment the overflown buffer was allocated. In REVEN, however, we can not only find the code that allocated the buffer, but also values of registers and content of memory in that precise moment. The simplest way (that I know of) would be to simply taint register r8 to see where its value came from - going back to the past far enough, at some point its value must have been determined by whoever allocated this buffer, or it would not have pointed to the end of this same buffer now.

However, tainting r8 backward produces too long a path that just keeps bouncing in the loop that we're in. While it eventually gets to where we want to be, it slows down the UI. So our first goal is to get to the beginning of our loop's execution. We copied the address of our access-violation-triggering instruction (0x7ff729ca224a) and went to the Search widget, where we searched for all uses of this address.



Results: this exact instruction has been executed 130705 times in our recording; in other words, it would take a significant chunk of one's lifetime to just scroll up to the start of the loop. However, it only took one press of the "Next" button to get to the first iteration of the loop - because we were already positioned on the last iteration.

 



 

This got us to the very first execution of our instruction in this recording, and we can see that it wrote 0x41 to memory (to the very same buffer it overflowed 130704 iterations later). The 0x41 left to it was written earlier by a similar instruction in the same loop. Now that we're out of the loop, so to speak, let's taint r8 and see where it got from.



We launched tainting for value of register r8 from the current transition, back to some function we selected high in the call stack; we chose function Zone_WriteBackDirtyZones. Why not taint all the way back to the very first transition? Because we're only interested in who has allocated the memory buffer that got overflown, but the taint would go way further back in time because the address of this allocation was influenced by earlier allocations (that's how the heap works) and that is just irrelevant to us.

Tainting identified a couple of hundred transitions, and we're interested in those at the very end of the list (i.e., the earliest ones). When one is often looking for memory allocations, some familiar Windows functions catch one's eye - and RtlAllocateHeap is one of them.



RtlAllocateHeap takes three arguments, allocation size being the third one - which in x64 calling convention means register r8. We can see that the value of r8 when this function was called was 0x80010. This means that the actual buffer allocated was of this size, but we still want to see if this is a hard-coded size or perhaps dynamically calculated. So we go further back in time through the taint results.



In the very first transition found with tainting, inside funtion File_WriteZoneToFile, we see a call was made to a function Mem_Alloc, which is not publicly documented but is clearly used for allocating the memory block we're after. Most importantly, we can see that a hard-coded value 0x80000 was provided to it, which is clearly the size of the buffer. (0x10 was subsequently added to it inside Mem_Alloc, which is nicely seen by walking through the taint.)

Now we know what happened: a fixed-sized memory block was allocated in function File_WriteZoneToFile, whose name implies that the DNS zone we've updated with our malformed requests was going to be written to file. At some point function Dns_SecurityKeyToBase64String was called that overflowed this buffer after base64-encoding the signature from our requests. Let's just see if function Dns_SecurityKeyToBase64String was called more than once, as we know that a single malformed request doesn't produce the crash.

To do that, we used the "Symbol call" search and provided the function name.



The search produced 6 hits, meaning that function Dns_SecurityKeyToBase64String was called 6 times. This indicates that it was our 6th DNS request that finally caused the writing to go beyond the buffer end. Some additional analysis showed that all these calls added their output to the same growing string inside the fixed-size buffer, which was supposed to be finally saved to the zone file.

Our REVEN analysis was done here. At this point we could have created a micropatch in various ways, making sure to prevent writing past the buffer end, but since we had Microsoft's official patches available, we wanted to see what they have done.

We used IDA and BinDiff to compare dns.exe from February 2021 and March 2021, and found 9 functions modified by the March update.



Technically, we could just search our recording for any execution inside these functions, and hopefully find that just one of them was executed - which would mean that the fix was included there. Actually, we did exactly that and found SigFileWrite to be the only one - but this would be very cumbersome if hundreds of functions were modified, especially if the recording included many of the modified functions that have nothing to do with our bug.

The most reliable method would be to inspect the entire taint list to see which of the modified functions affected the value we were tainting. Taint search is currently not supported by REVEN user interface, but we could undoubtedly use the API to achieve that (note to self: send a feature request to Tetrane). We're not that fluent in REVEN API yet so we took the third route: the call stack.

It is quite likely that one of the modified functions would appear in the call stack of our crash instruction. But in our case we don't see it (see the call stack on one of the screenshots above). We do see, however, that function Dns_SecurityKeyToBase64String seems to have been called by function LdrpDispatchUserCallTarget, which has a suspect name. Let's just click on that in REVEN and see what happened there.

 


We see that function RR_WriteToFile made a call to LdrpDispatchUserCallTarget, which then made a jump to SigFileWrite function. The latter is executed, but not seen in the call stack because LdrpDispatchUserCallTarget jumped to it instead of calling. Note that function LdrpDispatchUserCallTarget is part of Control Flow Guard as explained in this Trend Micro article. So whenever you see LdrpDispatchUserCallTarget in the call stack, you'll want to look for which function it jumped to.

Now that we have our "patch suspect", let's see how the original (February) and modified (March) versions of SigFileWrite function compare in BinDiff.



We see that four sanity checks were added, perhaps excessively but efficiently:

  1. If length of the SIG record is less than 0x12 (minimum possible), then exit.
  2. If length of the SIG record, subtracted by 0x12, is less than length of signer's name, then exit.
  3. If pointer to the end of the string (where it will be appended) is larger than end of buffer, then exit.
  4. The final, most complex-looking check, is compiler's "artistic rendition" of multiplication of signature length by 4/3, which is the number of characters that base64-encoding will require. If signature length multiplied by 4/3 is larger than the difference between string end pointer and end of buffer, then exit.

These checks make sure that the buffer will not get overflown, and will silently prevent DNS update records from being written to the zone file if end of buffer has been reached.

 

The Micropatch

Our micropatch does logically the same as Microsoft's, but it also adds an Exploit Blocked alert and log entry in case the buffer would have been overflown, as this would highly likely be a result of an exploitation attempt instead of something that would occur under normal circumstances.



MODULE_PATH "..\Affected_Modules\dns.exe_6.1.7601.24437_64bit\dns.exe"
PATCH_ID 597
PATCH_FORMAT_VER 2
VULN_ID 6993
PLATFORM win64

patchlet_start
    PATCHLET_ID 1
    PATCHLET_TYPE 2
    PATCHLET_OFFSET 0x7be32
    N_ORIGINALBYTES 5
    JUMPOVERBYTES 21            ; eliminate some original instructions that we'll add back in
    PIT dns.exe!0x7be6c
    
    code_start
        movzx eax, word[rdi+0Eh]    ; SIG request length
        cmp ax, 12h            ; is length below 12h?
        jb EXPLOIT_EXIT            ; packet is not valid - jump to exploit exit.
       
        movzx r8d, byte[rdi+42h]    ; get signer's name length (A.mal in the exploit example)
        sub ax, 12h            ; sub 12h from SIG request length
        lea ecx, [r8+2]            ; add 2 to signers name length
        cmp ax, cx            ; compare SIG request length and signer's name length
        jb EXPLOIT_EXIT            ; if signers name length is longer, jump to exploit exit.
       
        sub ax, cx            ; sub signer's name length from SIG request length
        cmp rsi, r11            ; r11 = sprintfSafeA return (end of string), rsi = end of buffer
        jb EXPLOIT_EXIT            ; jump to exploit exit if rsi < r11
       
        movzx r10d, ax            ; r10 = SIG request length - signer's name length
        mov eax, 0x55555556        ; eax = 0x100000000 / 3
        imul r10d            ; edx:eax = r10 * 0x100000000 / 3 (rdx = r10 / 3)
        mov eax, edx            ; eax = r10 / 3
        shr eax, 1Fh            ; eax = lowest bit of (r10 / 3)
        add edx, eax            ; add the lowest bit of eax ro edx (i.e., round up the division)
        lea eax, [4+rdx*4]        ; eax = 4*(rdx+1) - length of the base64-encoded string
        movsxd rcx, eax            ; rcx = length of the base64-encoded string
        mov rax, rsi            ; rax = end of buffer
        sub rax, r11            ; rax = end of buffer - end of existing string (i.e., space available)
        cmp rax, rcx            ; is space available less than length of base64-encoded string?
        jl EXPLOIT_EXIT            ; if so, go to exploit exit
       
        movzx eax, byte[rdi+42h]    ; original code preparing arguments for Dns_SecurityKeyToBase64String call
        lea rcx, [rax+rdi+44h]
        mov edx, r10d
        add rcx, r8
        mov r8, r11
        jmp END
       
    EXPLOIT_EXIT:
        mov r11, 0            ; r11 is later moved to rax which is the return of the function
        call PIT_ExploitBlocked        ; we call Exploit Blocked
        jmp PIT_0x7be6c            ; jump to last function block
    
    END:                    ; continue normal execution
    code_end
    
patchlet_end


Here's a video of the micropatch in action. You can see that without our micropatch, the POC, launched by a local non-admin user, successfully gets the DNS Service to crash (manually stopping the service just makes the crash happen earlier so we don't have to wait). This could be leveraged to remote arbitrary code execution of attacker's code as Local System. With our micropatch applied, the POC is blocked because the corrected code prevents writing beyond the allocated memory buffer.




We created this micropatch for the following Windows versions:

  1. Windows Server 2008 R2 without Extended Security Updates, updated to January 2020
  2. Windows Server 2008 R2 with year 1 of Extended Security Updates, updated to January 2021

According to our guidelines, this micropatch requires a 0patch PRO license. By the time you're reading this, the micropatch has already been distributed to all licensed online 0patch Agents and also automatically applied except where Enterprise policies prevented that. If you're not a 0patch user and would like to use this micropatch on your computer(s), create an account in 0patch Central, install 0patch Agent and register it to your account with appropriate amount of PRO licenses. Note that no computer restart is needed for installing the agent or applying/un-applying any 0patch micropatch.

We'd like to thank Eoin Carroll and Kevin McGrath for their analysis of the vulnerability, which allowed us to create a micropatch.



While you're here: If your organization has Windows 7 or Server 2008 R2 machines with Extended Security Updates and wouldn't mind saving lots of money on less expensive security patches in 2021 that don't even need your machines to be restarted, proceed to our New Year's Resolution. The same applies if you're still using Office 2010 and want to keep patching critical vulnerabilities now that support has ended.

To learn more about 0patch, please visit our Help Center.  




 


Thursday, February 11, 2021

Windows Print Spooler Keeps Delivering Vulnerabilities, And We Keep Patching Them (CVE-2020-1030)

 

 

by Mitja Kolsek, the 0patch Team

 

Security researcher Victor Mata of Accenture published a detailed analysis of a binary planting vulnerability in Windows Print Spooler (CVE-2020-1030), which they had previously reported to Microsoft in May 2020, and a fix for which was included in September 2020 Windows Updates.

The vulnerability (see proof-of-concept) lies - once more - in Print Spooler, this time indiscriminately creating a new "spooler" folder wherever a low-privileged local user instructed it to, doing so as a Local System account and giving said user powerful permissions on such folder. While this "feature" could probably be exploited in many other ways, there is a convenient exploitation target inside the Print Spooler service itself. Namely, the service tries to load a "point and print" driver from folder %SYSTEMROOT%\System32\spool\drivers\<ENVIRONMENT>\4, which does not exist, but can be created using this very "feature".

Microsoft's patch for this issue fixed the way a non-admin user can specify the spooler folder for a printer: Print Spooler service now checks (while impersonating the user) if said user has sufficient permissions to create such folder, including some symbolic link checks to thwart symlink-related shenanigans Print Spooler has been found to be riddled with.

Our micropatch does logically the same, and unfortunately is quite large for a micropatch (172 instructions) because the symlink checks just take a lot of code.

The micropatch was only written for Windows 7 and Windows Server 2008 R2 both (32bit and 64bit) without Extended Security Updates, because other supported systems can (and should) resolve it by applying Windows Updates.

This micropatch has already been distributed to all online 0patch Agents with a PRO license. To obtain the micropatch and have it applied on your computers along with other micropatches included with a PRO license, create an account in 0patch Central, install 0patch Agent and register it to your account. Note that no computer restart is needed for installing the agent or applying/un-applying any 0patch micropatch. 

And don't forget, if your organization has Windows 7 or Server 2008 R2 machines pending ESU subscription renewal and wouldn't mind saving lots of money and stress on security patching in 2021 that doesn't even make you restart computers, proceed to this New Year's Resolution.

To learn more about 0patch, please visit our Help Center

We'd like to thank Victor Mata of Accenture for publishing their analysis and providing a proof-of-concept that allowed us to reproduce the vulnerability and create a micropatch. We also encourage security researchers to privately share their analyses with us for micropatching.

Wednesday, February 10, 2021

Micropatches for CVE-2021-24074, CVE-2021-24086, and CVE-2021-24094?

by Mitja Kolsek, the 0patch Team

 

Users are asking about micropatches for CVE-2021-24074, CVE-2021-24086, and CVE-2021-24094, remotely exploitable vulnerabilities in Windows TCP/IP stack that were fixed by February 2021 Windows Updates (and left unpatched on Windows 7 and Server 2008 R2 machines without Extended Security Updates (year 2).

According to Microsoft's blog post on the matter, the two "arbitrary code execution" vulnerabilities are "complex which make it difficult to create functional exploits, so they are not likely in the short term," but that denial-of-service attacks could quickly be devised (from reverse-engineering of patches, we assume).

At the time of this writing (February 10, 2021) we're not developing patches for these vulnerabilities. The main reason is that in order to create a patch, we need to be able to reproduce the vulnerability, i.e., we need to have a proof-of-concept or an exploit that triggers it. None of these have been published or made otherwise available yet. (For the same reasons, they're also not available to attackers.) While we could reverse-engineer patches and try to create our own exploits, our time is better spent on fixing vulnerabilities we (and attackers) already can reliably reproduce, especially if official patches for them do not exist yet (such as this Internet Explorer 0day).

A likely second reason for not patching these vulnerabilities even if we were able to reproduce them would be that these vulnerabilities are likely entirely in Windows kernel, and Microsoft's Patch Guard prevents us from patching kernel code. While this is usually not a problem as most remotely exploitable vulnerabilities are in user space (where we can patch), in this case we recommend implementing Microsoft's workarounds described in respective KB articles, specifically, executing the following on all computers without February 2021 Windows Updates or later:

netsh int ipv4 set global sourceroutingbehavior=drop
netsh int ipv6 set global reassemblylimit=0

According to Microsoft's blog post, network packets that can be used for exploiting these vulnerabilities can also be blocked by firewall, but to protect yourself from internal attackers, making the above Windows systems settings will be more effective.

 

Thursday, January 28, 2021

Windows Installer Local Privilege Escalation 0day Gets a Micropatch

 


by Mitja Kolsek, the 0patch Team

 

[Update 2/9/2021: February 2021 Windows Updates included an official fix for this vulnerability and assigned it CVE-2021-1727. According to our guidelines, this micropatch is no longer FREE, but part of a PRO subscription.]

On December 26, security researcher Abdelhamid Naceri published a blog post with a number of 0days in various security products and a local privilege escalation 0day in Windows Installer. We were mostly interested in the latter.

Abdelhamid provided a proof-of-concept (the GitHub repository is disabled at the time of this writing) which allowed us to quickly reproduce the issue on Windows 10 v2004, but we were having difficulties reproducing it on other Windows 10 versions and older Windows systems. It took us a while to troubleshoot the underlying problem with reproduction and come January 2021 Patch Tuesday, it turned out this vulnerability wasn't patched by Microsoft. Having successfully reproduced the issue by then on all Windows versions back to Windows 7, we decided to create a micropatch to protect Windows users waiting for the official patch. (The micropatch would also be the only available patch for Windows 7 without Extended Security Updates (ESU), or Windows 7 with only the first year of ESU.)


The Vulnerability

This vulnerability is a bypass of Microsoft's fix for CVE-2020-16902 (described by Abdelhamid in detail here), which was itself a bypass of Microsoft's fixes for CVE-2020-0814 and CVE-2020-1302 (also found by Abdelhamid), both of which were a bypass of Microsoft's fix for CVE-2019-1415 (found by SandboxEscaper and described here).

Confusing? Well, some things aren't easy to fix, and Windows Installer is a pretty complex beast that can break a leg if you fix its arm, and then break its tail when you fix the leg. So you want to be careful when fixing.

The core of this vulnerability, and all others listed above, is in tricking Windows Installer into using attacker's own rollback script (a *.rbs file) instead of the rollback script created by msiexec.exe during the installation. See, when installing an MSI package, Windows Installer gradually builds up a rollback script in case the installation should fail at some point, and all changes made up to that point would have to be reverted. But if a local non-admin attacker manages to replace that rollback script with one that "reverts" some system registry value such that it will point to attacker's executable..., well, we get a local privilege escalation.

The proof-of-concept is using a rollback script that changes the value of  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\Fax\ImagePath to c:\Windows\temp\asmae.exe, which results in the Fax Service using attacker's asmae.exe when the service is launched. This service was used because any user is allowed to launch it, and it's running as Local System.

As far as this particular 0day goes, Microsoft's fix - which it bypasses - was attempting to block the planting of a malicious rollback script by first determining if it was safe to use the default c:\Config.Msi folder for storing the rollback script, and if not safe, using a different folder, c:\Windows\Installer\Config.Msi, instead. Abdelhamid noticed a logical flaw in this fix, forced Windows Installer to keep using c:\Config.Msi, and then performed the same steps as in his CVE-2020-16902 proof-of-concept to elevate himself to Local System.


Our Micropatch

We confess we do not understand why Microsoft decided to add more complexity with their fix for CVE-2020-16902 when they could have just unconditionally use the c:\Windows\Installer\Config.Msi folder for the callback script and completely avoid numerous attack vectors that c:\Config.Msi is exposed to. Maybe they didn't want to clutter the Windows folder.

Be it as it may, we decided that if Microsoft deemed c:\Windows\Installer\Config.Msi folder to be acceptable for hosting the rollback script under some attacker-controllable conditions, it shouldn't break anything if we forced Windows Installer to always use it for rollback scripts. It is running as Local System so permissions shouldn't be a problem, and a local attacker can't touch this folder in any relevant way.

And here it is, the single-instruction micropatch that fixes this 0day by changing the logic of Microsoft's fix for CVE-2020-16902 such that it now always decides to use c:\Windows\Installer\Config.Msi folder:



MODULE_PATH "..\Affected_Modules\msi.dll_5.0.19041.746_64bit\msi.dll"
PATCH_ID 538
PATCH_FORMAT_VER 2
VULN_ID 6912
PLATFORM win64

patchlet_start
    PATCHLET_ID 1
    PATCHLET_TYPE 2
    PATCHLET_OFFSET 0xc2bcc
    N_ORIGINALBYTES 5
    JUMPOVERBYTES 0
    
    code_start
        mov ebx,1    ; use C:\Windows\installer\Config.Msi instead of C:\Config.Msi
    code_end
    
patchlet_end


Here's a video of the micropatch in action. You can see that without our micropatch, the POC, launched by a local non-admin user, successfully modifies a registry value pointing to the Fax Service executable, leading to execution of attacker's code by Local System. With our micropatch applied, the POC is blocked because Windows Installer cannot be tricked into using c:\Config.Msi anymore.




We created this micropatch for the following Windows versions:

  1. Windows 10 v20H2, 32bit and 64bit, updated with January 2021 updates
  2. Windows 10 v2004, 32bit and 64bit, updated with January 2021 updates
  3. Windows 10 v1909, 32bit and 64bit, updated with January 2021 updates
  4. Windows 7, 32bit and 64bit, with ESU, updated with January 2021 updates
  5. Windows 7, 32bit and 64bit, without ESU, updated with January 2020 updates

What about Windows Servers? Fortunately, Windows Servers have a default security policy preventing non-admin users from launching any installations, which successfully prevents exploitation of this vulnerability. Nevertheless, our Windows 7 micropatches will also work on Windows Server 2008 R2, updated to January 2020 (without ESU), or to January 2021 (with ESU) should their system configuration allow non-admin installations.

According to our guidelines, this micropatch is free for everyone until Microsoft issues an official fix for it. By the time you're reading this the micropatch has already been distributed to all online 0patch Agents and also automatically applied except where Enterprise policies prevented that. If you're not a 0patch user and would like to use this micropatch on your computer(s), create an account in 0patch Central, install 0patch Agent and register it to your account. Note that no computer restart is needed for installing the agent or applying/un-applying any 0patch micropatch.
 
We'd like to thank Abdelhamid Naceri for their analysis of the vulnerability and an elegant proof-of-concept, which allowed us to create a micropatch.



While you're here: If your organization has Windows 7 or Server 2008 R2 machines with Extended Security Updates and wouldn't mind saving lots of money on less expensive security patches in 2021 that don't even need your machines to be restarted, proceed to our New Year's Resolution. The same applies if you're still using Office 2010 and want to keep patching critical vulnerabilities now that support has ended.

To learn more about 0patch, please visit our Help Center.  

Thursday, January 7, 2021

Local Privilege Escalation 0day in PsExec Gets a Micropatch

 

 

by Mitja Kolsek, the 0patch Team

[Update 3/25/2021: Seventy-seven days after we had issued a free micropatch for a local privilege escalation in Microsoft PsExec, a new PsExec version fixes the issue. 0patch users staying on version 2.32 remain protected by our micropatch. This issue was assigned CVE-2021-1733, although it was not properly fixed in version 2.32 as stated in Microsoft's advisory.]
 
[Update 2/17/2021: Corrected a sentence implying that PsExec may be part of various enterprise tools, while it's just commonly used in conjunction with such tools. (Thanks @wdormann)]

[Update 1/28/2021: Since our publication of micropatch for PsExec version 2.2, PsExec has been updated to versions 2.30, 2.31 and finally 2.32. where it still resides today. David was able to update his POC for each version so the current version 2.32. is still vulnerable to the same attack. Since this version seems to be here to stay for a while, we decided to port our micropatch to it to keep 0patch users with the latest PsExec version protected.]

Last month, security researcher David Wells of Tenable published an analysis of a local privilege escalation vulnerability in PsExec, a powerful management tool from SysInternals (acquired by Microsoft) that allows launching executables on remote computers.
 
It would be hard to find a Windows admin who hasn't used PsExec at some point in time, and just a tiny bit less hard to find one who isn't using it on a regular basis. Granted, some may not even know they're using PsExec because it's commonly used in conjunction with various enterprise tools - tools like JetBrains TeamCity, BMC Server Automation, Chocolatey and SolarWinds Orion.

The Vulnerability

 
The vulnerability is a pretty classic named pipe hijacking (a.k.a. named pipe squatting). When PsExec tries to launch an executable on the remote computer, it creates a temporary Windows service there using PSEXESVC.EXE which it extracts from its own body, launches that service under Local System user, and connects to its named pipe to provide it launch instructions. PSEXESVC.EXE creates the named pipe with permissions that don't allow a non-admin or non-system user to connect to it, which is good because otherwise any user could instruct the service to run arbitrary executable as Local System.
 
Now, the attack comprises a malicious local unprivileged process creating a named pipe with the same name as PSEXESVC.EXE uses, only before the service creates it. PSEXESVC.EXE, running as Local System, subsequently tries to create the same named pipe, but merely re-opens the existing one, leaving its permissions intact. At that point, attacker can connect to the named pipe and make the service run anything.
 
David has provided an elegant proof-of-concept for this vulnerability.
 
So which systems are at risk by this issue? Basically every Windows machine that admins remotely launch executables on using PsExec (or management tools utilizing PsExec) if the machine already has a non-admin attacker there trying to elevate their privileges.


Official Patch? It's... Complicated

 
At the time of this writing, there is no official patch available from Microsoft. PsExec.exe, and PsExec64.exe, which encapsulate the vulnerable PSEXESVC.EXE, are part of the PsTools suite, and were last updated in June 2016. According to Tenable's write-up, PsExec versions from 1.72 (built in 2006) to the latest version 2.2 (built in 2016) are all affected, meaning that the vulnerability has been there for about 14 years. [Update 1/28/2021: current version 2.32 is still affected.]

Note that PsExec is not part of Windows, and is also unlikely to be patchable with Windows Updates as it doesn't even have a designated installation location (one can just copy it anywhere and use it as a standalone executable). PsExec also doesn't have its own integrated update mechanism, meaning that while Microsoft can issue a new, patched version of it and put it on their website, all the vulnerable PsExec's out there will remain vulnerable until admins manually replace them with this new version.


Our Micropatch

 
Let's see how the relevant part of PSEXESVC.EXE looks where named pipe is created and connection requests accepted.



Function CreateNamedPipe is being called in a loop, each time waiting for an incoming request, spawning a new thread to process that request, and repeating the loop.
 
When fixing named pipe hijacking/squatting vulnerabilities, the obvious approach that comes to mind is using the FILE_FLAG_FIRST_PIPE_INSTANCE flag in the CreateNamedPipe call, which only allows the pipe to be created if it is the first instance of the pipe. We actually tried this approach but while it stopped the attack (as attacker's pipe was the first instance), it also broke PsExec because in the above loop, when the first request is accepted and sent for processing, a new instance of the pipe gets created - which is no longer the first instance.
 
So we went for the second best option - checking for existence of the named pipe immediately before the loop. We used a call to CreateNamedPipe with FILE_FLAG_FIRST_PIPE_INSTANCE to determine if a named pipe with this name already exists - and if so, we immediately terminate PSEXESVC.EXE, logging an "Exploit Blocked" event in the process.

Our micropatch has only 21 CPU instructions and should be easy to understand for anyone knowing x86 assembly and Windows API functions:
 


MODULE_PATH "..\Affected_Modules\PSEXESVC.exe_2.2_32bit\PSEXESVC.exe"
PATCH_ID 536
PATCH_FORMAT_VER 2
VULN_ID 6910
PLATFORM win32

patchlet_start
    PATCHLET_ID 1
    PATCHLET_TYPE 2
    PATCHLET_OFFSET 0x38ae
    N_ORIGINALBYTES 5
    JUMPOVERBYTES 12
    PIT Kernel32.dll!CreateNamedPipeW,PSEXESVC.exe!0x4e7a,Kernel32.dll!CloseHandle,Kernel32.dll!ExitProcess
    
    ; 0x4e7a -> __swprintf

    code_start   
        ; first three instructions repeated from original code to make
        ; room for the patch JMP
        lea eax, [ebp-414h]        ; buffer for pipe name
        push eax                   ; buffer on stack
        call PIT_0x4e7a            ; call __swprintf
        push 0                     ; lpSecurityAttributes
        push 0                     ; nDefaultTimeOut, A value of zero will result
                                   ; in a default time-out of 50 milliseconds.
        push 10000h                ; nInBufferSize
        push 10000h                ; nOutBufferSize
        push 0ffh                  ; nMaxInstances (the same number must be specified
                                   ; for other instances of the pipe.)
        push 6                     ; dwPipeMode (The same type mode must be specified
                                   ; for each instance of the pipe.)
        push 80003h                ; dwOpenMode - FILE_FLAG_FIRST_PIPE_INSTANCE
        lea eax, [ebp-414h]        ; buffer for pipe name
        push eax                   ; lpName
        call PIT_CreateNamedPipeW  ; Creates an instance of a named pipe and returns
                                   ; a handle for subsequent pipe operations.
        mov edi, eax
        cmp eax, 0xFFFFFFFF        ; check if handle exists
        jne CONTINUE               ; if Handle != -1 (INVALID_HANDLE_VALUE) continue
                                   ; with normal execution
       
        call PIT_ExploitBlocked    ; Exploit blocked pop up
        push -1                    ; uExitCode
        call PIT_ExitProcess       ; Ends the calling process and all its threads.   
    
    CONTINUE:
        push edi                   ; edi = pipe handle
        call PIT_CloseHandle       ; close the pipe handle.
    code_end
    
patchlet_end



Here's a video of the micropatch in action:





According to our guidelines, this micropatch is immediately available to ALL 0patch users for absolutely no cost. Note that no computer restart is needed for installing the agent or applying/un-applying any 0patch micropatches.

 

Frequently Asked Questions [Updated 1/28/2021]

 
Q: Which versions of PsExec does the micropatch fix?

A: Our micropatch currently applies to 32bit and 64bit PsExec versions 2.2 and 2.32. We might port it to older versions of PsExec as needed.
 
 
Q: How do we get the micropatch applied?
  1. Create a free 0patch account at https://central.0patch.com.
  2. Download and install 0patch Agent on all computers on which you're running executables with PsExec, then register it to your 0patch account.
  3. Make sure to use PsExec 2.2 or 2.32, or the micropatch won't get applied

Q: Does 0patch Agent have to be running on computers where we run PsExec, or remote computers where executables get launched using PsExec?
 
A: 0patch Agent needs to be running on the remote computers where executables get launched using PsExec. What PsExec does is copy PSEXESVC.EXE to the remote computer (into c:\Windows) and registers it remotely as a service on that computer, then launches that service. This remote PSEXESVC.EXE is what needs to be patched. Note that 0patch Agent can also safely be running on the computer where you run PsExec.
 
 
Q: Can we easily deploy this patch to multiple computers?
 
A: 0patch Agent supports silent (unattended) installation with auto-registration, and central management via 0patch Central. Please see User Manual for details and ask sales@0patch.com for an Enterprise trial.


Q: Will the micropatch also fix PsExec that is integrated into our enterprise product?

A: As long as PsExec used by the product is version 2.2 or 2.32, our micropatch will fix it. But again, 0patch Agent must be present on computers being managed by the enterprise product, not on the machine where said product is installed. If your enterprise product is using another version of PsExec and you cannot replace it, please contact support@0patch.com.


Q: Is there any other way to prevent exploitation of the described vulnerability?

A: Not to our knowledge. Until Microsoft issues a fixed version of PsExec, ours is the only patch that exists.


Q: Is this vulnerability a big deal?

A: Depends on your threat model. This vulnerability allows an attacker who can already run code on your remote computer as a non-admin (e.g., by logging in as a regular Terminal Server user, or establishing an RDP session as a domain user, or breaking into a vulnerable unprivileged service running on the remote computer) to elevate their privileges to Local System and completely take over the machine as soon as anyone uses PsExec against that machine. For home users and small businesses this is probably not a high-priority threat, while for large organizations it may be.



We'd like to thank  David Wells of Tenable for their excellent presentation of the vulnerability and an elegant proof-of-concept, which allowed us to create a micropatch.

While you're here: If your organization has Windows 7 or Server 2008 R2 machines with Extended Security Updates and wouldn't mind saving lots of money on less expensive security patches in 2021 that don't even need your machines to be restarted, proceed to our New Year's Resolution. The same applies if you're still using Office 2010 and want to keep patching critical vulnerabilities now that support has ended.

To learn more about 0patch, please visit our Help Center