Using clone will not undo features that have been enabled by libc. An example of this already happening is rseq, which could cause the kernel to read/write memory of the userspace process. In the future the standard library might also use mseal by default to protect itself, which would also thwart our attempts at unmapping everything. Solve all this by taking a step back and doing an execve into a tiny static binary that sets up the minimal environment required for the stub without using any standard library. That way we have a clean execution environment that is fully under the control of UML. Note that this changes things a bit as the FDs are not anymore shared with the kernel. Instead, we explicitly share the FDs for the physical memory and all existing iomem regions. Doing this is fine, as iomem regions cannot be added at runtime. Signed-off-by: Benjamin Berg <benjamin.berg@intel.com> Link: https://patch.msgid.link/20240919124511.282088-3-benjamin@sipsolutions.net [use pipe() instead of pipe2(), remove unneeded close() calls] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
88 lines
2.4 KiB
C
88 lines
2.4 KiB
C
#include <sys/ptrace.h>
|
|
#include <sys/prctl.h>
|
|
#include <asm/unistd.h>
|
|
#include <sysdep/stub.h>
|
|
#include <stub-data.h>
|
|
|
|
void _start(void);
|
|
|
|
noinline static void real_init(void)
|
|
{
|
|
struct stub_init_data init_data;
|
|
unsigned long res;
|
|
struct {
|
|
void *ss_sp;
|
|
int ss_flags;
|
|
size_t ss_size;
|
|
} stack = {
|
|
.ss_size = STUB_DATA_PAGES * UM_KERN_PAGE_SIZE,
|
|
};
|
|
struct {
|
|
void *sa_handler_;
|
|
unsigned long sa_flags;
|
|
void *sa_restorer;
|
|
unsigned long long sa_mask;
|
|
} sa = {
|
|
/* Need to set SA_RESTORER (but the handler never returns) */
|
|
.sa_flags = SA_ONSTACK | SA_NODEFER | SA_SIGINFO | 0x04000000,
|
|
/* no need to mask any signals */
|
|
.sa_mask = 0,
|
|
};
|
|
|
|
/* set a nice name */
|
|
stub_syscall2(__NR_prctl, PR_SET_NAME, (unsigned long)"uml-userspace");
|
|
|
|
/* read information from STDIN and close it */
|
|
res = stub_syscall3(__NR_read, 0,
|
|
(unsigned long)&init_data, sizeof(init_data));
|
|
if (res != sizeof(init_data))
|
|
stub_syscall1(__NR_exit, 10);
|
|
|
|
stub_syscall1(__NR_close, 0);
|
|
|
|
/* map stub code + data */
|
|
res = stub_syscall6(STUB_MMAP_NR,
|
|
init_data.stub_start, UM_KERN_PAGE_SIZE,
|
|
PROT_READ | PROT_EXEC, MAP_FIXED | MAP_SHARED,
|
|
init_data.stub_code_fd, init_data.stub_code_offset);
|
|
if (res != init_data.stub_start)
|
|
stub_syscall1(__NR_exit, 11);
|
|
|
|
res = stub_syscall6(STUB_MMAP_NR,
|
|
init_data.stub_start + UM_KERN_PAGE_SIZE,
|
|
STUB_DATA_PAGES * UM_KERN_PAGE_SIZE,
|
|
PROT_READ | PROT_WRITE, MAP_FIXED | MAP_SHARED,
|
|
init_data.stub_data_fd, init_data.stub_data_offset);
|
|
if (res != init_data.stub_start + UM_KERN_PAGE_SIZE)
|
|
stub_syscall1(__NR_exit, 12);
|
|
|
|
/* setup signal stack inside stub data */
|
|
stack.ss_sp = (void *)init_data.stub_start + UM_KERN_PAGE_SIZE;
|
|
stub_syscall2(__NR_sigaltstack, (unsigned long)&stack, 0);
|
|
|
|
/* register SIGSEGV handler */
|
|
sa.sa_handler_ = (void *) init_data.segv_handler;
|
|
res = stub_syscall4(__NR_rt_sigaction, SIGSEGV, (unsigned long)&sa, 0,
|
|
sizeof(sa.sa_mask));
|
|
if (res != 0)
|
|
stub_syscall1(__NR_exit, 13);
|
|
|
|
stub_syscall4(__NR_ptrace, PTRACE_TRACEME, 0, 0, 0);
|
|
|
|
stub_syscall2(__NR_kill, stub_syscall0(__NR_getpid), SIGSTOP);
|
|
|
|
stub_syscall1(__NR_exit, 14);
|
|
|
|
__builtin_unreachable();
|
|
}
|
|
|
|
void _start(void)
|
|
{
|
|
char *alloc;
|
|
|
|
/* Make enough space for the stub (including space for alignment) */
|
|
alloc = __builtin_alloca((1 + 2 * STUB_DATA_PAGES - 1) * UM_KERN_PAGE_SIZE);
|
|
asm volatile("" : "+r,m"(alloc) : : "memory");
|
|
|
|
real_init();
|
|
}
|