Factor-out x86_64 specific code

Libpulp is deeply tied on x86_64 architecture. This commit factors out
its code and generalize some of its concepts in order to work in other
architectures.

Signed-off-by: gbelinassi <[email protected]>

configure.ac

@@ -23,6 +23,9 @@ AC_INIT([libpulp],[0.3.6],[[email protected]])
 AC_CONFIG_AUX_DIR([config])
 AC_CONFIG_MACRO_DIRS([config])
 
+# For multiarch builds.
+AC_CANONICAL_TARGET
+
 # Check for the availability of macros from autoconf-archive.
 AC_MSG_CHECKING([autoconf-archive availability])
 m4_ifndef([AX_CHECK_COMPILE_FLAG], [ax_available="no"])
@@ -223,15 +226,10 @@ AC_SUBST([PAGE_SIZE], [$($GETCONF PAGE_SIZE)]))
 AC_SUBST([ULP_NOPS_LEN], [16])
 AC_SUBST([PRE_NOPS_LEN], [14])
 
-AC_SUBST([ULP_NOPS_LEN_ENDBR64], [$(expr $ULP_NOPS_LEN + 4)])
-AC_SUBST([PRE_NOPS_LEN_ENDBR64], [$(expr $PRE_NOPS_LEN + 4)])
-
 AC_DEFINE_UNQUOTED([ULP_NOPS_LEN], [$ULP_NOPS_LEN],
 [Total number of padding nops])
 AC_DEFINE_UNQUOTED([PRE_NOPS_LEN], [$PRE_NOPS_LEN],
 [Padding nops before the entry point of functions])
-AC_DEFINE_UNQUOTED([ULP_NOPS_LEN_ENDBR64], [$ULP_NOPS_LEN_ENDBR64],
-[Total number of padding nops when endbr64 is issued])
 
 AC_CONFIG_FILES([Makefile
 		 include/Makefile

include/arch/x86_64/arch_common.h

@@ -0,0 +1,40 @@
+#ifndef _ARCH_X86_64_H
+#define _ARCH_X86_64_H
+
+/** Intel endbr64 instruction optcode.  */
+#define INSN_ENDBR64 0xf3, 0x0f, 0x1e, 0xfa
+
+/** Offset of TLS pointer.  */
+#define TLS_DTV_OFFSET 0
+
+/** Struct used to store the registers in memory.  */
+typedef struct user_regs_struct registers_t;
+
+/** Register in which the function stores the return value.  */
+#define FUNCTION_RETURN_REG(reg) ((reg).rax)
+
+/** Register which acts as a program counter.  */
+#define PROGRAM_COUNTER_REG(reg) ((reg).rip)
+
+/** Register which acts as top of stack.  */
+#define STACK_TOP_REG(reg)       ((reg).rsp)
+
+/** Set the GLOBAL ENTRYPOINT REGISTER, which in x86_64 doesn't exist.  */
+#define SET_GLOBAL_ENTRYPOINT_REG(reg, val)
+
+/** Program load bias, which can be recovered by running `ld --verbose`.  */
+#define EXECUTABLE_START      0x400000UL
+
+/** The red zone.  */
+#define RED_ZONE_LEN          128
+
+/** File name of the dynamic loader.  */
+#define LD_LINUX              "ld-linux-x86-64.so.2"
+
+/**
+ * Number of bytes that the kernel subtracts from the program counter,
+ * when an ongoing syscall gets interrupted and must be restarted.
+ */
+#define RESTART_SYSCALL_SIZE  2
+
+#endif

include/ulp.h

@@ -81,6 +81,8 @@ int __ulp_apply_patch();
 void __ulp_print();
 
 /* functions */
+void *memwrite(void *dest, const void *src, size_t n);
+
 void free_metadata(struct ulp_metadata *ulp);
 
 int unload_handlers(struct ulp_metadata *ulp);

include/ulp_common.h

@@ -35,7 +35,6 @@
  * scenarios.  */
 #define ULP_METADATA_BUF_LEN (512 * 1024)
 #define ULP_PATH_LEN 256
-#define RED_ZONE_LEN 128
 
 #define ARRAY_LENGTH(v) (sizeof(v) / sizeof(*(v)))
 

lib/Makefile.am

@@ -24,8 +24,8 @@ libpulp_la_SOURCES = \
   interpose.c \
   msg_queue.c \
   error.c \
-  ulp_prologue.S \
-  ulp_interface.S
+  arch/$(target_cpu)/ulp_interface.S \
+  arch/$(target_cpu)/patch.c
 libpulp_la_DEPENDENCIES= libpulp.versions
 libpulp_la_LDFLAGS = \
   -ldl \
@@ -36,6 +36,6 @@ libpulp_la_LDFLAGS = \
 
 libpulp_la_LIBADD = $(top_builddir)/common/libcommon.la
 
-AM_CFLAGS += -I$(top_srcdir)/include
+AM_CFLAGS += -I$(top_srcdir)/include -I$(top_srcdir)/include/arch/$(target_cpu)
 
 EXTRA_DIST = libpulp.versions

lib/arch/x86_64/patch.c

@@ -0,0 +1,177 @@
+/*
+ *  libpulp - User-space Livepatching Library
+ *
+ *  Copyright (C) 2017-2024 SUSE Software Solutions GmbH
+ *
+ *  This file is part of libpulp.
+ *
+ *  libpulp is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU Lesser General Public
+ *  License as published by the Free Software Foundation; either
+ *  version 2.1 of the License, or (at your option) any later version.
+ *
+ *  libpulp is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *  Lesser General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with libpulp.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#define _GNU_SOURCE
+
+#include <stddef.h>
+#include <string.h>
+
+#include "config.h"
+#include "error.h"
+#include "msg_queue.h"
+#include "ulp.h"
+#include "arch/x86_64/arch_common.h"
+
+/** Intel endbr64 instruction optcode.  */
+static const uint8_t insn_endbr64[] = {INSN_ENDBR64};
+
+/* clang-format off */
+
+/** Offset of the data entry in the ulp_prologue.  */
+#define ULP_DATA_OFFSET 6           // ------------------------------+
+                                    //                               |
+static char ulp_prologue[ULP_NOPS_LEN] = { //                               |
+  // Preceding nops                                                  |
+  0xff, 0x25, 0, 0, 0, 0,           // jmp     0x0(%rip) <-------+   |
+  0, 0, 0, 0, 0, 0, 0, 0,           // <data>  &__ulp_prolog     | <-+
+  // Function entry is here                                      |
+  0xeb, -(PRE_NOPS_LEN + 2)         // jmp ----------------------+
+  // (+2 because the previous jump consumes 2 bytes.
+};
+
+#define ULP_NOPS_LEN_ENDBR64    (ULP_NOPS_LEN + 4)
+
+/** Offset of the data entry in the ulp_prologue.  */
+static char ulp_prologue_endbr64[ULP_NOPS_LEN_ENDBR64] = {
+  // Preceding nops
+  0xff, 0x25, 0, 0, 0, 0,           // jmp     0x0(%rip) <-------+
+  0, 0, 0, 0, 0, 0, 0, 0,           // <data>  &__ulp_prolog     |
+  // Function entry is here                                      |
+  INSN_ENDBR64,                     // endbr64                   |
+  0xeb, -(PRE_NOPS_LEN + 2 + 4)     // jmp ----------------------+
+  // (+2 because the previous jump consumes 2 bytes.
+};
+/* clang-format on */
+
+/** @brief Write new function address into data prologue of  `old_fentry`.
+ *
+ *  This function replaces the `<data>` section in prologue `old_fentry`
+ *  with a pointer to the new function given by `manager`, which will
+ *  replace the to be patched function.
+ *
+ *  @param old_fentry Pointer to prologue of to be replaced function
+ *  @param manager Address of new function.
+ */
+void
+ulp_patch_addr_absolute(void *old_fentry, void *manager)
+{
+  char *dst = (char *)old_fentry + ULP_DATA_OFFSET;
+  memwrite(dst, &manager, sizeof(void *));
+}
+
+/** @brief Copy the ulp proglogue layout into the function to be patched's
+ * prologue
+ *
+ * This function copies the new code prologue into the old function prologue
+ * in order to redirect the execution to the new function.
+ *
+ */
+static void
+ulp_patch_prologue_layout(void *old_fentry, const char *prologue, int len)
+{
+  memwrite(old_fentry, prologue, len);
+}
+
+/** @brief skip the ulp prologue.
+ *
+ * When a function gets live patch, the nops at its entry point get replaced
+ * with a backwards-jump to a small segment of code that redirects execution to
+ * the new version of the function. However, when all live patches to said
+ * function are deactivated (because the live patches have been reversed), the
+ * need for the backwards-jump is gone.
+ *
+ * The following function replaces the backwards-jump with nops, thus making
+ * the target function look like it did at the beginning of execution, i.e.
+ * without live patches.
+ *
+ * @param fentry        Address to write the prologue to.
+ */
+void
+ulp_skip_prologue(void *fentry)
+{
+  static const char insn_nop2[] = { 0x66, 0x90 };
+  int bias = 0;
+  if (memcmp(fentry, insn_endbr64, sizeof(insn_endbr64)) == 0)
+    bias += sizeof(insn_endbr64);
+
+  /* Do not jump backwards on function entry (0x6690 is a nop on x86). */
+  memwrite((char *)fentry + bias, insn_nop2, sizeof(insn_nop2));
+}
+
+/** @brief Actually patch the old function with the new function
+ *
+ * This function will finally patch the old function pointed by `old_faddr`
+ * with the one pointed by `new_faddr`, replacing the ulp NOP prologue with
+ * the intended content to redirect to the new function.
+ *
+ * @param old_faddr     Address of the old function.
+ * @param new_faddr     Address of the new function.
+ * @param enable        False to disable the redirection to the new function.
+ *
+ * @return              0 if success, error code otherwise.
+ */
+int
+ulp_patch_addr(void *old_faddr, void *new_faddr, int enable)
+{
+  void *addr;
+
+  int ulp_nops_len;
+  const char *prologue;
+
+  const unsigned char *as_bytes = old_faddr;
+
+  /* Check if the first instruction of old_function is endbr64.  In this
+     case, we have to handle things differently.  */
+  if (memcmp(old_faddr, insn_endbr64, sizeof(insn_endbr64)) == 0) {
+    ulp_nops_len = ULP_NOPS_LEN_ENDBR64;
+    prologue = ulp_prologue_endbr64;
+    as_bytes += sizeof(insn_endbr64);
+  }
+  else {
+    ulp_nops_len = ULP_NOPS_LEN;
+    prologue = ulp_prologue;
+  }
+
+  /* Check if we have the two NOP sequence or a JMP ref8 insn.  Else we might
+     be attempting to patch a non-livepatchable function.  */
+
+  if (!(as_bytes[0] == 0xEB ||
+        (as_bytes[1] == 0x90 &&
+         (as_bytes[0] == 0x90 || as_bytes[0] == 0x66)))) {
+    WARN("Function at addr %lx is not livepatchable",
+         (unsigned long)old_faddr);
+    return ENOPATCHABLE;
+  }
+
+  /* Find the starting address of the pages containing the nops. */
+  addr = old_faddr - PRE_NOPS_LEN;
+
+  /* Actually patch the prologue. */
+  if (enable) {
+    ulp_patch_prologue_layout(addr, prologue, ulp_nops_len);
+    ulp_patch_addr_absolute(addr, new_faddr);
+  }
+  else {
+    ulp_skip_prologue(old_faddr);
+  }
+
+  return 0;
+}

lib/ulp_interface.S → lib/arch/x86_64/ulp_interface.S

@@ -1,7 +1,7 @@
 /*
  *  libpulp - User-space Livepatching Library
  *
- *  Copyright (C) 2017-2021 SUSE Software Solutions GmbH
+ *  Copyright (C) 2017-2024 SUSE Software Solutions GmbH
  *
  *  This file is part of libpulp.
  *

lib/interpose.c

@@ -41,6 +41,7 @@
 #include "msg_queue.h"
 #include "ulp.h"
 #include "ulp_common.h"
+#include "arch_common.h"
 
 /* This header should be included last, as it poisons some symbols.  */
 #include "error.h"
@@ -393,8 +394,7 @@ get_ld_global_locks()
   const char *tok;
   int major, minor;
 
-  void *rtld_global =
-      get_loaded_symbol_addr("ld-linux-x86-64.so.2", "_rtld_global", NULL);
+  void *rtld_global = get_loaded_symbol_addr(LD_LINUX, "_rtld_global", NULL);
   if (!rtld_global) {
     libpulp_crash("symbol _rtld_global not found in ld-linux-x86_64.so\n");
   }