Re: [RFC PATCH v3 4/6] livepatch: reuse module loader code to write relocations

On Mon, Jan 11, 2016 at 05:56:13PM +0100, Petr Mladek wrote:

On Fri 2016-01-08 14:28:22, Jessica Yu wrote:

quoted

Reuse module loader code to write relocations, thereby eliminating the need
for architecture specific relocation code in livepatch. Namely, we reuse
apply_relocate_add() in the module loader to write relocations instead of
duplicating functionality in livepatch's klp_write_module_reloc(). To apply
relocation sections, remaining SHN_LIVEPATCH symbols referenced by relocs
are resolved and then apply_relocate_add() is called to apply those
relocations.

In addition, remove x86 livepatch relocation code. It is no longer needed
since symbol resolution and relocation work have been offloaded to module
loader.

--- a/kernel/livepatch/core.c
+++ b/kernel/livepatch/core.c

@@ -204,74 +207,70 @@ static int klp_find_object_symbol(const char *objname, const char *name,
 	return -EINVAL;
 }
 
-/*
- * external symbols are located outside the parent object (where the parent
- * object is either vmlinux or the kmod being patched).
- */
-static int klp_find_external_symbol(struct module *pmod, const char *name,
-				    unsigned long *addr)
+static int klp_resolve_symbols(Elf_Shdr *relsec, struct module *pmod)
 {
-	const struct kernel_symbol *sym;
+	int i, len, ret = 0;
+	Elf_Rela *relas;
+	Elf_Sym *sym;
+	char *symname, *sym_objname;
 
-	/* first, check if it's an exported symbol */
-	preempt_disable();
-	sym = find_symbol(name, NULL, NULL, true, true);
-	if (sym) {
-		*addr = sym->value;
-		preempt_enable();
-		return 0;
+	relas = (Elf_Rela *) relsec->sh_addr;
+	/* For each rela in this .klp.rel. section */
+	for (i = 0; i < relsec->sh_size / sizeof(Elf_Rela); i++) {
+		sym = pmod->core_symtab + ELF_R_SYM(relas[i].r_info);
+		symname = pmod->core_strtab + sym->st_name;
+
+		len = strcspn(symname + KLP_TAG_LEN, ".");

We should check that len is non-zero. Otherwise, sym_objname might
be empty string and symname might overflow below.

Also we should check that symname really starts with .klp.sym. to
avoid invalid memory access.

It would also be good to check for SHN_LIVEPATCH.

-- 
Josh