Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 26 Apr 2019 16:22:44 -0700
From: nadav.amit@...il.com
To: Peter Zijlstra <peterz@...radead.org>,
	Borislav Petkov <bp@...en8.de>,
	Andy Lutomirski <luto@...nel.org>,
	Ingo Molnar <mingo@...hat.com>
Cc: linux-kernel@...r.kernel.org,
	x86@...nel.org,
	hpa@...or.com,
	Thomas Gleixner <tglx@...utronix.de>,
	Nadav Amit <nadav.amit@...il.com>,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	linux_dti@...oud.com,
	linux-integrity@...r.kernel.org,
	linux-security-module@...r.kernel.org,
	akpm@...ux-foundation.org,
	kernel-hardening@...ts.openwall.com,
	linux-mm@...ck.org,
	will.deacon@....com,
	ard.biesheuvel@...aro.org,
	kristen@...ux.intel.com,
	deneen.t.dock@...el.com,
	Rick Edgecombe <rick.p.edgecombe@...el.com>,
	Nadav Amit <namit@...are.com>,
	Arnaldo Carvalho de Melo <acme@...nel.org>
Subject: [PATCH v6 05/24] uprobes: Initialize uprobes earlier

From: Nadav Amit <namit@...are.com>

In order to have a separate address space for text poking, we need to
duplicate init_mm early during start_kernel(). This, however, introduces
a problem since uprobes functions are called from dup_mmap(), but
uprobes is still not initialized in this early stage.

Since uprobes initialization is necassary for fork, and since all the
dependant initialization has been done when fork is initialized (percpu
and vmalloc), move uprobes initialization to fork_init(). It does not
seem uprobes introduces any security problem for the poking_mm.

Crash and burn if uprobes initialization fails, similarly to other early
initializations. Change the init_probes() name to probes_init() to match
other early initialization functions name convention.

Cc: Arnaldo Carvalho de Melo <acme@...nel.org>
Reported-by: kernel test robot <lkp@...el.com>
Signed-off-by: Nadav Amit <namit@...are.com>
---
 include/linux/uprobes.h | 5 +++++
 kernel/events/uprobes.c | 8 +++-----
 kernel/fork.c           | 1 +
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/include/linux/uprobes.h b/include/linux/uprobes.h
index 103a48a48872..12bf0b68ed92 100644
--- a/include/linux/uprobes.h
+++ b/include/linux/uprobes.h
@@ -115,6 +115,7 @@ struct uprobes_state {
 	struct xol_area		*xol_area;
 };
 
+extern void __init uprobes_init(void);
 extern int set_swbp(struct arch_uprobe *aup, struct mm_struct *mm, unsigned long vaddr);
 extern int set_orig_insn(struct arch_uprobe *aup, struct mm_struct *mm, unsigned long vaddr);
 extern bool is_swbp_insn(uprobe_opcode_t *insn);
@@ -154,6 +155,10 @@ extern void arch_uprobe_copy_ixol(struct page *page, unsigned long vaddr,
 struct uprobes_state {
 };
 
+static inline void uprobes_init(void)
+{
+}
+
 #define uprobe_get_trap_addr(regs)	instruction_pointer(regs)
 
 static inline int
diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index c5cde87329c7..e6a0d6be87e3 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -2294,16 +2294,14 @@ static struct notifier_block uprobe_exception_nb = {
 	.priority		= INT_MAX-1,	/* notified after kprobes, kgdb */
 };
 
-static int __init init_uprobes(void)
+void __init uprobes_init(void)
 {
 	int i;
 
 	for (i = 0; i < UPROBES_HASH_SZ; i++)
 		mutex_init(&uprobes_mmap_mutex[i]);
 
-	if (percpu_init_rwsem(&dup_mmap_sem))
-		return -ENOMEM;
+	BUG_ON(percpu_init_rwsem(&dup_mmap_sem));
 
-	return register_die_notifier(&uprobe_exception_nb);
+	BUG_ON(register_die_notifier(&uprobe_exception_nb));
 }
-__initcall(init_uprobes);
diff --git a/kernel/fork.c b/kernel/fork.c
index 9dcd18aa210b..44fba5e5e916 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -815,6 +815,7 @@ void __init fork_init(void)
 #endif
 
 	lockdep_init_task(&init_task);
+	uprobes_init();
 }
 
 int __weak arch_dup_task_struct(struct task_struct *dst,
-- 
2.17.1

Powered by blists - more mailing lists

Your e-mail address:

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.