lkml.org 
[lkml]   [2012]   [Jun]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[PATCH 3/3] uprobes: document uprobe_register() vs uprobe_mmap() race
From: Peter Zijlstra <a.p.zijlstra@chello.nl>

Because the mind is treacherous and makes us forget we need to write
stuff down.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
kernel/events/uprobes.c | 31 ++++++++++++++++++++++++++++---
1 files changed, 28 insertions(+), 3 deletions(-)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index c9836ae..87a47c6 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -44,6 +44,23 @@ static DEFINE_SPINLOCK(uprobes_treelock); /* serialize rbtree access */

#define UPROBES_HASH_SZ 13

+/*
+ * We need separate register/unregister and mmap/munmap lock hashes because
+ * of mmap_sem nesting.
+ *
+ * uprobe_register() needs to install probes on (potentially) all processes
+ * and thus needs to acquire multiple mmap_sems (consequtively, not
+ * concurrently), whereas uprobe_mmap() is called while holding mmap_sem
+ * for the particular process doing the mmap.
+ *
+ * uprobe_register()->register_for_each_vma() needs to drop/acquire mmap_sem
+ * because of lock order against i_mmap_mutex. This means there's a hole in
+ * the register vma iteration where a mmap() can happen.
+ *
+ * Thus uprobe_register() can race with uprobe_mmap() and we can try and
+ * install a probe where one is already installed.
+ */
+
/* serialize (un)register */
static struct mutex uprobes_mutex[UPROBES_HASH_SZ];

@@ -353,7 +370,9 @@ out:
int __weak set_swbp(struct arch_uprobe *auprobe, struct mm_struct *mm, unsigned long vaddr)
{
int result;
-
+ /*
+ * See the comment near uprobes_hash().
+ */
result = is_swbp_at_addr(mm, vaddr);
if (result == 1)
return -EEXIST;
@@ -850,6 +869,10 @@ static int register_for_each_vma(struct uprobe *uprobe, bool is_register)

if (is_register) {
err = install_breakpoint(uprobe, mm, vma, info->vaddr);
+ /*
+ * We can race against uprobe_register(), see the
+ * comment near uprobe_hash().
+ */
if (err == -EEXIST)
err = 0;
} else {
@@ -1058,8 +1081,10 @@ int uprobe_mmap(struct vm_area_struct *vma)
}

ret = install_breakpoint(uprobe, vma->vm_mm, vma, vaddr);
-
- /* Ignore double add: */
+ /*
+ * We can race against uprobe_register(), see the
+ * comment near uprobe_hash().
+ */
if (ret == -EEXIST) {
ret = 0;

--
1.5.5.1



\
 
 \ /
  Last update: 2012-06-04 17:41    [W:0.092 / U:4.236 seconds]
©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site