Fix kallsyms/insmod/rmmod race

PATCH In review David Howells 2 participants 6 messages Jan 17, 2005 – Jan 28, 2005

lore

Share & tools

lore

David Howells

author

2005-01-17 16:28 UTC #0 lore raw

PATCH 1/1 Fix kallsyms/insmod/rmmod race

The attached patch fixes a race between kallsyms and insmod/rmmod.

The problem is this:

(1) The various kallsyms functions poke around in the module list without any locking so that they can be called from the oops handler.

(2) Although insmod and rmmod use locks to exclude each other, these have no effect on the kallsyms function.

(3) Although rmmod modifies the module state with the machine "stopped", it hasn't removed the metadata from the module metadata list, meaning that as soon as the machine is "restarted", the metadata can be observed by kallsyms.

It's not possible to say that an item in that list should be ignored if it's state is marked as inactive - you can't get at the state information because you can't trust the metadata in which it is embedded.

Furthermore, list linkage information is embedded in the metadata too, so you can't trust that either...

(4) kallsyms may be walking the module list without a lock whilst either insmod or rmmod are busy changing it. insmod probably isn't a problem since nothing is going a way, but rmmod is as it's deleting an entry.

(5) Therefore nothing that uses these functions can in any way trust any pointers to "static" data (such as module symbol names or module names) that are returned.

(6) On ppc64 the problems are exacerbated since the hypervisor may reschedule bits of the kernel, making operations that appear adjacent occur a long time apart.

This patch fixes the race by only linking/unlinking modules into/from the master module list with the machine in the "stopped" state. This means that any "static" information can be trusted as far as the next kernel reschedule on any given CPU without the need to hold any locks.

However, I'm not sure how this is affected by preemption. I suspect more work may need to be done in that case, but I'm not entirely sure.

This also means that rmmod has to bump the machine into the stopped state twice... but since that shouldn't be a common operation, I don't think that's a problem.

Signed-off-by David Howells <[email protected]>

---
warthog>diffstat kallsyms-race-2611rc1.diff

 kallsyms.c |   16 ++++++++++++++--
 module.c   |   35 ++++++++++++++++++++++++++++-------
 2 files changed, 42 insertions(+), 9 deletions(-)

kernel/kallsyms.c +14 -2

diff -uNrp linux-2.6.11-rc1/kernel/kallsyms.c linux-2.6.11-rc1-kallsyms/kernel/kallsyms.c
@@ -139,13 +139,20 @@ unsigned long kallsyms_lookup_name(const
 	return module_kallsyms_lookup_name(name);
 }
 
-/* Lookup an address.  modname is set to NULL if it's in the kernel. */
+/*
+ * Lookup an address
+ * - modname is set to NULL if it's in the kernel
+ * - we guarantee that the returned name is valid until we reschedule even if
+ *   it resides in a module
+ * - we also guarantee that modname will be valid until rescheduled
+ */
 const char *kallsyms_lookup(unsigned long addr,
 			    unsigned long *symbolsize,
 			    unsigned long *offset,
 			    char **modname, char *namebuf)
 {
 	unsigned long i, low, high, mid;
+	const char *msym;
 
 	/* This kernel should never had been booted. */
 	BUG_ON(!kallsyms_addresses);

@@ -196,7 +203,12 @@ const char *kallsyms_lookup(unsigned lon
 		return namebuf;
 	}
 
-	return module_address_lookup(addr, symbolsize, offset, modname);
+	/* see if it's in a module */
+	msym = module_address_lookup(addr, symbolsize, offset, modname);
+	if (msym)
+		return strncpy(namebuf, msym, KSYM_NAME_LEN);
+
+	return NULL;
 }
 
 /* Replace "%s" in format with address, or returns -errno. */

kernel/module.c +28 -7

diff -uNrp linux-2.6.11-rc1/kernel/module.c linux-2.6.11-rc1-kallsyms/kernel/module.c
@@ -1072,14 +1072,24 @@ static void mod_kobject_remove(struct mo
 	kobject_unregister(&mod->mkobj.kobj);
 }
 
+/*
+ * unlink the module with the whole machine is stopped with interrupts off
+ * - this defends against kallsyms not taking locks
+ */
+static inline int __unlink_module(void *_mod)
+{
+	struct module *mod = _mod;
+	spin_lock(&modlist_lock);
+	list_del(&mod->list);
+	spin_unlock(&modlist_lock);
+	return 0;
+}
+
 /* Free a module, remove from lists, etc (must hold module mutex). */
 static void free_module(struct module *mod)
 {
 	/* Delete from various lists */
-	spin_lock_irq(&modlist_lock);
-	list_del(&mod->list);
-	spin_unlock_irq(&modlist_lock);
-
+	stop_machine_run(__unlink_module, mod, NR_CPUS);
 	remove_sect_attrs(mod);
 	mod_kobject_remove(mod);

@@ -1732,6 +1742,19 @@ static struct module *load_module(void _
 	goto free_hdr;
 }
 
+/*
+ * link the module with the whole machine is stopped with interrupts off
+ * - this defends against kallsyms not taking locks
+ */
+static inline int __link_module(void *_mod)
+{
+	struct module *mod = _mod;
+	spin_lock(&modlist_lock);
+	list_add(&mod->list, &modules);
+	spin_unlock(&modlist_lock);
+	return 0;
+}
+
 /* This is where the real work happens */
 asmlinkage long
 sys_init_module(void __user *umod,

@@ -1766,9 +1789,7 @@ sys_init_module(void __user *umod,
 
 	/* Now sew it into the lists.  They won't access us, since
            strong_try_module_get() will fail. */
-	spin_lock_irq(&modlist_lock);
-	list_add(&mod->list, &modules);
-	spin_unlock_irq(&modlist_lock);
+	stop_machine_run(__link_module, mod, NR_CPUS);
 
 	/* Drop lock so they can recurse */
 	up(&module_mutex);

Rusty Russell

2005-01-18 02:25 UTC #1 lore raw

Quoted 1 line · David Howells

The attached patch fixes a race between kallsyms and insmod/rmmod.

Hi David,

The more I looked at this, the more I warmed to it. I've known for a
while that people are using kallsyms not for OOPS (eg. /proc/$$/wchan), so we should provide a "grabs locks" version, but this solution gets around that nicely, while making life more certain for the oops case, too.

Good work!
Rusty.

Signature · 1 line

A bad analogy is like a leaky screwdriver -- Richard Braakman

David Howells

author

2005-01-18 19:50 UTC #2 lore raw

Quoted 5 lines · Rusty Russell

The more I looked at this, the more I warmed to it. I've known for a
while that people are using kallsyms not for OOPS (eg. /proc/$$/wchan), so we should provide a "grabs locks" version, but this solution gets around that nicely, while making life more certain for the oops case, too.

Hmmm... though it works on i386 SMP, it doesn't, however, seem to work on ppc64 SMP:-/

My pSeries box seems to think that it can't find any symbols from previously loaded modules, and my Power5 box is quite happy to load modules that depend on other modules but panics because it can't mount its root fs.

This is very odd, because the patch is simple enough. Is there anything obvious I've missed that you can see? Or maybe I'm just misunderstanding how stop_machine_run() works... maybe it can't be called during initialisation.

David

David Howells

author

2005-01-27 14:03 UTC #3 lore raw

Quoted 8 lines · David Howells

Quoted 5 lines · Rusty Russell
The more I looked at this, the more I warmed to it. I've known for a
while that people are using kallsyms not for OOPS (eg. /proc/$$/wchan), so we should provide a "grabs locks" version, but this solution gets around that nicely, while making life more certain for the oops case, too.

Quoted 6 lines

Hmmm... though it works on i386 SMP, it doesn't, however, seem to work on ppc64 SMP:-/
My pSeries box seems to think that it can't find any symbols from previously loaded modules, and my Power5 box is quite happy to load modules that depend on other modules but panics because it can't mount its root fs.

Turns out that the patch works. Userspace was being bad. The stripped down shell running as init (pid #1) wasn't taking into account that it would get notification of kernel threads exiting when it called wait(), and so ended up trying to load several modules at once, some of which required dependency modules loading first.

David

David Howells

author

2005-01-27 14:09 UTC #4 lore raw

PATCH 1/1 Fix kallsyms/insmod/rmmod race [try #2]

In reply to David Howells

The attached patch fixes a race between kallsyms and insmod/rmmod.

The problem is this:

(1) The various kallsyms functions poke around in the module list without any locking so that they can be called from the oops handler.

(2) Although insmod and rmmod use locks to exclude each other, these have no effect on the kallsyms function.

Furthermore, list linkage information is embedded in the metadata too, so you can't trust that either...

(5) Therefore nothing that uses these functions can in any way trust any pointers to "static" data (such as module symbol names or module names) that are returned.

(6) On ppc64 the problems are exacerbated since the hypervisor may reschedule bits of the kernel, making operations that appear adjacent occur a long time apart.

However, I'm not sure how this is affected by preemption. I suspect more work may need to be done in that case, but I'm not entirely sure.

This also means that rmmod has to bump the machine into the stopped state twice... but since that shouldn't be a common operation, I don't think that's a problem.

I've amended this patch to not get spinlocks whilst in the machine locked state - there's no point as nothing else can be holding spinlocks.

Signed-off-by David Howells <[email protected]>

---
warthog>diffstat kallsyms-race-2611rc1.diff

 kallsyms.c |   16 ++++++++++++++--
 module.c   |   31 ++++++++++++++++++++++++-------
 2 files changed, 38 insertions(+), 9 deletions(-)

kernel/kallsyms.c +14 -2

diff -uNrp linux-2.6.11-rc1/kernel/kallsyms.c linux-2.6.11-rc1-kallsyms/kernel/kallsyms.c
@@ -139,13 +139,20 @@ unsigned long kallsyms_lookup_name(const
 	return module_kallsyms_lookup_name(name);
 }
 
-/* Lookup an address.  modname is set to NULL if it's in the kernel. */
+/*
+ * Lookup an address
+ * - modname is set to NULL if it's in the kernel
+ * - we guarantee that the returned name is valid until we reschedule even if
+ *   it resides in a module
+ * - we also guarantee that modname will be valid until rescheduled
+ */
 const char *kallsyms_lookup(unsigned long addr,
 			    unsigned long *symbolsize,
 			    unsigned long *offset,
 			    char **modname, char *namebuf)
 {
 	unsigned long i, low, high, mid;
+	const char *msym;
 
 	/* This kernel should never had been booted. */
 	BUG_ON(!kallsyms_addresses);

@@ -196,7 +203,12 @@ const char *kallsyms_lookup(unsigned lon
 		return namebuf;
 	}
 
-	return module_address_lookup(addr, symbolsize, offset, modname);
+	/* see if it's in a module */
+	msym = module_address_lookup(addr, symbolsize, offset, modname);
+	if (msym)
+		return strncpy(namebuf, msym, KSYM_NAME_LEN);
+
+	return NULL;
 }
 
 /* Replace "%s" in format with address, or returns -errno. */

kernel/module.c +24 -7

diff -uNrp linux-2.6.11-rc1/kernel/module.c linux-2.6.11-rc1-kallsyms/kernel/module.c
@@ -1072,14 +1072,22 @@ static void mod_kobject_remove(struct mo
 	kobject_unregister(&mod->mkobj.kobj);
 }
 
+/*
+ * unlink the module with the whole machine is stopped with interrupts off
+ * - this defends against kallsyms not taking locks
+ */
+static inline int __unlink_module(void *_mod)
+{
+	struct module *mod = _mod;
+	list_del(&mod->list);
+	return 0;
+}
+
 /* Free a module, remove from lists, etc (must hold module mutex). */
 static void free_module(struct module *mod)
 {
 	/* Delete from various lists */
-	spin_lock_irq(&modlist_lock);
-	list_del(&mod->list);
-	spin_unlock_irq(&modlist_lock);
-
+	stop_machine_run(__unlink_module, mod, NR_CPUS);
 	remove_sect_attrs(mod);
 	mod_kobject_remove(mod);

@@ -1732,6 +1740,17 @@ static struct module *load_module(void _
 	goto free_hdr;
 }
 
+/*
+ * link the module with the whole machine is stopped with interrupts off
+ * - this defends against kallsyms not taking locks
+ */
+static inline int __link_module(void *_mod)
+{
+	struct module *mod = _mod;
+	list_add(&mod->list, &modules);
+	return 0;
+}
+
 /* This is where the real work happens */
 asmlinkage long
 sys_init_module(void __user *umod,

@@ -1766,9 +1785,7 @@ sys_init_module(void __user *umod,
 
 	/* Now sew it into the lists.  They won't access us, since
            strong_try_module_get() will fail. */
-	spin_lock_irq(&modlist_lock);
-	list_add(&mod->list, &modules);
-	spin_unlock_irq(&modlist_lock);
+	stop_machine_run(__link_module, mod, NR_CPUS);
 
 	/* Drop lock so they can recurse */
 	up(&module_mutex);

Rusty Russell

2005-01-28 01:12 UTC #5 lore raw

Re: [PATCH] Fix kallsyms/insmod/rmmod race [try #2]

Quoted 1 line · David Howells

Signed-off-by David Howells <[email protected]>

Excellent. Thanks David!

Rusty.

Signature · 1 line

A bad analogy is like a leaky screwdriver -- Richard Braakman