* [PATCH] Race between cat /proc/kallsyms and rmmod
@ 2016-01-30 10:23 Weilong Chen
2016-01-30 10:45 ` kbuild test robot
0 siblings, 1 reply; 5+ messages in thread
From: Weilong Chen @ 2016-01-30 10:23 UTC (permalink / raw)
To: linux-kernel, jcm, rusty
Iterating code of /proc/kallsyms calls module_get_kallsym() which grabs
and drops module_mutex internally and returns "struct module *",
module is removed, aforementioned "struct module *" is used in non-trivial
way.
So, grab module_mutex for entire operation like /proc/modules does.
Steps to reproduce:
while true; do modprobe xfs; rmmod xfs; done
vs
while true; do cat /proc/kallsyms >/dev/null; done
ref:http://lkml.iu.edu/hypermail/linux/kernel/0703.1/2582.html
https://bugzilla.kernel.org/show_bug.cgi?id=111541
Signed-off-by: Weilong Chen <chenweilong@huawei.com>
---
kernel/kallsyms.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c
index 3127ad5..aaecf19 100644
--- a/kernel/kallsyms.c
+++ b/kernel/kallsyms.c
@@ -515,6 +515,7 @@ static void *s_next(struct seq_file *m, void *p, loff_t *pos)
static void *s_start(struct seq_file *m, loff_t *pos)
{
+ mutex_lock(&module_mutex);
if (!update_iter(m->private, *pos))
return NULL;
return m->private;
@@ -522,6 +523,7 @@ static void *s_start(struct seq_file *m, loff_t *pos)
static void s_stop(struct seq_file *m, void *p)
{
+ mutex_unlock(&module_mutex);
}
static int s_show(struct seq_file *m, void *p)
--
1.8.3.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] Race between cat /proc/kallsyms and rmmod
2016-01-30 10:23 [PATCH] Race between cat /proc/kallsyms and rmmod Weilong Chen
@ 2016-01-30 10:45 ` kbuild test robot
0 siblings, 0 replies; 5+ messages in thread
From: kbuild test robot @ 2016-01-30 10:45 UTC (permalink / raw)
To: Weilong Chen; +Cc: kbuild-all, linux-kernel, jcm, rusty
[-- Attachment #1: Type: text/plain, Size: 2227 bytes --]
Hi Weilong,
[auto build test ERROR on v4.5-rc1]
[also build test ERROR on next-20160129]
[if your patch is applied to the wrong git tree, please drop us a note to help improving the system]
url: https://github.com/0day-ci/linux/commits/Weilong-Chen/Race-between-cat-proc-kallsyms-and-rmmod/20160130-182606
config: i386-randconfig-a0-201604 (attached as .config)
reproduce:
# save the attached .config to linux build tree
make ARCH=i386
All errors (new ones prefixed by >>):
In file included from include/linux/notifier.h:13:0,
from include/linux/memory_hotplug.h:6,
from include/linux/mmzone.h:728,
from include/linux/gfp.h:5,
from include/linux/kmod.h:22,
from include/linux/module.h:13,
from kernel/kallsyms.c:15:
kernel/kallsyms.c: In function 's_start':
>> kernel/kallsyms.c:519:14: error: 'module_mutex' undeclared (first use in this function)
mutex_lock(&module_mutex);
^
include/linux/mutex.h:146:44: note: in definition of macro 'mutex_lock'
#define mutex_lock(lock) mutex_lock_nested(lock, 0)
^
kernel/kallsyms.c:519:14: note: each undeclared identifier is reported only once for each function it appears in
mutex_lock(&module_mutex);
^
include/linux/mutex.h:146:44: note: in definition of macro 'mutex_lock'
#define mutex_lock(lock) mutex_lock_nested(lock, 0)
^
kernel/kallsyms.c: In function 's_stop':
kernel/kallsyms.c:527:16: error: 'module_mutex' undeclared (first use in this function)
mutex_unlock(&module_mutex);
^
vim +/module_mutex +519 kernel/kallsyms.c
513 return NULL;
514 return p;
515 }
516
517 static void *s_start(struct seq_file *m, loff_t *pos)
518 {
> 519 mutex_lock(&module_mutex);
520 if (!update_iter(m->private, *pos))
521 return NULL;
522 return m->private;
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/octet-stream, Size: 25004 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] Race between cat /proc/kallsyms and rmmod
2007-03-14 11:13 Alexey Dobriyan
2007-03-14 12:11 ` Andi Kleen
@ 2007-03-14 12:55 ` Paulo Marques
1 sibling, 0 replies; 5+ messages in thread
From: Paulo Marques @ 2007-03-14 12:55 UTC (permalink / raw)
To: Alexey Dobriyan; +Cc: akpm, linux-kernel, devel
Alexey Dobriyan wrote:
> Iterating code of /proc/kallsyms calls module_get_kallsym() which grabs
> and drops module_mutex internally and returns "struct module *",
> module is removed, aforementioned "struct module *" is used in non-trivial
> way.
> So, grab module_mutex for entire operation like /proc/modules does.
I would still prefer the other solution to avoid exposing "module_mutex"
outside of module.c like this :(
I'll try to send in a patch today for review.
--
Paulo Marques - www.grupopie.com
"As far as we know, our computer has never had an undetected error."
Weisert
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] Race between cat /proc/kallsyms and rmmod
2007-03-14 11:13 Alexey Dobriyan
@ 2007-03-14 12:11 ` Andi Kleen
2007-03-14 12:55 ` Paulo Marques
1 sibling, 0 replies; 5+ messages in thread
From: Andi Kleen @ 2007-03-14 12:11 UTC (permalink / raw)
To: Alexey Dobriyan; +Cc: akpm, linux-kernel, devel, pmarques
Alexey Dobriyan <adobriyan@sw.ru> writes:
>
> --- a/kernel/kallsyms.c
> +++ b/kernel/kallsyms.c
> @@ -40,6 +40,8 @@ extern const u16 kallsyms_token_index[]
>
> extern const unsigned long kallsyms_markers[] __attribute__((weak));
>
> +extern struct mutex module_mutex;
No externs in .c files
-Andi
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH] Race between cat /proc/kallsyms and rmmod
@ 2007-03-14 11:13 Alexey Dobriyan
2007-03-14 12:11 ` Andi Kleen
2007-03-14 12:55 ` Paulo Marques
0 siblings, 2 replies; 5+ messages in thread
From: Alexey Dobriyan @ 2007-03-14 11:13 UTC (permalink / raw)
To: akpm; +Cc: linux-kernel, devel, pmarques
Iterating code of /proc/kallsyms calls module_get_kallsym() which grabs
and drops module_mutex internally and returns "struct module *",
module is removed, aforementioned "struct module *" is used in non-trivial
way.
So, grab module_mutex for entire operation like /proc/modules does.
Steps to reproduce:
while true; do modprobe xfs; rmmod xfs; done
vs
while true; do cat /proc/kallsyms >/dev/null; done
[where xfs could be any module, I haven't tried]
BUG: unable to handle kernel paging request at virtual address e19f808c
printing eip:
c01dc361
*pde = 1ff5f067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
Modules linked in:
CPU: 0
EIP: 0060:[<c01dc361>] Not tainted VLI
EFLAGS: 00010297 (2.6.21-rc3-8b9909ded6922c33c221b105b26917780cfa497d #2)
EIP is at vsnprintf+0x2af/0x48c
eax: e19f808c ebx: ffffffff ecx: e19f808c edx: fffffffe
esi: dbe7aa84 edi: dbe2bf3c ebp: ffffffff esp: dbe2bec4
ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068
Process cat (pid: 7242, ti=dbe2b000 task=df5790b0 task.ti=dbe2b000)
Stack: e19d6fde 00000000 00000010 00000008 ffffffff 00000001 00000598 dbe7aa68
0002f362 00000010 dbe7b000 00000000 ffffffff c034bbe0 dbe7aa68 dfd31880
dfa31e80 00001000 c01586b0 dbe2bf2c dbe2bf2c dfd31880 dfd31880 c01289f6
Call Trace:
[<c01586b0>] seq_printf+0x2e/0x4b
[<c01289f6>] s_show+0x4b/0x7f
[<c0158c6e>] seq_read+0x196/0x278
[<c0158ad8>] seq_read+0x0/0x278
[<c0143c35>] vfs_read+0x72/0x93
[<c0143f1c>] sys_read+0x41/0x67
[<c0102486>] sysenter_past_esp+0x5f/0x85
=======================
Code: 74 24 28 73 03 c6 06 20 46 4d 85 ed 7f f1 e9 b9 00 00 00 8b 0f 81 f9 ff 0f 00 00 b8 ea 45 36 c0 0f 46 c8 8b 54 24 30 89 c8 eb 06 <80> 38 00 74 07 40 4a 83 fa ff 75 f4 29 c8 89 c3 89 e8 f6 44 24
EIP: [<c01dc361>] vsnprintf+0x2af/0x48c SS:ESP 0068:dbe2bec4
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---
kernel/kallsyms.c | 4 ++++
kernel/module.c | 5 +----
2 files changed, 5 insertions(+), 4 deletions(-)
--- a/kernel/kallsyms.c
+++ b/kernel/kallsyms.c
@@ -40,6 +40,8 @@ extern const u16 kallsyms_token_index[]
extern const unsigned long kallsyms_markers[] __attribute__((weak));
+extern struct mutex module_mutex;
+
static inline int is_kernel_inittext(unsigned long addr)
{
if (addr >= (unsigned long)_sinittext
@@ -369,6 +371,7 @@ static void *s_next(struct seq_file *m,
static void *s_start(struct seq_file *m, loff_t *pos)
{
+ mutex_lock(&module_mutex);
if (!update_iter(m->private, *pos))
return NULL;
return m->private;
@@ -376,6 +379,7 @@ static void *s_start(struct seq_file *m,
static void s_stop(struct seq_file *m, void *p)
{
+ mutex_unlock(&module_mutex);
}
static int s_show(struct seq_file *m, void *p)
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -62,7 +62,7 @@ #define INIT_OFFSET_MASK (1UL << (BITS_P
static DEFINE_SPINLOCK(modlist_lock);
/* List of modules, protected by module_mutex AND modlist_lock */
-static DEFINE_MUTEX(module_mutex);
+DEFINE_MUTEX(module_mutex);
static LIST_HEAD(modules);
static BLOCKING_NOTIFIER_HEAD(module_notify_list);
@@ -2124,19 +2124,16 @@ struct module *module_get_kallsym(unsign
{
struct module *mod;
- mutex_lock(&module_mutex);
list_for_each_entry(mod, &modules, list) {
if (symnum < mod->num_symtab) {
*value = mod->symtab[symnum].st_value;
*type = mod->symtab[symnum].st_info;
strlcpy(name, mod->strtab + mod->symtab[symnum].st_name,
namelen);
- mutex_unlock(&module_mutex);
return mod;
}
symnum -= mod->num_symtab;
}
- mutex_unlock(&module_mutex);
return NULL;
}
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2016-01-30 10:46 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-01-30 10:23 [PATCH] Race between cat /proc/kallsyms and rmmod Weilong Chen
2016-01-30 10:45 ` kbuild test robot
-- strict thread matches above, loose matches on Subject: below --
2007-03-14 11:13 Alexey Dobriyan
2007-03-14 12:11 ` Andi Kleen
2007-03-14 12:55 ` Paulo Marques
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).