BUG triggered in ptraceme
Hi Oleg! While been testing criu with linux-next we've triggered a BUG.
https://api.travis-ci.org/v3/job/430308998/log.txt
[ 2.461618] BUG: sleeping function called from invalid context at security/apparmor/include/cred.h:154
[ 2.461794] in_atomic(): 1, irqs_disabled(): 1, pid: 152, name: init
[ 2.461890] 1 lock held by init/152:
[ 2.461981] #0: 00000000f30c3fda (tasklist_lock){.+.+}, at: ptrace_traceme+0x1c/0x70
[ 2.462114] irq event stamp: 2524
[ 2.462242] hardirqs last enabled at (2523): [<ffffffff98002922>] do_syscall_64+0x12/0x190
[ 2.462363] hardirqs last disabled at (2524): [<ffffffff98b8b02f>] _raw_write_lock_irq+0xf/0x40
[ 2.462476] softirqs last enabled at (1904): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0
[ 2.462586] softirqs last disabled at (1902): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0
[ 2.462697] CPU: 1 PID: 152 Comm: init Not tainted 4.19.0-rc4-next-20180918+ #1Which is due to commit
commit 4b105cbbaf7c06e01c27391957dc3c446328d087
Author: Oleg Nesterov <[email protected]> Date: Wed Jun 17 16:27:33 2009 -0700
ptrace: do not use task_lock() for attach
because now after write_lock_irq(&tasklist_lock); apparmor calls for traceme and
static inline struct aa_label *begin_current_label_crit_section(void)
{
struct aa_label *label = aa_current_raw_label();--> might_sleep();
Take a look please, once time permit.
Cyrill
Quoted 25 lines · Cyrill Gorcunov
Hi Oleg! While been testing criu with linux-next we've triggered a BUG.
https://api.travis-ci.org/v3/job/430308998/log.txt[ 2.461618] BUG: sleeping function called from invalid context at security/apparmor/include/cred.h:154 [ 2.461794] in_atomic(): 1, irqs_disabled(): 1, pid: 152, name: init [ 2.461890] 1 lock held by init/152: [ 2.461981] #0: 00000000f30c3fda (tasklist_lock){.+.+}, at: ptrace_traceme+0x1c/0x70 [ 2.462114] irq event stamp: 2524 [ 2.462242] hardirqs last enabled at (2523): [<ffffffff98002922>] do_syscall_64+0x12/0x190 [ 2.462363] hardirqs last disabled at (2524): [<ffffffff98b8b02f>] _raw_write_lock_irq+0xf/0x40 [ 2.462476] softirqs last enabled at (1904): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0 [ 2.462586] softirqs last disabled at (1902): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0 [ 2.462697] CPU: 1 PID: 152 Comm: init Not tainted 4.19.0-rc4-next-20180918+ #1Which is due to commit
commit
4b105cbbaf7c06e01c27391957dc3c446328d087Author: Oleg Nesterov <[email protected]> Date: Wed Jun 17 16:27:33 2009 -0700ptrace: do not use
task_lock()for attachbecause now after
write_lock_irq(&tasklist_lock); apparmor calls for traceme and
Quoted 7 lines
static inline struct
aa_label*begin_current_label_crit_section(void){ struct aa_label *label = aa_current_raw_label();-->
might_sleep();Take a look please, once time permit.
Heh, actually not :) It is due to commit
commit 1f8266ff58840d698a1e96d2274189de1bdf7969
Author: Jann Horn <[email protected]> Date: Thu Sep 13 18:12:09 2018 +0200
which introduced might_sleep. Seems it is bad idea to send bug report without having a cup of coffee at the morning :)
Adding FS people to figure out whether GFP_KERNEL allocations with i_rwsem's held for writing are okay.
Quoted 34 lines · Cyrill Gorcunov
Quoted 25 lines · Cyrill Gorcunov
Hi Oleg! While been testing criu with linux-next we've triggered a BUG.
https://api.travis-ci.org/v3/job/430308998/log.txt[ 2.461618] BUG: sleeping function called from invalid context at security/apparmor/include/cred.h:154 [ 2.461794] in_atomic(): 1, irqs_disabled(): 1, pid: 152, name: init [ 2.461890] 1 lock held by init/152: [ 2.461981] #0: 00000000f30c3fda (tasklist_lock){.+.+}, at: ptrace_traceme+0x1c/0x70 [ 2.462114] irq event stamp: 2524 [ 2.462242] hardirqs last enabled at (2523): [<ffffffff98002922>] do_syscall_64+0x12/0x190 [ 2.462363] hardirqs last disabled at (2524): [<ffffffff98b8b02f>] _raw_write_lock_irq+0xf/0x40 [ 2.462476] softirqs last enabled at (1904): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0 [ 2.462586] softirqs last disabled at (1902): [<ffffffff98ac79ef>] unix_sock_destructor+0x4f/0xc0 [ 2.462697] CPU: 1 PID: 152 Comm: init Not tainted 4.19.0-rc4-next-20180918+ #1Which is due to commit
commit
4b105cbbaf7c06e01c27391957dc3c446328d087Author: Oleg Nesterov <[email protected]> Date: Wed Jun 17 16:27:33 2009 -0700ptrace: do not use
task_lock()for attachbecause now after
write_lock_irq(&tasklist_lock); apparmor calls for traceme andQuoted 7 lines
static inline struct
aa_label*begin_current_label_crit_section(void){ struct aa_label *label = aa_current_raw_label();-->
might_sleep();Take a look please, once time permit.
Quoted 8 lines
Heh, actually not :) It is due to commit
commit
1f8266ff58840d698a1e96d2274189de1bdf7969Author: Jann Horn <[email protected]> Date: Thu Sep 13 18:12:09 2018 +0200which introduced
might_sleep. Seems it is bad idea to send bug report without having a cup of coffee at the morning :)
Yeah, I fixed one sleep-in-atomic bug and figured I'd throw a might_sleep() in there for good measure... sigh.
I guess now I have to go through all the callers of begin_current_label_crit_section() to see what else looks wrong...
apparmor_ptrace_traceme() is wrong, as reported...
apparmor_path_link() looks icky, but I'm not sure - from what I can tell, it's called with an i_rwsem held for writing, and that probably makes calling back into filesystem context from there a bad idea? OTOH, it's just the i_rwsem of a newly-created path, so I don't know whether that's actually an issue...
security_path_rename() is called with two i_rwsem's held, but again, I'm not sure whether that's a problem.
On Wed, Sep 19, 2018 at 04:16:50PM +0200, Jann Horn wrote:
...
Quoted 7 lines
Quoted 7 lines
Heh, actually not :) It is due to commit
commit
1f8266ff58840d698a1e96d2274189de1bdf7969Author: Jann Horn <[email protected]> Date: Thu Sep 13 18:12:09 2018 +0200
Quoted 8 lines
Quoted 2 lines
which introduced
might_sleep. Seems it is bad idea to send bug report without having a cup of coffee at the morning :)Yeah, I fixed one sleep-in-atomic bug and figured I'd throw a
might_sleep()in there for good measure... sigh.
I guess now I have to go through all the callers ofbegin_current_label_crit_section()to see what else looks wrong...
Yes, I fear so. Need to check every caller just to be sure.
Quoted 10 lines
apparmor_ptrace_traceme()is wrong, as reported...
apparmor_path_link()looks icky, but I'm not sure - from what I can tell, it's called with ani_rwsemheld for writing, and that probably makes calling back into filesystem context from there a bad idea? OTOH, it's just thei_rwsemof a newly-created path, so I don't know whether that's actually an issue...
security_path_rename()is called with two i_rwsem's held, but again, I'm not sure whether that's a problem.
Lets wait for fs people opinions.
Quoted 29 lines · Cyrill Gorcunov
On Wed, Sep 19, 2018 at 04:16:50PM +0200, Jann Horn wrote:
...Quoted 7 lines
Quoted 7 lines
Heh, actually not :) It is due to commit
commit
1f8266ff58840d698a1e96d2274189de1bdf7969Author: Jann Horn <[email protected]> Date: Thu Sep 13 18:12:09 2018 +0200Quoted 8 lines
Quoted 2 lines
which introduced
might_sleep. Seems it is bad idea to send bug report without having a cup of coffee at the morning :)Yeah, I fixed one sleep-in-atomic bug and figured I'd throw a
might_sleep()in there for good measure... sigh.
I guess now I have to go through all the callers ofbegin_current_label_crit_section()to see what else looks wrong...Yes, I fear so. Need to check every caller just to be sure.
Quoted 8 lines
apparmor_ptrace_traceme()is wrong, as reported...
apparmor_path_link()looks icky, but I'm not sure - from what I can tell, it's called with ani_rwsemheld for writing, and that probably makes calling back into filesystem context from there a bad idea? OTOH, it's just thei_rwsemof a newly-created path, so I don't know whether that's actually an issue...
Quoted 4 lines
Quoted 2 lines
security_path_rename()is called with two i_rwsem's held, but again, I'm not sure whether that's a problem.Lets wait for fs people opinions.
No reply yet - I guess I should just fix up the traceme case for now.
On Thu, Sep 27, 2018 at 05:35:10PM +0200, Jann Horn wrote:
...
Quoted 4 lines
Quoted 2 lines
Lets wait for fs people opinions.
No reply yet - I guess I should just fix up the traceme case for now.
Yes please.