another use-after-free in ep_remove_wait_queue()

another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

Hi,

We found the following use-after-free with userspace code polling on
a pressure file in a non-root cgroup using epoll.

[   57.183661] BUG: KASAN: use-after-free in _raw_spin_lock_irqsave+0x76/0x130
[   57.186662] Write of size 4 at addr ffff888114976428 by task a.out/2426

[   57.190551] CPU: 0 PID: 2426 Comm: a.out Not tainted 6.2.0-rc2+ #16
[   57.193384] Hardware name: Amazon EC2 c5a.large/, BIOS 1.0 10/16/2017
[   57.196272] Call Trace:
[   57.197565]  <TASK>
[   57.198714]  dump_stack_lvl+0x8f/0xc0
[   57.200494]  print_report+0x16c/0x4e0
[   57.202084]  ? _raw_spin_lock_irqsave+0x76/0x130
[   57.204077]  kasan_report+0xc3/0xf0
[   57.205587]  ? entry_SYSCALL_64_after_hwframe+0x63/0xcd
[   57.207760]  ? _raw_spin_lock_irqsave+0x76/0x130
[   57.209685]  kasan_check_range+0x2d2/0x310
[   57.211477]  _raw_spin_lock_irqsave+0x76/0x130
[   57.213355]  remove_wait_queue+0x25/0x130
[   57.215102]  ep_free+0x12d/0x220
[   57.216506]  ep_eventpoll_release+0x3c/0x40
[   57.218254]  __fput+0x32b/0x700
[   57.221486]  task_work_run+0x1db/0x230
[   57.224885]  exit_to_user_mode_prepare+0xfd/0x100
[   57.228662]  syscall_exit_to_user_mode+0x20/0x40
[   57.232360]  do_syscall_64+0x52/0x90
[   57.235691]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
[   57.239572] RIP: 0033:0x7fadf96e1c44
[   57.242865] Code: 00 00 b8 ff ff ff ff eb 9c b8 ff ff ff ff eb 95 e8 01 e2 01 00 90 8b 05 2a ac 2c 00 48 63 ff 85 c0 75 11 b8 03 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 3a f3 c3 48 83 ec 18 48 89 7c 24 08 e8 e4 a0
[   57.255244] RSP: 002b:00007ffd1d1b7b98 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
[   57.261714] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007fadf96e1c44
[   57.266293] RDX: 0000000000000000 RSI: 00007ffd1d1b7b60 RDI: 0000000000000004
[   57.270979] RBP: 00007ffd1d1b7bf0 R08: 00000000004007e0 R09: 00007fadf9a0f240
[   57.275856] R10: 00000000000006ba R11: 0000000000000246 R12: 00000000004005e0
[   57.280478] R13: 00007ffd1d1b7cd0 R14: 0000000000000000 R15: 0000000000000000
[   57.285059]  </TASK>

[   57.290402] Allocated by task 2426:
[   57.293705]  kasan_set_track+0x3d/0x60
[   57.297102]  __kasan_kmalloc+0x85/0x90
[   57.300491]  psi_trigger_create+0x155/0x850
[   57.304040]  pressure_write+0x200/0x510
[   57.307508]  cgroup_file_write+0x1de/0x3e0
[   57.310949]  kernfs_fop_write_iter+0x27d/0x380
[   57.314601]  vfs_write+0x7d7/0xaa0
[   57.317891]  ksys_write+0xd7/0x1a0
[   57.321152]  do_syscall_64+0x43/0x90
[   57.324496]  entry_SYSCALL_64_after_hwframe+0x63/0xcd

[   57.330887] Freed by task 2429:
[   57.334053]  kasan_set_track+0x3d/0x60
[   57.337448]  kasan_save_free_info+0x27/0x40
[   57.340977]  ____kasan_slab_free+0x11d/0x170
[   57.344570]  slab_free_freelist_hook+0x87/0x150
[   57.348236]  __kmem_cache_free+0xcb/0x180
[   57.351710]  psi_trigger_destroy+0x458/0x550
[   57.355274]  cgroup_file_release+0x96/0x110
[   57.358779]  kernfs_drain_open_files+0x238/0x420
[   57.362519]  kernfs_drain+0x191/0x2a0
[   57.365901]  __kernfs_remove+0x3a6/0x600
[   57.369363]  kernfs_remove_by_name_ns+0xc2/0x120
[   57.373073]  cgroup_addrm_files+0x90f/0xcf0
[   57.376610]  cgroup_destroy_locked+0x48a/0x730
[   57.380260]  cgroup_rmdir+0x2b/0x130
[   57.383650]  kernfs_iop_rmdir+0x17a/0x230
[   57.387201]  vfs_rmdir+0x196/0x410
[   57.390442]  do_rmdir+0x1c7/0x3f0
[   57.393651]  __x64_sys_rmdir+0x45/0x50
[   57.397000]  do_syscall_64+0x43/0x90
[   57.400340]  entry_SYSCALL_64_after_hwframe+0x63/0xcd

[   57.406689] The buggy address belongs to the object at ffff888114976400
                which belongs to the cache kmalloc-128 of size 128
[   57.414907] The buggy address is located 40 bytes inside of
                128-byte region [ffff888114976400, ffff888114976480)

[   57.425474] The buggy address belongs to the physical page:
[   57.429541] page:000000008c5ecb31 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x114976
[   57.436725] flags: 0x2fffff80000200(slab|node=0|zone=2|lastcpupid=0x1fffff)
[   57.441235] raw: 002fffff80000200 ffff8881000418c0 dead000000000100 dead000000000122
[   57.447793] raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
[   57.454274] page dumped because: kasan: bad access detected

[   57.460990] Memory state around the buggy address:
[   57.464744]  ffff888114976300: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   57.471126]  ffff888114976380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   57.477447] >ffff888114976400: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   57.483833]                                   ^
[   57.487541]  ffff888114976480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   57.493976]  ffff888114976500: 00 00 00 00 00 00 00 00 00 00 00 00 00 01 fc fc

Here is the simple repro.

#include <fcntl.h>
#include <sys/epoll.h>
#include <unistd.h>

int main(void)
{
	const char trigger_str[] = "some 100000 1000000";
	int fd, epfd;
	struct epoll_event event;
	struct epoll_event events[1];

	fd = open("/cgroup2/test/cpu.pressure", O_RDWR);
	write(fd, trigger_str, sizeof(trigger_str));
	epfd = epoll_create(1);
	event.events = EPOLLPRI | EPOLLET;
	event.data.fd = fd;
	epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &event);
	epoll_wait(epfd, events, 1, -1); /* returns after rmdir */
	sleep(5);
	close(epfd);
	close(fd);

	return 0;
}

# mkdir /cgroup2
# mount -t cgroup2 none /cgroup2
# mkdir /cgroup2/test
# ./a.out &
# rmdir /cgroup2/test

Looks like calling wake_up_pollfree() in psi_trigger_destroy() can properly
clear the queue and then avoid this use-after-free, but POLLFREE wasn't
considered enough there for the past similar issue[1]. While
wake_up_pollfree() could *also* be called in psi_trigger_destroy(), it may
be awkward and there can be more appropriate solution. It would be great if
experts could have a look.

[1] https://lkml.org/lkml/2021/11/16/264


Regards,
Munehisa

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Fri, Jan 6, 2023 at 2:49 PM Munehisa Kamata <kamatam@amazon.com> wrote:
>
> Hi,
>
> We found the following use-after-free with userspace code polling on
> a pressure file in a non-root cgroup using epoll.
>
> [   57.183661] BUG: KASAN: use-after-free in _raw_spin_lock_irqsave+0x76/0x130
> [   57.186662] Write of size 4 at addr ffff888114976428 by task a.out/2426
>
> [   57.190551] CPU: 0 PID: 2426 Comm: a.out Not tainted 6.2.0-rc2+ #16
> [   57.193384] Hardware name: Amazon EC2 c5a.large/, BIOS 1.0 10/16/2017
> [   57.196272] Call Trace:
> [   57.197565]  <TASK>
> [   57.198714]  dump_stack_lvl+0x8f/0xc0
> [   57.200494]  print_report+0x16c/0x4e0
> [   57.202084]  ? _raw_spin_lock_irqsave+0x76/0x130
> [   57.204077]  kasan_report+0xc3/0xf0
> [   57.205587]  ? entry_SYSCALL_64_after_hwframe+0x63/0xcd
> [   57.207760]  ? _raw_spin_lock_irqsave+0x76/0x130
> [   57.209685]  kasan_check_range+0x2d2/0x310
> [   57.211477]  _raw_spin_lock_irqsave+0x76/0x130
> [   57.213355]  remove_wait_queue+0x25/0x130
> [   57.215102]  ep_free+0x12d/0x220
> [   57.216506]  ep_eventpoll_release+0x3c/0x40
> [   57.218254]  __fput+0x32b/0x700
> [   57.221486]  task_work_run+0x1db/0x230
> [   57.224885]  exit_to_user_mode_prepare+0xfd/0x100
> [   57.228662]  syscall_exit_to_user_mode+0x20/0x40
> [   57.232360]  do_syscall_64+0x52/0x90
> [   57.235691]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> [   57.239572] RIP: 0033:0x7fadf96e1c44
> [   57.242865] Code: 00 00 b8 ff ff ff ff eb 9c b8 ff ff ff ff eb 95 e8 01 e2 01 00 90 8b 05 2a ac 2c 00 48 63 ff 85 c0 75 11 b8 03 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 3a f3 c3 48 83 ec 18 48 89 7c 24 08 e8 e4 a0
> [   57.255244] RSP: 002b:00007ffd1d1b7b98 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
> [   57.261714] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007fadf96e1c44
> [   57.266293] RDX: 0000000000000000 RSI: 00007ffd1d1b7b60 RDI: 0000000000000004
> [   57.270979] RBP: 00007ffd1d1b7bf0 R08: 00000000004007e0 R09: 00007fadf9a0f240
> [   57.275856] R10: 00000000000006ba R11: 0000000000000246 R12: 00000000004005e0
> [   57.280478] R13: 00007ffd1d1b7cd0 R14: 0000000000000000 R15: 0000000000000000
> [   57.285059]  </TASK>
>
> [   57.290402] Allocated by task 2426:
> [   57.293705]  kasan_set_track+0x3d/0x60
> [   57.297102]  __kasan_kmalloc+0x85/0x90
> [   57.300491]  psi_trigger_create+0x155/0x850
> [   57.304040]  pressure_write+0x200/0x510
> [   57.307508]  cgroup_file_write+0x1de/0x3e0
> [   57.310949]  kernfs_fop_write_iter+0x27d/0x380
> [   57.314601]  vfs_write+0x7d7/0xaa0
> [   57.317891]  ksys_write+0xd7/0x1a0
> [   57.321152]  do_syscall_64+0x43/0x90
> [   57.324496]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
>
> [   57.330887] Freed by task 2429:
> [   57.334053]  kasan_set_track+0x3d/0x60
> [   57.337448]  kasan_save_free_info+0x27/0x40
> [   57.340977]  ____kasan_slab_free+0x11d/0x170
> [   57.344570]  slab_free_freelist_hook+0x87/0x150
> [   57.348236]  __kmem_cache_free+0xcb/0x180
> [   57.351710]  psi_trigger_destroy+0x458/0x550
> [   57.355274]  cgroup_file_release+0x96/0x110
> [   57.358779]  kernfs_drain_open_files+0x238/0x420
> [   57.362519]  kernfs_drain+0x191/0x2a0
> [   57.365901]  __kernfs_remove+0x3a6/0x600
> [   57.369363]  kernfs_remove_by_name_ns+0xc2/0x120
> [   57.373073]  cgroup_addrm_files+0x90f/0xcf0
> [   57.376610]  cgroup_destroy_locked+0x48a/0x730
> [   57.380260]  cgroup_rmdir+0x2b/0x130
> [   57.383650]  kernfs_iop_rmdir+0x17a/0x230
> [   57.387201]  vfs_rmdir+0x196/0x410
> [   57.390442]  do_rmdir+0x1c7/0x3f0
> [   57.393651]  __x64_sys_rmdir+0x45/0x50
> [   57.397000]  do_syscall_64+0x43/0x90
> [   57.400340]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
>
> [   57.406689] The buggy address belongs to the object at ffff888114976400
>                 which belongs to the cache kmalloc-128 of size 128
> [   57.414907] The buggy address is located 40 bytes inside of
>                 128-byte region [ffff888114976400, ffff888114976480)
>
> [   57.425474] The buggy address belongs to the physical page:
> [   57.429541] page:000000008c5ecb31 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x114976
> [   57.436725] flags: 0x2fffff80000200(slab|node=0|zone=2|lastcpupid=0x1fffff)
> [   57.441235] raw: 002fffff80000200 ffff8881000418c0 dead000000000100 dead000000000122
> [   57.447793] raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
> [   57.454274] page dumped because: kasan: bad access detected
>
> [   57.460990] Memory state around the buggy address:
> [   57.464744]  ffff888114976300: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> [   57.471126]  ffff888114976380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> [   57.477447] >ffff888114976400: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> [   57.483833]                                   ^
> [   57.487541]  ffff888114976480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> [   57.493976]  ffff888114976500: 00 00 00 00 00 00 00 00 00 00 00 00 00 01 fc fc
>
> Here is the simple repro.
>
> #include <fcntl.h>
> #include <sys/epoll.h>
> #include <unistd.h>
>
> int main(void)
> {
>         const char trigger_str[] = "some 100000 1000000";
>         int fd, epfd;
>         struct epoll_event event;
>         struct epoll_event events[1];
>
>         fd = open("/cgroup2/test/cpu.pressure", O_RDWR);
>         write(fd, trigger_str, sizeof(trigger_str));
>         epfd = epoll_create(1);
>         event.events = EPOLLPRI | EPOLLET;
>         event.data.fd = fd;
>         epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &event);
>         epoll_wait(epfd, events, 1, -1); /* returns after rmdir */
>         sleep(5);
>         close(epfd);
>         close(fd);
>
>         return 0;
> }
>
> # mkdir /cgroup2
> # mount -t cgroup2 none /cgroup2
> # mkdir /cgroup2/test
> # ./a.out &
> # rmdir /cgroup2/test
>
> Looks like calling wake_up_pollfree() in psi_trigger_destroy() can properly
> clear the queue and then avoid this use-after-free, but POLLFREE wasn't
> considered enough there for the past similar issue[1]. While
> wake_up_pollfree() could *also* be called in psi_trigger_destroy(), it may
> be awkward and there can be more appropriate solution. It would be great if
> experts could have a look.

Thanks for the report, Munehisa. I'll look into this over the weekend
or on Monday.
Suren.

>
> [1] https://lkml.org/lkml/2021/11/16/264
>
>
> Regards,
> Munehisa

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Fri, Jan 20, 2023 at 9:18 PM Hillf Danton <hdanton@sina.com> wrote:
>
> On Fri, 20 Jan 2023 08:28:25 -0800 Suren Baghdasaryan <surenb@google.com>
> > On Fri, Jan 20, 2023 at 1:00 AM Hillf Danton <hdanton@sina.com> wrote:
> > > +++ b/kernel/sched/psi.c
> > > @@ -1529,6 +1529,7 @@ static int psi_fop_release(struct inode
> > >  {
> > >         struct seq_file *seq = file->private_data;
> > >
> > > +       eventpoll_release_file(file);
> >
> > Be careful here and see the comment in
> > https://elixir.bootlin.com/linux/latest/source/fs/eventpoll.c#L912.
> > eventpoll_release_file() assumes that the last fput() was called and
> > nobody other than ep_free() will race with us. So, this will not be
> > that simple.
>
> The epmutex serializes eventpoll_release_file() and ep_free(). And this
> is in psi_fop_release(), so no chance is likely left for another release.
>
> > Besides if we really need to fix the order here, the fix
> > should be somewhere at the level of cgroup_file_release() or even
> > kernfs to work for other similar situations.
>
> Good point but cgroup and kernfs have no idea of psi trigger.

Yes, that's why I think if we really need to fix the order here and do
it properly, it won't be straightforward. IMHO wake_up_pollfree() is
an appropriate and simple fix for this.

>
> The bonus of the uaf is check polled file upon release in scenarios like
> the psi trigger.
>

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Fri, Jan 20, 2023 at 1:00 AM Hillf Danton <hdanton@sina.com> wrote:
>
> On Thu, 19 Jan 2023 17:37:11 -0800 Suren Baghdasaryan <surenb@google.com> wrote:
> > On Thu, Jan 19, 2023 at 5:31 PM Hillf Danton <hdanton@sina.com> wrote:
> > > On Thu, 19 Jan 2023 13:01:42 -0800 Suren Baghdasaryan <surenb@google.com> wrote:
> > > >
> > > > Hi Folks,
> > > > I spent some more time digging into the details and this is what's
> > > > happening. When we call rmdir to delete the cgroup with the pressure
> > > > file being epoll'ed, roughly the following call chain happens in the
> > > > context of the shell process:
> > > >
> > > > do_rmdir
> > > >   cgroup_rmdir
> > > >     kernfs_drain_open_files
> > > >       cgroup_file_release
> > > >         cgroup_pressure_release
> > > >           psi_trigger_destroy
> > > >
> > > > Later on in the context of our reproducer, the last fput() is called
> > > > causing wait queue removal:
> > > >
> > > > fput
> > > >   ep_eventpoll_release
> > > >     ep_free
> > > >       ep_remove_wait_queue
> > > >         remove_wait_queue
> > > >
> > > > By this time psi_trigger_destroy() already destroyed the trigger's
> > > > waitqueue head and we hit UAF.
> > > > I think the conceptual problem here (or maybe that's by design?) is
> > > > that cgroup_file_release() is not really tied to the file's real
> > > > lifetime (when the last fput() is issued). Otherwise fput() would call
> > > > eventpoll_release() before f_op->release() and the order would be fine
> > > > (we would remove the wait queue first in eventpoll_release() and then
> > > > f_op->release() would cause trigger's destruction).
> > >
> > >   eventpoll_release
> > >     eventpoll_release_file
> > >       ep_remove
> > >         ep_unregister_pollwait
> > >           ep_remove_wait_queue
> > >
> >
> > Yes but fput() calls eventpoll_release() *before* f_op->release(), so
> > waitqueue_head would be removed before trigger destruction.
>
> Then check if file is polled before destroying trigger.
>
> +++ b/kernel/sched/psi.c
> @@ -1529,6 +1529,7 @@ static int psi_fop_release(struct inode
>  {
>         struct seq_file *seq = file->private_data;
>
> +       eventpoll_release_file(file);

Be careful here and see the comment in
https://elixir.bootlin.com/linux/latest/source/fs/eventpoll.c#L912.
eventpoll_release_file() assumes that the last fput() was called and
nobody other than ep_free() will race with us. So, this will not be
that simple. Besides if we really need to fix the order here, the fix
should be somewhere at the level of cgroup_file_release() or even
kernfs to work for other similar situations.

>         psi_trigger_destroy(seq->private);
>         return single_release(inode, file);
>  }
>

Re: another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

On Fri, 2023-01-20 02:46:13 +0000, Munehisa Kamata <kamatam@amazon.com> wrote:
>
> On Fri, 2023-01-20 01:37:11 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > On Thu, Jan 19, 2023 at 5:31 PM Hillf Danton <hdanton@sina.com> wrote:
> > >
> > > On Thu, 19 Jan 2023 13:01:42 -0800 Suren Baghdasaryan <surenb@google.com> wrote:
> > > >
> > > > Hi Folks,
> > > > I spent some more time digging into the details and this is what's
> > > > happening. When we call rmdir to delete the cgroup with the pressure
> > > > file being epoll'ed, roughly the following call chain happens in the
> > > > context of the shell process:
> > > >
> > > > do_rmdir
> > > >   cgroup_rmdir
> > > >     kernfs_drain_open_files
> > > >       cgroup_file_release
> > > >         cgroup_pressure_release
> > > >           psi_trigger_destroy
> > > >
> > > > Later on in the context of our reproducer, the last fput() is called
> > > > causing wait queue removal:
> > > >
> > > > fput
> > > >   ep_eventpoll_release
> > > >     ep_free
> > > >       ep_remove_wait_queue
> > > >         remove_wait_queue
> > > >
> > > > By this time psi_trigger_destroy() already destroyed the trigger's
> > > > waitqueue head and we hit UAF.
> > > > I think the conceptual problem here (or maybe that's by design?) is
> > > > that cgroup_file_release() is not really tied to the file's real
> > > > lifetime (when the last fput() is issued). Otherwise fput() would call
> > > > eventpoll_release() before f_op->release() and the order would be fine
> > > > (we would remove the wait queue first in eventpoll_release() and then
> > > > f_op->release() would cause trigger's destruction).
> > >
> > >   eventpoll_release
> > >     eventpoll_release_file
> > >       ep_remove
> > >         ep_unregister_pollwait
> > >           ep_remove_wait_queue
> > >
> > 
> > Yes but fput() calls eventpoll_release() *before* f_op->release(), so
> > waitqueue_head would be removed before trigger destruction.
>  
> But pwq->whead is still pointing the freed head, then we just hit the same
> issue earlier?
 
Ah nevermind, that was just a hypothetical case if cgroup_file_release()
was tied to file's lifetime and assuming trigger destruction that frees    
the queue and clears pwq->whead would happen later in f_op->release();     
there is no such an implementation today.                                  
                                                                           
Sorry for noise.

> > > Different roads run into the same Roma city.
> > 
> > You butchered the phrase :)
> > 
> > >
> > > > Considering these findings, I think we can use the wake_up_pollfree()
> > > > without contradicting the comment at
> > > > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253
> > > > because indeed, cgroup_file_release() and therefore
> > > > psi_trigger_destroy() are not tied to the file's lifetime.
> > > >
> > > > I'm CC'ing Tejun to check if this makes sense to him and
> > > > cgroup_file_release() is working as expected in this case.
> > > >
> > > > Munehisha, if Tejun confirms this is all valid, could you please post
> > > > a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
> > > > don't need to worry about wake_up_all() because we have a limitation
> > > > of one trigger per file descriptor:
> > > > https://elixir.bootlin.com/linux/latest/source/kernel/sched/psi.c#L1419,
> > > > so there can be only one waiter.
> > > > Thanks,
> > > > Suren.
> > >
> > 
> > 
> 
> 

Re: another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

On Fri, 2023-01-20 01:37:11 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Thu, Jan 19, 2023 at 5:31 PM Hillf Danton <hdanton@sina.com> wrote:
> >
> > On Thu, 19 Jan 2023 13:01:42 -0800 Suren Baghdasaryan <surenb@google.com> wrote:
> > >
> > > Hi Folks,
> > > I spent some more time digging into the details and this is what's
> > > happening. When we call rmdir to delete the cgroup with the pressure
> > > file being epoll'ed, roughly the following call chain happens in the
> > > context of the shell process:
> > >
> > > do_rmdir
> > >   cgroup_rmdir
> > >     kernfs_drain_open_files
> > >       cgroup_file_release
> > >         cgroup_pressure_release
> > >           psi_trigger_destroy
> > >
> > > Later on in the context of our reproducer, the last fput() is called
> > > causing wait queue removal:
> > >
> > > fput
> > >   ep_eventpoll_release
> > >     ep_free
> > >       ep_remove_wait_queue
> > >         remove_wait_queue
> > >
> > > By this time psi_trigger_destroy() already destroyed the trigger's
> > > waitqueue head and we hit UAF.
> > > I think the conceptual problem here (or maybe that's by design?) is
> > > that cgroup_file_release() is not really tied to the file's real
> > > lifetime (when the last fput() is issued). Otherwise fput() would call
> > > eventpoll_release() before f_op->release() and the order would be fine
> > > (we would remove the wait queue first in eventpoll_release() and then
> > > f_op->release() would cause trigger's destruction).
> >
> >   eventpoll_release
> >     eventpoll_release_file
> >       ep_remove
> >         ep_unregister_pollwait
> >           ep_remove_wait_queue
> >
> 
> Yes but fput() calls eventpoll_release() *before* f_op->release(), so
> waitqueue_head would be removed before trigger destruction.
 
But pwq->whead is still pointing the freed head, then we just hit the same
issue earlier?

> > Different roads run into the same Roma city.
> 
> You butchered the phrase :)
> 
> >
> > > Considering these findings, I think we can use the wake_up_pollfree()
> > > without contradicting the comment at
> > > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253
> > > because indeed, cgroup_file_release() and therefore
> > > psi_trigger_destroy() are not tied to the file's lifetime.
> > >
> > > I'm CC'ing Tejun to check if this makes sense to him and
> > > cgroup_file_release() is working as expected in this case.
> > >
> > > Munehisha, if Tejun confirms this is all valid, could you please post
> > > a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
> > > don't need to worry about wake_up_all() because we have a limitation
> > > of one trigger per file descriptor:
> > > https://elixir.bootlin.com/linux/latest/source/kernel/sched/psi.c#L1419,
> > > so there can be only one waiter.
> > > Thanks,
> > > Suren.
> >
> 
> 

Re: another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

On Thu, 2023-01-19 21:01:42 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
>
> Munehisha, if Tejun confirms this is all valid, could you please post
> a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
> don't need to worry about wake_up_all() because we have a limitation
> of one trigger per file descriptor:
> https://elixir.bootlin.com/linux/latest/source/kernel/sched/psi.c#L1419,
> so there can be only one waiter.

Yes, will do (and will CC stable).


Regards,
Munehisa

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Thu, Jan 19, 2023 at 5:31 PM Hillf Danton <hdanton@sina.com> wrote:
>
> On Thu, 19 Jan 2023 13:01:42 -0800 Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > Hi Folks,
> > I spent some more time digging into the details and this is what's
> > happening. When we call rmdir to delete the cgroup with the pressure
> > file being epoll'ed, roughly the following call chain happens in the
> > context of the shell process:
> >
> > do_rmdir
> >   cgroup_rmdir
> >     kernfs_drain_open_files
> >       cgroup_file_release
> >         cgroup_pressure_release
> >           psi_trigger_destroy
> >
> > Later on in the context of our reproducer, the last fput() is called
> > causing wait queue removal:
> >
> > fput
> >   ep_eventpoll_release
> >     ep_free
> >       ep_remove_wait_queue
> >         remove_wait_queue
> >
> > By this time psi_trigger_destroy() already destroyed the trigger's
> > waitqueue head and we hit UAF.
> > I think the conceptual problem here (or maybe that's by design?) is
> > that cgroup_file_release() is not really tied to the file's real
> > lifetime (when the last fput() is issued). Otherwise fput() would call
> > eventpoll_release() before f_op->release() and the order would be fine
> > (we would remove the wait queue first in eventpoll_release() and then
> > f_op->release() would cause trigger's destruction).
>
>   eventpoll_release
>     eventpoll_release_file
>       ep_remove
>         ep_unregister_pollwait
>           ep_remove_wait_queue
>

Yes but fput() calls eventpoll_release() *before* f_op->release(), so
waitqueue_head would be removed before trigger destruction.

> Different roads run into the same Roma city.

You butchered the phrase :)

>
> > Considering these findings, I think we can use the wake_up_pollfree()
> > without contradicting the comment at
> > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253
> > because indeed, cgroup_file_release() and therefore
> > psi_trigger_destroy() are not tied to the file's lifetime.
> >
> > I'm CC'ing Tejun to check if this makes sense to him and
> > cgroup_file_release() is working as expected in this case.
> >
> > Munehisha, if Tejun confirms this is all valid, could you please post
> > a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
> > don't need to worry about wake_up_all() because we have a limitation
> > of one trigger per file descriptor:
> > https://elixir.bootlin.com/linux/latest/source/kernel/sched/psi.c#L1419,
> > so there can be only one waiter.
> > Thanks,
> > Suren.
>

Re: another use-after-free in ep_remove_wait_queue()

[Johannes Weiner]

On Thu, Jan 19, 2023 at 01:01:42PM -0800, Suren Baghdasaryan wrote:
> I spent some more time digging into the details and this is what's
> happening. When we call rmdir to delete the cgroup with the pressure
> file being epoll'ed, roughly the following call chain happens in the
> context of the shell process:
> 
> do_rmdir
>   cgroup_rmdir
>     kernfs_drain_open_files
>       cgroup_file_release
>         cgroup_pressure_release
>           psi_trigger_destroy
> 
> Later on in the context of our reproducer, the last fput() is called
> causing wait queue removal:
> 
> fput
>   ep_eventpoll_release
>     ep_free
>       ep_remove_wait_queue
>         remove_wait_queue
> 
> By this time psi_trigger_destroy() already destroyed the trigger's
> waitqueue head and we hit UAF.
> I think the conceptual problem here (or maybe that's by design?) is
> that cgroup_file_release() is not really tied to the file's real
> lifetime (when the last fput() is issued). Otherwise fput() would call
> eventpoll_release() before f_op->release() and the order would be fine
> (we would remove the wait queue first in eventpoll_release() and then
> f_op->release() would cause trigger's destruction).
> Considering these findings, I think we can use the wake_up_pollfree()
> without contradicting the comment at
> https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253
> because indeed, cgroup_file_release() and therefore
> psi_trigger_destroy() are not tied to the file's lifetime.
> 
> I'm CC'ing Tejun to check if this makes sense to him and
> cgroup_file_release() is working as expected in this case.
> 
> Munehisha, if Tejun confirms this is all valid, could you please post
> a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
> don't need to worry about wake_up_all() because we have a limitation
> of one trigger per file descriptor:

Solid analysis!

Indeed, wake_up_pollfree() should fix it.

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Wed, Jan 18, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Fri, Jan 13, 2023 at 9:52 AM Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > On Thu, Jan 12, 2023 at 6:26 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > >
> > > On Thu, 2023-01-12 22:01:24 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
> > > >
> > > > On Mon, Jan 9, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > > > >
> > > > > On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > > > > >
> > > > > > On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> > > > > > >
> > > > > > > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > > > > > > >
> > > > > > > > That patch survived the repro in my original post, however, the waker
> > > > > > > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > > > > > > the pressure file got closed. So, if the following modified repro runs
> > > > > > > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > > > > > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > > > > > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > > > > > > the queue, but wake_up_pollfree() does.
> > > > > > >
> > > > > > > Thanks for your testing. And the debugging completes.
> > > > > > >
> > > > > > > Mind sending a patch with wake_up_pollfree() folded?
> > > > > >
> > > > > > I finally had some time to look into this issue. I don't think
> > > > > > delaying destruction in psi_trigger_destroy() because there are still
> > > > > > users of the trigger as Hillf suggested is a good way to go. Before
> > > > > > [1] correct trigger destruction was handled using a
> > > > > > psi_trigger.refcount. For some reason I thought it's not needed
> > > > > > anymore when we placed one-trigger-per-file restriction in that patch,
> > > > > > so I removed it. Obviously that was a wrong move, so I think the
> > > > > > cleanest way would be to bring back the refcounting. That way the last
> > > > > > user of the trigger (either psi_trigger_poll() or psi_fop_release())
> > > > > > will free the trigger.
> > > > > > I'll check once more to make sure I did not miss anything and if there
> > > > > > are no objections, will post a fix.
> > > > >
> > > > > Uh, I recalled now why refcounting was not helpful here. I'm making
> > > > > the same mistake of thinking that poll_wait() blocks until the call to
> > > > > wake_up() which is not the case. Let me think if there is anything
> > > > > better than wake_up_pollfree() for this case.
> > > >
> > > > Hi Munehisa,
> > > > Sorry for the delay. I was trying to reproduce the issue but even
> > > > after adding a delay before ep_remove_wait_queue() it did not happen.
> > >
> > > Hi Suren,
> > >
> > > Thank you for your help here.
> > >
> > > Just in case, do you have KASAN enabled in your config? If not, this may
> > > just silently corrupt a certain memory location and not immediately
> > > followed by obvious messages or noticeable event like oops.
> >
> > Yes, KASAN was enabled in my build.
> >
> > >
> > > > One thing about wake_up_pollfree() solution that does not seem right
> > > > to me is this comment at
> > > > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253:
> > > >
> > > > `In the very rare cases where a ->poll() implementation uses a
> > > > waitqueue whose lifetime is tied to a task rather than to the 'struct
> > > > file' being polled, this function must be called before the waitqueue
> > > > is freed...`
> > > >
> > > > In our case we free the waitqueue from cgroup_pressure_release(),
> > > > which is the handler for `release` operation on cgroup psi files. The
> > > > other place calling psi_trigger_destroy() is psi_fop_release(), which
> > > > is also tied to the lifetime to the psi files.  Therefore the lifetime
> > > > of the trigger's waitqueue is tied to the lifetime of the files and
> > > > IIUC, we should not be required to use wake_up_pollfree().
> > > > Could you please post your .config file? I might be missing some
> > > > configuration which prevents the issue from happening on my side.
> > >
> > > Sure, here is my config.
> > >
> > >  https://gist.github.com/kamatam9/a078bdd9f695e7a0767b061c60e48d50
> > >
> > > I confirmed that it's reliably reproducible with v6.2-rc3 as shown below.
> > >
> > >  https://gist.github.com/kamatam9/096a79cf59d8ed8785c4267e917b8675
> >
> > Thanks! I'll try to figure out the difference.
>
> Sorry for the slow progress on this issue. I'm multiplexing between
> several tasks ATM but I did not forget about this one.
> Even though I still can't get the kasan UAF report, I clearly see the
> wrong order when tracing these functions and forcing the delay before
> ep_remove_wait_queue(). I don't think that should be happening, so
> something in the release process I think needs fixing. Will update
> once I figure out the root cause, hopefully sometime this week.

Hi Folks,
I spent some more time digging into the details and this is what's
happening. When we call rmdir to delete the cgroup with the pressure
file being epoll'ed, roughly the following call chain happens in the
context of the shell process:

do_rmdir
  cgroup_rmdir
    kernfs_drain_open_files
      cgroup_file_release
        cgroup_pressure_release
          psi_trigger_destroy

Later on in the context of our reproducer, the last fput() is called
causing wait queue removal:

fput
  ep_eventpoll_release
    ep_free
      ep_remove_wait_queue
        remove_wait_queue

By this time psi_trigger_destroy() already destroyed the trigger's
waitqueue head and we hit UAF.
I think the conceptual problem here (or maybe that's by design?) is
that cgroup_file_release() is not really tied to the file's real
lifetime (when the last fput() is issued). Otherwise fput() would call
eventpoll_release() before f_op->release() and the order would be fine
(we would remove the wait queue first in eventpoll_release() and then
f_op->release() would cause trigger's destruction).
Considering these findings, I think we can use the wake_up_pollfree()
without contradicting the comment at
https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253
because indeed, cgroup_file_release() and therefore
psi_trigger_destroy() are not tied to the file's lifetime.

I'm CC'ing Tejun to check if this makes sense to him and
cgroup_file_release() is working as expected in this case.

Munehisha, if Tejun confirms this is all valid, could you please post
a patch replacing wake_up_interruptible() with wake_up_pollfree()? We
don't need to worry about wake_up_all() because we have a limitation
of one trigger per file descriptor:
https://elixir.bootlin.com/linux/latest/source/kernel/sched/psi.c#L1419,
so there can be only one waiter.
Thanks,
Suren.


>
>
> > Suren.
> >
> > >
> > >
> > > Regards,
> > > Munehisa
> > >
> > >
> > > > Thanks,
> > > > Suren.
> > > >
> > > > >
> > > > >
> > > > > >
> > > > > > [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
> > > > > >
> > > > > > Thanks,
> > > > > > Suren.
> > > > > >
> > > > > > >
> > > > > > > Hillf
> > > >
> > > >

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Fri, Jan 13, 2023 at 9:52 AM Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Thu, Jan 12, 2023 at 6:26 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> >
> > On Thu, 2023-01-12 22:01:24 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
> > >
> > > On Mon, Jan 9, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > > >
> > > > On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > > > >
> > > > > On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> > > > > >
> > > > > > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > > > > > >
> > > > > > > That patch survived the repro in my original post, however, the waker
> > > > > > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > > > > > the pressure file got closed. So, if the following modified repro runs
> > > > > > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > > > > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > > > > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > > > > > the queue, but wake_up_pollfree() does.
> > > > > >
> > > > > > Thanks for your testing. And the debugging completes.
> > > > > >
> > > > > > Mind sending a patch with wake_up_pollfree() folded?
> > > > >
> > > > > I finally had some time to look into this issue. I don't think
> > > > > delaying destruction in psi_trigger_destroy() because there are still
> > > > > users of the trigger as Hillf suggested is a good way to go. Before
> > > > > [1] correct trigger destruction was handled using a
> > > > > psi_trigger.refcount. For some reason I thought it's not needed
> > > > > anymore when we placed one-trigger-per-file restriction in that patch,
> > > > > so I removed it. Obviously that was a wrong move, so I think the
> > > > > cleanest way would be to bring back the refcounting. That way the last
> > > > > user of the trigger (either psi_trigger_poll() or psi_fop_release())
> > > > > will free the trigger.
> > > > > I'll check once more to make sure I did not miss anything and if there
> > > > > are no objections, will post a fix.
> > > >
> > > > Uh, I recalled now why refcounting was not helpful here. I'm making
> > > > the same mistake of thinking that poll_wait() blocks until the call to
> > > > wake_up() which is not the case. Let me think if there is anything
> > > > better than wake_up_pollfree() for this case.
> > >
> > > Hi Munehisa,
> > > Sorry for the delay. I was trying to reproduce the issue but even
> > > after adding a delay before ep_remove_wait_queue() it did not happen.
> >
> > Hi Suren,
> >
> > Thank you for your help here.
> >
> > Just in case, do you have KASAN enabled in your config? If not, this may
> > just silently corrupt a certain memory location and not immediately
> > followed by obvious messages or noticeable event like oops.
>
> Yes, KASAN was enabled in my build.
>
> >
> > > One thing about wake_up_pollfree() solution that does not seem right
> > > to me is this comment at
> > > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253:
> > >
> > > `In the very rare cases where a ->poll() implementation uses a
> > > waitqueue whose lifetime is tied to a task rather than to the 'struct
> > > file' being polled, this function must be called before the waitqueue
> > > is freed...`
> > >
> > > In our case we free the waitqueue from cgroup_pressure_release(),
> > > which is the handler for `release` operation on cgroup psi files. The
> > > other place calling psi_trigger_destroy() is psi_fop_release(), which
> > > is also tied to the lifetime to the psi files.  Therefore the lifetime
> > > of the trigger's waitqueue is tied to the lifetime of the files and
> > > IIUC, we should not be required to use wake_up_pollfree().
> > > Could you please post your .config file? I might be missing some
> > > configuration which prevents the issue from happening on my side.
> >
> > Sure, here is my config.
> >
> >  https://gist.github.com/kamatam9/a078bdd9f695e7a0767b061c60e48d50
> >
> > I confirmed that it's reliably reproducible with v6.2-rc3 as shown below.
> >
> >  https://gist.github.com/kamatam9/096a79cf59d8ed8785c4267e917b8675
>
> Thanks! I'll try to figure out the difference.

Sorry for the slow progress on this issue. I'm multiplexing between
several tasks ATM but I did not forget about this one.
Even though I still can't get the kasan UAF report, I clearly see the
wrong order when tracing these functions and forcing the delay before
ep_remove_wait_queue(). I don't think that should be happening, so
something in the release process I think needs fixing. Will update
once I figure out the root cause, hopefully sometime this week.


> Suren.
>
> >
> >
> > Regards,
> > Munehisa
> >
> >
> > > Thanks,
> > > Suren.
> > >
> > > >
> > > >
> > > > >
> > > > > [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
> > > > >
> > > > > Thanks,
> > > > > Suren.
> > > > >
> > > > > >
> > > > > > Hillf
> > >
> > >

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Thu, Jan 12, 2023 at 6:26 PM Munehisa Kamata <kamatam@amazon.com> wrote:
>
> On Thu, 2023-01-12 22:01:24 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > On Mon, Jan 9, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > >
> > > On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > > >
> > > > On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> > > > >
> > > > > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > > > > >
> > > > > > That patch survived the repro in my original post, however, the waker
> > > > > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > > > > the pressure file got closed. So, if the following modified repro runs
> > > > > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > > > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > > > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > > > > the queue, but wake_up_pollfree() does.
> > > > >
> > > > > Thanks for your testing. And the debugging completes.
> > > > >
> > > > > Mind sending a patch with wake_up_pollfree() folded?
> > > >
> > > > I finally had some time to look into this issue. I don't think
> > > > delaying destruction in psi_trigger_destroy() because there are still
> > > > users of the trigger as Hillf suggested is a good way to go. Before
> > > > [1] correct trigger destruction was handled using a
> > > > psi_trigger.refcount. For some reason I thought it's not needed
> > > > anymore when we placed one-trigger-per-file restriction in that patch,
> > > > so I removed it. Obviously that was a wrong move, so I think the
> > > > cleanest way would be to bring back the refcounting. That way the last
> > > > user of the trigger (either psi_trigger_poll() or psi_fop_release())
> > > > will free the trigger.
> > > > I'll check once more to make sure I did not miss anything and if there
> > > > are no objections, will post a fix.
> > >
> > > Uh, I recalled now why refcounting was not helpful here. I'm making
> > > the same mistake of thinking that poll_wait() blocks until the call to
> > > wake_up() which is not the case. Let me think if there is anything
> > > better than wake_up_pollfree() for this case.
> >
> > Hi Munehisa,
> > Sorry for the delay. I was trying to reproduce the issue but even
> > after adding a delay before ep_remove_wait_queue() it did not happen.
>
> Hi Suren,
>
> Thank you for your help here.
>
> Just in case, do you have KASAN enabled in your config? If not, this may
> just silently corrupt a certain memory location and not immediately
> followed by obvious messages or noticeable event like oops.

Yes, KASAN was enabled in my build.

>
> > One thing about wake_up_pollfree() solution that does not seem right
> > to me is this comment at
> > https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253:
> >
> > `In the very rare cases where a ->poll() implementation uses a
> > waitqueue whose lifetime is tied to a task rather than to the 'struct
> > file' being polled, this function must be called before the waitqueue
> > is freed...`
> >
> > In our case we free the waitqueue from cgroup_pressure_release(),
> > which is the handler for `release` operation on cgroup psi files. The
> > other place calling psi_trigger_destroy() is psi_fop_release(), which
> > is also tied to the lifetime to the psi files.  Therefore the lifetime
> > of the trigger's waitqueue is tied to the lifetime of the files and
> > IIUC, we should not be required to use wake_up_pollfree().
> > Could you please post your .config file? I might be missing some
> > configuration which prevents the issue from happening on my side.
>
> Sure, here is my config.
>
>  https://gist.github.com/kamatam9/a078bdd9f695e7a0767b061c60e48d50
>
> I confirmed that it's reliably reproducible with v6.2-rc3 as shown below.
>
>  https://gist.github.com/kamatam9/096a79cf59d8ed8785c4267e917b8675

Thanks! I'll try to figure out the difference.
Suren.

>
>
> Regards,
> Munehisa
>
>
> > Thanks,
> > Suren.
> >
> > >
> > >
> > > >
> > > > [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
> > > >
> > > > Thanks,
> > > > Suren.
> > > >
> > > > >
> > > > > Hillf
> >
> >

Re: another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

On Thu, 2023-01-12 22:01:24 +0000, Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Mon, Jan 9, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
> > >
> > > On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> > > >
> > > > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > > > >
> > > > > That patch survived the repro in my original post, however, the waker
> > > > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > > > the pressure file got closed. So, if the following modified repro runs
> > > > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > > > the queue, but wake_up_pollfree() does.
> > > >
> > > > Thanks for your testing. And the debugging completes.
> > > >
> > > > Mind sending a patch with wake_up_pollfree() folded?
> > >
> > > I finally had some time to look into this issue. I don't think
> > > delaying destruction in psi_trigger_destroy() because there are still
> > > users of the trigger as Hillf suggested is a good way to go. Before
> > > [1] correct trigger destruction was handled using a
> > > psi_trigger.refcount. For some reason I thought it's not needed
> > > anymore when we placed one-trigger-per-file restriction in that patch,
> > > so I removed it. Obviously that was a wrong move, so I think the
> > > cleanest way would be to bring back the refcounting. That way the last
> > > user of the trigger (either psi_trigger_poll() or psi_fop_release())
> > > will free the trigger.
> > > I'll check once more to make sure I did not miss anything and if there
> > > are no objections, will post a fix.
> >
> > Uh, I recalled now why refcounting was not helpful here. I'm making
> > the same mistake of thinking that poll_wait() blocks until the call to
> > wake_up() which is not the case. Let me think if there is anything
> > better than wake_up_pollfree() for this case.
> 
> Hi Munehisa,
> Sorry for the delay. I was trying to reproduce the issue but even
> after adding a delay before ep_remove_wait_queue() it did not happen.

Hi Suren,

Thank you for your help here.

Just in case, do you have KASAN enabled in your config? If not, this may
just silently corrupt a certain memory location and not immediately
followed by obvious messages or noticeable event like oops.

> One thing about wake_up_pollfree() solution that does not seem right
> to me is this comment at
> https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253:
> 
> `In the very rare cases where a ->poll() implementation uses a
> waitqueue whose lifetime is tied to a task rather than to the 'struct
> file' being polled, this function must be called before the waitqueue
> is freed...`
> 
> In our case we free the waitqueue from cgroup_pressure_release(),
> which is the handler for `release` operation on cgroup psi files. The
> other place calling psi_trigger_destroy() is psi_fop_release(), which
> is also tied to the lifetime to the psi files.  Therefore the lifetime
> of the trigger's waitqueue is tied to the lifetime of the files and
> IIUC, we should not be required to use wake_up_pollfree().
> Could you please post your .config file? I might be missing some
> configuration which prevents the issue from happening on my side.

Sure, here is my config.

 https://gist.github.com/kamatam9/a078bdd9f695e7a0767b061c60e48d50

I confirmed that it's reliably reproducible with v6.2-rc3 as shown below.

 https://gist.github.com/kamatam9/096a79cf59d8ed8785c4267e917b8675


Regards,
Munehisa


> Thanks,
> Suren.
> 
> >
> >
> > >
> > > [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
> > >
> > > Thanks,
> > > Suren.
> > >
> > > >
> > > > Hillf
> 
> 

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Mon, Jan 9, 2023 at 7:06 PM Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
> >
> > On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> > >
> > > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > > >
> > > > That patch survived the repro in my original post, however, the waker
> > > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > > the pressure file got closed. So, if the following modified repro runs
> > > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > > the queue, but wake_up_pollfree() does.
> > >
> > > Thanks for your testing. And the debugging completes.
> > >
> > > Mind sending a patch with wake_up_pollfree() folded?
> >
> > I finally had some time to look into this issue. I don't think
> > delaying destruction in psi_trigger_destroy() because there are still
> > users of the trigger as Hillf suggested is a good way to go. Before
> > [1] correct trigger destruction was handled using a
> > psi_trigger.refcount. For some reason I thought it's not needed
> > anymore when we placed one-trigger-per-file restriction in that patch,
> > so I removed it. Obviously that was a wrong move, so I think the
> > cleanest way would be to bring back the refcounting. That way the last
> > user of the trigger (either psi_trigger_poll() or psi_fop_release())
> > will free the trigger.
> > I'll check once more to make sure I did not miss anything and if there
> > are no objections, will post a fix.
>
> Uh, I recalled now why refcounting was not helpful here. I'm making
> the same mistake of thinking that poll_wait() blocks until the call to
> wake_up() which is not the case. Let me think if there is anything
> better than wake_up_pollfree() for this case.

Hi Munehisa,
Sorry for the delay. I was trying to reproduce the issue but even
after adding a delay before ep_remove_wait_queue() it did not happen.
One thing about wake_up_pollfree() solution that does not seem right
to me is this comment at
https://elixir.bootlin.com/linux/latest/source/include/linux/wait.h#L253:

`In the very rare cases where a ->poll() implementation uses a
waitqueue whose lifetime is tied to a task rather than to the 'struct
file' being polled, this function must be called before the waitqueue
is freed...`

In our case we free the waitqueue from cgroup_pressure_release(),
which is the handler for `release` operation on cgroup psi files. The
other place calling psi_trigger_destroy() is psi_fop_release(), which
is also tied to the lifetime to the psi files.  Therefore the lifetime
of the trigger's waitqueue is tied to the lifetime of the files and
IIUC, we should not be required to use wake_up_pollfree().
Could you please post your .config file? I might be missing some
configuration which prevents the issue from happening on my side.
Thanks,
Suren.

>
>
> >
> > [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
> >
> > Thanks,
> > Suren.
> >
> > >
> > > Hillf

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Mon, Jan 9, 2023 at 5:33 PM Suren Baghdasaryan <surenb@google.com> wrote:
>
> On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
> >
> > On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > >
> > > That patch survived the repro in my original post, however, the waker
> > > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > > the pressure file got closed. So, if the following modified repro runs
> > > with the patch, the waker never returns (unless the sleeper gets killed)
> > > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > > see with the patch, does it? Even wake_up_all() does not appear to empty
> > > the queue, but wake_up_pollfree() does.
> >
> > Thanks for your testing. And the debugging completes.
> >
> > Mind sending a patch with wake_up_pollfree() folded?
>
> I finally had some time to look into this issue. I don't think
> delaying destruction in psi_trigger_destroy() because there are still
> users of the trigger as Hillf suggested is a good way to go. Before
> [1] correct trigger destruction was handled using a
> psi_trigger.refcount. For some reason I thought it's not needed
> anymore when we placed one-trigger-per-file restriction in that patch,
> so I removed it. Obviously that was a wrong move, so I think the
> cleanest way would be to bring back the refcounting. That way the last
> user of the trigger (either psi_trigger_poll() or psi_fop_release())
> will free the trigger.
> I'll check once more to make sure I did not miss anything and if there
> are no objections, will post a fix.

Uh, I recalled now why refcounting was not helpful here. I'm making
the same mistake of thinking that poll_wait() blocks until the call to
wake_up() which is not the case. Let me think if there is anything
better than wake_up_pollfree() for this case.


>
> [1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/
>
> Thanks,
> Suren.
>
> >
> > Hillf

Re: another use-after-free in ep_remove_wait_queue()

[Suren Baghdasaryan]

On Sun, Jan 8, 2023 at 3:49 PM Hillf Danton <hdanton@sina.com> wrote:
>
> On 8 Jan 2023 14:25:48 -0800 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> >
> > That patch survived the repro in my original post, however, the waker
> > (rmdir) was getting stuck until a file descriptor of the epoll instance or
> > the pressure file got closed. So, if the following modified repro runs
> > with the patch, the waker never returns (unless the sleeper gets killed)
> > while holding cgroup_mutex. This doesn't seem to be what you expected to
> > see with the patch, does it? Even wake_up_all() does not appear to empty
> > the queue, but wake_up_pollfree() does.
>
> Thanks for your testing. And the debugging completes.
>
> Mind sending a patch with wake_up_pollfree() folded?

I finally had some time to look into this issue. I don't think
delaying destruction in psi_trigger_destroy() because there are still
users of the trigger as Hillf suggested is a good way to go. Before
[1] correct trigger destruction was handled using a
psi_trigger.refcount. For some reason I thought it's not needed
anymore when we placed one-trigger-per-file restriction in that patch,
so I removed it. Obviously that was a wrong move, so I think the
cleanest way would be to bring back the refcounting. That way the last
user of the trigger (either psi_trigger_poll() or psi_fop_release())
will free the trigger.
I'll check once more to make sure I did not miss anything and if there
are no objections, will post a fix.

[1] https://lore.kernel.org/lkml/20220111232309.1786347-1-surenb@google.com/

Thanks,
Suren.

>
> Hillf

Re: another use-after-free in ep_remove_wait_queue()

[Munehisa Kamata]

Hi Hillf,

On Sat, 2023-01-07 08:07:02 +0000, Hillf Danton <hdanton@sina.com> wrote:
> On Fri, Jan 6, 2023 at 2:49 PM Munehisa Kamata <kamatam@amazon.com> wrote:
> > Hi,
> >
> > We found the following use-after-free with userspace code polling on
> > a pressure file in a non-root cgroup using epoll.
> >
> > [   57.183661] BUG: KASAN: use-after-free in _raw_spin_lock_irqsave+0x76/0x130
> > [   57.186662] Write of size 4 at addr ffff888114976428 by task a.out/2426
> >
> > [   57.190551] CPU: 0 PID: 2426 Comm: a.out Not tainted 6.2.0-rc2+ #16
> > [   57.193384] Hardware name: Amazon EC2 c5a.large/, BIOS 1.0 10/16/2017
> > [   57.196272] Call Trace:
> > [   57.197565]  <TASK>
> > [   57.198714]  dump_stack_lvl+0x8f/0xc0
> > [   57.200494]  print_report+0x16c/0x4e0
> > [   57.202084]  ? _raw_spin_lock_irqsave+0x76/0x130
> > [   57.204077]  kasan_report+0xc3/0xf0
> > [   57.205587]  ? entry_SYSCALL_64_after_hwframe+0x63/0xcd
> > [   57.207760]  ? _raw_spin_lock_irqsave+0x76/0x130
> > [   57.209685]  kasan_check_range+0x2d2/0x310
> > [   57.211477]  _raw_spin_lock_irqsave+0x76/0x130
> > [   57.213355]  remove_wait_queue+0x25/0x130
> > [   57.215102]  ep_free+0x12d/0x220
> > [   57.216506]  ep_eventpoll_release+0x3c/0x40
> > [   57.218254]  __fput+0x32b/0x700
> > [   57.221486]  task_work_run+0x1db/0x230
> > [   57.224885]  exit_to_user_mode_prepare+0xfd/0x100
> > [   57.228662]  syscall_exit_to_user_mode+0x20/0x40
> > [   57.232360]  do_syscall_64+0x52/0x90
> > [   57.235691]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> > [   57.239572] RIP: 0033:0x7fadf96e1c44
> > [   57.242865] Code: 00 00 b8 ff ff ff ff eb 9c b8 ff ff ff ff eb 95 e8 01 e2 01 00 90 8b 05 2a ac 2c 00 48 63 ff 85 c0 75 11 b8 03 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 3a f3 c3 48 83 ec 18 48 89 7c 24 08 e8 e4 a0
> > [   57.255244] RSP: 002b:00007ffd1d1b7b98 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
> > [   57.261714] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007fadf96e1c44
> > [   57.266293] RDX: 0000000000000000 RSI: 00007ffd1d1b7b60 RDI: 0000000000000004
> > [   57.270979] RBP: 00007ffd1d1b7bf0 R08: 00000000004007e0 R09: 00007fadf9a0f240
> > [   57.275856] R10: 00000000000006ba R11: 0000000000000246 R12: 00000000004005e0
> > [   57.280478] R13: 00007ffd1d1b7cd0 R14: 0000000000000000 R15: 0000000000000000
> > [   57.285059]  </TASK>
> >
> > [   57.290402] Allocated by task 2426:
> > [   57.293705]  kasan_set_track+0x3d/0x60
> > [   57.297102]  __kasan_kmalloc+0x85/0x90
> > [   57.300491]  psi_trigger_create+0x155/0x850
> > [   57.304040]  pressure_write+0x200/0x510
> > [   57.307508]  cgroup_file_write+0x1de/0x3e0
> > [   57.310949]  kernfs_fop_write_iter+0x27d/0x380
> > [   57.314601]  vfs_write+0x7d7/0xaa0
> > [   57.317891]  ksys_write+0xd7/0x1a0
> > [   57.321152]  do_syscall_64+0x43/0x90
> > [   57.324496]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> >
> > [   57.330887] Freed by task 2429:
> > [   57.334053]  kasan_set_track+0x3d/0x60
> > [   57.337448]  kasan_save_free_info+0x27/0x40
> > [   57.340977]  ____kasan_slab_free+0x11d/0x170
> > [   57.344570]  slab_free_freelist_hook+0x87/0x150
> > [   57.348236]  __kmem_cache_free+0xcb/0x180
> > [   57.351710]  psi_trigger_destroy+0x458/0x550
> > [   57.355274]  cgroup_file_release+0x96/0x110
> > [   57.358779]  kernfs_drain_open_files+0x238/0x420
> > [   57.362519]  kernfs_drain+0x191/0x2a0
> > [   57.365901]  __kernfs_remove+0x3a6/0x600
> > [   57.369363]  kernfs_remove_by_name_ns+0xc2/0x120
> > [   57.373073]  cgroup_addrm_files+0x90f/0xcf0
> > [   57.376610]  cgroup_destroy_locked+0x48a/0x730
> > [   57.380260]  cgroup_rmdir+0x2b/0x130
> > [   57.383650]  kernfs_iop_rmdir+0x17a/0x230
> > [   57.387201]  vfs_rmdir+0x196/0x410
> > [   57.390442]  do_rmdir+0x1c7/0x3f0
> > [   57.393651]  __x64_sys_rmdir+0x45/0x50
> > [   57.397000]  do_syscall_64+0x43/0x90
> > [   57.400340]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> >
> > [   57.406689] The buggy address belongs to the object at ffff888114976400
> >                 which belongs to the cache kmalloc-128 of size 128
> > [   57.414907] The buggy address is located 40 bytes inside of
> >                 128-byte region [ffff888114976400, ffff888114976480)
> >
> > [   57.425474] The buggy address belongs to the physical page:
> > [   57.429541] page:000000008c5ecb31 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x114976
> > [   57.436725] flags: 0x2fffff80000200(slab|node=0|zone=2|lastcpupid=0x1fffff)
> > [   57.441235] raw: 002fffff80000200 ffff8881000418c0 dead000000000100 dead000000000122
> > [   57.447793] raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
> > [   57.454274] page dumped because: kasan: bad access detected
> >
> > [   57.460990] Memory state around the buggy address:
> > [   57.464744]  ffff888114976300: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> > [   57.471126]  ffff888114976380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> > [   57.477447] >ffff888114976400: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> > [   57.483833]                                   ^
> > [   57.487541]  ffff888114976480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> > [   57.493976]  ffff888114976500: 00 00 00 00 00 00 00 00 00 00 00 00 00 01 fc fc
> >
> > Here is the simple repro.
> >
> > #include <fcntl.h>
> > #include <sys/epoll.h>
> > #include <unistd.h>
> >
> > int main(void)
> > {
> >         const char trigger_str[] = "some 100000 1000000";
> >         int fd, epfd;
> >         struct epoll_event event;
> >         struct epoll_event events[1];
> >
> >         fd = open("/cgroup2/test/cpu.pressure", O_RDWR);
> >         write(fd, trigger_str, sizeof(trigger_str));
> >         epfd = epoll_create(1);
> >         event.events = EPOLLPRI | EPOLLET;
> >         event.data.fd = fd;
> >         epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &event);
> >         epoll_wait(epfd, events, 1, -1); /* returns after rmdir */
> >         sleep(5);
> >         close(epfd);
> >         close(fd);
> >
> >         return 0;
> > }
> >
> > # mkdir /cgroup2
> > # mount -t cgroup2 none /cgroup2
> > # mkdir /cgroup2/test
> > # ./a.out &
> > # rmdir /cgroup2/test
> >
> > Looks like calling wake_up_pollfree() in psi_trigger_destroy() can properly
> > clear the queue and then avoid this use-after-free, but POLLFREE wasn't
> > considered enough there for the past similar issue[1]. While
> > wake_up_pollfree() could *also* be called in psi_trigger_destroy(), it may
> > be awkward and there can be more appropriate solution. It would be great if
> > experts could have a look.
> 
> Thanks for your report.
> 
> In the wakeup pattern below,
> 
> 	cpu0				cpu2
> 	---				---
> 	wake_up(&t->event_wait);
> 	HZ later
> 	kfree(t);
> 					HZ *3 later
> 					the last sleeper gets on CPU
> 					remove_wait_queue
> 
> waker should wait for every sleeper to go off waitqueue.
> 
> See if the diff below could survive your repro.

That patch survived the repro in my original post, however, the waker
(rmdir) was getting stuck until a file descriptor of the epoll instance or
the pressure file got closed. So, if the following modified repro runs
with the patch, the waker never returns (unless the sleeper gets killed)
while holding cgroup_mutex. This doesn't seem to be what you expected to
see with the patch, does it? Even wake_up_all() does not appear to empty
the queue, but wake_up_pollfree() does.

#include <fcntl.h>
#include <sys/epoll.h>
#include <unistd.h>

int main(void)
{
	const char trigger_str[] = "some 100000 1000000";
	int fd, epfd;
	struct epoll_event event;
	struct epoll_event events[1];

	fd = open("/cgroup2/test/cpu.pressure", O_RDWR);
	write(fd, trigger_str, sizeof(trigger_str));
	epfd = epoll_create(1);
	event.events = EPOLLPRI | EPOLLET;
	event.data.fd = fd;
	epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &event);
	epoll_wait(epfd, events, 1, -1);

	pause();

	return 0;
}

# mkdir /cgroup2
# mount -t cgroup2 none /cgroup2
# mkdir /cgroup2/test
# ./a.out &
# rmdir /cgroup2/test
# klilall a.out


Regards,
Munehisa

> Hillf
> 
> --- mainline/kernel/sched/psi.c
> +++ y/kernel/sched/psi.c
> @@ -1346,7 +1346,7 @@ void psi_trigger_destroy(struct psi_trig
>  	 * Wakeup waiters to stop polling. Can happen if cgroup is deleted
>  	 * from under a polling process.
>  	 */
> -	wake_up_interruptible(&t->event_wait);
> +	wake_up_all(&t->event_wait);
>  
>  	mutex_lock(&group->trigger_lock);
>  
> @@ -1394,6 +1394,10 @@ void psi_trigger_destroy(struct psi_trig
>  		kthread_stop(task_to_destroy);
>  		atomic_set(&group->poll_scheduled, 0);
>  	}
> +
> +	while (wq_has_sleeper(&t->event_wait))
> +		schedule_timeout_idle(HZ);
> +
>  	kfree(t);
>  }
>  
> --
> 
>