general protection fault in proc_kill_sb

general protection fault in proc_kill_sb

[syzbot]

Hello,

syzbot found the following crash on:

HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=16e12212100000
kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
dashboard link: https://syzkaller.appspot.com/bug?extid=4abac52934a48af5ff19
compiler:       gcc (GCC) 9.0.0 20181231 (experimental)

Unfortunately, I don't have any reproducer for this crash yet.

IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+4abac52934a48af5ff19@syzkaller.appspotmail.com

RBP: 000000000078bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000005
R13: 0000000000000751 R14: 00000000004ca3ea R15: 00007f8970ff76d4
general protection fault, probably for non-canonical address 0xdffffc0000000001: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000008-0x000000000000000f]
CPU: 0 PID: 6840 Comm: syz-executor.2 Not tainted 5.7.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:proc_kill_sb+0x4d/0x110 fs/proc/root.c:267
Code: c1 ea 03 80 3c 02 00 0f 85 ad 00 00 00 49 8b ac 24 68 06 00 00 48 b8 00 00 00 00 00 fc ff df 48 8d 7d 08 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 a1 00 00 00 4c 8b 6d 08 4d 85 ed 74 0d e8 cb 51
RSP: 0018:ffffc90007b67d48 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 0000000000000001 RCX: ffffc9000d6d1000
RDX: 0000000000000001 RSI: ffffffff81e37c9d RDI: 0000000000000008
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000001
R10: ffffffff8c347a2f R11: fffffbfff1868f45 R12: ffff888059e1a000
R13: 00000000fffffff4 R14: 0000000000000000 R15: ffff888059e1a068
FS:  00007f8970ff7700(0000) GS:ffff8880ae600000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000073c138 CR3: 00000000588dc000 CR4: 00000000001406f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 deactivate_locked_super+0x8c/0xf0 fs/super.c:335
 vfs_get_super+0x258/0x2d0 fs/super.c:1212
 vfs_get_tree+0x89/0x2f0 fs/super.c:1547
 do_new_mount fs/namespace.c:2874 [inline]
 do_mount+0x1306/0x1b40 fs/namespace.c:3199
 __do_sys_mount fs/namespace.c:3409 [inline]
 __se_sys_mount fs/namespace.c:3386 [inline]
 __x64_sys_mount+0x18f/0x230 fs/namespace.c:3386
 do_syscall_64+0xf6/0x7d0 arch/x86/entry/common.c:295
 entry_SYSCALL_64_after_hwframe+0x49/0xb3
RIP: 0033:0x45ca69
Code: 0d b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 db b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007f8970ff6c78 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
RAX: ffffffffffffffda RBX: 00000000004f62c0 RCX: 000000000045ca69
RDX: 00000000200000c0 RSI: 0000000020000200 RDI: 0000000020000340
RBP: 000000000078bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000005
R13: 0000000000000751 R14: 00000000004ca3ea R15: 00007f8970ff76d4
Modules linked in:
---[ end trace 7f26d357bd21d77e ]---
RIP: 0010:proc_kill_sb+0x4d/0x110 fs/proc/root.c:267
Code: c1 ea 03 80 3c 02 00 0f 85 ad 00 00 00 49 8b ac 24 68 06 00 00 48 b8 00 00 00 00 00 fc ff df 48 8d 7d 08 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 a1 00 00 00 4c 8b 6d 08 4d 85 ed 74 0d e8 cb 51
RSP: 0018:ffffc90007b67d48 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 0000000000000001 RCX: ffffc9000d6d1000
RDX: 0000000000000001 RSI: ffffffff81e37c9d RDI: 0000000000000008
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000001
R10: ffffffff8c347a2f R11: fffffbfff1868f45 R12: ffff888059e1a000
R13: 00000000fffffff4 R14: 0000000000000000 R15: ffff888059e1a068
FS:  00007f8970ff7700(0000) GS:ffff8880ae600000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f19326ab000 CR3: 00000000588dc000 CR4: 00000000001406f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400


---
This bug is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzkaller@googlegroups.com.

syzbot will keep track of this bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

Re: general protection fault in proc_kill_sb

[Tetsuo Handa]

On 2020/06/10 19:56, syzbot wrote:
> Hello,
> 
> syzbot found the following crash on:
> 
> HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=16e12212100000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> dashboard link: https://syzkaller.appspot.com/bug?extid=4abac52934a48af5ff19
> compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> 
> Unfortunately, I don't have any reproducer for this crash yet.

The report says proc_sb_info(sb) == NULL at proc_kill_sb() which was called via
fs->kill_sb(s) from deactivate_locked_super(). The console log says that memory
allocation for proc_sb_info(sb) failed due to memory allocation fault injection.

[ 1492.052802][ T6840] FAULT_INJECTION: forcing a failure.
[ 1492.052802][ T6840] name failslab, interval 1, probability 0, space 0, times 0
[ 1492.077153][ T6840] CPU: 0 PID: 6840 Comm: syz-executor.2 Not tainted 5.7.0-syzkaller #0
[ 1492.085449][ T6840] Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
[ 1492.095511][ T6840] Call Trace:
[ 1492.098811][ T6840]  dump_stack+0x188/0x20d
[ 1492.103157][ T6840]  should_fail.cold+0x5/0xa
[ 1492.107686][ T6840]  ? fault_create_debugfs_attr+0x140/0x140
[ 1492.107721][ T6840]  ? idr_replace+0xee/0x160
[ 1492.127210][ T6840]  should_failslab+0x5/0xf
[ 1492.131638][ T6840]  kmem_cache_alloc_trace+0x2d0/0x7d0
[ 1492.137020][ T6840]  ? up_write+0x148/0x470
[ 1492.141367][ T6840]  proc_fill_super+0x79/0x5c0
[ 1492.146052][ T6840]  ? proc_parse_param+0x8a0/0x8a0
[ 1492.151092][ T6840]  vfs_get_super+0x12e/0x2d0
[ 1492.155694][ T6840]  vfs_get_tree+0x89/0x2f0
[ 1492.160126][ T6840]  do_mount+0x1306/0x1b40
[ 1492.164467][ T6840]  ? copy_mount_string+0x40/0x40
[ 1492.169411][ T6840]  ? __might_fault+0x190/0x1d0
[ 1492.174188][ T6840]  ? _copy_from_user+0x13c/0x1a0
[ 1492.179138][ T6840]  ? memdup_user+0x7c/0xd0
[ 1492.183575][ T6840]  __x64_sys_mount+0x18f/0x230
[ 1492.188351][ T6840]  do_syscall_64+0xf6/0x7d0
[ 1492.192861][ T6840]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
[ 1492.198759][ T6840] RIP: 0033:0x45ca69

That is, proc_kill_sb() was assuming "s->s_fs_info = fs_info;" is always
called from proc_fill_super() which is called via fill_super(sb, fc); from
vfs_get_super().

Re: general protection fault in proc_kill_sb

[Alexey Gladkov]

On Wed, Jun 10, 2020 at 08:23:33PM +0900, Tetsuo Handa wrote:
> On 2020/06/10 19:56, syzbot wrote:
> > Hello,
> > 
> > syzbot found the following crash on:
> > 
> > HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=16e12212100000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> > dashboard link: https://syzkaller.appspot.com/bug?extid=4abac52934a48af5ff19
> > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > 
> > Unfortunately, I don't have any reproducer for this crash yet.
> 
> The report says proc_sb_info(sb) == NULL at proc_kill_sb() which was called via
> fs->kill_sb(s) from deactivate_locked_super(). The console log says that memory
> allocation for proc_sb_info(sb) failed due to memory allocation fault injection.
> 
> [ 1492.052802][ T6840] FAULT_INJECTION: forcing a failure.
> [ 1492.052802][ T6840] name failslab, interval 1, probability 0, space 0, times 0
> [ 1492.077153][ T6840] CPU: 0 PID: 6840 Comm: syz-executor.2 Not tainted 5.7.0-syzkaller #0
> [ 1492.085449][ T6840] Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
> [ 1492.095511][ T6840] Call Trace:
> [ 1492.098811][ T6840]  dump_stack+0x188/0x20d
> [ 1492.103157][ T6840]  should_fail.cold+0x5/0xa
> [ 1492.107686][ T6840]  ? fault_create_debugfs_attr+0x140/0x140
> [ 1492.107721][ T6840]  ? idr_replace+0xee/0x160
> [ 1492.127210][ T6840]  should_failslab+0x5/0xf
> [ 1492.131638][ T6840]  kmem_cache_alloc_trace+0x2d0/0x7d0
> [ 1492.137020][ T6840]  ? up_write+0x148/0x470
> [ 1492.141367][ T6840]  proc_fill_super+0x79/0x5c0
> [ 1492.146052][ T6840]  ? proc_parse_param+0x8a0/0x8a0
> [ 1492.151092][ T6840]  vfs_get_super+0x12e/0x2d0
> [ 1492.155694][ T6840]  vfs_get_tree+0x89/0x2f0
> [ 1492.160126][ T6840]  do_mount+0x1306/0x1b40
> [ 1492.164467][ T6840]  ? copy_mount_string+0x40/0x40
> [ 1492.169411][ T6840]  ? __might_fault+0x190/0x1d0
> [ 1492.174188][ T6840]  ? _copy_from_user+0x13c/0x1a0
> [ 1492.179138][ T6840]  ? memdup_user+0x7c/0xd0
> [ 1492.183575][ T6840]  __x64_sys_mount+0x18f/0x230
> [ 1492.188351][ T6840]  do_syscall_64+0xf6/0x7d0
> [ 1492.192861][ T6840]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
> [ 1492.198759][ T6840] RIP: 0033:0x45ca69
> 
> That is, proc_kill_sb() was assuming "s->s_fs_info = fs_info;" is always
> called from proc_fill_super() which is called via fill_super(sb, fc); from
> vfs_get_super().

Yes. If fill_super() fails before filling up fs_info,
deactivate_locked_super() will be called and sb->s_fs_info may be NULL.

-- 
Rgrds, legion

[PATCH] proc: s_fs_info may be NULL when proc_kill_sb is called

[Alexey Gladkov]

syzbot found that proc_fill_super() fails before filling up sb->s_fs_info,
deactivate_locked_super() will be called and sb->s_fs_info will be NULL.
The proc_kill_sb() does not expect fs_info to be NULL which is wrong.

Link: https://lore.kernel.org/lkml/0000000000002d7ca605a7b8b1c5@google.com
Reported-by: syzbot+4abac52934a48af5ff19@syzkaller.appspotmail.com
Fixes: fa10fed30f25 ("proc: allow to mount many instances of proc in one pid namespace")
Signed-off-by: Alexey Gladkov <gladkov.alexey@gmail.com>
---
 fs/proc/root.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/fs/proc/root.c b/fs/proc/root.c
index ffebed1999e5..a715eb9f196a 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -264,15 +264,18 @@ static void proc_kill_sb(struct super_block *sb)
 {
 	struct proc_fs_info *fs_info = proc_sb_info(sb);
 
-	if (fs_info->proc_self)
-		dput(fs_info->proc_self);
+	if (fs_info) {
+		if (fs_info->proc_self)
+			dput(fs_info->proc_self);
 
-	if (fs_info->proc_thread_self)
-		dput(fs_info->proc_thread_self);
+		if (fs_info->proc_thread_self)
+			dput(fs_info->proc_thread_self);
+
+		put_pid_ns(fs_info->pid_ns);
+		kfree(fs_info);
+	}
 
 	kill_anon_super(sb);
-	put_pid_ns(fs_info->pid_ns);
-	kfree(fs_info);
 }
 
 static struct file_system_type proc_fs_type = {
-- 
2.25.4

Re: [PATCH] proc: s_fs_info may be NULL when proc_kill_sb is called

[Eric W. Biederman]

Alexey Gladkov <gladkov.alexey@gmail.com> writes:

> syzbot found that proc_fill_super() fails before filling up sb->s_fs_info,
> deactivate_locked_super() will be called and sb->s_fs_info will be NULL.
> The proc_kill_sb() does not expect fs_info to be NULL which is wrong.

For the case where s_fs_info is never allocated this looks correct.
That is because generic_shutdown_super has a special for !sb->s_root.

However for the existing cases I can't convince myself that it is safe
to change the order we free the pid namespace and free fs_info.

There is a lot of code that can run while generic_shutdown_super is
running and purging all of the inodes.  We have crazy things like
proc_flush_pid that might care, as well proc_evict_inode.

I haven't found anything that actually references fs_info or actually
depends on the pid namespace living longer than the proc inode but it
would be really easy to miss something.

Can you send a v2 version does not change the order things are freed in
for the case where we do allocate fs_info.  That will make it trivially
safe to apply.

Otherwise this looks like a very good patch.

Thank you,
Eric


> Link: https://lore.kernel.org/lkml/0000000000002d7ca605a7b8b1c5@google.com
> Reported-by: syzbot+4abac52934a48af5ff19@syzkaller.appspotmail.com
> Fixes: fa10fed30f25 ("proc: allow to mount many instances of proc in one pid namespace")
> Signed-off-by: Alexey Gladkov <gladkov.alexey@gmail.com>
> ---
>  fs/proc/root.c | 15 +++++++++------
>  1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/fs/proc/root.c b/fs/proc/root.c
> index ffebed1999e5..a715eb9f196a 100644
> --- a/fs/proc/root.c
> +++ b/fs/proc/root.c
> @@ -264,15 +264,18 @@ static void proc_kill_sb(struct super_block *sb)
>  {
>  	struct proc_fs_info *fs_info = proc_sb_info(sb);
>  
> -	if (fs_info->proc_self)
> -		dput(fs_info->proc_self);
> +	if (fs_info) {
> +		if (fs_info->proc_self)
> +			dput(fs_info->proc_self);
>  
> -	if (fs_info->proc_thread_self)
> -		dput(fs_info->proc_thread_self);
> +		if (fs_info->proc_thread_self)
> +			dput(fs_info->proc_thread_self);
> +
> +		put_pid_ns(fs_info->pid_ns);
> +		kfree(fs_info);
> +	}
>  
>  	kill_anon_super(sb);
> -	put_pid_ns(fs_info->pid_ns);
> -	kfree(fs_info);
>  }
>  
>  static struct file_system_type proc_fs_type = {

Re: [PATCH] proc: s_fs_info may be NULL when proc_kill_sb is called

[Al Viro]

On Wed, Jun 10, 2020 at 12:12:54PM -0500, Eric W. Biederman wrote:

> >  {
> >  	struct proc_fs_info *fs_info = proc_sb_info(sb);
> >  
> > -	if (fs_info->proc_self)
> > -		dput(fs_info->proc_self);
> > +	if (fs_info) {
> > +		if (fs_info->proc_self)
> > +			dput(fs_info->proc_self);
> >  
> > -	if (fs_info->proc_thread_self)
> > -		dput(fs_info->proc_thread_self);
> > +		if (fs_info->proc_thread_self)
> > +			dput(fs_info->proc_thread_self);
> > +
> > +		put_pid_ns(fs_info->pid_ns);
> > +		kfree(fs_info);

While we are at it, dput(NULL) is an explicit no-op.

[PATCH v2] proc: s_fs_info may be NULL when proc_kill_sb is called

[Alexey Gladkov]

syzbot found that proc_fill_super() fails before filling up sb->s_fs_info,
deactivate_locked_super() will be called and sb->s_fs_info will be NULL.
The proc_kill_sb() does not expect fs_info to be NULL which is wrong.

Link: https://lore.kernel.org/lkml/0000000000002d7ca605a7b8b1c5@google.com
Reported-by: syzbot+4abac52934a48af5ff19@syzkaller.appspotmail.com
Fixes: fa10fed30f25 ("proc: allow to mount many instances of proc in one pid namespace")
Signed-off-by: Alexey Gladkov <gladkov.alexey@gmail.com>
---
 fs/proc/root.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/fs/proc/root.c b/fs/proc/root.c
index ffebed1999e5..5e444d4f9717 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -264,11 +264,13 @@ static void proc_kill_sb(struct super_block *sb)
 {
 	struct proc_fs_info *fs_info = proc_sb_info(sb);
 
-	if (fs_info->proc_self)
-		dput(fs_info->proc_self);
+	if (!fs_info) {
+		kill_anon_super(sb);
+		return;
+	}
 
-	if (fs_info->proc_thread_self)
-		dput(fs_info->proc_thread_self);
+	dput(fs_info->proc_self);
+	dput(fs_info->proc_thread_self);
 
 	kill_anon_super(sb);
 	put_pid_ns(fs_info->pid_ns);
-- 
2.25.4

Re: [PATCH v2] proc: s_fs_info may be NULL when proc_kill_sb is called

[Eric W. Biederman]

Alexey Gladkov <gladkov.alexey@gmail.com> writes:

> syzbot found that proc_fill_super() fails before filling up sb->s_fs_info,
> deactivate_locked_super() will be called and sb->s_fs_info will be NULL.
> The proc_kill_sb() does not expect fs_info to be NULL which is wrong.
>
> Link: https://lore.kernel.org/lkml/0000000000002d7ca605a7b8b1c5@google.com
> Reported-by: syzbot+4abac52934a48af5ff19@syzkaller.appspotmail.com
> Fixes: fa10fed30f25 ("proc: allow to mount many instances of proc in one pid namespace")
> Signed-off-by: Alexey Gladkov <gladkov.alexey@gmail.com>

applied

Eric