Mailing List Archive

[syzbot] KMSAN: uninit-value in from_kuid
Hello,

syzbot found the following issue on:

HEAD commit: 425295055ce6 kmsan: core: address comments to kmsan-checks.h
git tree: https://github.com/google/kmsan.git master
console output: https://syzkaller.appspot.com/x/log.txt?x=1640209ab00000
kernel config: https://syzkaller.appspot.com/x/.config?x=2d142cdf4204061
dashboard link: https://syzkaller.appspot.com/bug?extid=dfac92a50024b54acaa4
compiler: clang version 14.0.0 (git@github.com:llvm/llvm-project.git 0996585c8e3b3d409494eb5f1cad714b9e1f7fb5), GNU ld (GNU Binutils for Debian) 2.35.2
userspace arch: i386

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

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+dfac92a50024b54acaa4@syzkaller.appspotmail.com

=====================================================
BUG: KMSAN: uninit-value in map_id_up_base kernel/user_namespace.c:335 [inline]
BUG: KMSAN: uninit-value in map_id_up kernel/user_namespace.c:365 [inline]
BUG: KMSAN: uninit-value in from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
map_id_up_base kernel/user_namespace.c:335 [inline]
map_id_up kernel/user_namespace.c:365 [inline]
from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
p9pdu_vwritef+0x15ab/0x5120 net/9p/protocol.c:398
p9pdu_writef+0x23a/0x280 net/9p/protocol.c:539
p9pdu_vwritef+0x21f0/0x5120 net/9p/protocol.c:490
p9_client_prepare_req+0xa4b/0xff0 net/9p/client.c:709
p9_client_rpc+0x278/0x1410 net/9p/client.c:740
p9_client_setattr+0x115/0x2c0 net/9p/client.c:1899
v9fs_vfs_setattr_dotl+0x7e2/0xd70 fs/9p/vfs_inode_dotl.c:590
notify_change+0x1fe3/0x2170 fs/attr.c:410
vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
do_utimes_path fs/utimes.c:98 [inline]
do_utimes fs/utimes.c:144 [inline]
__do_sys_utime32 fs/utimes.c:247 [inline]
__se_sys_utime32+0x386/0x520 fs/utimes.c:235
__ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
__do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
entry_SYSENTER_compat_after_hwframe+0x4d/0x5c

Uninit was stored to memory at:
v9fs_vfs_setattr_dotl+0x58a/0xd70 fs/9p/vfs_inode_dotl.c:567
notify_change+0x1fe3/0x2170 fs/attr.c:410
vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
do_utimes_path fs/utimes.c:98 [inline]
do_utimes fs/utimes.c:144 [inline]
__do_sys_utime32 fs/utimes.c:247 [inline]
__se_sys_utime32+0x386/0x520 fs/utimes.c:235
__ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
__do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
entry_SYSENTER_compat_after_hwframe+0x4d/0x5c

Local variable newattrs created at:
vfs_utimes+0x69/0xc70 fs/utimes.c:22
do_utimes_path fs/utimes.c:98 [inline]
do_utimes fs/utimes.c:144 [inline]
__do_sys_utime32 fs/utimes.c:247 [inline]
__se_sys_utime32+0x386/0x520 fs/utimes.c:235
=====================================================


---
This report 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 issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
Re: [syzbot] KMSAN: uninit-value in from_kuid [ In reply to ]
On Sat, Nov 27, 2021 at 07:50:27AM -0800, syzbot wrote:
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit: 425295055ce6 kmsan: core: address comments to kmsan-checks.h
> git tree: https://github.com/google/kmsan.git master
> console output: https://syzkaller.appspot.com/x/log.txt?x=1640209ab00000
> kernel config: https://syzkaller.appspot.com/x/.config?x=2d142cdf4204061
> dashboard link: https://syzkaller.appspot.com/bug?extid=dfac92a50024b54acaa4
> compiler: clang version 14.0.0 (git@github.com:llvm/llvm-project.git 0996585c8e3b3d409494eb5f1cad714b9e1f7fb5), GNU ld (GNU Binutils for Debian) 2.35.2
> userspace arch: i386
>
> Unfortunately, I don't have any reproducer for this issue yet.
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+dfac92a50024b54acaa4@syzkaller.appspotmail.com
>
> =====================================================
> BUG: KMSAN: uninit-value in map_id_up_base kernel/user_namespace.c:335 [inline]
> BUG: KMSAN: uninit-value in map_id_up kernel/user_namespace.c:365 [inline]
> BUG: KMSAN: uninit-value in from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> map_id_up_base kernel/user_namespace.c:335 [inline]
> map_id_up kernel/user_namespace.c:365 [inline]
> from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> p9pdu_vwritef+0x15ab/0x5120 net/9p/protocol.c:398
> p9pdu_writef+0x23a/0x280 net/9p/protocol.c:539
> p9pdu_vwritef+0x21f0/0x5120 net/9p/protocol.c:490
> p9_client_prepare_req+0xa4b/0xff0 net/9p/client.c:709
> p9_client_rpc+0x278/0x1410 net/9p/client.c:740
> p9_client_setattr+0x115/0x2c0 net/9p/client.c:1899
> v9fs_vfs_setattr_dotl+0x7e2/0xd70 fs/9p/vfs_inode_dotl.c:590
> notify_change+0x1fe3/0x2170 fs/attr.c:410
> vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
> do_utimes_path fs/utimes.c:98 [inline]
> do_utimes fs/utimes.c:144 [inline]
> __do_sys_utime32 fs/utimes.c:247 [inline]
> __se_sys_utime32+0x386/0x520 fs/utimes.c:235
> __ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
> do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
> __do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
> do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
> do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
> entry_SYSENTER_compat_after_hwframe+0x4d/0x5c
>
> Uninit was stored to memory at:
> v9fs_vfs_setattr_dotl+0x58a/0xd70 fs/9p/vfs_inode_dotl.c:567

That's a bug in the 9P2000.L implementation of .setattr.
It copies struct iattr values without checking ia_valid. That's causing
uninitalized memory to be copied. I sent a fix to 9p for this.

Christian
Re: [syzbot] KMSAN: uninit-value in from_kuid [ In reply to ]
On Mon, Nov 29, 2021 at 12:47 PM Christian Brauner
<christian.brauner@ubuntu.com> wrote:
>
> On Sat, Nov 27, 2021 at 07:50:27AM -0800, syzbot wrote:
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit: 425295055ce6 kmsan: core: address comments to kmsan-checks.h
> > git tree: https://github.com/google/kmsan.git master
> > console output: https://syzkaller.appspot.com/x/log.txt?x=1640209ab00000
> > kernel config: https://syzkaller.appspot.com/x/.config?x=2d142cdf4204061
> > dashboard link: https://syzkaller.appspot.com/bug?extid=dfac92a50024b54acaa4
> > compiler: clang version 14.0.0 (git@github.com:llvm/llvm-project.git 0996585c8e3b3d409494eb5f1cad714b9e1f7fb5), GNU ld (GNU Binutils for Debian) 2.35.2
> > userspace arch: i386
> >
> > Unfortunately, I don't have any reproducer for this issue yet.
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+dfac92a50024b54acaa4@syzkaller.appspotmail.com
> >
> > =====================================================
> > BUG: KMSAN: uninit-value in map_id_up_base kernel/user_namespace.c:335 [inline]
> > BUG: KMSAN: uninit-value in map_id_up kernel/user_namespace.c:365 [inline]
> > BUG: KMSAN: uninit-value in from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > map_id_up_base kernel/user_namespace.c:335 [inline]
> > map_id_up kernel/user_namespace.c:365 [inline]
> > from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > p9pdu_vwritef+0x15ab/0x5120 net/9p/protocol.c:398
> > p9pdu_writef+0x23a/0x280 net/9p/protocol.c:539
> > p9pdu_vwritef+0x21f0/0x5120 net/9p/protocol.c:490
> > p9_client_prepare_req+0xa4b/0xff0 net/9p/client.c:709
> > p9_client_rpc+0x278/0x1410 net/9p/client.c:740
> > p9_client_setattr+0x115/0x2c0 net/9p/client.c:1899
> > v9fs_vfs_setattr_dotl+0x7e2/0xd70 fs/9p/vfs_inode_dotl.c:590
> > notify_change+0x1fe3/0x2170 fs/attr.c:410
> > vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
> > do_utimes_path fs/utimes.c:98 [inline]
> > do_utimes fs/utimes.c:144 [inline]
> > __do_sys_utime32 fs/utimes.c:247 [inline]
> > __se_sys_utime32+0x386/0x520 fs/utimes.c:235
> > __ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
> > do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
> > __do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
> > do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
> > do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
> > entry_SYSENTER_compat_after_hwframe+0x4d/0x5c
> >
> > Uninit was stored to memory at:
> > v9fs_vfs_setattr_dotl+0x58a/0xd70 fs/9p/vfs_inode_dotl.c:567
>
> That's a bug in the 9P2000.L implementation of .setattr.
> It copies struct iattr values without checking ia_valid. That's causing
> uninitalized memory to be copied. I sent a fix to 9p for this.
>
> Christian

Thanks!


--
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Paul Manicle, Halimah DeLaine Prado
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Re: [syzbot] KMSAN: uninit-value in from_kuid [ In reply to ]
On Mon, Nov 29, 2021 at 12:47 PM Christian Brauner
<christian.brauner@ubuntu.com> wrote:
>
> On Sat, Nov 27, 2021 at 07:50:27AM -0800, syzbot wrote:
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit: 425295055ce6 kmsan: core: address comments to kmsan-checks.h
> > git tree: https://github.com/google/kmsan.git master
> > console output: https://syzkaller.appspot.com/x/log.txt?x=1640209ab00000
> > kernel config: https://syzkaller.appspot.com/x/.config?x=2d142cdf4204061
> > dashboard link: https://syzkaller.appspot.com/bug?extid=dfac92a50024b54acaa4
> > compiler: clang version 14.0.0 (git@github.com:llvm/llvm-project.git 0996585c8e3b3d409494eb5f1cad714b9e1f7fb5), GNU ld (GNU Binutils for Debian) 2.35.2
> > userspace arch: i386
> >
> > Unfortunately, I don't have any reproducer for this issue yet.
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+dfac92a50024b54acaa4@syzkaller.appspotmail.com
> >
> > =====================================================
> > BUG: KMSAN: uninit-value in map_id_up_base kernel/user_namespace.c:335 [inline]
> > BUG: KMSAN: uninit-value in map_id_up kernel/user_namespace.c:365 [inline]
> > BUG: KMSAN: uninit-value in from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > map_id_up_base kernel/user_namespace.c:335 [inline]
> > map_id_up kernel/user_namespace.c:365 [inline]
> > from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > p9pdu_vwritef+0x15ab/0x5120 net/9p/protocol.c:398
> > p9pdu_writef+0x23a/0x280 net/9p/protocol.c:539
> > p9pdu_vwritef+0x21f0/0x5120 net/9p/protocol.c:490
> > p9_client_prepare_req+0xa4b/0xff0 net/9p/client.c:709
> > p9_client_rpc+0x278/0x1410 net/9p/client.c:740
> > p9_client_setattr+0x115/0x2c0 net/9p/client.c:1899
> > v9fs_vfs_setattr_dotl+0x7e2/0xd70 fs/9p/vfs_inode_dotl.c:590
> > notify_change+0x1fe3/0x2170 fs/attr.c:410
> > vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
> > do_utimes_path fs/utimes.c:98 [inline]
> > do_utimes fs/utimes.c:144 [inline]
> > __do_sys_utime32 fs/utimes.c:247 [inline]
> > __se_sys_utime32+0x386/0x520 fs/utimes.c:235
> > __ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
> > do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
> > __do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
> > do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
> > do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
> > entry_SYSENTER_compat_after_hwframe+0x4d/0x5c
> >
> > Uninit was stored to memory at:
> > v9fs_vfs_setattr_dotl+0x58a/0xd70 fs/9p/vfs_inode_dotl.c:567
>
> That's a bug in the 9P2000.L implementation of .setattr.
> It copies struct iattr values without checking ia_valid. That's causing
> uninitalized memory to be copied. I sent a fix to 9p for this.
>
> Christian

Christian,

Do you think it makes sense to request a CVE for this issue?
If so, were you going to request one? Otherwise I can do that.

--
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Paul Manicle, Halimah DeLaine Prado
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Re: [syzbot] KMSAN: uninit-value in from_kuid [ In reply to ]
On Mon, Nov 29, 2021 at 04:16:21PM +0100, Alexander Potapenko wrote:
> On Mon, Nov 29, 2021 at 12:47 PM Christian Brauner
> <christian.brauner@ubuntu.com> wrote:
> >
> > On Sat, Nov 27, 2021 at 07:50:27AM -0800, syzbot wrote:
> > > Hello,
> > >
> > > syzbot found the following issue on:
> > >
> > > HEAD commit: 425295055ce6 kmsan: core: address comments to kmsan-checks.h
> > > git tree: https://github.com/google/kmsan.git master
> > > console output: https://syzkaller.appspot.com/x/log.txt?x=1640209ab00000
> > > kernel config: https://syzkaller.appspot.com/x/.config?x=2d142cdf4204061
> > > dashboard link: https://syzkaller.appspot.com/bug?extid=dfac92a50024b54acaa4
> > > compiler: clang version 14.0.0 (git@github.com:llvm/llvm-project.git 0996585c8e3b3d409494eb5f1cad714b9e1f7fb5), GNU ld (GNU Binutils for Debian) 2.35.2
> > > userspace arch: i386
> > >
> > > Unfortunately, I don't have any reproducer for this issue yet.
> > >
> > > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > > Reported-by: syzbot+dfac92a50024b54acaa4@syzkaller.appspotmail.com
> > >
> > > =====================================================
> > > BUG: KMSAN: uninit-value in map_id_up_base kernel/user_namespace.c:335 [inline]
> > > BUG: KMSAN: uninit-value in map_id_up kernel/user_namespace.c:365 [inline]
> > > BUG: KMSAN: uninit-value in from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > > map_id_up_base kernel/user_namespace.c:335 [inline]
> > > map_id_up kernel/user_namespace.c:365 [inline]
> > > from_kuid+0x51d/0xbd0 kernel/user_namespace.c:413
> > > p9pdu_vwritef+0x15ab/0x5120 net/9p/protocol.c:398
> > > p9pdu_writef+0x23a/0x280 net/9p/protocol.c:539
> > > p9pdu_vwritef+0x21f0/0x5120 net/9p/protocol.c:490
> > > p9_client_prepare_req+0xa4b/0xff0 net/9p/client.c:709
> > > p9_client_rpc+0x278/0x1410 net/9p/client.c:740
> > > p9_client_setattr+0x115/0x2c0 net/9p/client.c:1899
> > > v9fs_vfs_setattr_dotl+0x7e2/0xd70 fs/9p/vfs_inode_dotl.c:590
> > > notify_change+0x1fe3/0x2170 fs/attr.c:410
> > > vfs_utimes+0x8aa/0xc70 fs/utimes.c:65
> > > do_utimes_path fs/utimes.c:98 [inline]
> > > do_utimes fs/utimes.c:144 [inline]
> > > __do_sys_utime32 fs/utimes.c:247 [inline]
> > > __se_sys_utime32+0x386/0x520 fs/utimes.c:235
> > > __ia32_sys_utime32+0x91/0xc0 fs/utimes.c:235
> > > do_syscall_32_irqs_on arch/x86/entry/common.c:114 [inline]
> > > __do_fast_syscall_32+0x96/0xf0 arch/x86/entry/common.c:180
> > > do_fast_syscall_32+0x34/0x70 arch/x86/entry/common.c:205
> > > do_SYSENTER_32+0x1b/0x20 arch/x86/entry/common.c:248
> > > entry_SYSENTER_compat_after_hwframe+0x4d/0x5c
> > >
> > > Uninit was stored to memory at:
> > > v9fs_vfs_setattr_dotl+0x58a/0xd70 fs/9p/vfs_inode_dotl.c:567
> >
> > That's a bug in the 9P2000.L implementation of .setattr.
> > It copies struct iattr values without checking ia_valid. That's causing
> > uninitalized memory to be copied. I sent a fix to 9p for this.
> >
> > Christian
>
> Christian,
>
> Do you think it makes sense to request a CVE for this issue?
> If so, were you going to request one? Otherwise I can do that.

I mean, it's neither my bug nor did I detect it, I just fixed it. :)
If you would like this to be a CVE then sure go ahead.

(I don't understand the rules around this well enough tbh. For example,
during the last merge window there were at least 3 or 4 NULL pointer
derefs or UAFs in newly added but already released code. Should all of
these get a CVE without a working exploit?)