mm: fault in __do_fault

81 views
Skip to first unread message

Dmitry Vyukov

unread,
Feb 28, 2017, 9:05:14 AM2/28/17
to Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.z...@linux.intel.com, Michal Hocko, linu...@kvack.org, LKML, Andrea Arcangeli, Mike Kravetz, syzkaller
Hello,

The following program triggers GPF in __do_fault:
https://gist.githubusercontent.com/dvyukov/27345737fca18d92ef761e7fa08aec9b/raw/d99d02511d0bf9a8d6f6bd9c79d373a26924e974/gistfile1.txt

general protection fault: 0000 [#1] SMP KASAN
Modules linked in:
CPU: 3 PID: 2955 Comm: a.out Not tainted 4.10.0+ #230
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
task: ffff88006b4e2480 task.stack: ffff88006b0b0000
RIP: 0010:__read_once_size include/linux/compiler.h:247 [inline]
RIP: 0010:compound_head include/linux/page-flags.h:146 [inline]
RIP: 0010:trylock_page include/linux/pagemap.h:442 [inline]
RIP: 0010:lock_page include/linux/pagemap.h:452 [inline]
RIP: 0010:__do_fault+0x247/0x3e0 mm/memory.c:2898
RSP: 0000:ffff88006b0b7470 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 1ffff1000d616e91 RCX: 0000000000000000
RDX: 0000000000000004 RSI: dffffc0000000000 RDI: 0000000000000020
RBP: ffff88006b0b7550 R08: 0000000000000001 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000004 R12: ffff88006b0b74e8
R13: 0000000000000000 R14: ffff88006b0b7528 R15: 0000000000000000
FS: 00007fef8bb55700(0000) GS:ffff88006d180000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020012fe0 CR3: 000000006a8dc000 CR4: 00000000001406e0
Call Trace:
do_read_fault mm/memory.c:3266 [inline]
do_fault+0x5ee/0x2080 mm/memory.c:3366
handle_pte_fault mm/memory.c:3596 [inline]
__handle_mm_fault+0x1062/0x2cb0 mm/memory.c:3710
handle_mm_fault+0x1e2/0x480 mm/memory.c:3747
__do_page_fault+0x4f6/0xb60 arch/x86/mm/fault.c:1396
trace_do_page_fault+0x141/0x6c0 arch/x86/mm/fault.c:1489
do_async_page_fault+0x72/0xc0 arch/x86/kernel/kvm.c:264
async_page_fault+0x28/0x30 arch/x86/entry/entry_64.S:1014
RIP: 0033:0x43dda0
RSP: 002b:00007fef8bb54878 EFLAGS: 00010202
RAX: 00007fef8bb548b0 RBX: 0000000000000000 RCX: 000000000000000e
RDX: 0000000000000400 RSI: 0000000020012fe0 RDI: 00007fef8bb548b0
RBP: 00007fef8bb54cd0 R08: 0000000000000400 R09: 00007fef8bb54d10
R10: 00007fef8bb559d0 R11: 0000000000000202 R12: 0000000000000000
R13: 0000000000000000 R14: 00007fef8bb559c0 R15: 00007fef8bb55700
Code: 00 e8 0e 35 b8 ff e8 79 09 a2 02 4c 89 e2 49 8d 7f 20 48 b8 00
00 00 00 00 fc ff df 48 c1 ea 03 c6 04 02 00 48 89 fa 48 c1 ea 03 <80>
3c 02 00 0f 85 ec 00 00 00 4c 89 e2 48 b8 00 00 00 00 00 fc
RIP: __read_once_size include/linux/compiler.h:247 [inline] RSP:
ffff88006b0b7470
RIP: compound_head include/linux/page-flags.h:146 [inline] RSP: ffff88006b0b7470
RIP: trylock_page include/linux/pagemap.h:442 [inline] RSP: ffff88006b0b7470
RIP: lock_page include/linux/pagemap.h:452 [inline] RSP: ffff88006b0b7470
RIP: __do_fault+0x247/0x3e0 mm/memory.c:2898 RSP: ffff88006b0b7470
---[ end trace d24250eecdbb8a34 ]---

On commit e5d56efc97f8240d0b5d66c03949382b6d7e5570

Kirill A. Shutemov

unread,
Feb 28, 2017, 10:40:15 AM2/28/17
to Dmitry Vyukov, Andrea Arcangeli, Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.z...@linux.intel.com, Michal Hocko, linu...@kvack.org, LKML, Mike Kravetz, syzkaller
__do_fault() expects to see a page if ->fault() returns zero, but
handle_userfault() mess this up, making shmem_fault() return 0 without
setting vmf->page.

I'm not entirely sure it's fully correct, but the patch below fixes the
crash for me.

Andrea, does it look okay for you?

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index 625b7285a37b..56f61f1a1dc1 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -489,7 +489,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
* in such case.
*/
down_read(&mm->mmap_sem);
- ret = 0;
+ ret = VM_FAULT_NOPAGE;
}
}

--
Kirill A. Shutemov

Andrea Arcangeli

unread,
Feb 28, 2017, 10:42:04 AM2/28/17
to Dmitry Vyukov, Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.z...@linux.intel.com, Michal Hocko, linu...@kvack.org, LKML, Mike Kravetz, syzkaller
Hello Dmitry,

On Tue, Feb 28, 2017 at 03:04:53PM +0100, Dmitry Vyukov wrote:
Can you verify this fix:

From a65381bc86d2963713b6a9c4a73cded7dd184282 Mon Sep 17 00:00:00 2001
From: Andrea Arcangeli <aarc...@redhat.com>
Date: Tue, 28 Feb 2017 16:36:59 +0100
Subject: [PATCH 1/1] userfaultfd: shmem: __do_fault requires VM_FAULT_NOPAGE

__do_fault assumes vmf->page has been initialized and is valid if
VM_FAULT_NOPAGE is not returned by vma->vm_ops->fault(vma, vmf).

handle_userfault() in turn should return VM_FAULT_NOPAGE if it doesn't
return VM_FAULT_SIGBUS or VM_FAULT_RETRY (the other two
possibilities).

This VM_FAULT_NOPAGE case is only invoked when signal are pending and
it didn't matter for anonymous memory before. It only started to
matter since shmem was introduced. hugetlbfs also takes a different
path and doesn't exercise __do_fault.

Signed-off-by: Andrea Arcangeli <aarc...@redhat.com>
---
fs/userfaultfd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index fb6d02b..de28f43 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -500,7 +500,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)

Andrea Arcangeli

unread,
Feb 28, 2017, 10:55:52 AM2/28/17
to Kirill A. Shutemov, Dmitry Vyukov, Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.z...@linux.intel.com, Michal Hocko, linu...@kvack.org, LKML, Mike Kravetz, syzkaller
On Tue, Feb 28, 2017 at 06:32:20PM +0300, Kirill A. Shutemov wrote:
> Andrea, does it look okay for you?
>
> diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
> index 625b7285a37b..56f61f1a1dc1 100644
> --- a/fs/userfaultfd.c
> +++ b/fs/userfaultfd.c
> @@ -489,7 +489,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
> * in such case.
> */
> down_read(&mm->mmap_sem);
> - ret = 0;
> + ret = VM_FAULT_NOPAGE;
> }
> }

Yes, I did the same fix.

Dmitry Vyukov

unread,
Mar 3, 2017, 2:42:29 AM3/3/17
to Andrea Arcangeli, Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.z...@linux.intel.com, Michal Hocko, linu...@kvack.org, LKML, Mike Kravetz, syzkaller
On Tue, Feb 28, 2017 at 4:42 PM, Andrea Arcangeli <aarc...@redhat.com> wrote:
> Hello Dmitry,
>
> On Tue, Feb 28, 2017 at 03:04:53PM +0100, Dmitry Vyukov wrote:
>> Hello,
>>
>> The following program triggers GPF in __do_fault:
>> https://gist.githubusercontent.com/dvyukov/27345737fca18d92ef761e7fa08aec9b/raw/d99d02511d0bf9a8d6f6bd9c79d373a26924e974/gistfile1.txt
>
> Can you verify this fix:


Applied the patch on bots.
> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages