* mm: fault in __do_fault
@ 2017-02-28 14:04 Dmitry Vyukov
2017-02-28 15:32 ` Kirill A. Shutemov
2017-02-28 15:42 ` Andrea Arcangeli
0 siblings, 2 replies; 5+ messages in thread
From: Dmitry Vyukov @ 2017-02-28 14:04 UTC (permalink / raw)
To: Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.zwisler,
Michal Hocko, linux-mm, LKML, Andrea Arcangeli, Mike Kravetz
Cc: 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
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: mm: fault in __do_fault
2017-02-28 14:04 mm: fault in __do_fault Dmitry Vyukov
@ 2017-02-28 15:32 ` Kirill A. Shutemov
2017-02-28 15:55 ` Andrea Arcangeli
2017-02-28 15:42 ` Andrea Arcangeli
1 sibling, 1 reply; 5+ messages in thread
From: Kirill A. Shutemov @ 2017-02-28 15:32 UTC (permalink / raw)
To: Dmitry Vyukov, Andrea Arcangeli
Cc: Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.zwisler,
Michal Hocko, linux-mm, LKML, Mike Kravetz, syzkaller
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
>
> 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
__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
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: mm: fault in __do_fault
2017-02-28 14:04 mm: fault in __do_fault Dmitry Vyukov
2017-02-28 15:32 ` Kirill A. Shutemov
@ 2017-02-28 15:42 ` Andrea Arcangeli
2017-03-03 7:42 ` Dmitry Vyukov
1 sibling, 1 reply; 5+ messages in thread
From: Andrea Arcangeli @ 2017-02-28 15:42 UTC (permalink / raw)
To: Dmitry Vyukov
Cc: Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.zwisler,
Michal Hocko, linux-mm, LKML, Mike Kravetz, syzkaller
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:
>From a65381bc86d2963713b6a9c4a73cded7dd184282 Mon Sep 17 00:00:00 2001
From: Andrea Arcangeli <aarcange@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 <aarcange@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)
* in such case.
*/
down_read(&mm->mmap_sem);
- ret = 0;
+ ret = VM_FAULT_NOPAGE;
}
}
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: mm: fault in __do_fault
2017-02-28 15:32 ` Kirill A. Shutemov
@ 2017-02-28 15:55 ` Andrea Arcangeli
0 siblings, 0 replies; 5+ messages in thread
From: Andrea Arcangeli @ 2017-02-28 15:55 UTC (permalink / raw)
To: Kirill A. Shutemov
Cc: Dmitry Vyukov, Andrew Morton, Kirill A. Shutemov, Jan Kara,
ross.zwisler, Michal Hocko, linux-mm, 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.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: mm: fault in __do_fault
2017-02-28 15:42 ` Andrea Arcangeli
@ 2017-03-03 7:42 ` Dmitry Vyukov
0 siblings, 0 replies; 5+ messages in thread
From: Dmitry Vyukov @ 2017-03-03 7:42 UTC (permalink / raw)
To: Andrea Arcangeli
Cc: Andrew Morton, Kirill A. Shutemov, Jan Kara, ross.zwisler,
Michal Hocko, linux-mm, LKML, Mike Kravetz, syzkaller
On Tue, Feb 28, 2017 at 4:42 PM, Andrea Arcangeli <aarcange@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.
> From a65381bc86d2963713b6a9c4a73cded7dd184282 Mon Sep 17 00:00:00 2001
> From: Andrea Arcangeli <aarcange@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 <aarcange@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)
> * in such case.
> */
> down_read(&mm->mmap_sem);
> - ret = 0;
> + ret = VM_FAULT_NOPAGE;
> }
> }
>
>
> --
> 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+unsubscribe@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2017-03-03 7:42 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-02-28 14:04 mm: fault in __do_fault Dmitry Vyukov
2017-02-28 15:32 ` Kirill A. Shutemov
2017-02-28 15:55 ` Andrea Arcangeli
2017-02-28 15:42 ` Andrea Arcangeli
2017-03-03 7:42 ` Dmitry Vyukov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).