The patch titled
Subject: mm: slub: fix slab walking for init_on_free
has been added to the -mm tree. Its filename is
mm-slub-fix-slab-walking-for-init_on_free.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/mm-slub-fix-slab-walking-for-init_on_free.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/mm-slub-fix-slab-walking-for-init_on_free.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Laura Abbott <labbott@xxxxxxxxxx>
Subject: mm: slub: fix slab walking for init_on_free
To properly clear the slab on free with slab_want_init_on_free, we walk
the list of free objects using get_freepointer/set_freepointer. The value
we get from get_freepointer may not be valid. This isn't an issue since
an actual value will get written later but this means there's a chance of
triggering a bug if we use this value with set_freepointer:
[ 4.478342] kernel BUG at mm/slub.c:306!
[ 4.482437] invalid opcode: 0000 [#1] PREEMPT PTI
[ 4.485750] CPU: 0 PID: 0 Comm: swapper Not tainted 5.2.0-05754-g6471384a #4
[ 4.490635] RIP: 0010:kfree+0x58a/0x5c0
[ 4.493679] Code: 48 83 05 78 37 51 02 01 0f 0b 48 83 05 7e 37 51 02 01 48 83 05 7e 37 51 02 01 48 83 05 7e 37 51 02 01 48 83 05 d6 37 51 02 01 <0f> 0b 48 83 05 d4 37 51 02 01 48 83 05 d4 37 51 02 01 48 83 05 d4
[ 4.506827] RSP: 0000:ffffffff82603d90 EFLAGS: 00010002
[ 4.510475] RAX: ffff8c3976c04320 RBX: ffff8c3976c04300 RCX: 0000000000000000
[ 4.515420] RDX: ffff8c3976c04300 RSI: 0000000000000000 RDI: ffff8c3976c04320
[ 4.520331] RBP: ffffffff82603db8 R08: 0000000000000000 R09: 0000000000000000
[ 4.525288] R10: ffff8c3976c04320 R11: ffffffff8289e1e0 R12: ffffd52cc8db0100
[ 4.530180] R13: ffff8c3976c01a00 R14: ffffffff810f10d4 R15: ffff8c3976c04300
[ 4.535079] FS: 0000000000000000(0000) GS:ffffffff8266b000(0000) knlGS:0000000000000000
[ 4.540628] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 4.544593] CR2: ffff8c397ffff000 CR3: 0000000125020000 CR4: 00000000000406b0
[ 4.549558] Call Trace:
[ 4.551266] apply_wqattrs_prepare+0x154/0x280
[ 4.554357] apply_workqueue_attrs_locked+0x4e/0xe0
[ 4.557728] apply_workqueue_attrs+0x36/0x60
[ 4.560654] alloc_workqueue+0x25a/0x6d0
[ 4.563381] ? kmem_cache_alloc_trace+0x1e3/0x500
[ 4.566628] ? __mutex_unlock_slowpath+0x44/0x3f0
[ 4.569875] workqueue_init_early+0x246/0x348
[ 4.573025] start_kernel+0x3c7/0x7ec
[ 4.575558] x86_64_start_reservations+0x40/0x49
[ 4.578738] x86_64_start_kernel+0xda/0xe4
[ 4.581600] secondary_startup_64+0xb6/0xc0
[ 4.584473] Modules linked in:
[ 4.586620] ---[ end trace f67eb9af4d8d492b ]---
Fix this by ensuring the value we set with set_freepointer is either NULL
or another value in the chain.
Link: http://lkml.kernel.org/r/20190731193240.29477-1-labbott@xxxxxxxxxx
Fixes: 6471384af2a6 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
Signed-off-by: Laura Abbott <labbott@xxxxxxxxxx>
Reported-by: kernel test robot <rong.a.chen@xxxxxxxxx>
Reviewed-by: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Alexander Potapenko <glider@xxxxxxxxxx>
Cc: Laura Abbott <labbott@xxxxxxxxxx>
Cc: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
Cc: "Serge E. Hallyn" <serge@xxxxxxxxxx>
Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
Cc: Kostya Serebryany <kcc@xxxxxxxxxx>
Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Cc: Sandeep Patil <sspatil@xxxxxxxxxxx>
Cc: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
Cc: Jann Horn <jannh@xxxxxxxxxx>
Cc: Mark Rutland <mark.rutland@xxxxxxx>
Cc: Marco Elver <elver@xxxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxx>
Cc: David Rientjes <rientjes@xxxxxxxxxx>
Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/slub.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
--- a/mm/slub.c~mm-slub-fix-slab-walking-for-init_on_free
+++ a/mm/slub.c
@@ -1432,7 +1432,9 @@ static inline bool slab_free_freelist_ho
void *old_tail = *tail ? *tail : *head;
int rsize;
- if (slab_want_init_on_free(s))
+ if (slab_want_init_on_free(s)) {
+ void *p = NULL;
+
do {
object = next;
next = get_freepointer(s, object);
@@ -1445,8 +1447,10 @@ static inline bool slab_free_freelist_ho
: 0;
memset((char *)object + s->inuse, 0,
s->size - s->inuse - rsize);
- set_freepointer(s, object, next);
+ set_freepointer(s, object, p);
+ p = object;
} while (object != old_tail);
+ }
/*
* Compiler cannot detect this function can be removed if slab_free_hook()
_
Patches currently in -mm which might be from labbott@xxxxxxxxxx are
mm-slub-fix-slab-walking-for-init_on_free.patch
