Re: [PATCH] fs/seq_file: fix out-of-bounds read
From: Vegard Nossum
Date: Tue Aug 23 2016 - 12:50:42 EST
On 08/17/2016 05:21 PM, Vegard Nossum wrote:
seq_read() is a nasty piece of work, not to mention buggy.
I was getting these:
BUG: KASAN: slab-out-of-bounds in seq_read+0xcd2/0x1480 at addr ffff880116889880
Read of size 2713 by task trinity-c2/1329
CPU: 2 PID: 1329 Comm: trinity-c2 Not tainted 4.8.0-rc1+ #96
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.9.3-0-ge2fc41e-prebuilt.qemu-project.org 04/01/2014
0000000000000000 ffff880110ebf8c0 ffffffff82344770 0000000041b58ab3
ffffffff84f97be8 ffffffff823446c4 0000000000000001 ffffffff00000028
ffff880110ebf8f8 ffff880110ebf8a8 ffff880110ebf998 0000000025843ea1
Call Trace:
[<ffffffff82344770>] dump_stack+0xac/0xfc
[<ffffffff823446c4>] ? _atomic_dec_and_lock+0xc4/0xc4
[<ffffffff817964dc>] kasan_object_err+0x1c/0x80
[<ffffffff8179687b>] kasan_report_error+0x2cb/0x7e0
[<ffffffff81796e3e>] kasan_report+0x4e/0x80
[<ffffffff81410e00>] ? down_trylock+0x40/0x90
[<ffffffff81886522>] ? seq_read+0xcd2/0x1480
[<ffffffff817952be>] check_memory_region+0x13e/0x1a0
[<ffffffff817954e1>] kasan_check_read+0x11/0x20
[<ffffffff81886522>] seq_read+0xcd2/0x1480
[<ffffffff81410e15>] ? down_trylock+0x55/0x90
[<ffffffff81435f51>] ? vprintk_emit+0x151/0x740
[<ffffffff81435fd4>] ? vprintk_emit+0x1d4/0x740
[<ffffffff81435172>] ? console_trylock+0x12/0x100
[<ffffffff845f7d4c>] ? _raw_spin_unlock+0x2c/0x50
[<ffffffff81885850>] ? seq_open+0x220/0x220
[<ffffffff8199ccc0>] ? proc_reg_write+0x260/0x260
[<ffffffff814368fa>] ? vprintk_default+0x1a/0x20
[<ffffffff81631e8d>] ? printk+0x94/0xb0
[<ffffffff81631df9>] ? cpumask_weight.constprop.26+0x34/0x34
[<ffffffff8199cdcb>] proc_reg_read+0x10b/0x260
[<ffffffff8199ccc0>] ? proc_reg_write+0x260/0x260
[<ffffffff817fca00>] do_loop_readv_writev.part.5+0x140/0x2c0
[<ffffffff817ffac9>] do_readv_writev+0x589/0x860
[<ffffffff817ff540>] ? rw_verify_area+0x370/0x370
[<ffffffff845eebef>] ? mutex_lock_nested+0x49f/0x8e0
[<ffffffff81871624>] ? __fdget_pos+0x94/0xe0
[<ffffffff845ee750>] ? ww_mutex_unlock+0x390/0x390
[<ffffffff8180480b>] vfs_readv+0x7b/0xd0
[<ffffffff81871624>] ? __fdget_pos+0x94/0xe0
[<ffffffff81804938>] do_readv+0xd8/0x2c0
[<ffffffff81804860>] ? vfs_readv+0xd0/0xd0
[<ffffffff823c7143>] ? __this_cpu_preempt_check+0x13/0x20
[<ffffffff81805220>] ? do_pwritev+0x1b0/0x1b0
[<ffffffff8180522b>] SyS_readv+0xb/0x10
[<ffffffff81007bd3>] do_syscall_64+0x1b3/0x4b0
[<ffffffff845f85aa>] entry_SYSCALL64_slow_path+0x25/0x25
Object at ffff880116889100, in cache kmalloc-4096 size: 4096
Allocated:
PID = 1329
[<ffffffff81235f26>] save_stack_trace+0x26/0x80
[<ffffffff81795366>] save_stack+0x46/0xd0
[<ffffffff81795b1d>] kasan_kmalloc+0xad/0xe0
[<ffffffff8179094a>] __kmalloc+0x1aa/0x4a0
[<ffffffff81884bc5>] seq_buf_alloc+0x35/0x40
[<ffffffff81886028>] seq_read+0x7d8/0x1480
[<ffffffff8199cdcb>] proc_reg_read+0x10b/0x260
[<ffffffff817fca00>] do_loop_readv_writev.part.5+0x140/0x2c0
[<ffffffff817ffac9>] do_readv_writev+0x589/0x860
[<ffffffff8180480b>] vfs_readv+0x7b/0xd0
[<ffffffff81804938>] do_readv+0xd8/0x2c0
[<ffffffff8180522b>] SyS_readv+0xb/0x10
[<ffffffff81007bd3>] do_syscall_64+0x1b3/0x4b0
[<ffffffff845f85aa>] return_from_SYSCALL_64+0x0/0x6a
Freed:
PID = 0
(stack is not available)
Memory state around the buggy address:
ffff88011688a000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ffff88011688a080: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff88011688a100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff88011688a180: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff88011688a200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Disabling lock debugging due to kernel taint
This seems to be the same thing that Dave Jones was seeing here:
https://lkml.org/lkml/2016/8/12/334
There are multiple issues here:
1) If we enter the function with a non-empty buffer, there is an attempt
to flush it. But it was not clearing m->from after doing so, which
means that if we try to do this flush twice in a row without any call
to traverse() in between, we are going to be reading from the wrong
place -- the splat above, fixed by this patch.
2) If there's a short write to userspace because of page faults, the
buffer may already contain multiple lines (i.e. pos has advanced by
more than 1), but we don't save the progress that was made so the
next call will output what we've already returned previously. Since
that is a much less serious issue (and I have a headache after
staring at seq_read() for the past 8 hours), I'll leave that for now.
Reported-by: Dave Jones <davej@xxxxxxxxxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Vegard Nossum <vegard.nossum@xxxxxxxxxx>
---
fs/seq_file.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 19f532e..6dc4296 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -223,8 +223,10 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
size -= n;
buf += n;
copied += n;
- if (!m->count)
+ if (!m->count) {
+ m->from = 0;
m->index++;
+ }
if (!size)
goto Done;
}
Al, ping? I assumed this would go through you but I didn't hear anything
back so just checking. Looks like akpm has been handling the more recent
patches in this area, however. Added some other Ccs in hope of a
review/ack too.
This is (I think) an old bug and it allows unprivileged userspace to
read beyond the end of m->buf.
Vegard