linux-kernel - Re: [PATCH] mm/mempolicy: Fix use-after-free of VMA iterator

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <yt9dildhke9l.fsf@linux.ibm.com>
Date:   Thu, 27 Apr 2023 08:52:54 +0200
From:   Sven Schnelle <svens@...ux.ibm.com>
To:     "Liam R. Howlett" <Liam.Howlett@...cle.com>
Cc:     linux-mm@...ck.org, linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        syzbot+a7c1ec5b1d71ceaa5186@...kaller.appspotmail.com,
        stable@...r.kernel.org, linux-s390@...r.kernel.org
Subject: Re: [PATCH] mm/mempolicy: Fix use-after-free of VMA iterator

"Liam R. Howlett" <Liam.Howlett@...cle.com> writes:

> set_mempolicy_home_node() iterates over a list of VMAs and calls
> mbind_range() on each VMA, which also iterates over the singular list of
> the VMA passed in and potentially splits the VMA.  Since the VMA
> iterator is not passed through, set_mempolicy_home_node() may now point
> to a stale node in the VMA tree.  This can result in a UAF as reported
> by syzbot.
>
> Avoid the stale maple tree node by passing the VMA iterator through to
> the underlying call to split_vma().
>
> mbind_range() is also overly complicated, since there are two calling
> functions and one already handles iterating over the VMAs.  Simplify
> mbind_range() to only handle merging and splitting of the VMAs.
>
> Align the new loop in do_mbind() and existing loop in
> set_mempolicy_home_node() to use the reduced mbind_range() function.
> This allows for a single location of the range calculation and avoids
> constantly looking up the previous VMA (since this is a loop over the
> VMAs).
>
> Link: https://lore.kernel.org/linux-mm/000000000000c93feb05f87e24ad@google.com/
> Reported-and-tested-by: syzbot+a7c1ec5b1d71ceaa5186@...kaller.appspotmail.com
> Fixes: 66850be55e8e ("mm/mempolicy: use vma iterator & maple state instead of vma linked list")
> Cc: <stable@...r.kernel.org>
> Signed-off-by: Liam R. Howlett <Liam.Howlett@...cle.com>
> ---

This breaks the vma02 testcase from ltp on s390:

 ~ # ./vma02
vma02       0  TINFO  :  pid = 617 addr = 0x3ff8f673000
vma02       0  TINFO  :  start = 0x3ff8f673000, end = 0x3ff8f674000
vma02       0  TINFO  :  start = 0x3ff8f674000, end = 0x3ff8f675000
vma02       0  TINFO  :  start = 0x3ff8f675000, end = 0x3ff8f676000
vma02       1  TFAIL  :  vma02.c:144: >1 unmerged VMAs.

When this happens the following VM_WARN_ON() is triggered:

[   25.628747] ------------[ cut here ]------------
[   25.628821] WARNING: CPU: 0 PID: 617 at mm/mmap.c:922 vma_merge+0x7ca/0x970
[   25.628834] Modules linked in:
[   25.628841] CPU: 0 PID: 617 Comm: vma02 Not tainted 6.3.0-09574-g285e1dccc348 #108
[   25.628846] Hardware name: IBM 3906 M04 704 (z/VM 7.1.0)
[   25.628850] Krnl PSW : 0704e00180000000 00000000004317b6 (vma_merge+0x7ce/0x970)
[   25.628859]            R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:2 PM:0 RI:0 EA:3
[   25.628866] Krnl GPRS: 000003ff8f62d000 000003ff8f674000 000000008e7cb908 000000000131f0f4
[   25.628871]            0000000000000001 0000038000bdba1c 000003ff8f675000 000000008e7cb908
[   25.628876]            000000008e7cb908 000003ff8f674000 000003ff8f675000 000000008e7ca308
[   25.628881]            0000000000000001 000000003ff8f674 0000000000431048 0000038000bdbb08
[   25.628891] Krnl Code: 00000000004317aa: a7f4fcea            brc     15,000000000043117e
[   25.628891]            00000000004317ae: b9040087            lgr     %r8,%r7
[   25.628891]           #00000000004317b2: af000000            mc      0,0
[   25.628891]           >00000000004317b6: a7f4fc5f            brc     15,0000000000431074
[   25.628891]            00000000004317ba: ec38fd30007c        cgij    %r3,0,8,000000000043121a
[   25.628891]            00000000004317c0: e31031d80004        lg      %r1,472(%r3)
[   25.628891]            00000000004317c6: e310f0f80024        stg     %r1,248(%r15)
[   25.628891]            00000000004317cc: a7f4fd27            brc     15,000000000043121a
[   25.628970] Call Trace:
[   25.628974]  [<00000000004317b6>] vma_merge+0x7ce/0x970
[   25.628979] ([<0000000000431048>] vma_merge+0x60/0x970)
[   25.628985]  [<0000000000482e5a>] mbind_range+0x13a/0x1c8
[   25.628990]  [<0000000000483448>] do_mbind+0x2c8/0x448
[   25.629013]  [<0000000000483662>] kernel_mbind+0x9a/0xb8
[   25.629018]  [<0000000000483724>] __s390x_sys_mbind+0x4c/0x58
[   25.629023]  [<0000000000f38c3a>] __do_syscall+0x1da/0x208
[   25.629030]  [<0000000000f4ed40>] system_call+0x70/0x98
[   25.629037] 1 lock held by vma02/617:
[   25.629042]  #0: 000000008c570ca8 (&mm->mmap_lock){++++}-{3:3}, at: do_mbind+0x102/0x448
[   25.629055] Last Breaking-Event-Address:
[   25.629058]  [<0000000000431070>] vma_merge+0x88/0x970
[   25.629066] irq event stamp: 12559
[   25.629070] hardirqs last  enabled at (12567): [<00000000001fcc6c>] __up_console_sem+0x8c/0xc0
[   25.629080] hardirqs last disabled at (12574): [<00000000001fcc4e>] __up_console_sem+0x6e/0xc0
[   25.629086] softirqs last  enabled at (10076): [<0000000000f516e2>] __do_softirq+0x512/0x618
[   25.629123] softirqs last disabled at (10053): [<0000000000162bbe>] __irq_exit_rcu+0x13e/0x170
[   25.629132] ---[ end trace 0000000000000000 ]---

which is:

       /* verify some invariant that must be enforced by the caller */
       VM_WARN_ON(prev && addr <= prev->vm_start);
-->    VM_WARN_ON(mid && end > mid->vm_end);
       VM_WARN_ON(addr >= end);

Any thoughts?

Thanks
Sven