diff mbox

[v8,2/6] arm64: ptrace: allow tracer to skip a system call

Message ID 546D7860.2010300@linaro.org
State New
Headers show

Commit Message

AKASHI Takahiro Nov. 20, 2014, 5:13 a.m. UTC 
On 11/20/2014 04:06 AM, Will Deacon wrote:
> On Wed, Nov 19, 2014 at 08:46:19AM +0000, AKASHI Takahiro wrote:
>> On 11/18/2014 11:04 PM, Will Deacon wrote:
>>> On Tue, Nov 18, 2014 at 01:10:34AM +0000, AKASHI Takahiro wrote:
>>>>
>>>> +	if (((int)regs->syscallno == -1) && (orig_syscallno == -1)) {
>>>> +		/*
>>>> +		 * user-issued syscall(-1):
>>>> +		 * RESTRICTION: We always return ENOSYS whatever value is
>>>> +		 *   stored in x0 (a return value) at this point.
>>>> +		 * Normally, with ptrace off, syscall(-1) returns -ENOSYS.
>>>> +		 * With ptrace on, however, if a tracer didn't pay any
>>>> +		 * attention to user-issued syscall(-1) and just let it go
>>>> +		 * without a hack here, it would return a value in x0 as in
>>>> +		 * other system call cases. This means that this system call
>>>> +		 * might succeed and see any bogus return value.
>>>> +		 * This should be definitely avoided.
>>>> +		 */
>>>> +		regs->regs[0] = -ENOSYS;
>>>> +	}
>>>
>>> I'm still really uncomfortable with this, and it doesn't seem to match what
>>> arch/arm/ does either.
>>
>> Yeah, I know but
>> as I mentioned before, syscall(-1) will be signaled on arm, and so we don't
>> have to care about a return value :)
>
> What does x86 do?

On x86, syscall(-1) returns -ENOSYS if not traced, and we can change a return
value if traced.

>>> Doesn't it also prevent a tracer from skipping syscall(-1)?
>>
>> Syscall(-1) will return -ENOSYS whether or not a syscallno is explicitly
>> replaced with -1 by a tracer, and, in this sense, it is *skipped*.
>
> Ok, but now userspace sees -ENOSYS for a skipped system call in that case,
> whereas it would usually see whatever the trace put in x0, right?

Yes.
If you don't really like this behavior, how about this patch instead of my [2/6] patch?


With this change, I believe, syscall(-1) returns -ENOSYS by default whether traced
or not, and still you can change a return value when tracing.
(But a drawback here is that a tracer will see -ENOSYS in x0 even at syscall entry
for syscall(-1).)


-Takahiro AKASHI



> Will
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Comments

AKASHI Takahiro Nov. 20, 2014, 5:52 a.m. UTC | #1 
On 11/20/2014 02:13 PM, AKASHI Takahiro wrote:
> On 11/20/2014 04:06 AM, Will Deacon wrote:
>> On Wed, Nov 19, 2014 at 08:46:19AM +0000, AKASHI Takahiro wrote:
>>> On 11/18/2014 11:04 PM, Will Deacon wrote:
>>>> On Tue, Nov 18, 2014 at 01:10:34AM +0000, AKASHI Takahiro wrote:
>>>>>
>>>>> +    if (((int)regs->syscallno == -1) && (orig_syscallno == -1)) {
>>>>> +        /*
>>>>> +         * user-issued syscall(-1):
>>>>> +         * RESTRICTION: We always return ENOSYS whatever value is
>>>>> +         *   stored in x0 (a return value) at this point.
>>>>> +         * Normally, with ptrace off, syscall(-1) returns -ENOSYS.
>>>>> +         * With ptrace on, however, if a tracer didn't pay any
>>>>> +         * attention to user-issued syscall(-1) and just let it go
>>>>> +         * without a hack here, it would return a value in x0 as in
>>>>> +         * other system call cases. This means that this system call
>>>>> +         * might succeed and see any bogus return value.
>>>>> +         * This should be definitely avoided.
>>>>> +         */
>>>>> +        regs->regs[0] = -ENOSYS;
>>>>> +    }
>>>>
>>>> I'm still really uncomfortable with this, and it doesn't seem to match what
>>>> arch/arm/ does either.
>>>
>>> Yeah, I know but
>>> as I mentioned before, syscall(-1) will be signaled on arm, and so we don't
>>> have to care about a return value :)
>>
>> What does x86 do?
>
> On x86, syscall(-1) returns -ENOSYS if not traced, and we can change a return
> value if traced.
>
>>>> Doesn't it also prevent a tracer from skipping syscall(-1)?
>>>
>>> Syscall(-1) will return -ENOSYS whether or not a syscallno is explicitly
>>> replaced with -1 by a tracer, and, in this sense, it is *skipped*.
>>
>> Ok, but now userspace sees -ENOSYS for a skipped system call in that case,
>> whereas it would usually see whatever the trace put in x0, right?
>
> Yes.
> If you don't really like this behavior, how about this patch instead of my [2/6] patch?
>
> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
> index 726b910..1ef57d0 100644
> --- a/arch/arm64/kernel/entry.S
> +++ b/arch/arm64/kernel/entry.S
> @@ -668,8 +668,15 @@ ENDPROC(el0_svc)
>           * switches, and waiting for our parent to respond.
>           */
>   __sys_trace:
> +       cmp     w8, #-1                         // default errno for invalid

I needed to correct the code here:
w8 should be w26, thinking of compat syscalls.

> +       b.ne    1f                              // system call
> +       mov     x0, #-ENOSYS
> +       str     x0, [sp, #S_X0]
> +1:

and this part might better be generalized like the following:

__sys_trace:
	cmp	w26, w25	// cannot use x26 and x25 here
	b.hs	1f		// scno > sc_nr || scno < 0
	b	2f
1:
	mov	x0, #-ENOSYS
	str	x0, [sp, #S_X0]
2:

If you will be comfortable, I will submit a new patch soon.

-Takahiro AKASHI


>          mov     x0, sp
>          bl      syscall_trace_enter
> +       cmp     w0, #-1                         // skip the syscall?
> +       b.eq    __sys_trace_return_skipped
>          adr     lr, __sys_trace_return          // return address
>          uxtw    scno, w0                        // syscall number (possibly new)
>          mov     x1, sp                          // pointer to regs
> @@ -684,6 +691,7 @@ __sys_trace:
>
>   __sys_trace_return:
>          str     x0, [sp]                        // save returned x0
> +__sys_trace_return_skipped:
>          mov     x0, sp
>          bl      syscall_trace_exit
>          b       ret_to_user
>
> With this change, I believe, syscall(-1) returns -ENOSYS by default whether traced
> or not, and still you can change a return value when tracing.
> (But a drawback here is that a tracer will see -ENOSYS in x0 even at syscall entry
> for syscall(-1).)
>
>
> -Takahiro AKASHI
>
>
>
>> Will
>>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
Will Deacon Nov. 20, 2014, 7:17 p.m. UTC | #2 
On Thu, Nov 20, 2014 at 05:13:04AM +0000, AKASHI Takahiro wrote:
> On 11/20/2014 04:06 AM, Will Deacon wrote:
> > On Wed, Nov 19, 2014 at 08:46:19AM +0000, AKASHI Takahiro wrote:
> >> Syscall(-1) will return -ENOSYS whether or not a syscallno is explicitly
> >> replaced with -1 by a tracer, and, in this sense, it is *skipped*.
> >
> > Ok, but now userspace sees -ENOSYS for a skipped system call in that case,
> > whereas it would usually see whatever the trace put in x0, right?
> 
> If you don't really like this behavior, how about this patch instead of my [2/6] patch?
> 
> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
> index 726b910..1ef57d0 100644
> --- a/arch/arm64/kernel/entry.S
> +++ b/arch/arm64/kernel/entry.S
> @@ -668,8 +668,15 @@ ENDPROC(el0_svc)
>           * switches, and waiting for our parent to respond.
>           */
>   __sys_trace:
> +       cmp     w8, #-1                         // default errno for invalid
> +       b.ne    1f                              // system call
> +       mov     x0, #-ENOSYS
> +       str     x0, [sp, #S_X0]
> +1:
>          mov     x0, sp
>          bl      syscall_trace_enter
> +       cmp     w0, #-1                         // skip the syscall?
> +       b.eq    __sys_trace_return_skipped
>          adr     lr, __sys_trace_return          // return address
>          uxtw    scno, w0                        // syscall number (possibly new)
>          mov     x1, sp                          // pointer to regs
> @@ -684,6 +691,7 @@ __sys_trace:
> 
>   __sys_trace_return:
>          str     x0, [sp]                        // save returned x0
> +__sys_trace_return_skipped:
>          mov     x0, sp
>          bl      syscall_trace_exit
>          b       ret_to_user
> 
> With this change, I believe, syscall(-1) returns -ENOSYS by default whether traced
> or not, and still you can change a return value when tracing.
> (But a drawback here is that a tracer will see -ENOSYS in x0 even at syscall entry
> for syscall(-1).)

But it's exactly these drawbacks that I'm objected to. syscall(-1) shouldn't
be treated any differently to syscall(42) with respect to restarting,
exactly like x86.

Will
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
AKASHI Takahiro Nov. 25, 2014, 7:42 a.m. UTC | #3 
On 11/21/2014 04:17 AM, Will Deacon wrote:
> On Thu, Nov 20, 2014 at 05:13:04AM +0000, AKASHI Takahiro wrote:
>> On 11/20/2014 04:06 AM, Will Deacon wrote:
>>> On Wed, Nov 19, 2014 at 08:46:19AM +0000, AKASHI Takahiro wrote:
>>>> Syscall(-1) will return -ENOSYS whether or not a syscallno is explicitly
>>>> replaced with -1 by a tracer, and, in this sense, it is *skipped*.
>>>
>>> Ok, but now userspace sees -ENOSYS for a skipped system call in that case,
>>> whereas it would usually see whatever the trace put in x0, right?
>>
>> If you don't really like this behavior, how about this patch instead of my [2/6] patch?
>>
>> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
>> index 726b910..1ef57d0 100644
>> --- a/arch/arm64/kernel/entry.S
>> +++ b/arch/arm64/kernel/entry.S
>> @@ -668,8 +668,15 @@ ENDPROC(el0_svc)
>>            * switches, and waiting for our parent to respond.
>>            */
>>    __sys_trace:
>> +       cmp     w8, #-1                         // default errno for invalid
>> +       b.ne    1f                              // system call
>> +       mov     x0, #-ENOSYS
>> +       str     x0, [sp, #S_X0]
>> +1:
>>           mov     x0, sp
>>           bl      syscall_trace_enter
>> +       cmp     w0, #-1                         // skip the syscall?
>> +       b.eq    __sys_trace_return_skipped
>>           adr     lr, __sys_trace_return          // return address
>>           uxtw    scno, w0                        // syscall number (possibly new)
>>           mov     x1, sp                          // pointer to regs
>> @@ -684,6 +691,7 @@ __sys_trace:
>>
>>    __sys_trace_return:
>>           str     x0, [sp]                        // save returned x0
>> +__sys_trace_return_skipped:
>>           mov     x0, sp
>>           bl      syscall_trace_exit
>>           b       ret_to_user
>>
>> With this change, I believe, syscall(-1) returns -ENOSYS by default whether traced
>> or not, and still you can change a return value when tracing.
>> (But a drawback here is that a tracer will see -ENOSYS in x0 even at syscall entry
>> for syscall(-1).)
>
> But it's exactly these drawbacks that I'm objected to. syscall(-1) shouldn't
> be treated any differently to syscall(42) with respect to restarting,
> exactly like x86.

Can you elaborate a bit more as to "restarting?"
We can't make any assumption about the number of arguments taken by *invalid* syscall(-1)
and so changing a value in x0 (or any other registers) doesn't make any difference.
()

-Takahiro AKASHI

> Will
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
diff mbox

Patch

diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 726b910..1ef57d0 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -668,8 +668,15 @@  ENDPROC(el0_svc)
          * switches, and waiting for our parent to respond.
          */
  __sys_trace:
+       cmp     w8, #-1                         // default errno for invalid
+       b.ne    1f                              // system call
+       mov     x0, #-ENOSYS
+       str     x0, [sp, #S_X0]
+1:
         mov     x0, sp
         bl      syscall_trace_enter
+       cmp     w0, #-1                         // skip the syscall?
+       b.eq    __sys_trace_return_skipped
         adr     lr, __sys_trace_return          // return address
         uxtw    scno, w0                        // syscall number (possibly new)
         mov     x1, sp                          // pointer to regs
@@ -684,6 +691,7 @@  __sys_trace:

  __sys_trace_return:
         str     x0, [sp]                        // save returned x0
+__sys_trace_return_skipped:
         mov     x0, sp
         bl      syscall_trace_exit
         b       ret_to_user