diff mbox

[v5,2/3] asm-generic: Add generic seccomp.h for secure computing mode 1

Message ID 1406020499-5537-3-git-send-email-takahiro.akashi@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

AKASHI Takahiro July 22, 2014, 9:14 a.m. UTC
Those values (__NR_seccomp_*) are used solely in secure_computing()
to identify mode 1 system calls. If compat system calls have different
syscall numbers, asm/seccomp.h may override them.

Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
---
 include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)
 create mode 100644 include/asm-generic/seccomp.h

Comments

Andy Lutomirski July 24, 2014, 3:40 a.m. UTC | #1
On 07/22/2014 02:14 AM, AKASHI Takahiro wrote:
> Those values (__NR_seccomp_*) are used solely in secure_computing()
> to identify mode 1 system calls. If compat system calls have different
> syscall numbers, asm/seccomp.h may override them.
>
> Acked-by: Arnd Bergmann <arnd@arndb.de>
> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
> ---
>   include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
>   1 file changed, 28 insertions(+)
>   create mode 100644 include/asm-generic/seccomp.h
>
> diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
> new file mode 100644
> index 0000000..5e97022
> --- /dev/null
> +++ b/include/asm-generic/seccomp.h
> @@ -0,0 +1,28 @@
> +/*
> + * include/asm-generic/seccomp.h
> + *
> + * Copyright (C) 2014 Linaro Limited
> + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + */
> +#ifndef _ASM_GENERIC_SECCOMP_H
> +#define _ASM_GENERIC_SECCOMP_H
> +
> +#include <asm-generic/unistd.h>
> +
> +#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
> +#define __NR_seccomp_read_32		__NR_read
> +#define __NR_seccomp_write_32		__NR_write
> +#define __NR_seccomp_exit_32		__NR_exit
> +#define __NR_seccomp_sigreturn_32	__NR_rt_sigreturn
> +#endif /* CONFIG_COMPAT && ! already defined */
> +
> +#define __NR_seccomp_read		__NR_read
> +#define __NR_seccomp_write		__NR_write
> +#define __NR_seccomp_exit		__NR_exit
> +#define __NR_seccomp_sigreturn		__NR_rt_sigreturn

I don't like these names.  __NR_seccomp_read sounds like the number of a 
syscall called seccomp_read.

Also, shouldn't something be including this header?  I'm confused.

--Andy
Kees Cook July 24, 2014, 4:41 a.m. UTC | #2
On Wed, Jul 23, 2014 at 8:40 PM, Andy Lutomirski <luto@amacapital.net> wrote:
> On 07/22/2014 02:14 AM, AKASHI Takahiro wrote:
>>
>> Those values (__NR_seccomp_*) are used solely in secure_computing()
>> to identify mode 1 system calls. If compat system calls have different
>> syscall numbers, asm/seccomp.h may override them.
>>
>> Acked-by: Arnd Bergmann <arnd@arndb.de>
>> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
>> ---
>>   include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
>>   1 file changed, 28 insertions(+)
>>   create mode 100644 include/asm-generic/seccomp.h
>>
>> diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
>> new file mode 100644
>> index 0000000..5e97022
>> --- /dev/null
>> +++ b/include/asm-generic/seccomp.h
>> @@ -0,0 +1,28 @@
>> +/*
>> + * include/asm-generic/seccomp.h
>> + *
>> + * Copyright (C) 2014 Linaro Limited
>> + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + */
>> +#ifndef _ASM_GENERIC_SECCOMP_H
>> +#define _ASM_GENERIC_SECCOMP_H
>> +
>> +#include <asm-generic/unistd.h>
>> +
>> +#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
>> +#define __NR_seccomp_read_32           __NR_read
>> +#define __NR_seccomp_write_32          __NR_write
>> +#define __NR_seccomp_exit_32           __NR_exit
>> +#define __NR_seccomp_sigreturn_32      __NR_rt_sigreturn
>> +#endif /* CONFIG_COMPAT && ! already defined */
>> +
>> +#define __NR_seccomp_read              __NR_read
>> +#define __NR_seccomp_write             __NR_write
>> +#define __NR_seccomp_exit              __NR_exit
>> +#define __NR_seccomp_sigreturn         __NR_rt_sigreturn
>
>
> I don't like these names.  __NR_seccomp_read sounds like the number of a
> syscall called seccomp_read.
>
> Also, shouldn't something be including this header?  I'm confused.

Ah! Good catch. These names are correct (see kernel/seccomp.c's
mode1_syscalls and mode1_syscalls_32 arrays), but the location of this
change was unexpected. I was expecting this file to live in
arch/*/include/asm/seccomp.h, not in include/asm-generic/seccomp.h.

However, since it's always the same list, it might make sense to
consolidate them into a single place as a default to make arch porting
easier. However, I think that should be a separate patch.

-Kees
AKASHI Takahiro July 24, 2014, 5:17 a.m. UTC | #3
On 07/24/2014 01:41 PM, Kees Cook wrote:
> On Wed, Jul 23, 2014 at 8:40 PM, Andy Lutomirski <luto@amacapital.net> wrote:
>> On 07/22/2014 02:14 AM, AKASHI Takahiro wrote:
>>>
>>> Those values (__NR_seccomp_*) are used solely in secure_computing()
>>> to identify mode 1 system calls. If compat system calls have different
>>> syscall numbers, asm/seccomp.h may override them.
>>>
>>> Acked-by: Arnd Bergmann <arnd@arndb.de>
>>> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>> ---
>>>    include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
>>>    1 file changed, 28 insertions(+)
>>>    create mode 100644 include/asm-generic/seccomp.h
>>>
>>> diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
>>> new file mode 100644
>>> index 0000000..5e97022
>>> --- /dev/null
>>> +++ b/include/asm-generic/seccomp.h
>>> @@ -0,0 +1,28 @@
>>> +/*
>>> + * include/asm-generic/seccomp.h
>>> + *
>>> + * Copyright (C) 2014 Linaro Limited
>>> + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>> + *
>>> + * This program is free software; you can redistribute it and/or modify
>>> + * it under the terms of the GNU General Public License version 2 as
>>> + * published by the Free Software Foundation.
>>> + */
>>> +#ifndef _ASM_GENERIC_SECCOMP_H
>>> +#define _ASM_GENERIC_SECCOMP_H
>>> +
>>> +#include <asm-generic/unistd.h>
>>> +
>>> +#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
>>> +#define __NR_seccomp_read_32           __NR_read
>>> +#define __NR_seccomp_write_32          __NR_write
>>> +#define __NR_seccomp_exit_32           __NR_exit
>>> +#define __NR_seccomp_sigreturn_32      __NR_rt_sigreturn
>>> +#endif /* CONFIG_COMPAT && ! already defined */
>>> +
>>> +#define __NR_seccomp_read              __NR_read
>>> +#define __NR_seccomp_write             __NR_write
>>> +#define __NR_seccomp_exit              __NR_exit
>>> +#define __NR_seccomp_sigreturn         __NR_rt_sigreturn
>>
>>
>> I don't like these names.  __NR_seccomp_read sounds like the number of a
>> syscall called seccomp_read.
>>
>> Also, shouldn't something be including this header?  I'm confused.
>
> Ah! Good catch. These names are correct (see kernel/seccomp.c's
> mode1_syscalls and mode1_syscalls_32 arrays), but the location of this
> change was unexpected. I was expecting this file to live in
> arch/*/include/asm/seccomp.h, not in include/asm-generic/seccomp.h.
>
> However, since it's always the same list, it might make sense to
> consolidate them into a single place as a default to make arch porting
> easier.

Yeah, that is why I put this file under include/asm-generic.

> However, I think that should be a separate patch.

Do you mean that the code for all the existing archs should also be changed
to use this (common) header?

-Takahiro AKASHI
>
> -Kees
>
Andy Lutomirski July 24, 2014, 2:57 p.m. UTC | #4
On Jul 23, 2014 10:17 PM, "AKASHI Takahiro" <takahiro.akashi@linaro.org> wrote:
>
> On 07/24/2014 01:41 PM, Kees Cook wrote:
>>
>> On Wed, Jul 23, 2014 at 8:40 PM, Andy Lutomirski <luto@amacapital.net> wrote:
>>>
>>> On 07/22/2014 02:14 AM, AKASHI Takahiro wrote:
>>>>
>>>>
>>>> Those values (__NR_seccomp_*) are used solely in secure_computing()
>>>> to identify mode 1 system calls. If compat system calls have different
>>>> syscall numbers, asm/seccomp.h may override them.
>>>>
>>>> Acked-by: Arnd Bergmann <arnd@arndb.de>
>>>> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>>> ---
>>>>    include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
>>>>    1 file changed, 28 insertions(+)
>>>>    create mode 100644 include/asm-generic/seccomp.h
>>>>
>>>> diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
>>>> new file mode 100644
>>>> index 0000000..5e97022
>>>> --- /dev/null
>>>> +++ b/include/asm-generic/seccomp.h
>>>> @@ -0,0 +1,28 @@
>>>> +/*
>>>> + * include/asm-generic/seccomp.h
>>>> + *
>>>> + * Copyright (C) 2014 Linaro Limited
>>>> + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>>> + *
>>>> + * This program is free software; you can redistribute it and/or modify
>>>> + * it under the terms of the GNU General Public License version 2 as
>>>> + * published by the Free Software Foundation.
>>>> + */
>>>> +#ifndef _ASM_GENERIC_SECCOMP_H
>>>> +#define _ASM_GENERIC_SECCOMP_H
>>>> +
>>>> +#include <asm-generic/unistd.h>
>>>> +
>>>> +#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
>>>> +#define __NR_seccomp_read_32           __NR_read
>>>> +#define __NR_seccomp_write_32          __NR_write
>>>> +#define __NR_seccomp_exit_32           __NR_exit
>>>> +#define __NR_seccomp_sigreturn_32      __NR_rt_sigreturn
>>>> +#endif /* CONFIG_COMPAT && ! already defined */
>>>> +
>>>> +#define __NR_seccomp_read              __NR_read
>>>> +#define __NR_seccomp_write             __NR_write
>>>> +#define __NR_seccomp_exit              __NR_exit
>>>> +#define __NR_seccomp_sigreturn         __NR_rt_sigreturn
>>>
>>>
>>>
>>> I don't like these names.  __NR_seccomp_read sounds like the number of a
>>> syscall called seccomp_read.
>>>
>>> Also, shouldn't something be including this header?  I'm confused.
>>
>>
>> Ah! Good catch. These names are correct (see kernel/seccomp.c's
>> mode1_syscalls and mode1_syscalls_32 arrays), but the location of this
>> change was unexpected. I was expecting this file to live in
>> arch/*/include/asm/seccomp.h, not in include/asm-generic/seccomp.h.
>>
>> However, since it's always the same list, it might make sense to
>> consolidate them into a single place as a default to make arch porting
>> easier.
>
>
> Yeah, that is why I put this file under include/asm-generic.

It seems odd that the header would be added without any users.  I
guess it's okay, since arm64 uses it in the followup patch.

>
>> However, I think that should be a separate patch.
>
>
> Do you mean that the code for all the existing archs should also be changed
> to use this (common) header?

If that works, yes.

--Andy

>
> -Takahiro AKASHI
>>
>>
>> -Kees
>>
AKASHI Takahiro July 25, 2014, 8:52 a.m. UTC | #5
On 07/24/2014 11:57 PM, Andy Lutomirski wrote:
> On Jul 23, 2014 10:17 PM, "AKASHI Takahiro" <takahiro.akashi@linaro.org> wrote:
>>
>> On 07/24/2014 01:41 PM, Kees Cook wrote:
>>>
>>> On Wed, Jul 23, 2014 at 8:40 PM, Andy Lutomirski <luto@amacapital.net> wrote:
>>>>
>>>> On 07/22/2014 02:14 AM, AKASHI Takahiro wrote:
>>>>>
>>>>>
>>>>> Those values (__NR_seccomp_*) are used solely in secure_computing()
>>>>> to identify mode 1 system calls. If compat system calls have different
>>>>> syscall numbers, asm/seccomp.h may override them.
>>>>>
>>>>> Acked-by: Arnd Bergmann <arnd@arndb.de>
>>>>> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>>>> ---
>>>>>     include/asm-generic/seccomp.h |   28 ++++++++++++++++++++++++++++
>>>>>     1 file changed, 28 insertions(+)
>>>>>     create mode 100644 include/asm-generic/seccomp.h
>>>>>
>>>>> diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
>>>>> new file mode 100644
>>>>> index 0000000..5e97022
>>>>> --- /dev/null
>>>>> +++ b/include/asm-generic/seccomp.h
>>>>> @@ -0,0 +1,28 @@
>>>>> +/*
>>>>> + * include/asm-generic/seccomp.h
>>>>> + *
>>>>> + * Copyright (C) 2014 Linaro Limited
>>>>> + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
>>>>> + *
>>>>> + * This program is free software; you can redistribute it and/or modify
>>>>> + * it under the terms of the GNU General Public License version 2 as
>>>>> + * published by the Free Software Foundation.
>>>>> + */
>>>>> +#ifndef _ASM_GENERIC_SECCOMP_H
>>>>> +#define _ASM_GENERIC_SECCOMP_H
>>>>> +
>>>>> +#include <asm-generic/unistd.h>
>>>>> +
>>>>> +#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
>>>>> +#define __NR_seccomp_read_32           __NR_read
>>>>> +#define __NR_seccomp_write_32          __NR_write
>>>>> +#define __NR_seccomp_exit_32           __NR_exit
>>>>> +#define __NR_seccomp_sigreturn_32      __NR_rt_sigreturn
>>>>> +#endif /* CONFIG_COMPAT && ! already defined */
>>>>> +
>>>>> +#define __NR_seccomp_read              __NR_read
>>>>> +#define __NR_seccomp_write             __NR_write
>>>>> +#define __NR_seccomp_exit              __NR_exit
>>>>> +#define __NR_seccomp_sigreturn         __NR_rt_sigreturn
>>>>
>>>>
>>>>
>>>> I don't like these names.  __NR_seccomp_read sounds like the number of a
>>>> syscall called seccomp_read.
>>>>
>>>> Also, shouldn't something be including this header?  I'm confused.
>>>
>>>
>>> Ah! Good catch. These names are correct (see kernel/seccomp.c's
>>> mode1_syscalls and mode1_syscalls_32 arrays), but the location of this
>>> change was unexpected. I was expecting this file to live in
>>> arch/*/include/asm/seccomp.h, not in include/asm-generic/seccomp.h.
>>>
>>> However, since it's always the same list, it might make sense to
>>> consolidate them into a single place as a default to make arch porting
>>> easier.
>>
>>
>> Yeah, that is why I put this file under include/asm-generic.
>
> It seems odd that the header would be added without any users.  I
> guess it's okay, since arm64 uses it in the followup patch.
>
>>
>>> However, I think that should be a separate patch.
>>
>>
>> Do you mean that the code for all the existing archs should also be changed
>> to use this (common) header?
>
> If that works, yes.

As is often the case, the patch itself is quite simple, but I can't
test it on other architectures.

-Takahiro AKASHI

> --Andy
>
>>
>> -Takahiro AKASHI
>>>
>>>
>>> -Kees
>>>
diff mbox

Patch

diff --git a/include/asm-generic/seccomp.h b/include/asm-generic/seccomp.h
new file mode 100644
index 0000000..5e97022
--- /dev/null
+++ b/include/asm-generic/seccomp.h
@@ -0,0 +1,28 @@ 
+/*
+ * include/asm-generic/seccomp.h
+ *
+ * Copyright (C) 2014 Linaro Limited
+ * Author: AKASHI Takahiro <takahiro.akashi@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_GENERIC_SECCOMP_H
+#define _ASM_GENERIC_SECCOMP_H
+
+#include <asm-generic/unistd.h>
+
+#if defined(CONFIG_COMPAT) && !defined(__NR_seccomp_read_32)
+#define __NR_seccomp_read_32		__NR_read
+#define __NR_seccomp_write_32		__NR_write
+#define __NR_seccomp_exit_32		__NR_exit
+#define __NR_seccomp_sigreturn_32	__NR_rt_sigreturn
+#endif /* CONFIG_COMPAT && ! already defined */
+
+#define __NR_seccomp_read		__NR_read
+#define __NR_seccomp_write		__NR_write
+#define __NR_seccomp_exit		__NR_exit
+#define __NR_seccomp_sigreturn		__NR_rt_sigreturn
+
+#endif /* _ASM_GENERIC_SECCOMP_H */