Re: [PATCH v2 2/3] sched: add a macro to define bitops for task atomic flags

From: Kees Cook
Date: Tue Sep 23 2014 - 02:37:25 EST


On Mon, Sep 22, 2014 at 11:04 PM, Zefan Li <lizefan@xxxxxxxxxx> wrote:
> This will simplify code when we add new flags.
>
> v2:
> - updated scripts/tags.sh, suggested by Peter
>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Miao Xie <miaox@xxxxxxxxxxxxxx>
> Cc: Kees Cook <keescook@xxxxxxxxxxxx>
> Cc: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Zefan Li <lizefan@xxxxxxxxxx>
> ---
> include/linux/sched.h | 20 +++++++++++---------
> scripts/tags.sh | 6 ++++++
> 2 files changed, 17 insertions(+), 9 deletions(-)
>
> diff --git a/include/linux/sched.h b/include/linux/sched.h
> index 4557765..04a2ae2 100644
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -1959,15 +1959,17 @@ static inline void memalloc_noio_restore(unsigned int flags)
> /* Per-process atomic flags. */
> #define PFA_NO_NEW_PRIVS 0 /* May not gain new privileges. */
>
> -static inline bool task_no_new_privs(struct task_struct *p)
> -{
> - return test_bit(PFA_NO_NEW_PRIVS, &p->atomic_flags);
> -}
> -
> -static inline void task_set_no_new_privs(struct task_struct *p)
> -{
> - set_bit(PFA_NO_NEW_PRIVS, &p->atomic_flags);
> -}
> +#define TASK_PFA_BITOPS(name, func) \
> +static inline bool task_##func(struct task_struct *p) \
> +{ return test_bit(PFA_##name, &p->atomic_flags); } \
> + \
> +static inline void task_set_##func(struct task_struct *p) \
> +{ set_bit(PFA_##name, &p->atomic_flags); } \
> + \
> +static inline void task_clear_##func(struct task_struct *p) \
> +{ clear_bit(PFA_##name, &p->atomic_flags); }
> +
> +TASK_PFA_BITOPS(NO_NEW_PRIVS, no_new_privs)

One thing I don't like about this is that task_clear_no_new_privs()
ends up getting defined, and it should absolutely never be used. NNP
should never be cleared or there could be security issues. I realize
this isn't a very useful nit-pick, but I'd rather the function wasn't
even available for someone to accidentally use. Maybe break up the
macro with some kind of "write only" version like:

#define TASK_PFA_BITOPS_WO(name, func) \
static inline bool task_##func(struct task_struct *p) \
{ return test_bit(PFA_##name, &p->atomic_flags); } \
static inline void task_set_##func(struct task_struct *p) \
{ set_bit(PFA_##name, &p->atomic_flags); }

#define TASK_PFA_BITOPS(name, func) \
TASK_PFA_BITOPS_WO(name, func); \
static inline void task_clear_##func(struct task_struct *p) \
{ clear_bit(PFA_##name, &p->atomic_flags); }

TASK_PFA_BITOPS_WO(NO_NEW_PRIVS, no_new_privs)

And then all the new users can use TASK_PFA_BITOPS() normally since
they expect to use "clear"?

-Kees

>
> /*
> * task->jobctl flags
> diff --git a/scripts/tags.sh b/scripts/tags.sh
> index cbfd269..8591b57 100755
> --- a/scripts/tags.sh
> +++ b/scripts/tags.sh
> @@ -197,6 +197,9 @@ exuberant()
> --regex-c++='/SETPCGFLAG\(([^,)]*).*/SetPageCgroup\1/' \
> --regex-c++='/CLEARPCGFLAG\(([^,)]*).*/ClearPageCgroup\1/' \
> --regex-c++='/TESTCLEARPCGFLAG\(([^,)]*).*/TestClearPageCgroup\1/' \
> + --regex-c++='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_\1/' \
> + --regex-c++='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_set_\1/' \
> + --regex-c++='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_clear_\1/' \
> --regex-c='/PCI_OP_READ\((\w*).*[1-4]\)/pci_bus_read_config_\1/' \
> --regex-c='/PCI_OP_WRITE\((\w*).*[1-4]\)/pci_bus_write_config_\1/' \
> --regex-c='/DEFINE_(MUTEX|SEMAPHORE|SPINLOCK)\((\w*)/\2/v/' \
> @@ -260,6 +263,9 @@ emacs()
> --regex='/SETPCGFLAG\(([^,)]*).*/SetPageCgroup\1/' \
> --regex='/CLEARPCGFLAG\(([^,)]*).*/ClearPageCgroup\1/' \
> --regex='/TESTCLEARPCGFLAG\(([^,)]*).*/TestClearPageCgroup\1/' \
> + --regex='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_\1/' \
> + --regex='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_set_\1/' \
> + --regex='/TASK_PFA_BITOPS\([^,]*,\s*([^)]*)\)/task_clear_\1/' \
> --regex='/_PE(\([^,)]*\).*/PEVENT_ERRNO__\1/' \
> --regex='/PCI_OP_READ(\([a-z]*[a-z]\).*[1-4])/pci_bus_read_config_\1/' \
> --regex='/PCI_OP_WRITE(\([a-z]*[a-z]\).*[1-4])/pci_bus_write_config_\1/'\
> --
> 1.8.0.2
>



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