On 01/22/2016 07:02 PM, Dave Hansen wrote:
> One of Vlastimil's comments made me go dig back in to the uprobes
> code's use of get_user_pages(). I decided to change both of them
> to be "foreign" accesses.
>
> This also fixes the nommu breakage that Vlastimil noted last time.
>
> Srikar, I'd appreciate if you can have a look at the uprobes.c
> modifications, especially the comment. I don't think this will
> change any behavior, but I want to make sure the comment is
> accurate.
>
> ---
>
> From: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
>
> For protection keys, we need to understand whether protections
> should be enforced in software or not. In general, we enforce
> protections when working on our own task, but not when on others.
> We call these "current" and "foreign" operations.
>
> This patch introduces a new get_user_pages() variant:
>
> get_user_pages_foreign()
>
> We modify the vanilla get_user_pages() so it can no longer be
> used on mm/tasks other than 'current/current->mm', which is by
> far the most common way it is called. Using it makes a few of
> the call sites look a bit nicer.
>
> In other words, get_user_pages_foreign() is a replacement for
> when get_user_pages() is called on non-current tsk/mm.
>
> This also switches get_user_pages_(un)locked() over to be like
> get_user_pages() and not take a tsk/mm. There is no
> get_user_pages_foreign_(un)locked(). If someone wants that
> behavior they just have to use "__" variant and pass in
> FOLL_FOREIGN explicitly.
>
> The uprobes is_trap_at_addr() location holds mmap_sem and
> calls get_user_pages(current->mm) on an instruction address. This
> makes it a pretty unique gup caller. Being an instruction access
> and also really originating from the kernel (vs. the app), I opted
> to consider this a 'foreign' access where protection keys will not
> be enforced.
>
> Signed-off-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
> Cc: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> Cc: Srikar Dronamraju <srikar@xxxxxxxxxxxxxxxxxx>
> Cc: vbabka@xxxxxxx
Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
But,
> long __get_user_pages_unlocked(struct task_struct *tsk, struct mm_struct *mm,
> unsigned long start, unsigned long nr_pages,
> int write, int force, struct page **pages,
> unsigned int gup_flags)
> {
> long ret;
> - down_read(&mm->mmap_sem);
> - ret = get_user_pages(tsk, mm, start, nr_pages, write, force,
> - pages, NULL);
> - up_read(&mm->mmap_sem);
> + down_read(¤t->mm->mmap_sem);
> + ret = get_user_pages(start, nr_pages, write, force, pages, NULL);
> + up_read(¤t->mm->mmap_sem);
I understand your reply to lkp report also means that this no longer locks
current's mmap_sem? :)
Vlastimil
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>