Re: [PATCH 1/3] uprobes: install_breakpoint() should fail ifis_swbp_insn() == T

From: Oleg Nesterov
Date: Fri Jun 01 2012 - 14:40:25 EST


On 06/01, Srikar Dronamraju wrote:
>
> * Oleg Nesterov <oleg@xxxxxxxxxx> [2012-05-31 20:53:39]:
>
> > __find_next_vma_info() checks tmpvi->mm == vma->vm_mm to detect the
> > already visited mm/vma. However, afaics this can be false positive?
> >
> > The caller, register_for_each_vma(), does mmput() and after that
> > this memory can be freed and re-used as another mm_struct.
> >
>
> Before we do a mmput(), we take the mmap_sem for that mm. So this mm
> cannot be freed until we complete insertion/removal. If the mm gets
> reused after the insertion/removal, and maps the inode,

Yes, see the previous email from me, we can rely on uprobe_mmap().

> Or are you hinting at some other problem?

Perhaps this deserves a comment. I mean, to explain that yes,
tmpvi->mm == vma->vm_mm can be wrong but this is fine.

However, I hope we simply can kill this code. See build_map_info()
I sent.

Oleg.

--
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/