Re: architecture paper draft

["Stephen C. Tweedie" <sct at redhat dot com>]

Hi,

On Fri, 2005-02-11 at 22:58, Frank Ch. Eigler wrote:

> The big doubt in my mind is what to do about user-space.  We seem
> to lack a reliable way of accessing memory there from the full
> variety of contexts into which probes could be placed.

I've asked, but nobody has been able to give me a list of what you're
wanting to actually *do* here.  What are the important cases?  Have we
got examples of the sorts of things we expect the users to want to do?

The fact that you can't access user memory from an interrupt is
important from a correctness point of view, but it's unlikely to affect
probes that users are going to come up with all that often!

Do you want probes to be able to access arbitrary process contexts?  Or
just the current one?

What about locks --- do you want to be able to access the memory deep
inside locked sections of the kernel?

If you're willing to go through a somewhat more heavyweight mechanism
than simple direct memory reference, there's always access_process_vm(),
which is what ptrace() uses to get at an arbitrary process's memory
contents.  That deals with manually faulting any pages that it needs,
but it can't be called from any context where sleeping and blocking
locks are forbidden.

--Stephen