Michael> Roland, I think VM_DONTCOPY is needed here.
Michael> If a process forks, we must prevent the child from
Michael> accessing the parent's hardware page. Otherwise the child
Michael> can corrupt the parent's queues since the hardware wont
Michael> be able to distinguish b
Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> Subject: Re: [PATCH 11/16] IB uverbs: add mthca mmap support
>
> Andrew> What's the thinking behind the VM_DONTCOPY there?
>
> As I said before, I don't think the thinking behind VM_DONTCOPY was
> correct thinking. Let's take it out.
Roland, I
Andrew> What's the thinking behind the VM_DONTCOPY there?
As I said before, I don't think the thinking behind VM_DONTCOPY was
correct thinking. Let's take it out.
I've now answered all your questions on this patchset (or at least
written something in response to all your questions ;). What'
Andrew> What's the thinking behind the VM_DONTCOPY there?
I think that was my paranoia about something like a process doing a
fork, the original process exiting, and the new process having page
still mapped even though the file has been released. This is bad
because then we could map the same
Roland Dreier <[EMAIL PROTECTED]> wrote:
>
> Add support for mmap() method to mthca, so that userspace can get
> access to doorbell registers. This allows userspace to get direct
> access to the HCA for data path operations.
>
> Each userspace context gets its own copy of the doorbell registers a