[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH RFC 01/44] passthrough/vtd: Don't DMA to the stack in queue_invalidate_wait()

On 05/01/2018 09:21, Jan Beulich wrote:
>>>> On 04.01.18 at 21:21, <andrew.cooper3@xxxxxxxxxx> wrote:
>> DMA-ing to the stack is generally considered bad practice.  In this case, if 
>> a
>> timeout occurs because of a sluggish device which is processing the request,
>> the completion notification will corrupt the stack of a subsequent deeper 
>> call
>> tree.
>> Place the poll_slot in a percpu area and DMA to that instead.
>> Note: This change does not address other issues with the current
>> implementation, such as once a timeout has been suffered, subsequent
>> completions can't be correlated with their requests.
>> Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
>> Julien: This wants backporting to all releases, and therefore should be
>> considered for 4.10 at this point.
> Interesting remark at this point in time ;-)

Oops yes.  This might leak the point at which I shelved the plan.

With this all out in the open now, observant people might notice now
many of my 4.10 patches are relevant to the issues at hand.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.