> From: Tian, Kevin > Sent: Monday, October 23, 2017 3:05 PM > > > From: Andrew Cooper [mailto:andrew.coop...@citrix.com] > > Sent: Saturday, October 21, 2017 1:55 AM > > > > On 20/10/17 08:12, Jan Beulich wrote: > > >>>> On 19.10.17 at 18:22, <andrew.coop...@citrix.com> 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. > > >> > > >> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com> > > > Please could you extend the commit message to state the issue > > > remaining with using a single per-CPU slot? With that > > > Reviewed-by: Jan Beulich <jbeul...@suse.com> > > > > How about this? > > > > 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. > > > > Can you increase the poll data +1 every time when a new wait > comes? >
btw curious whether you observed a real issue or just eye-caught the problem? _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel
