Hi Shaohua, One question, did you still see the network performance penalty when Linux kernel cmdline intel_iommu was set to off ( intel_iommu=off) ?
Thanks, -ning -----Original Message----- From: Joerg Roedel [mailto:jroe...@suse.de] Sent: Friday, April 07, 2017 3:09 AM To: Shaohua Li <s...@fb.com> Cc: linux-kernel@vger.kernel.org; Wei, Gang <gang....@intel.com>; h...@linux.intel.com; mi...@kernel.org; kernel-t...@fb.com; Sun, Ning <ning....@intel.com>; sri...@fb.com; Eydelberg, Alex <alex.eydelb...@intel.com> Subject: Re: [RFC] x86/tboot: add an option to disable iommu force on On Mon, Apr 03, 2017 at 12:19:28PM -0700, Shaohua Li wrote: > On Wed, Mar 22, 2017 at 07:50:55AM -0400, Shaohua Li wrote: > > On Wed, Mar 22, 2017 at 11:49:00AM +0100, Joerg Roedel wrote: > > > Hi Shaohua, > > > > > > On Tue, Mar 21, 2017 at 11:37:51AM -0700, Shaohua Li wrote: > > > > IOMMU harms performance signficantly when we run very fast > > > > networking workloads. This is a limitation in hardware based on > > > > our observation, so we'd like to disable the IOMMU force on, but > > > > we do want to use TBOOT and we can sacrifice the DMA security > > > > bought by IOMMU. I must admit I know nothing about TBOOT, but > > > > TBOOT guys (cc-ed) think not eabling IOMMU is totally ok. > > > > > > Can you elaborate a bit more on the setup where the IOMMU still > > > harms network performance? With the recent scalability > > > improvements I measured only a minimal impact on 10GBit networking. > > Hi, > > > > It's 40GB networking doing XDP test. Software overhead is almost > > unaware, but it's the IOTLB miss (based on our analysis) which kills > > the performance. We observed the same performance issue even with > > software passthrough (identity mapping), only the hardware > > passthrough survives. The pps with iommu (with software passthrough) is > > only about ~30% of that without it. > > Any update on this? An explicit Ack from the tboot guys would be good to have. Joerg
