search for: ba275241030c

...IZE + > > virtio-net header len to PAGE_SIZE. This improves throughput for > > large packet workloads, as any workload with average packet size >= > > PAGE_SIZE will use PAGE_SIZE buffers. > > > > These optimizations interact positively with recent commit > > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > > optimizations benefit buffers of any size. > > > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs &g...

...for the receive queue. Packet buffer sizes range from MTU_SIZE + virtio-net header len to PAGE_SIZE. This improves throughput for large packet workloads, as any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE buffers. These optimizations interact positively with recent commit ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), which coalesces adjacent RX SKB fragments in virtio_net. The coalescing optimizations benefit buffers of any size. Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs between two QEMU VMs on a single physical...

skb_page_frag_refill currently permits only order-0 page allocs unless GFP_WAIT is used. Change skb_page_frag_refill to attempt higher-order page allocations whether or not GFP_WAIT is used. If memory cannot be allocated, the allocator will fall back to successively smaller page allocs (down to order-0 page allocs). This change brings skb_page_frag_refill in line with the existing page allocation

skb_page_frag_refill currently permits only order-0 page allocs unless GFP_WAIT is used. Change skb_page_frag_refill to attempt higher-order page allocations whether or not GFP_WAIT is used. If memory cannot be allocated, the allocator will fall back to successively smaller page allocs (down to order-0 page allocs). This change brings skb_page_frag_refill in line with the existing page allocation

...for the receive queue. Packet buffer sizes range from MTU_SIZE + virtio-net header len to PAGE_SIZE. This improves throughput for large packet workloads, as any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE buffers. These optimizations interact positively with recent commit ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), which coalesces adjacent RX SKB fragments in virtio_net. The coalescing optimizations benefit buffers of any size. Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs between two QEMU VMs on a single physical...

...her is 64K. Looks like it can result we > allocate PAGE_SIZE buffer for 256 (which is bad since the > payload/truesize is low) bytes or 1500+ for 64K buffer (which is ok > since we can do coalescing). > > > > These optimizations interact positively with recent commit > > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > > optimizations benefit buffers of any size. > > > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs &gt...

...ith 256 bytes each packet another is 64K. Looks like it can result we allocate PAGE_SIZE buffer for 256 (which is bad since the payload/truesize is low) bytes or 1500+ for 64K buffer (which is ok since we can do coalescing). > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

...ith 256 bytes each packet another is 64K. Looks like it can result we allocate PAGE_SIZE buffer for 256 (which is bad since the payload/truesize is low) bytes or 1500+ for 64K buffer (which is ok since we can do coalescing). > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

...for the receive queue. Packet buffer sizes range from MTU_SIZE + virtio-net header len to PAGE_SIZE. This improves throughput for large packet workloads, as any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE buffers. These optimizations interact positively with recent commit ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), which coalesces adjacent RX SKB fragments in virtio_net. The coalescing optimizations benefit buffers of any size. Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs between two QEMU VMs on a single physical...

...t buffer sizes range from MTU_SIZE + virtio-net header > len to PAGE_SIZE. This improves throughput for large packet workloads, as > any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE > buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

...t buffer sizes range from MTU_SIZE + virtio-net header > len to PAGE_SIZE. This improves throughput for large packet workloads, as > any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE > buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

...for the receive queue. Packet buffer sizes range from MTU_SIZE + virtio-net header len to PAGE_SIZE. This improves throughput for large packet workloads, as any workload with average packet size >= PAGE_SIZE will use PAGE_SIZE buffers. These optimizations interact positively with recent commit ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), which coalesces adjacent RX SKB fragments in virtio_net. The coalescing optimizations benefit buffers of any size. Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs between two QEMU VMs on a single physical...

...t buffer sizes range from MTU_SIZE + > virtio-net header len to PAGE_SIZE. This improves throughput for > large packet workloads, as any workload with average packet size >= > PAGE_SIZE will use PAGE_SIZE buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

...t buffer sizes range from MTU_SIZE + > virtio-net header len to PAGE_SIZE. This improves throughput for > large packet workloads, as any workload with average packet size >= > PAGE_SIZE will use PAGE_SIZE buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU V...

Commit 2613af0ed18a ("virtio_net: migrate mergeable rx buffers to page frag allocators") changed the mergeable receive buffer size from PAGE_SIZE to MTU-size. However, the merge buffer size does not take into account the size of the virtio-net header. Consequently, packets that are MTU-size will take two buffers intead of one (to store the virtio-net header), substantially decreasing the

Commit 2613af0ed18a ("virtio_net: migrate mergeable rx buffers to page frag allocators") changed the mergeable receive buffer size from PAGE_SIZE to MTU-size. However, the merge buffer size does not take into account the size of the virtio-net header. Consequently, packets that are MTU-size will take two buffers intead of one (to store the virtio-net header), substantially decreasing the

skb_page_frag_refill currently permits only order-0 page allocs unless GFP_WAIT is used. Change skb_page_frag_refill to attempt higher-order page allocations whether or not GFP_WAIT is used. If memory cannot be allocated, the allocator will fall back to successively smaller page allocs (down to order-0 page allocs). This change brings skb_page_frag_refill in line with the existing page allocation

skb_page_frag_refill currently permits only order-0 page allocs unless GFP_WAIT is used. Change skb_page_frag_refill to attempt higher-order page allocations whether or not GFP_WAIT is used. If memory cannot be allocated, the allocator will fall back to successively smaller page allocs (down to order-0 page allocs). This change brings skb_page_frag_refill in line with the existing page allocation

...buffer sizes range from MTU_SIZE + > virtio-net header len to PAGE_SIZE. This improves throughput for > large packet workloads, as any workload with average packet size >= > PAGE_SIZE will use PAGE_SIZE buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU...

...buffer sizes range from MTU_SIZE + > virtio-net header len to PAGE_SIZE. This improves throughput for > large packet workloads, as any workload with average packet size >= > PAGE_SIZE will use PAGE_SIZE buffers. > > These optimizations interact positively with recent commit > ba275241030c ("virtio-net: coalesce rx frags when possible during rx"), > which coalesces adjacent RX SKB fragments in virtio_net. The coalescing > optimizations benefit buffers of any size. > > Benchmarks taken from an average of 5 netperf 30-second TCP_STREAM runs > between two QEMU...