search for: total_sg

...inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq, > return next; > } > > -static inline int virtqueue_add_split(struct virtqueue *_vq, > - struct scatterlist *sgs[], > - unsigned int total_sg, > - unsigned int out_sgs, > - unsigned int in_sgs, > - void *data, > - void *ctx, > - gfp_t gfp) > +...

...ers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -520,57 +520,38 @@ static inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq, return next; } -static inline int virtqueue_add_split(struct virtqueue *_vq, - struct scatterlist *sgs[], - unsigned int total_sg, - unsigned int out_sgs, - unsigned int in_sgs, - void *data, - void *ctx, - gfp_t gfp) +/* note: return NULL means no indirect that is valid. */ +static struct vring_desc *virtqueue_get_desc_split(struct vring_virtqueue *vq, + unsigned int total...

...9,10 @@ struct vring_virtqueue #define to_vvq(_vq) container_of(_vq, struct vring_virtqueue, vq) -/* Set up an indirect table of descriptors and add it to the queue. */ -static inline int vring_add_indirect(struct vring_virtqueue *vq, - struct scatterlist *sgs[], - unsigned int total_sg, - unsigned int out_sgs, - unsigned int in_sgs, - gfp_t gfp) +static struct vring_desc *alloc_indirect(unsigned int total_sg, gfp_t gfp) { - struct vring_desc *desc; - unsigned head; - struct scatterlist *sg; - int i, n; + struct vring_desc *desc; + unsigned int i; /*...

XDP socket(AF_XDP) is an excellent bypass kernel network framework. The zero copy feature of xsk (XDP socket) needs to be supported by the driver. The performance of zero copy is very good. ENV: Qemu with vhost. vhost cpu | Guest APP CPU |Guest Softirq CPU | PPS -----------------------------|---------------|------------------|------------ xmit by sockperf: 90% | 100%

Earlier discussion: https://lkml.org/lkml/2017/8/4/601 "Increased memory usage with scsi-mq" Downstream bug report: https://bugzilla.redhat.com/show_bug.cgi?id=1478201

Earlier discussion: https://lkml.org/lkml/2017/8/4/601 "Increased memory usage with scsi-mq" Downstream bug report: https://bugzilla.redhat.com/show_bug.cgi?id=1478201

...include/uapi/linux/virtio_config.h | 12 +- > > include/uapi/linux/virtio_ring.h | 61 ++ > > 4 files changed, 980 insertions(+), 195 deletions(-) [...] > > +static struct vring_packed_desc *alloc_indirect_packed(struct virtqueue *_vq, > > + unsigned int total_sg, > > + gfp_t gfp) > > +{ > > + struct vring_packed_desc *desc; > > + > > + /* > > + * We require lowmem mappings for the descriptors because > > + * otherwise virt_to_phys will give us bogus addresses in the > > + * virtqueue. > >...

...include/uapi/linux/virtio_config.h | 12 +- > > include/uapi/linux/virtio_ring.h | 61 ++ > > 4 files changed, 980 insertions(+), 195 deletions(-) [...] > > +static struct vring_packed_desc *alloc_indirect_packed(struct virtqueue *_vq, > > + unsigned int total_sg, > > + gfp_t gfp) > > +{ > > + struct vring_packed_desc *desc; > > + > > + /* > > + * We require lowmem mappings for the descriptors because > > + * otherwise virt_to_phys will give us bogus addresses in the > > + * virtqueue. > >...

I resurrected these patches after prompting from Andy Lutomirski's recent patches. I put them on the back-burner because vring_bench had a 15% slowdown on my laptop: pktgen testing revealed a speedup, if anything, so I've cleaned them up. Rusty Russell (3): virtio_net: pass well-formed sgs to virtqueue_add_*() virtio_ring: assume sgs are always well-formed. virtio_ring: unify

I resurrected these patches after prompting from Andy Lutomirski's recent patches. I put them on the back-burner because vring_bench had a 15% slowdown on my laptop: pktgen testing revealed a speedup, if anything, so I've cleaned them up. Rusty Russell (3): virtio_net: pass well-formed sgs to virtqueue_add_*() virtio_ring: assume sgs are always well-formed. virtio_ring: unify

...Previously vring_add_indirect() did the allocation and the simple > linear layout. We replace that with alloc_indirect() which allocates > the indirect table then chains it like the normal descriptor table so > we can reuse the core logic. > > + if (vq->indirect && total_sg > 1 && vq->vq.num_free) > + desc = alloc_indirect(total_sg, gfp); > + else > + desc = NULL; > + > + if (desc) { > + /* Use a single buffer which doesn't continue */ > + vq->vring.desc[head]...

...Previously vring_add_indirect() did the allocation and the simple > linear layout. We replace that with alloc_indirect() which allocates > the indirect table then chains it like the normal descriptor table so > we can reuse the core logic. > > + if (vq->indirect && total_sg > 1 && vq->vq.num_free) > + desc = alloc_indirect(total_sg, gfp); > + else > + desc = NULL; > + > + if (desc) { > + /* Use a single buffer which doesn't continue */ > + vq->vring.desc[head]...

XDP socket(AF_XDP) is an excellent bypass kernel network framework. The zero copy feature of xsk (XDP socket) needs to be supported by the driver. The performance of zero copy is very good. ENV: Qemu with vhost. vhost cpu | Guest APP CPU |Guest Softirq CPU | PPS -----------------------------|---------------|------------------|------------ xmit by sockperf: 90% | 100%

If using indirect descriptors, you can make the total_sg as large as you want. If not, BUG is too serious because the function later returns -ENOSPC. Thanks Paolo Bonzini, Christoph Hellwig. Signed-off-by: Richard W.M. Jones <rjones at redhat.com> --- drivers/virtio/virtio_ring.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff...

...gt; gfp_t gfp) > { > struct vring_virtqueue *vq = to_vvq(_vq); > @@ -767,9 +772,11 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > goto end; > } > > - err = virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs); > - if (err) > - goto err; > + if (dma_map) { Could we simply check sg->dma_addr in this case? Then we don't need to introduce the dma_map flag. Thanks > + err = virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs);...

Hi, I'm hitting this bug with both ext4 and btrfs. Here's an example of the backtrace: https://gist.github.com/vzctl/e888a821333979120932 I tried raising this BUG only for direct ring and it solved the problem: - BUG_ON(total_sg > vq->vring.num); + BUG_ON(total_sg > vq->vring.num && !vq->indirect); Shall I submit the patch or is a more elaborate fix required? -- Alexey On Mon, May 27, 2013 at 7:38 AM, Rusty Russell <rusty at rustcorp.com.au> wrote: > Dave Airlie <airlied at gm...

Hi, I'm hitting this bug with both ext4 and btrfs. Here's an example of the backtrace: https://gist.github.com/vzctl/e888a821333979120932 I tried raising this BUG only for direct ring and it solved the problem: - BUG_ON(total_sg > vq->vring.num); + BUG_ON(total_sg > vq->vring.num && !vq->indirect); Shall I submit the patch or is a more elaborate fix required? -- Alexey On Mon, May 27, 2013 at 7:38 AM, Rusty Russell <rusty at rustcorp.com.au> wrote: > Dave Airlie <airlied at gm...

.../virtio_config.h | 12 +- >>> include/uapi/linux/virtio_ring.h | 61 ++ >>> 4 files changed, 980 insertions(+), 195 deletions(-) > [...] >>> +static struct vring_packed_desc *alloc_indirect_packed(struct virtqueue *_vq, >>> + unsigned int total_sg, >>> + gfp_t gfp) >>> +{ >>> + struct vring_packed_desc *desc; >>> + >>> + /* >>> + * We require lowmem mappings for the descriptors because >>> + * otherwise virt_to_phys will give us bogus addresses in the >>> +...

...-- a/drivers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -520,6 +520,73 @@ static inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq, return next; } +static int virtqueue_map_sgs(struct vring_virtqueue *vq, + struct scatterlist *sgs[], + unsigned int total_sg, + unsigned int out_sgs, + unsigned int in_sgs) +{ + struct scatterlist *sg; + unsigned int n; + + for (n = 0; n < out_sgs; n++) { + for (sg = sgs[n]; sg; sg = sg_next(sg)) { + dma_addr_t addr = vring_map_one_sg(vq, sg, DMA_TO_DEVICE); + + if (vring_mapping_error(vq, addr)) +...

...vs. non-dma is relevant. I tried -net user a -net tap >> with iperf running in both directions. I also tried switching >> virtio_pci into non-DMA-API mode. No errors. >> >> Is there any chance that you could instrument that BUG_ON to print n, >> i, in_sgs, out_sgs, total_sg, total_in, and total_out? I assume this >> is some oddity with patch 1, but I'm mystified. > > Yes, its triggered by patch 1. I must be doing something wrong, since virtio_net isn't sending me through that code path at all. This may be related to ethtool refusing to enable sc...