search for: df51a35cf537

...ael S. Tsirkin <mst at redhat.com> Signed-off-by: David S. Miller <davem at davemloft.net> Signed-off-by: Sasha Levin <sashal at kernel.org> --- drivers/vhost/net.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index df51a35cf537..8beacbee2553 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -36,7 +36,7 @@ #include "vhost.h" -static int experimental_zcopytx = 1; +static int experimental_zcopytx = 0; module_param(experimental_zcopytx, int, 0444); MODULE_PARM_DESC(experimental_zcopytx, "Enab...

...think is, we want the weight to be checked in both fast path and slow path. > E.g. like the below. > Warning: completely untested. > > Signed-off-by: Michael S. Tsirkin <mst at redhat.com> > > --- > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > index df51a35cf537..a0f89a504cd9 100644 > --- a/drivers/vhost/net.c > +++ b/drivers/vhost/net.c > @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct vhost_net_virtqueue *nvq, > return 0; > } > > +/* Returns true if caller needs to go back and re-read the ring. */ > +static bool...

...think is, we want the weight to be checked in both fast path and slow path. > E.g. like the below. > Warning: completely untested. > > Signed-off-by: Michael S. Tsirkin <mst at redhat.com> > > --- > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > index df51a35cf537..a0f89a504cd9 100644 > --- a/drivers/vhost/net.c > +++ b/drivers/vhost/net.c > @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct vhost_net_virtqueue *nvq, > return 0; > } > > +/* Returns true if caller needs to go back and re-read the ring. */ > +static bool...

...E.g. like the below. > > > Warning: completely untested. > > > > > > Signed-off-by: Michael S. Tsirkin <mst at redhat.com> > > > > > > --- > > > > > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > > > index df51a35cf537..a0f89a504cd9 100644 > > > --- a/drivers/vhost/net.c > > > +++ b/drivers/vhost/net.c > > > @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct > > > vhost_net_virtqueue *nvq, > > > ????? return 0; > > > ? } > > > ? +/* Returns tr...

...E.g. like the below. > > > Warning: completely untested. > > > > > > Signed-off-by: Michael S. Tsirkin <mst at redhat.com> > > > > > > --- > > > > > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > > > index df51a35cf537..a0f89a504cd9 100644 > > > --- a/drivers/vhost/net.c > > > +++ b/drivers/vhost/net.c > > > @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct > > > vhost_net_virtqueue *nvq, > > > ????? return 0; > > > ? } > > > ? +/* Returns tr...

When the rx buffer is too small for a packet, we will discard the vq descriptor and retry it for the next packet: while ((sock_len = vhost_net_rx_peek_head_len(net, sock->sk, &busyloop_intr))) { ... /* On overrun, truncate and discard */ if (unlikely(headcount > UIO_MAXIOV)) { iov_iter_init(&msg.msg_iter, READ, vq->iov, 1, 1); err = sock->ops->recvmsg(sock,

When the rx buffer is too small for a packet, we will discard the vq descriptor and retry it for the next packet: while ((sock_len = vhost_net_rx_peek_head_len(net, sock->sk, &busyloop_intr))) { ... /* On overrun, truncate and discard */ if (unlikely(headcount > UIO_MAXIOV)) { iov_iter_init(&msg.msg_iter, READ, vq->iov, 1, 1); err = sock->ops->recvmsg(sock,

...se checks in 3 places are exactly the same on all paths and they are slow path. Why don't we put this in a function? E.g. like the below. Warning: completely untested. Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index df51a35cf537..a0f89a504cd9 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct vhost_net_virtqueue *nvq, return 0; } +/* Returns true if caller needs to go back and re-read the ring. */ +static bool empty_ring(struct vhost_net *net, struct v...

...h fast > path and slow path. > > >> E.g. like the below. >> Warning: completely untested. >> >> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> >> >> --- >> >> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c >> index df51a35cf537..a0f89a504cd9 100644 >> --- a/drivers/vhost/net.c >> +++ b/drivers/vhost/net.c >> @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct >> vhost_net_virtqueue *nvq, >> ????? return 0; >> ? } >> ? +/* Returns true if caller needs to go back and re-read...

...e below. >>>> Warning: completely untested. >>>> >>>> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> >>>> >>>> --- >>>> >>>> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c >>>> index df51a35cf537..a0f89a504cd9 100644 >>>> --- a/drivers/vhost/net.c >>>> +++ b/drivers/vhost/net.c >>>> @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct >>>> vhost_net_virtqueue *nvq, >>>> ????? return 0; >>>> ? } >>>>...

...; > > > > > > > > > Signed-off-by: Michael S. Tsirkin <mst at redhat.com> > > > > > > > > > > --- > > > > > > > > > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > > > > > index df51a35cf537..a0f89a504cd9 100644 > > > > > --- a/drivers/vhost/net.c > > > > > +++ b/drivers/vhost/net.c > > > > > @@ -761,6 +761,23 @@ static int vhost_net_build_xdp(struct > > > > > vhost_net_virtqueue *nvq, > > > > > ????? return 0;...

....write = vfio_device_fops_write, .unlocked_ioctl = vfio_device_fops_unl_ioctl, -#ifdef CONFIG_COMPAT - .compat_ioctl = vfio_device_fops_compat_ioctl, -#endif + .compat_ioctl = compat_ptr_ioctl, .mmap = vfio_device_fops_mmap, }; diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index df51a35cf537..1642b3573230 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -1765,14 +1765,6 @@ static long vhost_net_ioctl(struct file *f, unsigned int ioctl, } } -#ifdef CONFIG_COMPAT -static long vhost_net_compat_ioctl(struct file *f, unsigned int ioctl, - unsigned long arg) -{ - retu...

....write = vfio_device_fops_write, .unlocked_ioctl = vfio_device_fops_unl_ioctl, -#ifdef CONFIG_COMPAT - .compat_ioctl = vfio_device_fops_compat_ioctl, -#endif + .compat_ioctl = compat_ptr_ioctl, .mmap = vfio_device_fops_mmap, }; diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index df51a35cf537..1642b3573230 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -1765,14 +1765,6 @@ static long vhost_net_ioctl(struct file *f, unsigned int ioctl, } } -#ifdef CONFIG_COMPAT -static long vhost_net_compat_ioctl(struct file *f, unsigned int ioctl, - unsigned long arg) -{ - retu...

...ed_ioctl = vfio_device_fops_unl_ioctl, > -#ifdef CONFIG_COMPAT > - .compat_ioctl = vfio_device_fops_compat_ioctl, > -#endif > + .compat_ioctl = compat_ptr_ioctl, > .mmap = vfio_device_fops_mmap, > }; > > diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c > index df51a35cf537..1642b3573230 100644 > --- a/drivers/vhost/net.c > +++ b/drivers/vhost/net.c > @@ -1765,14 +1765,6 @@ static long vhost_net_ioctl(struct file *f, unsigned int ioctl, > } > } > > -#ifdef CONFIG_COMPAT > -static long vhost_net_compat_ioctl(struct file *f, unsigned int ioc...

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. This is done through setup kernel address through direct mapping and co-opreate VM management with MMU notifiers. Test shows about 23% improvement on TX PPS. TCP_STREAM doesn't see obvious

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling like SMAP. This is done through setup kernel address through direct mapping and co-opreate VM management with MMU notifiers. Test shows about 23% improvement on TX PPS. TCP_STREAM

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling like SMAP. This is done through setup kernel address through direct mapping and co-opreate VM management with MMU notifiers. Test shows about 23% improvement on TX PPS. TCP_STREAM