| Message ID | 181f994e13c816116fa69a1e92c2f69e6330f749.1658746417.git.lorenzo@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | BPF |
| Headers | show |
| Series | [bpf-next] xdp: report rx queue index in xdp_frame | expand |
> Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info > pointer. xdp_frame queue_index is currently used in cpumap code to covert > the xdp_frame into a xdp_buff. > xdp_frame size is not increased adding queue_index since an alignment padding > in the structure is used to insert queue_index field. > > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> > --- > include/net/xdp.h | 2 ++ > kernel/bpf/cpumap.c | 2 +- > 2 files changed, 3 insertions(+), 1 deletion(-) Hi Alexei and Daniel, this patch is marked as 'new, archived' in patchwork. Do I need to rebase and repost it? Regards, Lorenzo > > diff --git a/include/net/xdp.h b/include/net/xdp.h > index 04c852c7a77f..3567866b0af5 100644 > --- a/include/net/xdp.h > +++ b/include/net/xdp.h > @@ -172,6 +172,7 @@ struct xdp_frame { > struct xdp_mem_info mem; > struct net_device *dev_rx; /* used by cpumap */ > u32 flags; /* supported values defined in xdp_buff_flags */ > + u32 queue_index; > }; > > static __always_inline bool xdp_frame_has_frags(struct xdp_frame *frame) > @@ -301,6 +302,7 @@ struct xdp_frame *xdp_convert_buff_to_frame(struct xdp_buff *xdp) > > /* rxq only valid until napi_schedule ends, convert to xdp_mem_info */ > xdp_frame->mem = xdp->rxq->mem; > + xdp_frame->queue_index = xdp->rxq->queue_index; > > return xdp_frame; > } > diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c > index f4860ac756cd..09a792d088b3 100644 > --- a/kernel/bpf/cpumap.c > +++ b/kernel/bpf/cpumap.c > @@ -228,7 +228,7 @@ static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu, > > rxq.dev = xdpf->dev_rx; > rxq.mem = xdpf->mem; > - /* TODO: report queue_index to xdp_rxq_info */ > + rxq.queue_index = xdpf->queue_index; > > xdp_convert_frame_to_buff(xdpf, &xdp); > > -- > 2.37.1 >
On 8/16/22 11:45 AM, Lorenzo Bianconi wrote: >> Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info >> pointer. xdp_frame queue_index is currently used in cpumap code to covert >> the xdp_frame into a xdp_buff. >> xdp_frame size is not increased adding queue_index since an alignment padding >> in the structure is used to insert queue_index field. >> >> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> >> --- >> include/net/xdp.h | 2 ++ >> kernel/bpf/cpumap.c | 2 +- >> 2 files changed, 3 insertions(+), 1 deletion(-) > > > Hi Alexei and Daniel, > > this patch is marked as 'new, archived' in patchwork. > Do I need to rebase and repost it? Yes, please rebase and resend. Perhaps also improve the commit description a bit in terms of what it fixes, it's a bit terse to the reader above on what effect it has. Thanks, Daniel
> On 8/16/22 11:45 AM, Lorenzo Bianconi wrote: > > > Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info > > > pointer. xdp_frame queue_index is currently used in cpumap code to covert > > > the xdp_frame into a xdp_buff. > > > xdp_frame size is not increased adding queue_index since an alignment padding > > > in the structure is used to insert queue_index field. > > > > > > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> > > > --- > > > include/net/xdp.h | 2 ++ > > > kernel/bpf/cpumap.c | 2 +- > > > 2 files changed, 3 insertions(+), 1 deletion(-) > > > > > > Hi Alexei and Daniel, > > > > this patch is marked as 'new, archived' in patchwork. > > Do I need to rebase and repost it? > > Yes, please rebase and resend. Perhaps also improve the commit description > a bit in terms of what it fixes, it's a bit terse to the reader above on > what effect it has. ack thx, will do. Regards, Lorenzo > > Thanks, > Daniel
On 25/07/2022 12.56, Lorenzo Bianconi wrote: > Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info > pointer. xdp_frame queue_index is currently used in cpumap code to covert > the xdp_frame into a xdp_buff. Hmm, I'm unsure about this change, because the XDP-hints will also contain the rx_queue number. I do think it is relevant for the BPF-prog to get access to the rx_queue index, because it can be used for scaling the workload. > xdp_frame size is not increased adding queue_index since an alignment padding > in the structure is used to insert queue_index field. The rx_queue could be reduced from u32 to u16, but it might be faster to keep it u32, and reduce it when others need the space. > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> > --- > include/net/xdp.h | 2 ++ > kernel/bpf/cpumap.c | 2 +- > 2 files changed, 3 insertions(+), 1 deletion(-) > > diff --git a/include/net/xdp.h b/include/net/xdp.h > index 04c852c7a77f..3567866b0af5 100644 > --- a/include/net/xdp.h > +++ b/include/net/xdp.h > @@ -172,6 +172,7 @@ struct xdp_frame { > struct xdp_mem_info mem; > struct net_device *dev_rx; /* used by cpumap */ > u32 flags; /* supported values defined in xdp_buff_flags */ > + u32 queue_index; > }; > > static __always_inline bool xdp_frame_has_frags(struct xdp_frame *frame) > @@ -301,6 +302,7 @@ struct xdp_frame *xdp_convert_buff_to_frame(struct xdp_buff *xdp) > > /* rxq only valid until napi_schedule ends, convert to xdp_mem_info */ > xdp_frame->mem = xdp->rxq->mem; > + xdp_frame->queue_index = xdp->rxq->queue_index; > > return xdp_frame; > } > diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c > index f4860ac756cd..09a792d088b3 100644 > --- a/kernel/bpf/cpumap.c > +++ b/kernel/bpf/cpumap.c > @@ -228,7 +228,7 @@ static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu, > > rxq.dev = xdpf->dev_rx; > rxq.mem = xdpf->mem; > - /* TODO: report queue_index to xdp_rxq_info */ > + rxq.queue_index = xdpf->queue_index; > > xdp_convert_frame_to_buff(xdpf, &xdp); >
> > On 25/07/2022 12.56, Lorenzo Bianconi wrote: > > Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info > > pointer. xdp_frame queue_index is currently used in cpumap code to covert > > the xdp_frame into a xdp_buff. > > Hmm, I'm unsure about this change, because the XDP-hints will also > contain the rx_queue number. ack, but in this way each driver needs to fill this info, right? Maybe we can keep rx_queue number here and remove it from XDP-hints metadata? > > I do think it is relevant for the BPF-prog to get access to the rx_queue > index, because it can be used for scaling the workload. > > > xdp_frame size is not increased adding queue_index since an alignment padding > > in the structure is used to insert queue_index field. > > The rx_queue could be reduced from u32 to u16, but it might be faster to > keep it u32, and reduce it when others need the space. ack, agree. Regards, Lorenzo > > > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> > > --- > > include/net/xdp.h | 2 ++ > > kernel/bpf/cpumap.c | 2 +- > > 2 files changed, 3 insertions(+), 1 deletion(-) > > > > diff --git a/include/net/xdp.h b/include/net/xdp.h > > index 04c852c7a77f..3567866b0af5 100644 > > --- a/include/net/xdp.h > > +++ b/include/net/xdp.h > > @@ -172,6 +172,7 @@ struct xdp_frame { > > struct xdp_mem_info mem; > > struct net_device *dev_rx; /* used by cpumap */ > > u32 flags; /* supported values defined in xdp_buff_flags */ > > + u32 queue_index; > > }; > > static __always_inline bool xdp_frame_has_frags(struct xdp_frame *frame) > > @@ -301,6 +302,7 @@ struct xdp_frame *xdp_convert_buff_to_frame(struct xdp_buff *xdp) > > /* rxq only valid until napi_schedule ends, convert to xdp_mem_info */ > > xdp_frame->mem = xdp->rxq->mem; > > + xdp_frame->queue_index = xdp->rxq->queue_index; > > return xdp_frame; > > } > > diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c > > index f4860ac756cd..09a792d088b3 100644 > > --- a/kernel/bpf/cpumap.c > > +++ b/kernel/bpf/cpumap.c > > @@ -228,7 +228,7 @@ static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu, > > rxq.dev = xdpf->dev_rx; > > rxq.mem = xdpf->mem; > > - /* TODO: report queue_index to xdp_rxq_info */ > > + rxq.queue_index = xdpf->queue_index; > > xdp_convert_frame_to_buff(xdpf, &xdp); >
diff --git a/include/net/xdp.h b/include/net/xdp.h index 04c852c7a77f..3567866b0af5 100644 --- a/include/net/xdp.h +++ b/include/net/xdp.h @@ -172,6 +172,7 @@ struct xdp_frame { struct xdp_mem_info mem; struct net_device *dev_rx; /* used by cpumap */ u32 flags; /* supported values defined in xdp_buff_flags */ + u32 queue_index; }; static __always_inline bool xdp_frame_has_frags(struct xdp_frame *frame) @@ -301,6 +302,7 @@ struct xdp_frame *xdp_convert_buff_to_frame(struct xdp_buff *xdp) /* rxq only valid until napi_schedule ends, convert to xdp_mem_info */ xdp_frame->mem = xdp->rxq->mem; + xdp_frame->queue_index = xdp->rxq->queue_index; return xdp_frame; } diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c index f4860ac756cd..09a792d088b3 100644 --- a/kernel/bpf/cpumap.c +++ b/kernel/bpf/cpumap.c @@ -228,7 +228,7 @@ static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu, rxq.dev = xdpf->dev_rx; rxq.mem = xdpf->mem; - /* TODO: report queue_index to xdp_rxq_info */ + rxq.queue_index = xdpf->queue_index; xdp_convert_frame_to_buff(xdpf, &xdp);
Report rx queue index in xdp_frame according to the xdp_buff xdp_rxq_info pointer. xdp_frame queue_index is currently used in cpumap code to covert the xdp_frame into a xdp_buff. xdp_frame size is not increased adding queue_index since an alignment padding in the structure is used to insert queue_index field. Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> --- include/net/xdp.h | 2 ++ kernel/bpf/cpumap.c | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-)