From: "Björn Töpel" <bjorn.topel@intel.com> To: "Jesper Dangaard Brouer" <brouer@redhat.com>, "Björn Töpel" <bjorn.topel@gmail.com> Cc: ast@kernel.org, daniel@iogearbox.net, netdev@vger.kernel.org, bpf@vger.kernel.org, magnus.karlsson@intel.com, davem@davemloft.net, kuba@kernel.org, john.fastabend@gmail.com, intel-wired-lan@lists.osuosl.org Subject: Re: [PATCH bpf-next 6/6] ixgbe, xsk: finish napi loop if AF_XDP Rx queue is full Date: Fri, 4 Sep 2020 17:54:42 +0200 [thread overview] Message-ID: <59da4aa6-dbc5-b366-e84e-0030f6010e55@intel.com> (raw) In-Reply-To: <20200904173540.3a617eee@carbon> On 2020-09-04 17:35, Jesper Dangaard Brouer wrote: > On Fri, 4 Sep 2020 15:53:31 +0200 > Björn Töpel <bjorn.topel@gmail.com> wrote: > >> From: Björn Töpel <bjorn.topel@intel.com> >> >> Make the AF_XDP zero-copy path aware that the reason for redirect >> failure was due to full Rx queue. If so, exit the napi loop as soon as >> possible (exit the softirq processing), so that the userspace AF_XDP >> process can hopefully empty the Rx queue. This mainly helps the "one >> core scenario", where the userland process and Rx softirq processing >> is on the same core. >> >> Note that the early exit can only be performed if the "need wakeup" >> feature is enabled, because otherwise there is no notification >> mechanism available from the kernel side. >> >> This requires that the driver starts using the newly introduced >> xdp_do_redirect_ext() and xsk_do_redirect_rx_full() functions. >> >> Signed-off-by: Björn Töpel <bjorn.topel@intel.com> >> --- >> drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c | 23 ++++++++++++++------ >> 1 file changed, 16 insertions(+), 7 deletions(-) >> >> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> index 3771857cf887..a4aebfd986b3 100644 >> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> @@ -93,9 +93,11 @@ int ixgbe_xsk_pool_setup(struct ixgbe_adapter *adapter, >> >> static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, >> struct ixgbe_ring *rx_ring, >> - struct xdp_buff *xdp) >> + struct xdp_buff *xdp, >> + bool *early_exit) >> { >> int err, result = IXGBE_XDP_PASS; >> + enum bpf_map_type map_type; >> struct bpf_prog *xdp_prog; >> struct xdp_frame *xdpf; >> u32 act; >> @@ -116,8 +118,13 @@ static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, >> result = ixgbe_xmit_xdp_ring(adapter, xdpf); >> break; >> case XDP_REDIRECT: >> - err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog); >> - result = !err ? IXGBE_XDP_REDIR : IXGBE_XDP_CONSUMED; >> + err = xdp_do_redirect_ext(rx_ring->netdev, xdp, xdp_prog, &map_type); >> + if (err) { >> + *early_exit = xsk_do_redirect_rx_full(err, map_type); > > Have you tried calling xdp_do_flush (that calls __xsk_map_flush()) and > (I guess) xsk_set_rx_need_wakeup() here, instead of stopping the loop? > (Or doing this in xsk core). > Moving the need_wake logic to the xsk core/flush would be a very nice cleanup. The driver would still need to pass information from the driver though. Still, much cleaner. I'll take a stab at that. Thanks! > Looking at the code, the AF_XDP frames are "published" in the queue > rather late for AF_XDP. Maybe in an orthogonal optimization, have you > considered "publishing" the ring producer when e.g. the queue is > half-full? > Hmm, I haven't. You mean instead of yielding, you publish/submit? I *think* I still prefer stopping the processing. I'll play with this a bit! Very nice suggestions, Jesper! Thanks! Björn
WARNING: multiple messages have this Message-ID (diff)
From: =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= <bjorn.topel@intel.com> To: intel-wired-lan@osuosl.org Subject: [Intel-wired-lan] [PATCH bpf-next 6/6] ixgbe, xsk: finish napi loop if AF_XDP Rx queue is full Date: Fri, 4 Sep 2020 17:54:42 +0200 [thread overview] Message-ID: <59da4aa6-dbc5-b366-e84e-0030f6010e55@intel.com> (raw) In-Reply-To: <20200904173540.3a617eee@carbon> On 2020-09-04 17:35, Jesper Dangaard Brouer wrote: > On Fri, 4 Sep 2020 15:53:31 +0200 > Bj?rn T?pel <bjorn.topel@gmail.com> wrote: > >> From: Bj?rn T?pel <bjorn.topel@intel.com> >> >> Make the AF_XDP zero-copy path aware that the reason for redirect >> failure was due to full Rx queue. If so, exit the napi loop as soon as >> possible (exit the softirq processing), so that the userspace AF_XDP >> process can hopefully empty the Rx queue. This mainly helps the "one >> core scenario", where the userland process and Rx softirq processing >> is on the same core. >> >> Note that the early exit can only be performed if the "need wakeup" >> feature is enabled, because otherwise there is no notification >> mechanism available from the kernel side. >> >> This requires that the driver starts using the newly introduced >> xdp_do_redirect_ext() and xsk_do_redirect_rx_full() functions. >> >> Signed-off-by: Bj?rn T?pel <bjorn.topel@intel.com> >> --- >> drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c | 23 ++++++++++++++------ >> 1 file changed, 16 insertions(+), 7 deletions(-) >> >> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> index 3771857cf887..a4aebfd986b3 100644 >> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c >> @@ -93,9 +93,11 @@ int ixgbe_xsk_pool_setup(struct ixgbe_adapter *adapter, >> >> static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, >> struct ixgbe_ring *rx_ring, >> - struct xdp_buff *xdp) >> + struct xdp_buff *xdp, >> + bool *early_exit) >> { >> int err, result = IXGBE_XDP_PASS; >> + enum bpf_map_type map_type; >> struct bpf_prog *xdp_prog; >> struct xdp_frame *xdpf; >> u32 act; >> @@ -116,8 +118,13 @@ static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, >> result = ixgbe_xmit_xdp_ring(adapter, xdpf); >> break; >> case XDP_REDIRECT: >> - err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog); >> - result = !err ? IXGBE_XDP_REDIR : IXGBE_XDP_CONSUMED; >> + err = xdp_do_redirect_ext(rx_ring->netdev, xdp, xdp_prog, &map_type); >> + if (err) { >> + *early_exit = xsk_do_redirect_rx_full(err, map_type); > > Have you tried calling xdp_do_flush (that calls __xsk_map_flush()) and > (I guess) xsk_set_rx_need_wakeup() here, instead of stopping the loop? > (Or doing this in xsk core). > Moving the need_wake logic to the xsk core/flush would be a very nice cleanup. The driver would still need to pass information from the driver though. Still, much cleaner. I'll take a stab at that. Thanks! > Looking at the code, the AF_XDP frames are "published" in the queue > rather late for AF_XDP. Maybe in an orthogonal optimization, have you > considered "publishing" the ring producer when e.g. the queue is > half-full? > Hmm, I haven't. You mean instead of yielding, you publish/submit? I *think* I still prefer stopping the processing. I'll play with this a bit! Very nice suggestions, Jesper! Thanks! Bj?rn
next prev parent reply other threads:[~2020-09-04 15:54 UTC|newest] Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top 2020-09-04 13:53 [PATCH bpf-next 0/6] xsk: exit NAPI loop when AF_XDP Rx ring is full Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:53 ` [PATCH bpf-next 1/6] xsk: improve xdp_do_redirect() error codes Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:53 ` [PATCH bpf-next 2/6] xdp: introduce xdp_do_redirect_ext() function Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:53 ` [PATCH bpf-next 3/6] xsk: introduce xsk_do_redirect_rx_full() helper Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 15:11 ` Jesper Dangaard Brouer 2020-09-04 15:11 ` [Intel-wired-lan] " Jesper Dangaard Brouer 2020-09-04 15:39 ` Björn Töpel 2020-09-04 15:39 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-07 12:45 ` Jesper Dangaard Brouer 2020-09-07 12:45 ` [Intel-wired-lan] " Jesper Dangaard Brouer 2020-09-04 13:53 ` [PATCH bpf-next 4/6] i40e, xsk: finish napi loop if AF_XDP Rx queue is full Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:53 ` [PATCH bpf-next 5/6] ice, " Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:53 ` [PATCH bpf-next 6/6] ixgbe, " Björn Töpel 2020-09-04 13:53 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 15:35 ` Jesper Dangaard Brouer 2020-09-04 15:35 ` [Intel-wired-lan] " Jesper Dangaard Brouer 2020-09-04 15:54 ` Björn Töpel [this message] 2020-09-04 15:54 ` =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 13:59 ` [PATCH bpf-next 0/6] xsk: exit NAPI loop when AF_XDP Rx ring " Björn Töpel 2020-09-04 13:59 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-08 10:32 ` Maxim Mikityanskiy 2020-09-08 10:32 ` [Intel-wired-lan] " Maxim Mikityanskiy 2020-09-08 11:37 ` Magnus Karlsson 2020-09-08 11:37 ` [Intel-wired-lan] " Magnus Karlsson 2020-09-08 12:21 ` Björn Töpel 2020-09-08 12:21 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-09 15:37 ` Jesper Dangaard Brouer 2020-09-09 15:37 ` [Intel-wired-lan] " Jesper Dangaard Brouer 2020-09-04 14:27 ` Jesper Dangaard Brouer 2020-09-04 14:27 ` [Intel-wired-lan] " Jesper Dangaard Brouer 2020-09-04 14:32 ` Björn Töpel 2020-09-04 14:32 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-04 23:58 ` Jakub Kicinski 2020-09-04 23:58 ` [Intel-wired-lan] " Jakub Kicinski 2020-09-07 13:37 ` Björn Töpel 2020-09-07 13:37 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-07 18:40 ` Jakub Kicinski 2020-09-07 18:40 ` [Intel-wired-lan] " Jakub Kicinski 2020-09-08 6:58 ` Björn Töpel 2020-09-08 6:58 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-08 17:24 ` Jakub Kicinski 2020-09-08 17:24 ` [Intel-wired-lan] " Jakub Kicinski 2020-09-08 18:28 ` Björn Töpel 2020-09-08 18:28 ` [Intel-wired-lan] " =?unknown-8bit?q?Bj=C3=B6rn_T=C3=B6pel?= 2020-09-08 18:34 ` Jakub Kicinski 2020-09-08 18:34 ` [Intel-wired-lan] " Jakub Kicinski
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=59da4aa6-dbc5-b366-e84e-0030f6010e55@intel.com \ --to=bjorn.topel@intel.com \ --cc=ast@kernel.org \ --cc=bjorn.topel@gmail.com \ --cc=bpf@vger.kernel.org \ --cc=brouer@redhat.com \ --cc=daniel@iogearbox.net \ --cc=davem@davemloft.net \ --cc=intel-wired-lan@lists.osuosl.org \ --cc=john.fastabend@gmail.com \ --cc=kuba@kernel.org \ --cc=magnus.karlsson@intel.com \ --cc=netdev@vger.kernel.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.