From: Tan Tee Min <tee.min.tan@intel.com> To: Giuseppe Cavallaro <peppe.cavallaro@st.com>, Alexandre Torgue <alexandre.torgue@st.com>, Jose Abreu <joabreu@synopsys.com>, "David S . Miller" <davem@davemloft.net>, Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>, Maxime Coquelin <mcoquelin.stm32@gmail.com>, Rayagond Kokatanur <rayagond@vayavyalabs.com>, Richard Cochran <richardcochran@gmail.com> Cc: netdev@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, stable@vger.kernel.org, Voon Wei Feng <weifeng.voon@intel.com>, Wong Vee Khee <vee.khee.wong@intel.com>, Song Yoong Siang <yoong.siang.song@intel.com>, Tan Tee Min <tee.min.tan@intel.com> Subject: [PATCH net 1/1] net: stmmac: add fsleep() in HW Rx timestamp checking loop Date: Wed, 13 Apr 2022 12:01:15 +0800 [thread overview] Message-ID: <20220413040115.2351987-1-tee.min.tan@intel.com> (raw) There is a possibility that the context descriptor still owned by the DMA even the previous normal descriptor own bit is already cleared. Checking the context descriptor readiness without delay might be not enough time for the DMA to update the descriptor field, which causing failure in getting HW Rx timestamp. This patch introduces a 1us fsleep() in HW Rx timestamp checking loop to give time for DMA to update/complete the context descriptor. ptp4l Timestamp log without this patch: ----------------------------------------------------------- $ echo 10000 > /sys/class/net/enp0s30f4/gro_flush_timeout $ echo 10000 > /sys/class/net/enp0s30f4/napi_defer_hard_irqs $ ptp4l -P2Hi enp0s30f4 --step_threshold=1 -m ptp4l: selected /dev/ptp2 as PTP clock ptp4l: port 1: INITIALIZING to LISTENING on INIT_COMPLETE ptp4l: selected local clock 901210.fffe.b57df7 as best master ptp4l: port 1: new foreign master 22bb22.fffe.bb22bb-1 ptp4l: selected best master clock 22bb22.fffe.bb22bb ptp4l: port 1: LISTENING to UNCALIBRATED on RS_SLAVE ptp4l: port 1: UNCALIBRATED to SLAVE on MASTER_CLOCK_SELECTED ptp4l: port 1: received SYNC without timestamp ptp4l: rms 49 max 63 freq -9573 +/- 34 delay 71 +/- 1 ptp4l: rms 15 max 25 freq -9553 +/- 20 delay 72 +/- 0 ptp4l: port 1: received SYNC without timestamp ptp4l: rms 9 max 18 freq -9540 +/- 11 delay 70 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ptp4l: rms 16 max 29 freq -9519 +/- 12 delay 72 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ptp4l: rms 9 max 18 freq -9527 +/- 12 delay 72 +/- 0 ptp4l: rms 5 max 9 freq -9530 +/- 7 delay 70 +/- 0 ptp4l: rms 11 max 20 freq -9530 +/- 16 delay 72 +/- 0 ptp4l: rms 5 max 11 freq -9530 +/- 7 delay 74 +/- 0 ptp4l: rms 6 max 9 freq -9522 +/- 7 delay 72 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ----------------------------------------------------------- ptp4l Timestamp log with this patch: ----------------------------------------------------------- $ echo 10000 > /sys/class/net/enp0s30f4/gro_flush_timeout $ echo 10000 > /sys/class/net/enp0s30f4/napi_defer_hard_irqs $ ptp4l -P2Hi enp0s30f4 --step_threshold=1 -m ptp4l: selected /dev/ptp2 as PTP clock ptp4l: port 1: INITIALIZING to LISTENING on INIT_COMPLETE ptp4l: selected local clock 901210.fffe.b57df7 as best master ptp4l: port 1: new foreign master 22bb22.fffe.bb22bb-1 ptp4l: selected best master clock 22bb22.fffe.bb22bb ptp4l: port 1: LISTENING to UNCALIBRATED on RS_SLAVE ptp4l: port 1: UNCALIBRATED to SLAVE on MASTER_CLOCK_SELECTED ptp4l: rms 30 max 45 freq -9400 +/- 23 delay 72 +/- 0 ptp4l: rms 7 max 16 freq -9414 +/- 10 delay 70 +/- 0 ptp4l: rms 6 max 9 freq -9422 +/- 6 delay 72 +/- 0 ptp4l: rms 13 max 20 freq -9436 +/- 13 delay 74 +/- 0 ptp4l: rms 12 max 27 freq -9446 +/- 11 delay 72 +/- 0 ptp4l: rms 9 max 12 freq -9453 +/- 6 delay 74 +/- 0 ptp4l: rms 9 max 15 freq -9438 +/- 11 delay 74 +/- 0 ptp4l: rms 10 max 16 freq -9435 +/- 12 delay 74 +/- 0 ptp4l: rms 8 max 18 freq -9428 +/- 8 delay 72 +/- 0 ptp4l: rms 8 max 18 freq -9423 +/- 8 delay 72 +/- 0 ptp4l: rms 9 max 16 freq -9431 +/- 12 delay 70 +/- 0 ptp4l: rms 9 max 18 freq -9441 +/- 9 delay 72 +/- 0 ----------------------------------------------------------- Fixes: ba1ffd74df74 ("stmmac: fix PTP support for GMAC4") Cc: <stable@vger.kernel.org> # 5.4.x Signed-off-by: Song Yoong Siang <yoong.siang.song@intel.com> Signed-off-by: Tan Tee Min <tee.min.tan@intel.com> --- drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c index d3b4765c1a5b..289bf26a6105 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c @@ -279,10 +279,11 @@ static int dwmac4_wrback_get_rx_timestamp_status(void *desc, void *next_desc, /* Check if timestamp is OK from context descriptor */ do { ret = dwmac4_rx_check_timestamp(next_desc); - if (ret < 0) + if (ret <= 0) goto exit; i++; + fsleep(1); } while ((ret == 1) && (i < 10)); if (i == 10) -- 2.25.1
WARNING: multiple messages have this Message-ID (diff)
From: Tan Tee Min <tee.min.tan@intel.com> To: Giuseppe Cavallaro <peppe.cavallaro@st.com>, Alexandre Torgue <alexandre.torgue@st.com>, Jose Abreu <joabreu@synopsys.com>, "David S . Miller" <davem@davemloft.net>, Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>, Maxime Coquelin <mcoquelin.stm32@gmail.com>, Rayagond Kokatanur <rayagond@vayavyalabs.com>, Richard Cochran <richardcochran@gmail.com> Cc: netdev@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, stable@vger.kernel.org, Voon Wei Feng <weifeng.voon@intel.com>, Wong Vee Khee <vee.khee.wong@intel.com>, Song Yoong Siang <yoong.siang.song@intel.com>, Tan Tee Min <tee.min.tan@intel.com> Subject: [PATCH net 1/1] net: stmmac: add fsleep() in HW Rx timestamp checking loop Date: Wed, 13 Apr 2022 12:01:15 +0800 [thread overview] Message-ID: <20220413040115.2351987-1-tee.min.tan@intel.com> (raw) There is a possibility that the context descriptor still owned by the DMA even the previous normal descriptor own bit is already cleared. Checking the context descriptor readiness without delay might be not enough time for the DMA to update the descriptor field, which causing failure in getting HW Rx timestamp. This patch introduces a 1us fsleep() in HW Rx timestamp checking loop to give time for DMA to update/complete the context descriptor. ptp4l Timestamp log without this patch: ----------------------------------------------------------- $ echo 10000 > /sys/class/net/enp0s30f4/gro_flush_timeout $ echo 10000 > /sys/class/net/enp0s30f4/napi_defer_hard_irqs $ ptp4l -P2Hi enp0s30f4 --step_threshold=1 -m ptp4l: selected /dev/ptp2 as PTP clock ptp4l: port 1: INITIALIZING to LISTENING on INIT_COMPLETE ptp4l: selected local clock 901210.fffe.b57df7 as best master ptp4l: port 1: new foreign master 22bb22.fffe.bb22bb-1 ptp4l: selected best master clock 22bb22.fffe.bb22bb ptp4l: port 1: LISTENING to UNCALIBRATED on RS_SLAVE ptp4l: port 1: UNCALIBRATED to SLAVE on MASTER_CLOCK_SELECTED ptp4l: port 1: received SYNC without timestamp ptp4l: rms 49 max 63 freq -9573 +/- 34 delay 71 +/- 1 ptp4l: rms 15 max 25 freq -9553 +/- 20 delay 72 +/- 0 ptp4l: port 1: received SYNC without timestamp ptp4l: rms 9 max 18 freq -9540 +/- 11 delay 70 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ptp4l: rms 16 max 29 freq -9519 +/- 12 delay 72 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ptp4l: rms 9 max 18 freq -9527 +/- 12 delay 72 +/- 0 ptp4l: rms 5 max 9 freq -9530 +/- 7 delay 70 +/- 0 ptp4l: rms 11 max 20 freq -9530 +/- 16 delay 72 +/- 0 ptp4l: rms 5 max 11 freq -9530 +/- 7 delay 74 +/- 0 ptp4l: rms 6 max 9 freq -9522 +/- 7 delay 72 +/- 0 ptp4l: port 1: received PDELAY_REQ without timestamp ----------------------------------------------------------- ptp4l Timestamp log with this patch: ----------------------------------------------------------- $ echo 10000 > /sys/class/net/enp0s30f4/gro_flush_timeout $ echo 10000 > /sys/class/net/enp0s30f4/napi_defer_hard_irqs $ ptp4l -P2Hi enp0s30f4 --step_threshold=1 -m ptp4l: selected /dev/ptp2 as PTP clock ptp4l: port 1: INITIALIZING to LISTENING on INIT_COMPLETE ptp4l: selected local clock 901210.fffe.b57df7 as best master ptp4l: port 1: new foreign master 22bb22.fffe.bb22bb-1 ptp4l: selected best master clock 22bb22.fffe.bb22bb ptp4l: port 1: LISTENING to UNCALIBRATED on RS_SLAVE ptp4l: port 1: UNCALIBRATED to SLAVE on MASTER_CLOCK_SELECTED ptp4l: rms 30 max 45 freq -9400 +/- 23 delay 72 +/- 0 ptp4l: rms 7 max 16 freq -9414 +/- 10 delay 70 +/- 0 ptp4l: rms 6 max 9 freq -9422 +/- 6 delay 72 +/- 0 ptp4l: rms 13 max 20 freq -9436 +/- 13 delay 74 +/- 0 ptp4l: rms 12 max 27 freq -9446 +/- 11 delay 72 +/- 0 ptp4l: rms 9 max 12 freq -9453 +/- 6 delay 74 +/- 0 ptp4l: rms 9 max 15 freq -9438 +/- 11 delay 74 +/- 0 ptp4l: rms 10 max 16 freq -9435 +/- 12 delay 74 +/- 0 ptp4l: rms 8 max 18 freq -9428 +/- 8 delay 72 +/- 0 ptp4l: rms 8 max 18 freq -9423 +/- 8 delay 72 +/- 0 ptp4l: rms 9 max 16 freq -9431 +/- 12 delay 70 +/- 0 ptp4l: rms 9 max 18 freq -9441 +/- 9 delay 72 +/- 0 ----------------------------------------------------------- Fixes: ba1ffd74df74 ("stmmac: fix PTP support for GMAC4") Cc: <stable@vger.kernel.org> # 5.4.x Signed-off-by: Song Yoong Siang <yoong.siang.song@intel.com> Signed-off-by: Tan Tee Min <tee.min.tan@intel.com> --- drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c index d3b4765c1a5b..289bf26a6105 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c @@ -279,10 +279,11 @@ static int dwmac4_wrback_get_rx_timestamp_status(void *desc, void *next_desc, /* Check if timestamp is OK from context descriptor */ do { ret = dwmac4_rx_check_timestamp(next_desc); - if (ret < 0) + if (ret <= 0) goto exit; i++; + fsleep(1); } while ((ret == 1) && (i < 10)); if (i == 10) -- 2.25.1 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next reply other threads:[~2022-04-13 3:55 UTC|newest] Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-04-13 4:01 Tan Tee Min [this message] 2022-04-13 4:01 ` [PATCH net 1/1] net: stmmac: add fsleep() in HW Rx timestamp checking loop Tan Tee Min 2022-04-13 12:59 ` Richard Cochran 2022-04-13 12:59 ` Richard Cochran 2022-04-14 7:29 ` Tan Tee Min 2022-04-14 7:29 ` Tan Tee Min 2022-04-14 8:42 ` Jakub Kicinski 2022-04-14 8:42 ` Jakub Kicinski 2022-04-19 0:52 ` Tan Tee Min 2022-04-19 0:52 ` Tan Tee Min 2022-04-19 13:28 ` Richard Cochran 2022-04-19 13:28 ` Richard Cochran 2022-04-20 5:15 ` Tan Tee Min 2022-04-20 5:15 ` Tan Tee Min 2022-04-20 12:54 ` Richard Cochran 2022-04-20 12:54 ` Richard Cochran
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=20220413040115.2351987-1-tee.min.tan@intel.com \ --to=tee.min.tan@intel.com \ --cc=alexandre.torgue@st.com \ --cc=davem@davemloft.net \ --cc=joabreu@synopsys.com \ --cc=kuba@kernel.org \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-stm32@st-md-mailman.stormreply.com \ --cc=mcoquelin.stm32@gmail.com \ --cc=netdev@vger.kernel.org \ --cc=pabeni@redhat.com \ --cc=peppe.cavallaro@st.com \ --cc=rayagond@vayavyalabs.com \ --cc=richardcochran@gmail.com \ --cc=stable@vger.kernel.org \ --cc=vee.khee.wong@intel.com \ --cc=weifeng.voon@intel.com \ --cc=yoong.siang.song@intel.com \ /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.