stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org,
	Sergej Benilov <sergej.benilov@googlemail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Sasha Levin <sashal@kernel.org>
Subject: [PATCH 4.4 33/40] sis900: fix TX completion
Date: Thu, 18 Jul 2019 12:02:29 +0900	[thread overview]
Message-ID: <20190718030049.943975637@linuxfoundation.org> (raw)
In-Reply-To: <20190718030039.676518610@linuxfoundation.org>

[ Upstream commit 8ac8a01092b2added0749ef937037bf1912e13e3 ]

Since commit 605ad7f184b60cfaacbc038aa6c55ee68dee3c89 "tcp: refine TSO autosizing",
outbound throughput is dramatically reduced for some connections, as sis900
is doing TX completion within idle states only.

Make TX completion happen after every transmitted packet.

Test:
netperf

before patch:
> netperf -H remote -l -2000000 -- -s 1000000
MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to 95.223.112.76 () port 0 AF_INET : demo
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380 327680 327680    253.44      0.06

after patch:
> netperf -H remote -l -10000000 -- -s 1000000
MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to 95.223.112.76 () port 0 AF_INET : demo
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380 327680 327680    5.38       14.89

Thx to Dave Miller and Eric Dumazet for helpful hints

Signed-off-by: Sergej Benilov <sergej.benilov@googlemail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/net/ethernet/sis/sis900.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/sis/sis900.c b/drivers/net/ethernet/sis/sis900.c
index fd812d2e5e1c..dff5b56738d3 100644
--- a/drivers/net/ethernet/sis/sis900.c
+++ b/drivers/net/ethernet/sis/sis900.c
@@ -1058,7 +1058,7 @@ sis900_open(struct net_device *net_dev)
 	sis900_set_mode(sis_priv, HW_SPEED_10_MBPS, FDX_CAPABLE_HALF_SELECTED);
 
 	/* Enable all known interrupts by setting the interrupt mask. */
-	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE);
+	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE | TxDESC);
 	sw32(cr, RxENA | sr32(cr));
 	sw32(ier, IE);
 
@@ -1581,7 +1581,7 @@ static void sis900_tx_timeout(struct net_device *net_dev)
 	sw32(txdp, sis_priv->tx_ring_dma);
 
 	/* Enable all known interrupts by setting the interrupt mask. */
-	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE);
+	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE | TxDESC);
 }
 
 /**
@@ -1621,7 +1621,7 @@ sis900_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
 			spin_unlock_irqrestore(&sis_priv->lock, flags);
 			return NETDEV_TX_OK;
 	}
-	sis_priv->tx_ring[entry].cmdsts = (OWN | skb->len);
+	sis_priv->tx_ring[entry].cmdsts = (OWN | INTR | skb->len);
 	sw32(cr, TxENA | sr32(cr));
 
 	sis_priv->cur_tx ++;
@@ -1677,7 +1677,7 @@ static irqreturn_t sis900_interrupt(int irq, void *dev_instance)
 	do {
 		status = sr32(isr);
 
-		if ((status & (HIBERR|TxURN|TxERR|TxIDLE|RxORN|RxERR|RxOK)) == 0)
+		if ((status & (HIBERR|TxURN|TxERR|TxIDLE|TxDESC|RxORN|RxERR|RxOK)) == 0)
 			/* nothing intresting happened */
 			break;
 		handled = 1;
@@ -1687,7 +1687,7 @@ static irqreturn_t sis900_interrupt(int irq, void *dev_instance)
 			/* Rx interrupt */
 			sis900_rx(net_dev);
 
-		if (status & (TxURN | TxERR | TxIDLE))
+		if (status & (TxURN | TxERR | TxIDLE | TxDESC))
 			/* Tx interrupt */
 			sis900_finish_xmit(net_dev);
 
@@ -1899,8 +1899,8 @@ static void sis900_finish_xmit (struct net_device *net_dev)
 
 		if (tx_status & OWN) {
 			/* The packet is not transmitted yet (owned by hardware) !
-			 * Note: the interrupt is generated only when Tx Machine
-			 * is idle, so this is an almost impossible case */
+			 * Note: this is an almost impossible condition
+			 * in case of TxDESC ('descriptor interrupt') */
 			break;
 		}
 
@@ -2476,7 +2476,7 @@ static int sis900_resume(struct pci_dev *pci_dev)
 	sis900_set_mode(sis_priv, HW_SPEED_10_MBPS, FDX_CAPABLE_HALF_SELECTED);
 
 	/* Enable all known interrupts by setting the interrupt mask. */
-	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE);
+	sw32(imr, RxSOVR | RxORN | RxERR | RxOK | TxURN | TxERR | TxIDLE | TxDESC);
 	sw32(cr, RxENA | sr32(cr));
 	sw32(ier, IE);
 
-- 
2.20.1




  parent reply	other threads:[~2019-07-18  3:16 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-18  3:01 [PATCH 4.4 00/40] 4.4.186-stable review Greg Kroah-Hartman
2019-07-18  3:01 ` [PATCH 4.4 01/40] Input: elantech - enable middle button support on 2 ThinkPads Greg Kroah-Hartman
2019-07-18  3:01 ` [PATCH 4.4 02/40] samples, bpf: fix to change the buffer size for read() Greg Kroah-Hartman
2019-07-18  3:01 ` [PATCH 4.4 03/40] mac80211: mesh: fix RCU warning Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 04/40] mwifiex: Fix possible buffer overflows at parsing bss descriptor Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 05/40] dt-bindings: can: mcp251x: add mcp25625 support Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 06/40] can: mcp251x: add support for mcp25625 Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 07/40] Input: imx_keypad - make sure keyboard can always wake up system Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 08/40] ARM: davinci: da850-evm: call regulator_has_full_constraints() Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 09/40] ARM: davinci: da8xx: specify dma_coherent_mask for lcdc Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 10/40] md: fix for divide error in status_resync Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 11/40] bnx2x: Check if transceiver implements DDM before access Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 12/40] udf: Fix incorrect final NOT_ALLOCATED (hole) extent length Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 13/40] x86/ptrace: Fix possible spectre-v1 in ptrace_get_debugreg() Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 14/40] x86/tls: Fix possible spectre-v1 in do_get_thread_area() Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 15/40] mwifiex: Abort at too short BSS descriptor element Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 16/40] mwifiex: Fix heap overflow in mwifiex_uap_parse_tail_ies() Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 17/40] fscrypt: dont set policy for a dead directory Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 18/40] mwifiex: Dont abort on small, spec-compliant vendor IEs Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 19/40] USB: serial: ftdi_sio: add ID for isodebug v1 Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 20/40] USB: serial: option: add support for GosunCn ME3630 RNDIS mode Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 21/40] usb: gadget: ether: Fix race between gether_disconnect and rx_submit Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 22/40] usb: renesas_usbhs: add a workaround for a race condition of workqueue Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 23/40] staging: comedi: dt282x: fix a null pointer deref on interrupt Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 24/40] staging: comedi: amplc_pci230: fix " Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 25/40] carl9170: fix misuse of device driver API Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 26/40] VMCI: Fix integer overflow in VMCI handle arrays Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 27/40] MIPS: Remove superfluous check for __linux__ Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 28/40] e1000e: start network tx queue only when link is up Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 29/40] perf/core: Fix perf_sample_regs_user() mm check Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 30/40] ARM: omap2: remove incorrect __init annotation Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 31/40] be2net: fix link failure after ethtool offline test Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 32/40] ppp: mppe: Add softdep to arc4 Greg Kroah-Hartman
2019-07-18  3:02 ` Greg Kroah-Hartman [this message]
2019-07-18  3:02 ` [PATCH 4.4 34/40] dm verity: use message limit for data block corruption message Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 35/40] kvm: x86: avoid warning on repeated KVM_SET_TSS_ADDR Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 36/40] ARC: hide unused function unw_hdr_alloc Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 37/40] s390: fix stfle zero padding Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 38/40] s390/qdio: (re-)initialize tiqdio list entries Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 39/40] s390/qdio: dont touch the dsci in tiqdio_add_input_queues() Greg Kroah-Hartman
2019-07-18  3:02 ` [PATCH 4.4 40/40] KVM: x86: protect KVM_CREATE_PIT/KVM_CREATE_PIT2 with kvm->lock Greg Kroah-Hartman
2019-07-18  7:33 ` [PATCH 4.4 00/40] 4.4.186-stable review kernelci.org bot
2019-07-18  9:19 ` Jon Hunter
2019-07-18 15:26 ` Naresh Kamboju
2019-07-18 19:47 ` Guenter Roeck
2019-07-18 20:56 ` Kelsey Skunberg
2019-07-19  4:40 ` Bharath Vedartham

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=20190718030049.943975637@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=davem@davemloft.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sashal@kernel.org \
    --cc=sergej.benilov@googlemail.com \
    --cc=stable@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: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).